Datasets:

mimiklee

/

masterthesis-longt5-sum-8k

like 0

INTRODUCTION Sexual and reproductive health is defined by WHO as 'a state of physical, emotional, mental and social well-being in relation to all aspects of sexuality and reproduction, not merely absence of disease, dysfunction or infirmity'. 1 In terms of sexually transmitted infections (STIs), younger people aged 16-24 bear the heaviest burden of chlamydia and gonorrhoea with long-term adverse health effects including ectopic pregnancy and subfertility. [2][3][4] Inequalities in sexual health persist; STIs are positively associated with lower educational levels and living in more deprived areas. 2 5-7 High STI rates among young people also reflect broader aspects of poor sexual health, such as lack of --- STRENGTHS AND LIMITATIONS OF THIS STUDY <unk> Qualitative research has an important role in gaining greater in-depth insight and complementing data of randomised controlled trials, especially if there are unanticipated results, as in the case of the safetxt trial. <unk> Two sexual and reproductive health researchers not involved in the design and implementation of the safetxt trial, independently analysed 3526 open feedback comments from trial participants and conducted 18 semistructured interviews. <unk> Obtaining results from different sources, including qualitative data from the open feedback comments and interviews reported here, in addition to the quantitative trial data allowed for triangulation of results. <unk> Limitations are that many of these optional open feedback comments were only brief, and that we had to end the interview study slightly earlier due to the COVID-19 pandemic. Open access knowledge, skills or confidence in how to carry out safer sex behaviours and how to communicate with partners about sex and desired precautions. 8 We developed the safetxt intervention delivered by text message to reduce STI infection by increasing condom use, partner notification and STI testing before sex with new partners. 9 The intervention development was informed by behaviour change theory, 9 including the 'capability, opportunity and motivation model of behaviour' (COM). 10 This model is incorporated into the comprehensive 'behaviour change wheel' model, which aims to capture the full range of intervention functions involved in behaviour change; these include education, persuasion, environmental restructuring (encouraging people to change their environment to support the behaviour), training and enablement. Each intervention function can be implemented by a wide range of evidence-based behaviour change techniques. 11 In the case of sexual behaviour, knowledge, skills, beliefs, selfefficacy, and social and interpersonal influences have important effects on COM. 8 12 Our intervention aimed to influence these factors to reduce sexual risk behaviour and encourage STI preventive behaviour. The intervention text messages were developed based on the content of effective face-to-face safer sex interventions targeting condom use, [13][14][15] the factors known to influence safer sex behaviours 16 and the views of over 200 people aged 16-24 collected in focus groups, a questionnaire and qualitative interviews. 9 The latter included telephone interviews conducted with 16 young people 2-3 weeks after enrolling in a feasibility trial in 2013. 17 The findings were used to adapt the intervention. Intervention text messages were sent with decreasing frequency over the period of 12 months (online supplemental file 1). 18 Our randomised controlled trial to establish the effects of the intervention on STI, condom use, partner notification and STI testing before sex with new partners was conducted among 6250 people aged 16-24 diagnosed at UK sexual health clinics with chlamydia, gonorrhoea or non-specific urethritis. 19 20 Control group participants received a monthly untailored text message asking for information about changes in postal or email addresses. The safetxt intervention did not reduce STIs, there were slightly more infections in the intervention group with 22.2% (693/3123) versus 20.3% (633/3125) in the control group (OR 1.13, 95% CI 0.98 to 1.31). 19 20 There were some increases in self-reported precautionary behaviours such as condom use at last sex, OR 1.14, 95% CI 1.01 to 1.28. 19 20 Although our intervention did not target sexual partnerships, we assessed at 1-year follow-up the proportion of people who had two or more partners since joining the trial and found that it was slightly higher in the intervention vs control group (56.9% vs 54.8%, OR 1.11, 95% CI 1.00 to 1.24). This result, however, was not statistically significant (p=0.06) but could have contributed to the unexpected trial outcome. Other quantitative results, including on intermediate outcomes, did not clarify either, why a statistically significant effect was shown for the condom use outcome, but not for the biologic trial outcome. 19 20 To shed further light on this, we analysed and triangulated qualitative data from two sources, including open feedback from the last follow-up questionnaires at the end of the trial (at 12 months) 20 and semistructured interviews conducted after 12 months. In this paper, we present and discuss qualitative data on participant perceptions of the safetxt intervention and of 12 months trial participation with a view to exploring for whom, how and why the intervention worked or not and what improvements could be made in the future. --- METHODS We conducted qualitative research (as part of a mixedmethods approach integrated through an advanced intervention framework with embedded methods and narrative staged reporting 21 ) including the analysis of open feedback comments collected in the 12-month questionnaire 20 of the safetxt trial and semistructured qualitative interviews with participants after completing their involvement in the trial. The research team members are mixed-methods researchers within the areas of sexual and reproductive health. During the research, we followed a 'critical realist' (CR) paradigm, 22 23 as in terms of ontology, epistemology and methodology we position ourselves in the middle of a continuum between positivism, na<unk>ve realism and objectivism [23][24][25] on the one hand and interpretivism, relativism and constructionism 23 24 on the other. According to CR, there is a reality that exists independent of our thoughts about it, and while we can become more confident about what exists by observing, existence itself is not dependent on observation. CR also sees the social world as layered, complex and an open system and characterised by change. CRs often try to answer the question 'what works for whom, when and why?' and are typically pragmatic in their approach to methodology and methods. 23 26 Below, we provide details on the two data sources used for our qualitative analysis. --- Data source 1: free-text comments The final page of the 12-month questionnaire given to all trial participants (who had provided written informed consent on enrolment) included an open-ended question: 'Did anything good or bad happen as a result of being involved in the study or receiving the text messages? Please describe'. This question was followed by a blank space that participants had the option of completing themselves. Two researchers (AG and SB), who had not been involved in the design, implementation and quantitative evaluation of the intervention, independently coded the free-text comments and categorised data by theme, using Excel 2019 and NVivo v.12 respectively. AG and SB initially took a purposive sample of 12% (n=390) of free-text comments. They ensured that participants from --- Open access different gender and sexuality groups were represented (by adding random samples of comments from each group) and included all comments from participants reporting that someone else had read their messages or reporting partner violence. This was to ensure that the feedback from participants who might have experienced unforeseen intervention side effects was coded in detail. AG and SB then independently coded these comments inductively line-by-line, considering all content (almost all of it was relevant to our research question). They then collated codes into potential themes and compared these to check for consistency of analysis and to reduce the risk of imposing own assumptions and predefined theories onto participants' narratives. Subsequently, AG and SB independently analysed all remaining free-text comments, thereby adding newly generated themes, reviewing and naming themes. AG and SB then compared their findings again (which were consistent) and discussed them within the team. The findings were compared with data from the semi-structured interviews (data source 2). --- Data source 2: semistructured interviews We purposively recruited from safetxt participants based on trial allocation and sociodemographic characteristics (age, sexuality, ethnicity, index of multiple deprivation) to encompass a variety of experiences. Eligible were participants who indicated during trial follow-up that they agreed to be contacted for further research. We sent text messages about the interview study to those who had recently (<unk>6 months) completed trial follow-up. AG and SB then approached and provided verbal and written information to those who were interested in the study. After receiving written informed consent, SB and AG conducted interviews by video conferencing (including Teams, Zoom or WhatsApp) or telephone. We initially focused on the recruitment of intervention participants and found that after 14 interviews data saturation for key themes relevant to our research question regarding the intervention had been reached (based on reflective notes, concurrent data analysis, triangulation with results of open feedback analysis and team discussions). After completing four interviews with control group participants, we had to stop study activities due to the COVID-19 pandemic and related personal circumstances and were unable to resume the work at a later stage as funding could not be extended. Interviews lasted between 30 and 90 min (average about 60 min). The interviewers (AG and SB, both female) introduced themselves as public health researchers who had not been involved in the design of the safetxt trial. Both kept reflective journals throughout the research process and engaged in self-reflexivity not only during interviews, but also during analysis to recognise and avoid imposing own assumptions and predefined theories onto participants' narratives. The interviews followed a semistructured topic guide, which aimed to explore participants' experiences regarding trial participation, whether or not they had been able to carry out the behaviours targeted by the intervention, and (for those from the intervention group) the intervention and how and why the messages did or did not help. We first explored which intervention messages participants recalled without being prompted. We then showed, sent or read to participants some of the messages and asked which, if any, they found particularly helpful or not. We also asked participants to make suggestions for improvements of the interventions. (Topic guides and example intervention messages in online supplemental file 2). New topics not included in the guide were further explored during subsequent interviews. These topics and summaries of reflective field notes were also discussed with RF and/or CF during team meetings. After completing the interviews, participants were offered a £20 voucher as a thank you for their time. --- Analysis Interviews were audiorecorded, transcribed verbatim by a professional transcription service (bound to a confidentiality agreement), and reviewed for anonymity and accuracy of transcription by SB and AG, while listening to the audiorecordings. This was also part of the first step of the thematic analysis approach that we used, including (1) familiarising ourselves with the data, (2) generating initial codes, (3) searching for themes, (4) defining and naming themes and (5) producing the report. 27 This process was iterative as analyses were conducted alongside data collection. During the early stages, SB and AG first independently developed thematic codes from the same four interview transcripts, two of which were also coded by RF, to ensure consistency of coding. Thereafter, SB and AG independently coded their interview transcripts and categorised data by theme using NVivo v.12 and Microsoft Word 2019, respectively. At the later stages of thematic analysis, Microsoft Word 2019 was used to integrate and triangulate themes developed by both researchers from both data sources, based on comparisons and team discussions. During analysis meetings with the research team, results from open feedback comments (source 1) and interviews (source 2) were triangulated with quantitative trial data (including primary, secondary and intermediate outcomes) 19 20 and data from telephone interviews conducted as part of the 2013 feasibility trial 2-3 weeks after starting messages 17 looking for consistencies and inconsistencies across the different data sources and searching for deviant cases. --- Patient and public involvement Patients and members of the public were involved in all phases of the safetxt intervention development and trial, including part of the qualitative components of the safetxt evaluation reported here. Prior to development of the safetxt intervention, possible safer sex interventions were discussed with young people in five discussion groups (25 participants). Subsequently, patients who participated in formal focus group discussions, helped to design the content of the intervention, 9 --- Open access patient representative was included in the trial steering committee. In addition, 14 patient representatives from the King's College Hospital Sexual and Reproductive Health user group helped design the patient information, consent and follow-up procedures and all trial questionnaires, including the open feedback question. Due to time restrictions, we did not seek help from patients for the design and pilot-testing of the interview topic guides, but instead gained input from four young colleagues. After the interviews, most participants indicated that they would be happy to help with the dissemination of results once published. --- RESULTS Fifty-six per cent (n=3526/6248, intervention: n=1745, control: n=1781) of participants provided comments in the open feedback section of the 12-month questionnaire, 72% of those who completed a 12-month questionnaire (table 1). Participants across all sociodemographic backgrounds provided open feedback comments, and the characteristics of respondents were similar to the characteristics of safetxt trial participants. 19 About 27% (intervention: 24%, control: 29%) of those who provided open feedback on whether anything good or bad had happened (see the Methods section for exact question) merely stated 'no', 'n/a', 'don't know', 'nothing', 'neutral', 'no difference' or a brief statement saying either they were unsure or did not notice any change as a result of participating in the study, for example, 'I carried on as usual, nothing good or bad happened'. A further 3% of comments from control group participants merely stated that they were in the control group or did not receive any intervention messages or similar. The remaining free text comments (intervention: 76%, control: 70%) provided another free text response (beyond the aforementioned statements) that was generally mostly only a few sentences long, with some participants providing longer feedback (8% of intervention and 5% of control group comments were >50 words long). We completed 18 interviews between February and May 2020. Respondent characteristics are in table 2. Open feedback was overwhelmingly positive both about the intervention text messages and being in the trial. Many intervention and control group participants commented on the usefulness and convenience of having an STI test kit sent to their home for primary outcome assessment. Intervention group participants commented positively on the tone of the intervention text messages finding them friendly, reassuring, helpful and written in a nonjudgmental manner. Participants also found that mobile phone delivery was a trusted, appropriate and convenient way to access information. Conversely, a few people in open feedback had concerns about keeping their messages private or reported that messages were annoying and many in both, the intervention and control arm, indicated that there was no change and nothing good or bad had happened as a result of being in the study. Findings from open feedback and the interviews were consistent, but interviews allowed to gain greater insight into themes that had had been generated during analysis of open feedback comments. Results from both sources are summarised by major theme below with example quotes provided in box 1 (intervention group) and box 2 (control group). --- Knowledge and awareness of safer sex Intervention group participants reported the messages were 'clear', 'concise' and 'informative'. Participants reported impact on their general knowledge of practising safer sex including new ways to protect themselves, how STIs are contracted, the risks and consequences of unprotected sex and the need to go for regular testing. Some participants appreciated intervention messages as a 'proper' source of information with links to trustworthy internet sites that clarified which information from other less reliable sources was correct. A few participants in the open feedback reported messages only said things they already knew. Many intervention participants, but also some control group participants, reported increased awareness of the importance of safer sex behaviours. Control group participants were 'indirectly' reminded of safer sex importance, because the regular texts reminded them their previous STI and/or because trial participation raised their awareness and motivation. This greater awareness reportedly influenced some intervention and control group participants in being more 'careful' in their choice of sexual partners and/or having less casual sex. --- Confidence, agency, well-being and communication Intervention group participants reported an increased confidence and agency in asserting their needs, for example, greater agency in only having sex when they wanted to. Some participants reported benefits in their sexual well-being such as, 'feeling positive' about their sex lives, respecting their body more or greater sexual pleasure through feeling more in control of their sex lives. In both the intervention and control group, sexual health was reported to be a 'difficult' and a 'taboo' subject to talk about. Sharing intervention text messages with partners, friends, housemates and siblings was a catalyst for facilitating open and honest dialogues about sexual health and helped many participants feel less embarrassed raising the topic. Showing partners messages was also used to reinforce requests to use condoms. One person reported the intervention gave them the confidence to start a new relationship after their STI. --- Changes in condom use Many intervention, but also some control group respondents reported having been'more cautious' after receiving messages and that texts were good reminders to use condoms. Several participants explicitly reported increased condom use especially with casual or new partners. Intervention group participants attributed this to increases in their Open access confidence and knowledge of how to stay protected from STIs as well as greater confidence in being able to bring up the topic of condom use. Practical tips, including to prevent condoms to break or slip off, had been particularly helpful. One participant, however, said it would be helpful to have more advice on what to do if a partner refuses to use a condom. Those who used condoms did not necessarily use them on every occasion. Reportedly, the messages also led some to encourage their peers to use protection. --- Effects on partner notification Participants in the intervention group commonly reported that the text messages enabled them to speak Open access more confidently (calmly and sooner) to their partners about their infection, impacting on how they told partners. Intervention content that chlamydia was common and easy to treat helped facilitate conversations with partners about infection. This content also reduced concerns about getting chlamydia. There were reports that the messages motivated some participants to tell partners. Some stated that the text message examples they received arrived after they had notified partners, and regretted that they had not received them earlier. Some reported only learning from the messages that the clinic could have informed their partner. Two comments referred to unknown partner contact details. --- Increased STI testing Participants from both groups reported they sought further STI testing as a result of being in the study. Messages made some participants feel it was 'Ok to get tested' and directly or indirectly reminded intervention and control group participants to test or test more frequently than they normally would have. Participants reported going for testing after having sex with a new partner (none mentioned testing before first sex). A few intervention participants reported frequent STI testing rather than condom use as a way of managing STI risk and to 'keep track of partners'. --- Box 1 Continued 'I now have regular texts and have only not had sexual intercourse without a condom once, which was as a result of me and my partner both having alcohol. (21, WSM, OF) 'No, not much has changed in regards to me because I like to, I consider myself quite a safe person so I do wear protection where I can.' (26, MSM, I) 'I have been better at using a condom-but this may be just because of getting chlamydia last year, not because of the texts.' (18, WSM, intervention, OF) 'I look back now and I realise that it definitely was a form of definitely like some sort of self-harming, of like I was just, the only way, you know, I'd have (unprotected) sex with so many people, to make myself feel bad about myself almost.(...)I just didn't really care, I had no self-respect, I didn't really care about myself, my body really,... so I think the study definitely made me realise that' (23, WSM, I) Effects on partner notification 'I think I would have gone a very different way about doing it (notifying partner), I think I would have sort of hid it away and taken, it would have taken me a lot longer to do it because I would have been embarrassed, but the text messages, like I said, they really do make you realise that you're not the only one in this situation, so...' (18, WSM, I) 'I think there was one....that made me realise that actually it's normal to not want to tell someone, and it's normal to feel really uncomfortable about it, but actually I need to tell them, and(...)the texts inspired me to reach out to my friends, and then my friends help me create a message that I then sent to people, so yeah.' (23, WSM, I) 'The text study was really helpful and insightful it helped me to be able to tell my sexual partner that I had been given a positive result for chlamydia and it helped me understand how to speak to him and tell him.' (23, WSM, intervention, OF) 'I think where it gives examples of how to tell, I think that helps, because... you don't really know how to put it, or how to start it,... a lot of people are actually quite embarrassed or they're scared of what the other person might say or they just don't know what to say so some people actually leave it, which is how other people get infected' (20, WSM, I) 'I remember thinking like 'oh this is so annoying that I got it now and not like on the day when I actually had to like tell them'.(...)Because I was thinking 'oh I've really like gone through all that like internal stress of being like how do I tell him?' and all that stuff like before and like telling him and then getting the text after.'(25, WSM, I) --- Increased STI testing 'I'd say the text messages made me get checked more often but I would have got checked anyway, but probably not as much as I did without the text messages. (21, WSMW, I) 'I got tested sooner after having had unprotected sex than I probably would have done had I not received a safer sex message text.' (21, WSM, OF) '...the texts definitely were probably part of it, but I think just sort of the maturity side of it, and sort of getting in a better frame of mind where I could ask somebody, after I had sex with them, when were you last tested, because I really didn't want to get it again.' (21, WSM, I) --- Reduction of isolation and stigma '...very helpful to feel less like you were the only one.' (21, MSW, OF) 'it was just reassuring to know that it wasn't just me getting them...' (21, WSM, I) --- Continued --- Box 1 Continued 'I think having regular texts written in the way that they were, it's really sort of like reassuring that you're not alone.(...)I'm not ashamed of my sexual health anymore, I don't think, I think before I was, I sort of thought that STIs were something to be ashamed of, but now definitely I know that they are more common than I thought they were, and they can be treated, easier than I thought they could be as well.' (18, WSM, I) 'Good for reminding you to keep getting tested and removes the stigma.' (24, WSM, OF) 'Thanks to studies like these, there is less shame relating to STI testing so I received the help I needed to get right away.' (23 years, WSM, OF) '...when you have that sort of thought at the back of your mind that it could go wrong, what if it does go wrong, I'm scared, it's, you feel sort of alone, but then with the text messages it really did help me sort of come out of that corner... I think it's... the way they were worded, it wasn't sort of, they weren't ordering me to do anything, they weren't demanding us to do anything, they were just suggesting, they were just informing, and I think that's a lot better than being sort of too firm with things.' (18, WSM, I). '...the stigma is still very much there so it's so easy to feel like 'oh I'm the only one, I can't tell anyone, I don't want people to think... because it could be one time but people assume just you're very promiscuous to get an STI... So I think it's really good... it's not just the physical treatment of it in regards to your body but like the mental treatment of it. It's like it's a common bacterial infection, just saying the word common makes people feel less alone so it could help their emotional wellbeing as well.' ( 26 --- Open access For some participants, the STI home testing kit was perceived as a central positive aspect of the study, and knowing that another'screening' test would be done made one control group participant'more inclined to use condoms'. --- Reduction of isolation and stigma Many intervention participants said that taking part in the study reassured them and reduced their feeling of being 'the only one', a common feeling after being diagnosed with an STI. Participants frequently commented on the reduction of'stigma','shame' and feeling 'less embarrassed' about having had an STI which was perceived to be reassuring and to have benefits for emotional and mental well-being. In addition, learning that STIs could be easily treated reportedly reassured participants. Some control group participants also noted feeling 'less alone' as they 'belonged to a group of people that have had chlamydia or gonorrhoea', and one reported that being in the study reduced their embarrassment about having an STI. Another control group participant, however, emphasised that the study made her 'feel less alone', but not 'feel less ashamed' and she would have liked to be in the group that received texts with support and information. --- STI diagnosis and trial participation effects Some participants from both groups reported that changes in their behaviour were a consequence of having an STI rather than receiving intervention messages. Additionally, in open feedback many in the control group commented that participating in the study enabled them to make a commitment to changing their behaviour, and a few said that it prompted them to seek help, for example, about abusive relationships. As mentioned in the relevant sections above, the control group texts simply about trial participation had reminded many to adopt precautions such as using condoms, STI testing and asking partners about their last test for STIs. A few participants mentioned that they joined safetxt when they had been at a 'turning point', and would have Box 2 Control group extracts illustrating perceived impact of having a sexually transmitted infection and trial participation 'I'm very happy to have participated and hope that you get some conclusive results.' (24, MSM, OF) 'I have been a lot more insistent of using condoms during sex. This could have been due to contracting chlamydia last summer which was treated and not wanting to get it again. I was part of the placebo group in the study but still got a text every month or so to keep my details updated. This made me thought of the study so could have reminded me anyway.' (21, WSM, OF) 'I guess I've been more inclined to use condoms and have less unprotected sex as a screening was always in the back of my mind.' (24, MSM, OF) 'Made me more aware of my sexual health by receiving the texts, it was almost like a reminder as sometimes sexual health can be at the back of your mind whereas when receiving the texts it was like a reminder and kept it at the forefront of your mind' (18, WSM, OF) '...receiving these texts made me feel good about taking steps towards being more aware and a part of something bigger that helped me be a better adult (18, MSM, OF) 'I didn't receive many messages. However, I became more conscious of my sexual health. I take precautions when I remember although, I haven't always used anything. I have been more conscious of sleeping with new people I don't know that well and have avoided this. (19, WSM, OF) 'I was sort of more wary about who I slept with, it's like I didn't sleep with as many people that I was before, I don't know if that was just because of my age or if... I don't know.(...) Like I went through a bit of a rough patch when I was younger and I feel like that sort of did include sleeping around a bit more and then I came out of it(...) and I was more like, I didn't want to just sleep with anyone, I was sort of more picky.(...) I feel like it did play a little role [joining the study), like agreeing to be part of the safetxt I think was like a turning point as well in its own right. (21, WSM, I) 'Through the whole process of being diagnosed with an STI has made me consider my life choices.... I am reluctant to have a 'one-night stand' as I have previously experienced the consequences of unprotected sex with unfamiliar people. Overall, I have thought more about my actions, not so much as a result of the texts I receive, but instead because of what has happened with my health.' (19, MSW, OF) 'I was made far more aware of how unsafe I was being, when in the past I would make more decisions in the moment which were unsafe and unthoughtful about the consequences. Having regular texts made me far more conscious about safe sex-it was a great reminder; as it is easy to forget.' (19, WSM, OF) '... I was in the group that didn't receive texts about safe sex, however just being involved in the study and completing the questionnaires gave me a greater awareness of the benefits of practicing safe sex even after the shock from my initial diagnosis wore off... (18, WSM, OF) 'Made me more cautious of who to sleep with. Due to constant reminders.' (19, MSW, OF) 'The only kind of messages I was receiving were the ones about confirming my address and contact information. In spite of that, I was still more aware to be cautious and ask people if they were getting tested etc.' ( Open access changed their behaviour anyway, but appreciated the safetxt support during this time of change. One control group participant, who had reportedly meanwhile changed due to the STI and becoming more mature, thought that safetxt support if targeted at younger people could help them avoid having to go through the same 'quite big stressful event' of having an STI. --- Recommendations for future interventions Recipients felt the intervention was especially helpful for younger people such as late secondary school/first year post school (online supplemental file 3). Many interview participants and free-text comments reported that not enough was taught in schools and the texts were much more useful than what they were taught at school. Participants mentioned additional topics that would be helpful to include, such as peer pressure to have sex, further content on dealing with people who do not want to use protection, and pleasurable aspects of condom use; a few women who have sex with men and women requested more information on safer sex between women and two men who have sex with men (MSM) wanted the intervention to cover 'chem sex' (stimulant enhanced and prolonged 'no-strings' sexual sessions between MSM connecting through apps 28 ). A few participants suggested further personalisation of safetxt messages and an option to choose from a wider range of topics from the outset (in addition to the 'text 2 to hear more' option). Some requested better mental health support to explore why people have unprotected sex. Suggestions from participants for changes in the timing and frequency of messages often focused on having some form of control of message frequency, with some wanting less messages (especially at the beginning) and others more (especially towards the end). Although many participants said that certain intervention message content would'stick' with them, some would have liked to continue receiving texts, as they served as reminders. --- DISCUSSION According to recipients, the safetxt intervention increased awareness of the importance of avoiding STIs and related knowledge about ways to prevent them. Participants reported improved confidence, agency, sexual well-being and communication about sexual health with partners, friends and family members. They attributed to these improvements, increases in condom use, STI testing, more confident partner notification and (for a few) disclosure of diagnoses. There was a reduced sense of isolation, stigma, shame and embarrassment about having an STI which reportedly reassured some participants and improved their emotional well-being. Participants from both the intervention and control group reported that having an STI influenced their safer sex behaviours. Control group participants reported that taking part in the study had influenced their commitment to safer sex behaviours. The control group text message about trial participation reminded many about the importance of safer sex and acted as a trigger for STI testing and condom use. Our qualitative analyses of interviews and open feedback are mainly consistent with the trial results. However, recipients' reports suggest larger differences in behaviour than were demonstrated in the trials results. Possible reasons include social desirability bias, an incorrect attribution of changes in behaviour to the intervention rather than the experience of STI and a strong Hawthorne effect 29, including the trial participation messages sent to the control group reportedly acting as a prompt for safer sex behaviours. Our findings suggest that young people felt positive impacts of the safetxt intervention on their sexual and reproductive well-being. These benefits include increase in confidence, agency, communication and precautionary behaviours. 1 The perceived value of safetxt from recipients' accounts accords with the trial results showing higher condom use at 12 months. The'spill-over' effect resulting from participants reportedly encouraging their peers to use condoms, was not quantitatively assessed during the trial. Recipients accounts that the main perceived benefit of the intervention was in 'how' to tell partners rather than 'whether' to tell them about their STI accords with the only slightly higher levels of partner notification in the intervention group. The results are not in line with the public health impact of safetxt as the trial results found STIs were not reduced, with slightly more infections in the intervention group. The trial results suggested (although not statistically significant) that there were slightly more participants with two or more partners and a new partner in the intervention group compared with the control group during the course of the trial (altering partnerships was not an intervention aim). The findings from this study involving interviews and feedback obtained after the 12-month intervention of the safetxt trial were in keeping with the findings from telephone interviews conducted in 2013 during the intervention development 2-3 weeks after receiving the first messages, but included longer-term impacts. 17 A strength of the interviews and the open feedback analysis was that it was conducted by two researchers not previously involved in the intervention development or trial.

Objectives People aged 16-24 are more likely than other age groups to acquire sexually transmitted infections (STI). Safetxt was a randomised controlled trial of a theorybased digital health intervention to reduce STIs among 16-24 year-old people in the UK. We report results of qualitative research regarding participants' perceptions and experiences of the intervention and trial participation. Design Qualitative thematic analysis following a critical realist paradigm of written open feedback comments provided in the 12-month follow-up questionnaire and semistructured interviews. Setting Safetxt trial participants were recruited from UK sexual health clinics. Participants Trial inclusion criteria: people aged 16-24 diagnosed with or treated for chlamydia, gonorrhoea or non-specific urethritis. Optional open feedback provided by 3526 of 6248 safetxt participants at 12 months and interviews with a purposive sample of 18 participants after the trial. Results We summarise and report results in seven broad themes. According to recipients, the safetxt intervention increased awareness of the importance of avoiding STIs and ways to prevent them. Participants reported improved confidence, agency, sexual well-being and communication about sexual health with partners, friends and family. Recipients attributed increased condom use, increased STI testing after (rather than before) sex with new partners, and more confident partner notification to the intervention. Recipients described a reduced sense of isolation and stigma in having an STI. Control group participants reported that having had an STI and receiving control texts asking them to report any changes in contact details acted as reminders to use condoms and get tested. We also summarise participant recommendations for future interventions and studies. Conclusions While control group participants reported precautionary behaviours were 'triggered' by trial participation, intervention recipients reported additional benefits of the intervention in increasing precautionary behaviours and in broader aspects of sexual health such as confidence, communication, emotional well-being and agency. Trial registration ISRCTN registry ISRCTN64390461.

'spill-over' effect resulting from participants reportedly encouraging their peers to use condoms, was not quantitatively assessed during the trial. Recipients accounts that the main perceived benefit of the intervention was in 'how' to tell partners rather than 'whether' to tell them about their STI accords with the only slightly higher levels of partner notification in the intervention group. The results are not in line with the public health impact of safetxt as the trial results found STIs were not reduced, with slightly more infections in the intervention group. The trial results suggested (although not statistically significant) that there were slightly more participants with two or more partners and a new partner in the intervention group compared with the control group during the course of the trial (altering partnerships was not an intervention aim). The findings from this study involving interviews and feedback obtained after the 12-month intervention of the safetxt trial were in keeping with the findings from telephone interviews conducted in 2013 during the intervention development 2-3 weeks after receiving the first messages, but included longer-term impacts. 17 A strength of the interviews and the open feedback analysis was that it was conducted by two researchers not previously involved in the intervention development or trial. We analysed all of the open feedback comments left by over 3500 trial participants (56% of participants who enrolled into the trial and 72% of those who completed the 12-month questionnaire). The experience of those not leaving a free-text comment may be different from those who did. However, the characteristics of respondents were similar to the characteristics of trial participants including those from diverse sociodemographic and ethnic groups. It is not possible to blind participants receiving a behavioural intervention, which could introduce bias when obtaining feedback. All open feedback comments were brief, optional and completed at the end Open access of their involvement in the trial, so it was not possible to explore participant views in depth or follow-up on feedback. During interviews, however, (and despite having to stop the interview study slightly earlier due to the COVID-19 pandemic), we were able to gain greater insight into themes that had been generated during analysis of open feedback comments. Our qualitative analyses provide little direct evidence to explain the unanticipated quantitative trial findings, but raise some plausible explanations. Both qualitative analyses and quantitative analyses of intermediate and secondary trial outcomes showed increased correct condom use self-efficacy and increased condom use. This effect did not seem big enough to translate into reduced reinfection rates in the intervention group, given that those who reported increasingly using condoms did not necessarily use them on every occasion. In addition, a few intervention participants seemed to prefer a secondary prevention approach with frequent STI testing over a primary prevention approach with consistent condom use. In both intervention and control groups, there were large reductions between baseline and follow-up in the number of partners in the preceding year, as would be expected if high-risk people were reverting to the norm. However, there was a marginally smaller reduction in the intervention group. Previous trials of group interventions targeting those at high risk for STI, have had unanticipated effects in normalising risk behaviours and increasing STI. 30 The'shock' of having had an STI and receiving control group messages reminding them of their STI might have deterred control group participants for a longer period from engaging in new relationships than intervention group participants. Some intervention recipients reported feeling less ashamed about their STI, generally more confident in discussing sexual health and/or reassured that their infection could be easily treated. Lower stigma about having an STI carried benefits in emotional well-being and reportedly gave a few the confidence to start a new relationship following their STI. While this was a positive outcome from recipient's perspectives, starting a new relationship confers some additional STI risk. Whether that risk is worth taking depends on what people are getting out of new relationships. Our analysis suggest intervention recipients were better equipped to get the sex they want (and to avoid the sex they do not want). The trial suggests that sex was no less a risk for STIs, but it may have had more value to them. Our qualitative analysis also suggested that testing after sex with new partners was increased, but not before first sex. The safetxt trial indicator assessed STI testing 'prior' to first sex with new partners (showing no difference between groups), whereas the few previous mHealth trials we identified in a systematic review that showed an effect on STI/HIV-testing, only enquired about whether participants had an STI test within a specified time period. 31 Secondary analysis of the safetxt trial data looking at overall testing data in clinics (rather than self-reported tests 'prior' to first sex) is consistent with this with slightly higher clinic testing for STI in the intervention groups (1549/3123, 50%) vs the control group (1477/3125, 47%). --- Conclusion This research has described the perceived impacts of receiving the intervention and control group messages on participants. A randomised controlled trial was needed to identify slightly higher STI diagnoses in the intervention group. The qualitative findings and trial results both show that the components of the safetxt intervention promoting condom use were effective. Since this is a unique finding not seen in any previous similar mHealth interventions, 31 service providers could consider delivering this content. Further research could consider recipients recommendations for future interventions and explore how to achieve and measure positive impacts of reduced stigma about having an STI and increase sexual well-being as well as reduce subsequent STI. Twitter Sima Berendes @BerendesSima --- Competing interests None declared. Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details. Patient consent for publication Not applicable. Provenance and peer review Not commissioned; externally peer reviewed. --- Ethics approval Data availability statement Data are available on reasonable request. Deidentified data collected as part of the randomised controlled trial will be made available via the data sharing portal FreeBIRD after publication of the primary and secondary analyses as outlined in the Data sharing statement of the trial publication Free et al (2022). Part of the study materials and anonymised extracts of the interview study conducted after trial completion, have been included in online supplemental files of this article. Sharing of further anonymised qualitative data extracts on reasonable request would have to be in line with data protection laws and subject to appropriate ethics committee approval.

Objectives People aged 16-24 are more likely than other age groups to acquire sexually transmitted infections (STI). Safetxt was a randomised controlled trial of a theorybased digital health intervention to reduce STIs among 16-24 year-old people in the UK. We report results of qualitative research regarding participants' perceptions and experiences of the intervention and trial participation. Design Qualitative thematic analysis following a critical realist paradigm of written open feedback comments provided in the 12-month follow-up questionnaire and semistructured interviews. Setting Safetxt trial participants were recruited from UK sexual health clinics. Participants Trial inclusion criteria: people aged 16-24 diagnosed with or treated for chlamydia, gonorrhoea or non-specific urethritis. Optional open feedback provided by 3526 of 6248 safetxt participants at 12 months and interviews with a purposive sample of 18 participants after the trial. Results We summarise and report results in seven broad themes. According to recipients, the safetxt intervention increased awareness of the importance of avoiding STIs and ways to prevent them. Participants reported improved confidence, agency, sexual well-being and communication about sexual health with partners, friends and family. Recipients attributed increased condom use, increased STI testing after (rather than before) sex with new partners, and more confident partner notification to the intervention. Recipients described a reduced sense of isolation and stigma in having an STI. Control group participants reported that having had an STI and receiving control texts asking them to report any changes in contact details acted as reminders to use condoms and get tested. We also summarise participant recommendations for future interventions and studies. Conclusions While control group participants reported precautionary behaviours were 'triggered' by trial participation, intervention recipients reported additional benefits of the intervention in increasing precautionary behaviours and in broader aspects of sexual health such as confidence, communication, emotional well-being and agency. Trial registration ISRCTN registry ISRCTN64390461.

Participants The survey was publicised among doctors, nurses, laboratory staff and allied health professionals. 3109 people answered the survey, of whom 472 (15%) were considering going to work in the epidemic but had not yet volunteered. 1791 (57.6%) had not considered going, 704 (22.6%) had considered going but decided not to, 53 (1.7%) had volunteered to go and 14 (0.45%) had already been and worked in the epidemic. --- Results For those considering going to West Africa, the most important factor preventing them from volunteering was a lack of information to help them decide; fear of getting Ebola and partners' concerns came next. Uncertainty about their potential role, current work commitments and inability to get agreement from their employer were also important barriers, whereas clarity over training would be an important enabler. In contrast, for those who were not considering going, or who had decided against going, family considerations and partner concerns were the most important factors. --- Introduction On 21st March 2014, the World Health Organisation was officially notified of an outbreak of Ebola virus disease due to Zaire ebolavirus in Guinea, Liberia and Sierra Leone. The outbreak was declared a "public health emergency of international concern" on 8th August 2014 [1]. As of January 9 th 2015, a total of 21086 cases (13376 laboratory confirmed) and 8289 deaths have been reported [2]. The epidemic is currently doubling approximately every 4 weeks and the case fatality rate, when based on the most accurate available information, is around 70 percent [1]. Small numbers of cases have occurred in Nigeria, Senegal and Mali. In late September the outbreak became transcontinental with the importation of a previously subclinical case to Texas, USA from Liberia [3]. Onward transmission occurred in the healthcare facility in Texas [3], and also occurred in Spain, following the repatriation of an infected healthcare worker [4]; more countries in West Africa may also be at risk of Ebola [5]. This is an exceptional situation with the potential for spread to almost any country in the world [6]. The global response to the outbreak has been slow. As early as April 2014 Médecins Sans Frontières (MSF) warned that this outbreak was "unprecedented" [7]. MSF has criticised the speed of response on several occasions [8,9] and on 5th September 2014, the number of deaths reported to WHO in this outbreak surpassed those in all other known outbreaks combined [1,[10][11][12][13][14][15]. In October 2014 Oxfam suggested that the world had only two months to get the epidemic under control [16]. Tackling the current Ebola virus outbreak requires a global response in terms of money, infrastructure and people. On the 21st October 2014 MSF had only 270 international staff and 3018 local staff working in Guinea, Liberia and Sierra Leone [17]. The World Bank have called for at least 5000 more medical and support staff [18]. In addition to the World Bank, organisations such as MSF, WHO and UNICEF have called for more qualified staff to help [19]. In the UK, approximately 1000 healthcare workers have so far volunteered to go to West Africa to help in the response [20]. Many more have considered going, but not yet volunteered. There are likely to be many factors that influence a person's decision regarding whether or not to volunteer in a situation like this. We wanted to understand what these factors are, and in particular whether any of them might be amenable to intervention or influence. Knowledge of the relative contributions of the different enablers and barriers might guide UK policymakers as to what is needed to ensure more healthcare workers volunteer to help control the Ebola outbreak. Therefore, we conducted a survey of UK health professionals to understand their attitudes towards going to work in the Ebola epidemic in West Africa. --- Methods To understand what some of the potential barriers and enablers might be that influence the decision of healthcare workers over volunteering to go to West Africa we examined social media, blogs and online comments (see S1 File for information sources). We also conducted small focus group discussions of healthcare workers. Based on these we produced a draft questionnaire which we piloted on a small number of different healthcare workers before modifying into the final version (S3 File). Briefly, the questionnaire asked whether respondents had considered going to work in the Ebola outbreak and what decision they had come to. Two questions investigated what the barriers and enabling factors were according to a 5 point Likert scale from "strongly agree" to "strongly disagree." The fourth question concerned where respondents got their information on Ebola from, and subsequent questions gathered demographic information such as profession, age, sex and level of experience. Free text boxes were included to pick up any other concepts not initially identified in the questionnaire and to enable participants to elaborate on their responses. We used the web based Surveymonkey to create and distribute the questionnaire. --- Ethics statement The questionnaire and study protocol were approved by the University of Liverpool research ethics committee (RETH 000774). The survey went live on Wednesday 15th October 2014 and was disseminated using multiple means including various professional colleges, societies, training bodies, letters to the BMJ [21] and the nursing press. A list of the organisations that disseminated the questionnaire is shown in the S3 File. It was also advertised informally by word of mouth and using social media. The survey, which can be found at www.surveymonkey/s/HPRUebola, was entirely anonymous. Initial data were reviewed after one week, the free text comments were analysed and recurrent concepts identified (1450 respondents). These initial responses were used to modify the questionnaire through inclusion of 2 additional barriers and 4 enabling factors (S4 File). The revised questionnaire went live on Wednesday 22 nd October 2014. Responses were downloaded as comma separated values at 9:20pm on 4 th November 2014 (S1 Dataset). Analysis was conducted using R software version 2.15.3 (R core team 2013). Responses were analysed descriptively and proportions of respondents giving certain answers were calculated. In order to rank the relative importance of barriers and enablers, values were assigned to responses on the Likert scale ("Strongly agree", +2; "Agree", +1"; "Neither disagree nor agree", 0; "Disagree", -1; "Strongly disagree" -2) and the total score for each barrier or enabler was calculated. We wished to explore whether demographic or other factors accounted for any responses observed. We were also interested in identifying whether particular barriers may cluster together. This is important because these barriers would all need to be addressed to affect an individual's decision; barriers that did not cluster could be dealt with in isolation. In order to address these questions simultaneously, we used redundancy analysis, a form of multivariate analysis that combines a principal component analysis, to identify clusters, with regression to identify significant explanatory variables; this analysis used the R package "vegan" [22], according to the methods described by Borcard [23]. A matrix of explanatory variables was constructed for the redundancy analysis, on which the response to each barrier was regressed. A forward selection process was used to select significant variables which explained the greatest proportion of variance in the response data, and permutation tests used to test significance of RDA axes. Triplots were produced according to correlations between variables (scaling 2 in the vegan package). --- Results A total of 3109 people completed the survey between 15 th October and 4 th November 2014. Two thousand and ninety eight (68%) respondents were doctors, 674 (22%) were nurses and the remainder were a mixture of armed forces health professionals, paramedics, pharmacists and a wide range of other allied health professionals (3% did not give their profession). The largest group of respondents, 943 (31%), worked in acute specialties such as acute medicine, emergency medicine and intensive care (Table 1). Medical specialties came next with 728 (24%) respondents, followed by others, including primary care, infection specialties, paediatrics, surgery and obstetrics & gynaecology. Respondents were generally experienced, 77% having more than 5 years of experience since their primary health care qualification and 55% with more than 10 years. Fifty-one percent of those answering had children or other dependents at home. Four hundred and seventy-two (15%) respondents were considering going to West Africa to help with the Ebola virus epidemic, but had not yet volunteered ("Considering"); 1791 (58%) had not considered going ("Not Considered"); 704 (23%) had considered it and decided not to go ("Decided Against"); 53 (1.7%) had made definite plans to go and 14 (0.4%) had already been to West Africa to help in the outbreak. Our analysis focussed on the 472 people in the Considering group, because this is the group who may be willing to go to West Africa. For people in this group, the most important barrier identified for not yet having volunteered was insufficient information to reach a decision (Fig. 1). Some of the areas of what information is required can be summed up in this quote: "Lack of information is my main barrier. I have no idea whether my skill-set would be useful there, if I am needed there, how to go about joining the efforts, or how to negotiate the time off with my trust. Any information would be greatly appreciated." The main areas where information could be targeted are shown in Table 2. In particular, although there is a dedicated website where NHS employees can express their interest (http://ukmed.humanities.manchester.ac.uk/), it is clear from these responses that this website is not widely known by many people (including those who would like to go and help). Additionally the information people are seeking is not available, either on that website or elsewhere. People would appear to welcome a more direct appeal for help. Amongst doctors in training in particular there was a need for clarity on how it would affect their training programmes. There were also many comments regarding the lack of information concerning exactly what skills or experience would be useful. Finally the main area of clarity required is getting timely responses from the organisations sending people out to help. Areas of lesser concern that nevertheless caused respondents to comment were the risk of contracting Ebola and medical evacuation. The next two barriers were the fear of getting Ebola, and a partner's concerns about them going. Further important issues for the Considering group included uncertainty around what their role would be in the epidemic, work commitments at home and not being able to get sufficient time off from their employer. The free text comments exemplified many of these issues: "Lack of information is my main barrier. I have no idea whether my skill-set would be useful there, if I am needed there, how to go about joining the efforts, or how to negotiate the time off with my trust." (Male, gastroenterology registrar.) "As a junior doctor I want to help but don't know if I would be eligible or useful. It would be really helpful if UKIEMR/other organisations published a list of necessary/desirable qualities."(Male, F2, acute medicine,) "My host trust refuses to let me go due to significant staff shortages of middle grade doctors of appropriate training to fill my place, and also currently they do not have the funds to Table 1. Demographic information for all the respondents who completed the survey, divided according to the answer to whether they have considered going to West Africa to help in the current Ebola virus epidemic. replace me for 3 months. I offered to take unpaid leave and for my salary to be used by the Trust to fund a locum". (Female, emergency medicine registrar.) --- Considering The barriers for people in the Considering group were very different to those in the Not Considered and Decided Against groups; for these two groups, many of the issues were similar. Thus family commitments and partners concerns were the two most important issues, and insufficient information was much less important (Fig. 1). Of note, nurses in the group that Decided against going to work in the epidemic were more likely that other groups of respondents to answer that their employer prevented them from volunteering. The demographics of the groups also differed; 43% of those considering going were in the 26-35 age group, compared with 36% of those not considering going, and 39% of those who had decided against going (Table 1; <unk> 2 test p = 0.002). Those considering going were also less likely to have children (<unk> 2 test p <unk> 0.0001). They also less frequently "Agreed" or "Strongly Agreed" with the questions that emphasised the barriers to going, for example they were significantly less afraid of contracting Ebola (<unk> 2 test p <unk> 0.0001). Interestingly, fear of Ebola was associated with getting information predominantly from the media in the group considering going but who had not yet volunteered. Conversely, across all respondents to the survey, Barriers and enablers to going to West Africa to help with the Ebola outbreak for four groups of respondents. The importance of each issue is indicated on a 5 point Likert scale from strongly disagree to strongly agree, for those who were considering going but had not yet decided ("Considering"); those who had not considered going ("Not Considered"); those who had considered it and decided not to go ("Decided Against"); those who had volunteered and were waiting to go ("Volunteered"), and those who had already been ("Already Been"). Issues marked * were introduced in the second version of the questionnaire from 22 nd October onwards (1450 responses). Data are the percentage of respondents giving the answers indicated; and the rank is indicated showing how important that issue was for that group. The values from which the figure is derived are given in S1 and S2 Tables. --- Area of information required Example responses Exactly how people can volunteer "I am frustrated. I don't know how to get out there. I want to go" "I really want to go, but I don't know how and with which organisation." "I would love to go and help, but am unsure of how to go about doing it" "I would be keen to go but I don't know how I would get involved" "Don't know how or to whom I can talk to to make this happen" "I would love to go and help out in West Africa but I have no idea how I would go about volunteering" "I haven't come across a 'one-stop' site for info re potential NHS volunteers..........is there one I've missed?" "Need a simple and well publicised sign up procedure" "I would have liked the opportunity to attend an event for potential volunteers to get an idea of whether I could do something useful there without any commitment at this stage" Need for information to be more directly disseminated to front line healthcare workers "I would have liked the opportunity to attend an event for potential volunteers to get an idea of whether I could do something useful there without any commitment at this stage" "If you need NHS staff to volunteer to help with the Ebola outbreak, you need to approach staff more directly. I cannot remember having any email requesting that I consider it." "Until I read the letters in the BMJ I wasn't aware that volunteers were being proactively recruited" "I have not been made aware of any official drive for UK doctors to travel to the outbreak, but would certainly be interested in doing so. Do get in touch." "Bring information to us, i.e. a recruitment drive within the NHS to give us all the info" "I personally haven't been aware of any campaigns/information given to health care workers asking for volunteers. Only aware of it due to media coverage" getting information predominantly from the medical literature (irrespective of how much information was from the media) was associated with a roughly fourfold decrease in fear of contracting Ebola (<unk> 2 test p <unk> 0.0001). Need For those considering going to West Africa who had not yet volunteered, the most important "enabling" factor, which would make them more likely to go was more information and training (Fig. 1). The availability of effective treatments and/or a vaccine against Ebola also featured highly as factors that would reassure the survey respondents. Interestingly, reassurance about repatriation in the event of contracting Ebola did not feature highly as a barrier, except in the small group of respondents who had actually been to work in the epidemic, where it ranked as the number one concern. To determine how demographic and other factors influenced the barriers, and whether any particular factors clustered together, we performed redundancy analyses, both on the whole dataset and for the three main groups: Considering, Not Considered, and Decided Against. For this analysis the responses to the question regarding barriers to going were the independent variables, and the demographic (or other) data were the explanatory variables (e.g. age, previous experience, profession, specialty). All variables that showed a significant association with responses regarding barriers are shown on the redundancy analysis plots, Fig. 2 and S1 Fig. The demographic characteristics for the Considering group were very diverse, but in general, those variables that might be expected to cluster together did so. For example, having children was closely associated with having family commitments and reporting a partner's concerns as barriers. In this analysis lack of sufficient information, which was the most important barrier, was not strongly associated with any other variable, but was loosely associated with two other clusters of barriers: fears of unrest, and worries about not being repatriated if unwell with Ebola and concerns about what the role would be, and insufficient experience and fear of catching Ebola. These factors were less of a concern for the subgroup of people who were considering going and already had experience of working in sub-Saharan Africa; such healthcare workers were more likely to be male and to have longer professional experience. Across all respondents to the survey, concerns about leaving families or partners were more frequently reported by people with children, and particularly in the 36-45 age group. People with previous experience in sub-Saharan Africa were less concerned that they did not know what they would actually do, or that they did not have the right experience; they were also less concerned about civil unrest or not being repatriated. People who reported that they obtained most of their information about the Ebola crisis through the media were more likely to feel --- Area of information required Example responses Risk of infection "Although I am potentially willing to help there is no clear information about current infection rates amongst staff volunteering, and about methods of keeping safe...... I need this information to weigh the risks to myself and family against the urge to assist. I do not know where this information could be found." Need for information regarding evacuation "I contacted NGOs working in the West Africa and I once agreed to work in the field with one of these NGOs. However, I could not obtain appropriate information on risk management such as emergency evacuation; thus my employer did not agree for me to work in the west Africa." "... there is no clear guidance about what would happen if one of us contracted the virus." "The main concern is the uncertainty regarding medical evacuation" "Once.....the rules for medical evacuation are defined I would most likely be willing to go..." "I would find it easier to decide if I had more information re options for participating, such as job descriptions, location, set-up, training, medical insurance, medical evacuation if needed." doi:10.1371/journal.pone.0120013.t002 they did not have enough information to inform their decision, and had less idea of what would be expected of them. Doctors with increasing professional experience were most concerned about the impact on their colleagues and families. The younger age groups, and those in allied health professions (pharmacists, biomedical scientists, paramedics and nurses) had more concerns that their career would be adversely affected, or reported that their employer would not allow them to go. These categories of respondents also tended to get more of their information from the media. Those working in infection specialties were the least likely to report having insufficient information. --- Discussion In this study, 15% of more than 3000 healthcare workers who responded said they are considering going to West Africa to help with the Ebola outbreak; the primary reason why they had not yet volunteered was a lack of information to help them decide. In addition, concerns about what their role would be, and the attitudes of their employer also contributed. All of these are factors could potentially be modified. Fear of contracting Ebola also featured highly among the reasons for not yet volunteering, as well as the concerns of a partner. Our redundancy analyses indicated that, as a barrier, lack of information did not cluster tightly with any of the other barriers, though it did associate loosely with respondents' concerns about what their role would be, and having insufficient experience, fears of civil unrest, of catching Ebola and of not being repatriated if unwell with the disease. Many of these concerns would likely be allayed with appropriate information, underscoring the importance of this one factor. The absence of adequate information may leave health workers getting more information from the media, which we have shown is associated with greater fear of contracting Ebola compared with obtaining information from more definitive sources such as the medical literature. Nearly 85% of respondents had either not considered, or had decided against going to help in the epidemic. The overwhelming reasons for this were a partner's concerns, and family commitments, especially having children or other dependents. Interestingly the responses of those two groups were very similar, suggesting that even those who reported they had not considered going, may have considered it at some level, and decided not to pursue it. Compared with these two groups, those considering going to West Africa were less likely to have children or dependents, and generally perceived all of the barriers to be less of a hindrance, except for lack of access to information. Because lack of information seemed to be a key factor, we examined the websites of the main organisations who are sending volunteers to work in the Ebola epidemic (The British Red Cross, International Medical Corps, Médecins Sans Frontières, Save the Children, and UK-Med) to see what information was available (Table 3). In general, these organisations are calling for doctors and nurses who have a reasonable amount of experience (over 3 years post registration), preferably including experience in low resource settings. The desirable specialties were emergency medicine, infectious diseases, critical care and paediatrics. Another issue highlighted by those who were considering going was the need for training specific to the tasks that will be carried out. Many potential volunteers seemed unaware that rigorous training is included in the typical 4-6 week deployment. This again highlights the provision of reliable information as a key missing component of the current response. The redundancy analyses for all the groups of respondents showed a strong relationship between reporting there was insufficient information, and obtaining most of their information from the media. Some respondents commented on the fact that if they were asked directly to help out (e. g. via an e-mail), they would be more likely to consider it, rather than just hearing about the need for volunteers via the media. Additional barriers identified by our respondents in free text comments were uncertainty about whether pay would continue as normal and lack of clarity about whether employers would release NHS staff members. A limitation of our study is that it used a convenience sample and is unlikely to be completely representative of all UK healthcare workers. Given the fast moving nature of the Ebola epidemic the time necessary to fully assess response bias in the UK health worker population would preclude the study results being sufficiently useful to inform policy. A total of 3109 people completed the survey, equivalent to 0.33% of the 937,000 registered doctors, nurses and midwives in the UK [29,30]. Thirty-one percent of respondents worked in acute care specialties, 68% were doctors and 22% were nurses. These figures do not reflect the UK health care worker population, in which nurses outnumber doctors more than two to one. This is perhaps the most significant limitation of our study, as there is a greater need for skilled nursing care than medical expertise in the epidemic. It is possible that this imbalance in respondents has led to us wrongly identifying significant barriers. For example, nurses who had decided against going to work in the epidemic were more likely to cite their employer as a reason not to go than doctors were. However, it is of note that, in the Undecided group, there were no significant differences between the nurses responding and any other group. Two percent of our respondents have either made definite plans to go to West Africa, or have actually been and worked in the epidemic. Generalised across the whole health worker population, this would equate to nearly 19,000 volunteers. In reality, the number is nearer 1000. This confirms the expected bias in our responding population in favour of people who are more interested in helping with the epidemic. However, given that this is primarily the group of interest, particularly the barriers and enablers to them going to West Africa, we do not think this is a major limitation. A further limitation is the fact we changed the questionnaire during the study. Nevertheless, the central conclusion of our study, that lack of information is hindering potential volunteers, is based on questions that were present in all versions of the questionnaire and was not influenced by the alteration a week into the survey period. A final limitation is that we did not specifically ask if the destination country would influence willingness to volunteer. The majority of British health workers volunteering in the Ebola epidemic are being deployed to Sierra Leone because of Britain's historic links with that country. We detected no evidence from free-text comments that the country of destination would influence the decision to volunteer. In summary, our study has shown that many more people are considering going to West Africa than have actually signed up, and one of the major factors holding them back is lack of information. Policies which were aimed specifically at addressing this, such as a well-publicised, high quality portal of reliable information, would likely result in more UK healthcare workers volunteering to help tackle Ebola in West Africa. --- Data are available as a supplementary file. This file has been modified very slightly to remove any potentially personal identifiable information.

To understand the barriers and enablers for UK healthcare workers who are considering going to work in the current Ebola outbreak in West Africa, but have not yet volunteered.After focus group discussions, and a pilot questionnaire, an anonymous survey was conducted using SurveyMonkey to determine whether people had considered going to West Africa, what factors might make them more or less likely to volunteer, and whether any of these were modifiable factors.

INTRODUCTION Networked information technologies have changed the way people use and reuse creative -and frequently copyrightedmaterials. This change has generated excitement, and heated Best Paper Honorable Mention at CHI 2011. debate, among content-creators, technologists, legal academics, and media scholars. Media theorist Lev Manovich argues that remixing is an ancient cultural tradition (e.g., he has suggested that ancient Rome was a "remix" of ancient Greece) but that information technologies have accelerated these processes and made remixing more salient [14]. Sinnreich et al. argue that "configurable culture" has been significantly transformed by networked technologies which introduce perfect copying and allow people not only to be inspired by extant creations but to remix the original works themselves [25]. Legal scholars have stressed the importance of remixing in cultural creation broadly and warned that current copyright and intellectual property laws may hinder creativity and innovation [11,1]. Several of the most influential scholarly explorations of remixing as a cultural phenomenon have focused on youth's remixing practices. For example, work on remixing by Jenkins [10] and Ito [9] has focused on young people's use and re-use of media. Palfrey and Gasser have suggested that the cultural practices of "digital native" youth have had a significant transformative effect on our culture [17]. Throughout his book "Remix," Lessig uses youth's reuse practices to support an argument against what he considers excessive copyright legal protection [12]. Yet, despite a wide interest in remixing and authorship, researchers have only recently engaged in empirical research on the subject [4]. Several recent treatments have presented studies of video remixing communities [5,24], music remixing communities [4], collaborative video game communities [13] and social network sites [18]. There is also another quantitative study of our empirical setting [8] focused on characterizing the variety of responses to remixing. These studies have tended to be general and largely descriptive ex-arXiv:1507.01285v1 [cs.HC] 5 Jul 2015 aminations of remixing practice. This work has pointed to the existence of norms [5] and the territoriality of digital creators [26] and has considered issues of motivation [4]. However, empirical work has yet to unpack in detail the key social mechanisms that scholars have suggested drive behavior, norms, and motivation in remixing communities. Perhaps no mechanism has been more frequently cited as critical for remixing activity than attribution and the related phenomena of plagiarism, reputation, and status. For example, recent survey-based work has suggested that the "authenticity and legitimacy" of creative work "are premised on the explicit acknowledgment of the source materials or 'original creator"' and that such acknowledgment is a key component of how adults assess the fairness or ethical nature of content reuse [25]. Attribution, in this sense, can be seen as an important way that people distinguish remixing from "theft." Judge and law professor Richard Posner stresses the importance of attribution and explains that this is important even when there is no monetary benefit to being attributed. For example, he explains that European copyright law is based on a doctrine of "moral rights" that "entitles a writer or other artist to be credited for his original work and this 'attribution right', as it is called, would give him a legal claim against a plagiarist." Posner also explains that "acknowledgment" of another's contributions to a derivative negates any charge of plagiarism, although it may not establish originality [20]. Attribution plays such an important role in remix culture that Creative Commons made a requirement for attribution a component of all their licenses after more than 97% of licensors opted to require attribution when it was offered as a choice [2]. Young people's perceptions of attribution and complications around copying have also been examined. An article by Friedman describes that adolescents who allowed "computer pirating" -the unauthorized copying of computer programs -did so because technological affordances made it difficult for adolescents to identify "harmful or unjust consequences of computer-mediated actions" [7]. In a second study, psychologists Olson and Shaw have found that by five years old, "children understand that others have ideas and dislike the copying of these ideas" [16]. Yet, despite the fact that researchers in human computer interaction have begun to explore the complexity of attribution and cited its importance to remixing [13], many designers of online communities pay little attention to issues of attribution in their designs -a fact that is reflected in user behavior. For example, research on the use of photos from the photo sharing site Flickr [22], as well as a number of other usergenerated content communities [23], suggests that most reusers fail to attribute re-used content in ways that public-use licenses require. Although theory and survey based work points to a need to design for attribution in user-generated content communities, we still know very little about how attribution works or how designers might go about doing so. Indeed, our study suggests that the most obvious efforts to design for attribution are likely to be ineffective. In this paper, we employ a mixed methods approach that combines qualitative and quantitative analyses to explore users' reactions to attribution and its absence in a large remixing community. First, we introduce our empirical setting; using qualitative data from users forums and comments, we present a rich description of remixing and evidence to support our core proposition that credit plays a central role in remixing in our environment. Second, we contextualize and describe a technological intervention in our setting, responding directly to several user suggestions, that automated the attribution of creators of antecedent projects when content was remixed. Third, we present a tentative quantitative analysis of the effect of this intervention along with a parallel analysis of the practice of manual creditgiving. We find that credit-giving, done manually, is associated with more positive reactions but that automatic attribution by the system is not associated with a similar effect. Fourth, we present analysis of a set of in-depth interviews with twelve users which helps confirm, and add nuance and depth to, our quantitative findings. Our results suggest that young users see an important, if currently under-appreciated and under-theorized, difference between credit and attribution. Credit represents more than a public reference to an "upstream" user's contributions. Coming from another human, credit can involve an explicit acknowledgment, an expression of gratitude, and an expression of deference, in a way that simple attribution can not. Our results suggest that identical attribution information means something very different to users when it comes from a computer, and when it comes from a human -and that users often feel that acknowledgment is worth much less when it comes from a system. We conclude that designers should create affordances that make it easier for users to credit each other, rather than to merely pursue automated means of acknowledgment. Our study offers two distinct contributions for social scientists and for technology designers. The first is an improved understanding of the way that attribution and credit work in user-generated content communities. The second is a broader contribution to the literature on design that suggests an important limitation to technologists' ability to support community norms and a suggestion for how designers might create affordances. Functionality that allows users to express information that a system might otherwise show automatically may play an important role in successful design for social media environments. --- SCRATCH: A COMMUNITY OF YOUNG REMIXERS The Scratch online community is a free and publicly available website where young people share their own video games, animated stories, interactive art, and simulations [15]. Participants use the Scratch programming environment [21], a desktop application, to create these interactive projects by putting together images, music and sounds with programming command blocks (See Figure 1). The Scratch website was officially announced in 2007 and, as of September 2010, has more than 600,000 user accounts who have shared 1.3 million projects. At the time of writing, Scratch users share on average one new project per minute. Examples of projects range from an interactive virtual cake maker, to a simulation of an operating system, to a Pokemon-inspired video game, to an animation about climate change, to tutorials on how to draw cartoons. Like other user-generated content websites, such as YouTube or Flickr, Scratch projects are displayed on a webpage (See Figure 2) where people can interact with them, read metadata and give feedback. Visitors can use their mouse and/or keyboard to control a video game or other type of interactive projects or simply observe an animation play out in a web browser. Metadata displayed next to projects includes a text-based description of the project, the creator's name, the number of views, downloads, "love its," remixes, and galleries (i.e., sets of projects) that the project belongs to. Users can interact with projects by giving feedback in the form of tags, comments, or clicks on the "love it" button, and can flag a project as "inappropriate" for review by site administrators. Participants' self-reported ages range primarily from 8 to 17 years-old with 12 being the median. Thirty-six percent of users self-report as female. A large minority of users are from the United States (41%) while other countries prominently represented include the United Kingdom, Thailand, Australia, Canada, Brazil, South Korea, Taiwan, Colombia and Mexico. About 28% of all users -more than 170,000have uploaded at least one project. --- Remixing in Scratch Scratch users can download any project shared on the website, open it up in the Scratch authoring environment, learn how it was made, and "remix" it. In Scratch, the term "remixing" refers to the creation of any new version of a Scratch program by adding, removing or changing the programming blocks, images or sounds. In this section we use qualitative data from the Scratch website to provide social context for remixing and to suggest that credit plays an important role in how users conceive of appropriate remixing practice. Remixing in Scratch is not only technically possible, it is something that the administrators of the website encourage and try to foster as a way for people to learn from others and collaborate. On every project page, the Scratch website displays a hyperlink with the text "Some rights reserved" that points to a child-friendly interpretation of the Creative Commons Attribution-Share Alike license under which all Scratch projects are licensed. 1 Even the name Scratch is a reference to hip hop DJs' practice of mixing records. A large portion of all projects shared on the Scratch website (28%) are remixes of other projects. That said, remixing is not universally unproblematic in Scratch. Previous quantitative analysis of the the Scratch community showed that Scratch participants react both positively and negatively to the remixing of their projects and found that of those users who viewed a remix of their project, about one-fifth left positive comments while the same proportion of users accused the remixer of plagiarism [8]. This ambivalent reaction to remixing is echoed, and given additional texture, in the comments and complaints left by users on the Scratch website and sent to Scratch administrators. For example, even before the Scratch website was publicly announced, a number of early adopters became upset when they found remixes of their projects on the website. Indeed, one of the very first complaints about Scratch occurred on the discussion forums where a 13 year-old boy asked: Is it allowed if someone uses your game, changes the theme, then calls it 'their creation'? Because I created a game called "Paddling 1.5" and a few days later, a user called "julie" redid the background, and called it 'her creation' and I am really annoyed with her for taking credit for MY project!!2 A similar complaint was sent to the website administrators by a 14-year old boy: I think there should be a way to report plagiarized projects I've been seeing a lot of people's projects taken and renamed. This member, named kings651, has 44 projects, and most of them are made by other people. He even has one that I saw my friend make so I know he actually made it. In other cases, the disagreements over remixing were more public and involved communication via projects and com-ments. For example, user koolkid15 wrote the following message in a comment which was left is response to a remix that shows a cat frowning: Hi i'm koolkid15 the original creator of luigi disco jay-man41 copied me!! and didn't even aknowladge me he didn't change anything!! I wrote or drew!! and jayman...if your reading this think about other people!!!! Despite the fact that Scratch was conceived, designed, and launched as a platform for remixing, these users expressed their displeasure at remixing. That said, none of these users complained directly about the reuse of their project in general, but in terms of unfair "taking credit", plagiarism, and a lack of acknowledgment. Remixing was seen as problematic for koolkid15, for example, because of the nontransformative nature of reuse, the lack of acknowledgment of antecedent contributors, and the confusion about credit that would result. Of course, other, more positive, scenarios around remixing also played out in Scratch. For example, jellogaliboo created a remix of Catham's project and wrote the following in the project notes: "i kinda copied Catham's "jetpackcat" game. i used the kitty, the blocks (i added and changed some), and the fuel thingy." Catham later posted his approval of the remix saying, "I like what you changed about my project!" Like this example, many of these positive experiences involved explicit credit-giving by a remixer to the creator of the antecedent project. --- DESIGN INTERVENTION: AUTOMATING ATTRIBUTION Several user complaints about remixing and plagiarism also included suggestions for how Scratch's designers might address them. For example, in response to the forum thread mentioned in the previous section, a 16 year-old proposed two potential design-based solutions: Make it so you can only download a view of how your game/story/animation works. Or make it so downloadable Scratch files have read only protection. Maybe downloaded Scratch files, after being uploaded, are marked with the creators name at the bottom, and then any DIFFERENT people who edit it after are put on the list. Influenced by these comments, Scratch administrators came to believe that negative responses towards remixing were often due to the fact that Scratch users did not acknowledge the sources of their remixes. As a result, these administrators implemented an architectural design change to the Scratch community along the lines suggested by the user in the second half of the quotation above. The design change in question involved the introduction of a new technological facility that automatically identified and labeled remixes and inserted hyperlink pointers under each remix to the remix's antecedent and the antecedent's author (see Figure 3). Two days after the introduction of this feature, functionality was added to link to a comprehensive list of derivative works from the pages of antecedent projects (see Figure 4). The new feature was announced in the discussion forums by an administrator of the website and user responses were positive. User terminator99 suggested that the change was, "Awesome." Another user, marsUp, posted a comment saying, "That's a very useful feature! I like that we can do pingpong like modding in Scratch." Users who did not visit the discussion forums also responded well to the the new feature. For example, user greekPlus posted a comment on a remix he created saying, "i remixed it for you but i do not know how to ad credit to you for thinking of it in the first place." A few minutes later he realized that the remix automatically displayed the attribution and posted the a comment saying, "never mind it did it for me. cool!" --- STUDY 1: HUMAN AND MACHINE ATTRIBUTION Although initial user feedback to the automatic attribution feature was positive, users continued to complain about remixing. In Study 1a, we present a quantitative analysis to more fully evaluate the effect of the technological design change described in the previous section. In Study 1b, we offer a parallel analysis of the relationship between manual crediting-giving by users and users' reactions to being remixed. Both studies build on a dataset used in previous work by Hill, Monroy-Hernández, and Olson [8]. This dataset includes remix-pairs determined by an algorithm using detailed project metadata tracked by the Scratch online community. The dataset is limited in that it does not include projects whose concepts were copied by a user who had seen another's work but who did not actually copy code, graphics or sound. Similarly, the dataset contains no measure of the "originality" of projects or an indicator based on ideas that were taken from a source outside Scratch (e.g., a user may have created a Pacman clone which would not be considered a remix in our analysis). The data presented here includes each coded reactions of the author of antecedent projects (i.e., originators) on remixes of their projects shared by other users in the site during a twelve week period after Scratch's launch from May 15 through October 28, 2007. Although 2,543 remixes were shared in this period, we limit our analysis to the 932 projects (37% of the total) that had been viewed at the time of data collection by the project originator -a necessary prerequisite to any response. Of these 932 remixes that were viewed by a project originator, 388 originators (42%) left comments on the remixes in question. The remaining were coded as "silence." Comments left by originators were coded by two coders, blind to the hypotheses of the study and who were found to be reliable [8], as being positive, neutral, or negative. They were also coded as containing accusations of plagiarism (projects in which the the originator directly accused the remixer of copying, e.g., "Hello mr plagiarist", "Copycat!") or hinting plagiarism (projects in which the originator implied that the remixer had copied but did not state this explicitly, e.g., "I mostly pretty much made this whole entire game"). Unless it also contained an explicitly negative reaction, an accusation of plagiarism was not coded as "negative." However, because plagiarism tends to be viewed as negative within Scratch (as suggested by the quotations in the previous section) and more broadly in society [20], we re-coded accusations of plagiarism (both direct and hinting) as "negative" except, as was the case in several comments coded as "hinting plagiarism," when these accusations were in comments that were also coded as positive. Previous published work using this dataset, and subsequent robustness checks, show that our results are substantively unchanged if we exclude these explicit charges of plagiarism from the "negative" category or exclude only the weaker "hinting plagiarism" accusations. --- Study 1a: Automatic Attribution To test the effectiveness of automatic attribution, we consider the effect of the design intervention described in the previous section. The design change took place six weeks after the public launch of the Scratch community and at the precise midpoint in our data collection window. The intervention affected all projects hosted on the Scratch online community including projects shared before the automatic attribution functionality was activated. As a result, we classify originators' reactions as occurring outside a technological regime of automatic attribution when a project was both uploaded and viewed by a project's originator before automatic attribution functionality was activated. A comparison of the distribution of coded comments between positive, neutral, negative, and silent in the periods before and after the intervention suggests that the introduction of automatic attribution had little effect on the distribution of reaction types (See Figure 5). Although the period after the intervention saw a larger proportion of users remaining silent and a smaller proportion of both positive and negative comments, <unk> 2 tests suggest that there is no statistically significant difference in originator reactions between remixes viewed before or after the introduction of automatic attribution (<unk> 2 = 3.94; df = 3; p = 0.27). As a result, we cannot conclude that the there is any relationship between the presence, or absence, of an automatic attribution system in Scratch and the distribution of different types of reactions. These results suggest that automatic attribution systems may have limited effectiveness in communities like Scratch. Of course, our analysis is not without important limitations. For example, the existence of an automatic attribution regime may also affect the behavior of users preparing remixes. A remixer might avoid making perfect copies of projects if they know that their copies will be attributed and are more likely to be discovered. --- Study 1b: Manual Crediting While the introduction of an automatic attribution feature to Scratch appears to have had a limited effect on originators responses to remixes of their projects, the presence or absence of credit was a recurring theme in discussions on Scratch online forums -as shown in the quotes in the previous section -and in many of the coded reactions from the periods both before and after the introduction of automatic attribution. Indeed, in project descriptions or notes from the periods both before and after the change, remixers frequently "manually" gave credit to the originators of their work. Even after remixes were automatically attributed to originators, remixers who did not also give credit manually -essentially producing information redundant to what was already being displayed by the system -were criticized. For example, after the introduction of automatic attribution functionality, a user left the following comment on a remix of their project: Bryan, you need to give me Pumaboy credit for this wonderful game that I mostly pretty much kinda totally made this whole entire game... and that you need to give me some credit for it For this user, automatic attribution by the system did not represent a sufficient or valid form of credit-giving. In the following study, we test for this effect of "manual" creditgiving by remixers on coded response types using a method that parallels the analysis in Study 1a and that uses the same dataset. Manual crediting can happen in multiple ways. Exploratory coding of 133 randomly selected projects showed that 35 (36%) of each remix pair gave credit. Of these 35 projects, 34 gave credit in the project description field while 1 project only gave credit in a "credits" screen inside the game. As a result, the authors of this study split the sample of projects used in the Study 1a and coded each of of the user-created descriptions for the presence or absence of explicit or manual credit-giving. To first establish that we are examining distinct behaviors, we attempted to establish that automatic and manual attribution do not act as substitutes for each other. As suggested by our qualitative findings and our results in Study 1a, we found little difference in the rate of explicit credit giving between projects created in the presence or absence of automatic attribution. Overall, 276 (about 30%) of the 932 projects in our sample offered explicit credit in the description field of the project. Manual crediting-giving was a widespread practice both before automatic attribution, when 31% of projects in our sample offered explicit credit, and after, when 27% did so. The difference between these two periods was not statistically significant (<unk> 2 = 1.41; df = 1; p = 0.24). Previous work studying Jumpcut, a video remixing website, supports the idea that automatic and manual credit giving are not interchangable phenomena. One Jumpcut user with permission to creative derivative works commented that they, "still feel a moral obligation to people as creators who have a moral right to be attributed (and notified) despite the physical design which accomplishes this automatically" [5]. We measured effectiveness of manual credit giving using a parallel analysis to Study 1a. As in Study 1a, we compared the distribution of originator reactions in the presence, and absence, of manual credit-giving by remixers. We found that negative reactions are less common in the presence of manual credit but that this difference is very small (from 16% without manual credit to 14% with it). However, we see that the proportion of users who react positively almost doubles in the presence of credit-giving (from 16% with no crediting to 31% in its presence). A graph of these results are shown in Figure 6. Tests show that we can confidently reject the null hypothesis that these differences in the distribution of reactions are due to random variation (<unk> 2 = 27.60; df = 3; p <unk> 0.001). Also important to note is a difference in the number of users who are silent after viewing a project (62% in the absence of manual credit versus 49% in its presence). This larger proportion of commenting in general may have an important substantive effect on the discourse and behavior on the site because silent originators may, for obvious reasons, have a more limited effect on attitudes toward remixing and user experience than vocal users do. As a robustness check, we considered the reaction of only originators who left comments (n = 388) and found that even with a smaller sample, our result were stronger. In the restricted sample, 41% reacted negatively when they were not given credit. However, only 27% did so when they were credited. Similarly, 42% of users who left comments on projects that did not give credit manually left positive messages. Nearly two thirds of comments (61%) were positive when credit was given. These differences, in the reduced sample that includes only explicit reactions, were also statistically significantly different (<unk> 2 = 14.09; df = 2; p <unk> 0.001). We include the large number of silent participants because we believe that nonresponse is an important type of reaction with real effects on the community. Understanding the reasons behind nonresponse and the effect of silence in response to different types of credit giving remains an opportunity for further research. Although not presented here due to limited space, we followed the general model of previous work using this dataset [8] and tested logistic regression models on dichotomous variables indicating the presence of negative and positive reactions and found that basic relationships described above were robust to the introduction of a control for the the intervention, to an interaction between these two variables, and to controls for the gender and age of originators and to the antecedent project's complexity. Both before or after the intervention, manual crediting resulted in more positive comments by the originators of remixed projects. Of course, the results presented here are uncontrolled, bivariate, relationships and we caution that these results, while provocative, should still be viewed as largely tentative. As we show in the subsequent qualitative analysis, attribution and credit-giving are complex social processes. We do not claim that the preceding analyses capture it fully. --- STUDY 2: INTERVIEWS WITH PARTICIPANTS In order to explore the reasoning behind young people's remixing behavior and attitudes toward attribution as we observed it in Study 1, we engaged in a second qualitative study and directly asked kids what role attribution and credit plays in their moral evaluations of remixing. --- Methodology We conducted twelve one-hour semi-structured interviews with kids aged 8 to 17 years old. All of the interviewees had experience using computers and had access to the Internet at home. All the interviewees live in the United States except for one who lives in New Zealand. The participants were recruited via the Scratch website and during meet-ups with educators, teachers and young Scratch users. Eight of the interviews were conducted in person, in the Boston area, and the rest over the phone or voice over IP. The interviews were audio-recorded and transcribed before fully analyzing them. Nine of the interviewees were members of the Scratch community. The remaining three did not use Scratch but were included as a way to check if people who do not use Scratch have similar views about remixing, attribution, and credit. We found no substantive difference between the Scratch users and non-users in their answers to questions related to the hypothetical automatic and manual mechanism for attribution. Before each interview, subjects completed a survey that elicited demographic information and posed questions about their familiarity with other technologies and which was primarily designed to get a sense of the interviewees' social and technical background. Interviews were structured around a protocol that included a set of nine fictional remixing cases intended to elicit conversations about remixing. 3 The cases were inspired by Sinnreich et al.'s theoretical work and from three years of experience moderating the Scratch community. They were designed to present cases where remixing could be controversial but where there is no clear "correct" answer. The goal of the cases was to offer a concrete, and common, set of dilemmas to stimulate broad conversations about attitudes toward remixing. The cases were presented in the form of printed screenshots of different project pages from the Scratch website (anonymized to avoid referring to real cases that users might 3 Our interview protocol including example cases is available at http://www.media.mit.edu/<unk>andresmh/chi2011/interview.html. have seen). The print outs were shown to the interviewees (or discussed over the phone) while explaining each case. All the cases included a remix and its corresponding antecedent project. The cases varied in the presence of automatic attribution, manual credit, and the degree of similarity between the remix and its antecedent. For example, the first three cases were: 1. A remix and its antecedent are identical. The project notes only describe how to play the video game. The remix shows the automatic attribution but no manual credit on the notes. 2. A remix and its antecedent are different (as seen visually and in project metadata) but one can clearly see the influence of its antecedent project. The project notes of the remix show manual credit but no automatic attribution. The interviewee was told to imagine the site had a glitch that prevented it from connecting it to its antecedent. 3. The same set of remix and antecedent projects as in (2) but this time automatic attribution is displayed but manual credit is not. Each of the interview logs was coded using inductive codes and grounded theory [3]. The coded responses were analyzed based on categories related to how interviewees answered specific questions about the distinction between automatic attribution and manual credit. --- Results Confirming the results of Study 1, for users of Scratch, automatic attribution was generally seen as insincere and insufficient. Throughout the interviews, we found that for most of the kids, getting explicit credit from another person was preferred over attribution given automatically by the system. When asked why, kids often responded that knowing that another person had cared enough to give credit was valued more than what the computer system would do on its own. The fact that it takes some work, albeit minimal, to write an acknowledgment statement, sends a signal of empathy, authenticity and good intentions [6]. Amy articulated this when explaining why she preferred getting credit from another person: I would like it even more if the person did it [gave credit] on their own accord, because it would mean that [...] they weren't trying to copy it, pirate it. Similarly, Jon explained, "No [the "Based on" is not enough], because he [the remixer] didn't put that, it always says that." For Jon, automatic attribution is not authentic because it is always there and, as a result, it is clear that is not coming from the person doing the remix. Most of the interviewees seemed to have a clear notion of what they think a moral remix should be. For some, it is all about making something different. Jake for example, defines a "good" remix as, "if it has a bunch of differences then it's a good remix. If it has like two, then it's bad." In addition to the differences between the remix and its antecedent project, for some, manual credit is part of what makes it moral. Charles said, "[remixing] is taking somebody else's project and then changing a lot of it and sharing it and giving credit." Continuing, Charles explained: If Green had actually said in the project notes, "This is a remix of Red's project, full credit goes to him," then I would consider it a remix. But this [pointing at a remix without manual credit] is definitely a copy. Likewise, Ryan mentions that a fictional remix was, "perfectly fine because they gave credit in the project notes." Interviewees suggested that manual credit also allows users to be more expressive. For example, Susie explained that expressiveness is the reason that she prefers manual credit through the project notes saying, "I think the manual one is better because you can say 'thank you' and things like that. The automatic one just says 'it's based on."' Susie also notes that for her, the project notes are a space where a creator can express her wishes in regards to her intellectual property, independent, and even in contradiction to, the license of the projects: If I do a project that has music that I really like, I often download the project, take the music. Unless it says in the project notes, "Do not take the music." For Susie and other users of Scratch, the project notes are a space for more than just instructions on how to interact with one's project; they are an expressive space where one can communicate with an audience without having to encumber the creative piece of work with it. Others point at the fact that people do not pay as much attention to automatic attribution statement as much they do to the manual credit left in project descriptions. Jake, for example, explains that, while he agrees there is some usefulness to having both, project notes still are more important, "because, you know, sometimes people just like skim through a project and you don't see it 'til the end." Jake continued to say that creators that do not have both should get a "warning." Even though interviewees value manual credit, they still see the usefulness of the automatic mechanism as some sort of community-building prosthetic device -an explanation for the positive reactions to the feature's initial introduction. For example, Nicole argues that while manual credit on the notes has more value for her, the automatic attribution is useful as a backup and because it provides a link: Well, I think that they should probably write in the notes that -then it should also say "Based on blank's project," just in case they forget, and also because it gives a link to the original project and it gives a link to the user so you don't have to search for it. A similar explanation was articulated on a comment exchange on one the website's galleries. A teenage girl that actively participates in Scratch explained the pragmatic value of automatic attribution saying, "the 'based on' thingy, it gives a link, and we all luv links, less typing," before reiterating that manual credit is more valuable: at the beginning i thought that you don't have to give credit when the "based on" thingy is in there, but i realized a lot of people don't look at that, and i noticed people confused the remix with the original. Creating a Scratch project is a complicated task. A project's sources can be diverse and the creator can easily forget to acknowledge some, as Paul explains, when asked to choose between a system of manual credit or automatic attribution: The thing is, it would be a lot better if they had both. Because, sometimes people probably just forget to do that. And then people would not know. There are also situations where interviewees recognize what Posner calls the "awkwardness of acknowledgment," that is, situations where credit is not really needed and it can be an unnecessary burden or go against the aesthetics of the work [20]. For example, Paul mentioned that sometimes, there are some projects in Scratch that are remixed so much -like the sample projects that come with Scratch or some "remix chains"4 -where credit is not necessary

In this paper, we explore the role that attribution plays in shaping user reactions to content reuse, or remixing, in a large user-generated content community. We present two studies using data from the Scratch online communitya social media platform where hundreds of thousands of young people share and remix animations and video games. First, we present a quantitative analysis that examines the effects of a technological design intervention introducing automated attribution of remixes on users' reactions to being remixed. We compare this analysis to a parallel examination of "manual" credit-giving. Second, we present a qualitative analysis of twelve in-depth, semi-structured, interviews with Scratch participants on the subject of remixing and attribution. Results from both studies suggest that automatic attribution done by technological systems (i.e., the listing of names of contributors) plays a role that is distinct from, and less valuable than, credit which may superficially involve identical information but takes on new meaning when it is given by a human remixer. We discuss the implications of these findings for the designers of online communities and social media platforms.

erating that manual credit is more valuable: at the beginning i thought that you don't have to give credit when the "based on" thingy is in there, but i realized a lot of people don't look at that, and i noticed people confused the remix with the original. Creating a Scratch project is a complicated task. A project's sources can be diverse and the creator can easily forget to acknowledge some, as Paul explains, when asked to choose between a system of manual credit or automatic attribution: The thing is, it would be a lot better if they had both. Because, sometimes people probably just forget to do that. And then people would not know. There are also situations where interviewees recognize what Posner calls the "awkwardness of acknowledgment," that is, situations where credit is not really needed and it can be an unnecessary burden or go against the aesthetics of the work [20]. For example, Paul mentioned that sometimes, there are some projects in Scratch that are remixed so much -like the sample projects that come with Scratch or some "remix chains"4 -where credit is not necessary: There's this one called "perfect platformer base" which a lot of people remix. So I don't think that needs any credit. It's not actually a real game. It's all the levels and stuff are just demonstrations. Since manual crediting has a higher emotional value, some kids mentioned that conflicts over remixing could be addressed by the administrators of the site by editing the project of the remix in question, as a way to enforce credit without transforming it into attribution. Doing so would make it appear that a remixer had credited an antecedent when they had not. Susie offers a suggestion along these lines when asked about how the administrators of the website should deal with a case of a complaint over a remix that is a parody of someone else's project. Susie suggested that, "I might remove the project but I might not, you know, maybe I would edit the notes to to give credit Although not designed to be a random sample, these interviews support the proposition that both Scratch participants and other young people share a set of norms about characteristics that determine what a "good" or moral remix is. Among these norms, acknowledging one's sources seems to play a central role. However, participants also seem to share the opinion that this norm is not satisfied through an automated process. They clearly understand the pragmatic value of automating acknowledgment-giving, but they do not see it as a substitute for adherence to the social norm of creditgiving. They also see it as void of emotion and expressiveness. For Scratch users, normative constraints are separate from architectural constraints and one cannot replace the other. These findings support and enrich the results from our first study and help us understand better how Scratch participants, and perhaps kids in general, experience authorship norms and automation in online spaces. --- CONCLUSIONS Our results from Study 1a called into the question the effectiveness of automatic attribution functionality in encouraging more positive user reactions in Scratch. We build on these results in Study 1b to suggest that manual crediting may do the work that Scratch's designers had hoped automatic attribution would. Results from the analysis of user interviews presented in Study 2 help to answer the question of "why?" and suggest that users find manual credit to be more authentic and more meaningful to users because it takes more time and effort. Usually, UI improvements are designed to help reduce the time and effort involved in using a system. But in trying to help users by attributing automatically, Scratch's designers misunderstood the way that attribution as a social mechanism worked for Scratch's users. Our fundamental insight is that while both attribution and credit may be important, they are distinct concepts and that credit is, socially, worth more. A system can attribute the work of a user but credit, which is seen as much more important by users and which has a greater effect on user behavior, cannot be done automatically. Computers can attribute. Crediting, however, takes a human. As we suggested in our introduction, this fundamental result leads to two distinct contributions. First, and more specifically, our analysis offers an improved understanding of the way that attribution and credit works in user-generated content communities over what has been available in previous work. Our two studies suggest that scholars are correct to argue that credit plays an important role in social media communities and offer empirical confirmation for the important role that authenticity plays in how users conceptualize credit. In our in-depth interviews, we explain some of the reasons why this may be the case. Second, through our evaluation of an unsuccessful technological design, our work offers a broader, if more preliminary, contribution in suggesting an important limit of designers' ability to support community norms in social media systems. As the literature on design and social media grows, the importance of good support for communities with healthy norms promoting positive interactions is likely to increase. In attempting to design for these norms, we suspect that researchers will increasingly encounter similar challenges. We argue that designers should approach interventions iteratively. This design approach can be understood through the theoretical lens of the social construction of technology [19]: designers can't control technological outcomes which must be built through a close relationship between designers and users. Designers must move away from seeing their profession as providing solutions. They must channel users, work closely with them, and iterate together, to negotiate and achieve a set of shared goals. The prevalence of user-generated content sites stresses the importance of how online social spaces should deal with issues of attribution and our results are likely to be immediately relevant to designers. For example, the Semantic Clipboard is a tool built as a system of automatic attribution for content reuse [22]. Developed by researchers who found a high degree of Creative Commons license violations around the re-use of Flickr images, the tool is a Firefox plugin that provides, "license awareness of Web media," and enables people to automatically, "copy [media] along with the appropriate license metadata." Our results suggest one way that this approach may fall short. However, automatic attribution is not the only way that technologists can design to acknowledge others' contributions. Indeed, our results suggest that there may be gains from design changes which encourage credit-giving without simply automating attribution. For example, Scratch's designers might present users with a metadata field that prompts users to credit others and suggests antecedent authors whose work the system has determined may have played a role. This affordance might remind users to credit others, and might increase the amount of crediting, while maintaining a human role in the process and the extra effort that, our research has suggested, imbues manual credit giving with its value. We suggest that in other social media communities, similar affordances that help prompt or remind users to do things that a system might do automatically represent a class of increasingly important design patterns and a template for successful design interventions in support of community norms.

In this paper, we explore the role that attribution plays in shaping user reactions to content reuse, or remixing, in a large user-generated content community. We present two studies using data from the Scratch online communitya social media platform where hundreds of thousands of young people share and remix animations and video games. First, we present a quantitative analysis that examines the effects of a technological design intervention introducing automated attribution of remixes on users' reactions to being remixed. We compare this analysis to a parallel examination of "manual" credit-giving. Second, we present a qualitative analysis of twelve in-depth, semi-structured, interviews with Scratch participants on the subject of remixing and attribution. Results from both studies suggest that automatic attribution done by technological systems (i.e., the listing of names of contributors) plays a role that is distinct from, and less valuable than, credit which may superficially involve identical information but takes on new meaning when it is given by a human remixer. We discuss the implications of these findings for the designers of online communities and social media platforms.

INTRODUCTION The evidence is clear that public health emergencies (PHEs) can dramatically impact the substantial gains made in primary health care initiatives (1), with estimates suggesting that each year one out of five World Health Organization (WHO) member states experiences a PHE (2). According to the Model State Emergency Health Powers Act, a PHE can be defined as: "an occurrence or imminent threat of an illness or health condition, caused by bioterrorism, epidemic or pandemic disease, or novel and highly fatal infectious agent or biological toxins, that poses a substantial risk of a significant number of human fatalities or incidents of permanent or long-term disability" (3). The effects of PHEs are further exacerbated within continents that are comprised of fragile states-such as Africa-with inadequate health care systems (4). The African continent exhibits a unique set of characteristics including a large ethnic diversity (5), distinct physiographic patterns (6), vast mineral wealth (7) and a burgeoning youth population (almost 60% of the continent are aged below 25) (8). While intriguing, several countries across the continent continue to experience complex emergencies and health crises ranging from civil conflict and infectious diseases (e.g., HIV/AIDS and malaria) to severe drought and malnutrition. These humanitarian crises place significant strain on personal lives leading to socioeconomic instability, forced migration, and long-term refugee problems (9,10), that in turn have an adverse effect on attainment of the United Nations (UN) Sustainable Development Goals. Given the inadequate funding to PHE and disaster preparedness in many African countries (11), there remains some degree of uncertainty regarding the ability of the continent's countries to adequately respond to these concerns (12)-although the last decade has seen a significant increase in funding for research capacity in Africa (13). Due to the unpredictable nature of PHEs, it is perhaps unsurprising that PHE preparedness (PHEP) is an inherently complex process that involves a range of prevention, mitigation, and recovery activities that extend beyond just enabling a response to emergencies (14). Notably, PHEP is conceptualized as comprising 3 broad elements including: (1) pre-planned and coordinated rapid-response capability; (2) strengthening expertise and building a fully staffed workforce; and (3) ensuring accountability and quality improvement (14). Here we draw attention to African governments and institutions in relation to the handling of 3 distinct PHEs-namely, Ebola, COVID-19, and locust swarms-and how scientific and sustainable technology implementation, along with other effective and innovative responses, have played a role in their managing of these humanitarian crises. --- EBOLA OUTBREAK IN WEST-AFRICA Of the 34 documented Ebola outbreaks that have occurred since the first description of the virus in the Democratic Republic of Congo (DRC) in 1976 (15), the 2013-2016 West Africa epidemic was the largest and most widespread in historyculminating in more than 28,000 cases and over 11,000 deaths (16). Epidemiologists identified the index case of the outbreak in Meliandou, Guinea, and fruit bats are believed to have served as a reservoir of the virus and to be involved in the zoonotic spillover effect that led to a cascade of contagion that spurred the high number of Ebola cases and case fatality rate (CFR) in West Africa (15,17). An estimated US$ 2.2 billion in gross domestic product (GDP) was lost in 2015 by the 3 most affected countries (Liberia, Sierra Leona, and Guinea) (18). The epidemic also resulted in lower investment and a substantial loss in private sector growth, declining agricultural production that led to concerns about food security, and a decrease in cross-border trade as restrictions on movement of goods and services increased (19). Another consequence of the virus relates to rural-urban gradients of transmission and population-level beliefs and practices (i.e., shifts in where care was sought) (20)(21)(22), with a pervasive stigmatization of survivors playing a significant role (23). The stigma attached to Ebola has been reported to have led to social inequalities and mental health problems with a large portion of individuals afflicted by the disease suffering hostility and economic hardship (24,25). The Ebola outbreak in West Africa also highlighted various barriers to coordinated rapid-response capacity and the need for more robust global health security, particularly in settings with limited public health capacity (16,26). Indeed, the 3 countries most affected by the outbreak exhibited similar characteristics including: inadequate financial resources and health care systems (as reported by low numbers of nurses and doctors) in addition to a scarcity of medicines and personal protective equipment (PPE)-each of which represent a unique threat to containing the spread of infectious disease, and in turn provide a hurdle toward implementation of the International Health Regulations (IHR) (27). Notably, porous borders meant that the Ebola outbreak was not just restricted to Guinea, Liberia and Sierra Leone as cases were also reported in Nigeria, Senegal, and Mali (15,28). The slow recognition and delayed response to control the Ebola outbreak by West African governments, exposed defective containment strategies and poor crisis management in the countries' worst affected by the virus. Inadequate contact tracing and detection of suspected cases coupled with poor surveillance and gaps in the community's knowledge about the Ebola virus contributed to a rampant spread of the disease (29). Unfortunately, during the outbreak, governments in Liberia, Guinea and Sierra Leone failed to communicate effectively with citizens. In Liberia, this resulted in frustrations and riots in the capital city, Monrovia. --- Improving Pre-planned and Coordinated Rapid-Response Capacity As the effects of the Ebola virus continued to unfold in West Africa a key strategy shown to moderate the crippling effects of the epidemic involved an integrated and calibrated response strategy that included: (1) bolstering standardized supportive care of survivors via treatment for the symptoms and complications of Ebola (e.g., mental health and psychosocial support), (2) leveraging and deploying aid through international organizations such as the WHO and Médecins Sans Frontières; (3) funding for emergency Ebola treatment; (4) rapid and accurate Ebola diagnostics testing through platforms such as real-time polymerase chain reaction; (5) scaling-up of national disease surveillance activities (e.g., digital health/apps via mobile devices); (6) a licensed Ebola vaccine (Merck's VSV-ZEBOV vaccine) and (7) focusing on social science and community engagement through aspects of risk perception, tackling vaccine hesitancy and education as a means to minimize confusion and to empower individuals to adopt preventative behavior (30). This response strategy is estimated to have taken over a year to implement (31,32), with follow-up visits provided to each survivor every month over a period of 6 months and then every 3 months for a year (30). Six years on from the beginning of the Western Africa Ebola epidemic, the DRC has been grappling with its 12th Ebola outbreak. Active conflict, a severe measles outbreak and insecurity make this epidemic one of the most complex ever encountered (33). However, the Ministry of Health of the DRC has mounted an impressive response strategy to the outbreak with international support from the United States Centers for Disease Control and Prevention (CDC), WHO and Gavi, the vaccine alliance. Beyond placing an emphasis on developing local health care systems in the most affected areas (i.e., Kivu and Ituri), a strong aspect of the DRC Ebola response has been applying lessons learned from the outbreak in West Africa (30). These lessons include community engagement, better support of survivors, use of mobile phone data to inform the dynamics of Ebola transmission (via travel patterns and contact tracing) (34), and licensed approval for two vaccines (Merck's single dose VSV-ZEBOV vaccine and Jansen's two-dose vaccine regimen of Zabdeno and Mvabea), with recent estimates suggesting that more than 300,000 people have been immunized against Ebola through vaccination in the DRC (35). Despite the significance of the current vaccines against Ebola, there still remains the need to develop an effective strategy for optimal impact of vaccination (36). In this respect, Coltart et al. emphasize that prophylactic vaccination of health care workers (HCWs) could have a substantial epidemic-reducing effect on the spread of Ebola (37). Other evidence from mathematical and statistical models suggests that engaging HCWs to deliver vaccinations represents both a feasible and effective strategy that may be implemented in a future Ebola outbreak (38,39). --- Building Expertise and Fully Staffed Workforce Several notable initiatives have been developed to enhance capacity (viz., expertise) and build leadership within the local workforce at national and local levels in various countries, as a means of strengthening sustainable PHEP. For instance, the CDC's Surveillance Training for Ebola Preparedness (STEP) initiative was shown to be a successful mentorship and competence-based initiative that collaborated with various local training institutes and organizations to rapidly build the surveillance capacity of district surveillance officers in Mali, Guinea-Bissau, Senegal and Ivory Coast during the Ebola outbreak in West Africa (40). Along with the STEP initiative, implementation of laboratory capacity building programs to strengthen bio-risk and quality management systems, diagnostics and facility engineering, and bio-surveillance capacity was able to bolster emergency preparedness and response (41). Perhaps most notably, infection prevention and control capacity building programs for HCWs registered positive benefits on knowledge and practices of HCWs in the fight against Ebola outbreak in DRC (42) and Western Africa (43). Also, foreign medical worker deployment from the African Union Support to the Ebola Outbreak in West Africa and medical personnel from Cuba played a central role to fill the gap for skilled HCWs, as well as co-learning for skills development (44). --- Accountability and Quality Improvement In West Africa, community monitoring-which involves providing patients with information and enabling a public forum to monitor frontline workers-was beneficial in generating some form of social accountability and trust. An example of an effective community monitoring program was the Liberian government's door-to-door canvassing campaign during the Ebola epidemic (45). Through the Financial Tracking Service (FTS) of the United Nations Office for the Coordination of Humanitarian Affairs (UNOCHA), curated financial data (e.g., funding needs, commitments, pledges and projected funding) on Ebola virus outbreak were continuously updated and accessible in downloadable format on the UNOCHA website to facilitate accountability and transparency (46). --- COVID-19 OUTBREAK IN AFRICA The first case of coronavirus disease 2019 (COVID-19), the disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was reported on the African continent on February 14, 2020 in Egypt, with Sub-Saharan Africa (SSA) detecting its first case in Nigeria on February 27, 2020 (47). In most countries, the initial response to the COVID-19 pandemic was strong and proactive. Despite these measures, many public health experts predicted that the pandemic would severely overwhelm Africa's largely fragile and underfunded health systems. Of the 34 African countries surveyed in the WHO COVID-19 readiness status report, only 10 countries reported adequate capacity to respond to the epidemic, including with PPE for the population (48). The UN Economic Commission for Africa estimated that, in the worst-case scenario, 3.3 million Africans would die from the disease (49). Concerns over the combination of overstretched, underfunded health systems and the existing load of infectious and non-infectious diseases often led to scenarios being talked about in apocalyptic terms. More than a year into the pandemic, the continent has however thwarted most predictions regarding the spread of the virus. The health and social measures to contain the COVID-19 epidemic implemented by most countries are likely to have slowed the spread of the virus, and the number of confirmed cases and deaths in Africa remained lower than initially forecast. As of October 18, 2021, confirmed cases of COVID-19 from 55 African countries reached 8.4 million with a CFR of 2.6% (i.e., 215,784 deaths) (50). By early August 2021, it was estimated that only 3.5% of the global COVID-19 cases and 4.1% of the global COVID-19 related deaths were from Africa (50, 51)-a continent that accounts for 17% of the global population (52). Nevertheless, the magnitude of the challenge and the continent's underlying vulnerabilities should never be underestimated. The weak PHE management systems in most countries, have rendered it difficult to discern accurate transmission, hospitalization and mortality rates (53). For example, currently, the continent's testing rate is one of the lowest in the world. Therefore, the full scope of the pandemic remains uncertain. In addition, several countries are experiencing a second wave of the pandemic and some, such as Kenya, Egypt and Tunisia, have seen a third wave (54). This new wave of infections is thought to be associated with the emergence of variants that are more transmissible. Unfortunately, only a few countries have the capacity to carry out the specialized genomic sequencing required to detect coronavirus variants. Further, the health and economic shocks occasioned by the pandemic threaten to wipe out decades of economic progress and development gains in Africa. These risks put some countries on an unsustainable debt path (55). The pandemic has also laid bare structural shortcomings such as inadequate health, educational and technological infrastructure, limited social protection, gender inequality, large informal economies, lack of access to basic services, and constrained fiscal policy space (56). For example, the contraction in per capita GDP growth caused by the pandemic may have pushed an additional 26.2 million to 40 million (i.e., 2-3%) people into extreme poverty in SSA by the end of 2020 (57). Fighting a pandemic and its economic aftershocks requires enormous amounts of money. In higher income countries, governments have stepped forward with trillions in economic stimulus packages. But most developing countries do not have the money to cover the full costs of this pandemic. --- Improving Pre-planned and Coordinated Rapid-Response Capacity To forestall the COVID-19 health and economic crisis, most African countries developed response plans. Specifically, most African governments rapidly implemented public health and social measures to contain the pandemic, including closing borders, mandatory general lockdown, physical distancing measures, and establishing centers for quarantining of cases (58). Furthermore, response plans have also majorly centered around four main areas simultaneously, including: (1) saving lives; (2) protecting poor and vulnerable citizens and responding to the impact on their livelihood; (3) protecting and creating jobs through support to private sector and (4) building back better systems (59). For example, to save lives, several African countries focused on intensive surveillance and case-finding, leveraging the Integrated Disease Surveillance and Response framework (IDSR) (60). The Partnership to Accelerate COVID-19 Testing (PACT) Initiative, for instance, is an initiative launched by the African Union Commission and the Africa CDC to boost and coordinate procurement and supply chains for medical supplies and to support protracted testing for COVID-19 within the African setting (61). The continent has also been able to draw on previous experience in dealing with PHEs such as the Ebola crisis to make better decisions on public health and social measures. For example, several countries focused their response efforts toward community engagement, risk communication, and locally adapted innovations in tracing, treatment and isolation (56). With the vaccine roll-out underway in many African countries, ensuring an adequate supply of vaccines is a priority for the region. Countries have mainly accessed vaccines through the COVAX Facility, bilateral deals, and donations. Nonetheless, concerns regarding disparities in vaccine access and distribution remain widespread (62). Many developed countries have displayed a very high degree of "vaccine nationalism, " locking up most supplies and prioritizing the vaccination of their entire populations before releasing surpluses to protect even the most vulnerable populations in low-and middle-income countries (LAMICs) (63). As of October 16, 2021, Africa had administered 12.5 doses of COVID-19 vaccines per 100 people. The vaccination rate of the continent was far slower than the world average measured at 84.5 vaccines per 100 individuals on the same date (64). Further, while concerted global efforts are working to accelerate equitable access, vaccine hesitancy-driven in part by a trust deficit between communities and the actors leading vaccine rollout-risks prolonging the pandemic and its secondary waves of conflict and economic devastation. --- Building Expertise and Fully Staffed Workforce Across the African continent, HCWs are boosting their emergency response skills in tackling COVID-19, for example through virtual and in-person trainings organized by Ministries of Health and health organizations or research institutions (65). The PACT initiative for example, also focuses on the support for training and deployment of one million community HCWs to support contact tracing within the African setting (61). Government Ministries of Health have learned to harmonize research activities, through leveraging research laboratory capacity (both personnel and equipment) of academic research institutions and other in-country laboratories for community COVID-19 testing (66); as well as building foreign/international research partnership to improve testing or medical product development capacity (67). It follows that efforts to build research capacity to conduct good quality collaborative international COVID-19 vaccine trials in Africa will allow for better protection against this devastating infectious disease (68,69). --- Accountability and Quality Improvement With regard to economic response, Africa's fiscal realities limit what most countries can do to alleviate pressures on citizens (56). Several countries have undertaken measures to address the economic fallout of the pandemic. For example, some countries announced remedial fiscal and monetary measures, as well as food distribution and financial support to the most vulnerable groups. However, less has been done across countries to cushion against lost income and export earnings, dwindling remittances, and decreased government revenue. In addition, relatively few countries have articulated initiatives to mitigate the socio-economic impacts of COVID-19 in the long-term. Therefore, the road to recovery will be long and vary significantly across countries (48). Most African countries continue to rely on foreign aid in response to the impact of the pandemic. Since the start of the pandemic in March 2020, the World Bank has made available nearly US$ 24.7 billion to respond to the COVID-19 crisis through a combination of new operations in health, social protection, economic stimulus and other sectors, as well as redeployment of existing resources (70). Several African countries have also received foreign assistance from bilateral partners to help prevent, detect, and respond to the COVID-19 pandemic and strengthen their public health preparedness. For instance, France mobilized e1.2 billion to fight the spread of COVID-19 in the most vulnerable countries, most of which are in Africa (71). As with Ebola, the FTS of the UNOCHA has curated financial data on COVID-19 emergency funds as part of the Global Humanitarian Response Plan that are continuously updated and accessible through the UNOCHA website (72). Initiatives, for example by the African Union, have enabled increased dialogue and opportunity for learning and sharing among government officials, audit institutions, procurement oversight bodies, and civil society organizations on the African continent in relation to innovative accountability mechanisms and crisis budget support operations (73). --- LOCUST SWARMS IN EAST-AFRICA The impacts of the PHEs presented above are devastating enough without additional social and economic dislocation caused by non-disease outbreaks. In 2020, East Africa faced just such a situation when a surge of desert locusts invaded the Horn of Africa. In order to fast-track effective response toward the attack, on January 17, 2020, the UN's highest level of emergency (L3 protocols) was activated by the Director-General of the Food and Agriculture Organization (FAO) (74). Beyond widespread hatching, band and swarm formation in north East Ethiopia (57,450 hectares had been treated), immature swarms prevailed in Somalia (17,477 hectares) and to a lesser extent in North West Kenya (2100 hectares) (75). Desert locust infestation was also reported in 24 districts in Uganda around a similar timeframe (76). In Kenya, the 2020 desert locust invasion is considered the worst in 70 years (77). The combination of the ongoing COVID-19 pandemic and a desert locust outbreak has exerted an enormous economic toll and an even greater burden on the health systems in East Africa. The meager financial resources, which would have been fully vested into the COVID-19 programmes within East African governments, had to be rationed so that some resources are used to combat desert locusts, and this called for more borrowing. For instance, more than US$ 160 million was loaned to the East African countries of Kenya, Uganda, Djibouti and Ethiopia by the World Bank, to combat desert locusts ( 78) and yet additional loans for COVID-19 had also been secured by some of these governments [for instance US$ 1 billion for Kenya (79) and more than US$ 15 million for Uganda (80)] which further deepens their debt crisis. --- Improving Pre-planned and Coordinated Rapid-Response Capacity Through support from the World Bank, national response programs, i.e., Uganda Emergency Desert Locust Response Project (81), Kenya emergency locust response program (82) and Ethiopia Emergency Locust Response Project (83) were set up in early 2020. As part of the commitment plans, actions were stipulated through which desert locust control programs would be implemented in accordance with social and environmental standards, for example: environmental and social assessment of risks arising from the projects, occupational health and safety measures, and pollution prevention and management strategies (81)(82)(83). These programs also were set up with incountry coordination plans. For instance, in Kenya, a multiinstitutional technical team on desert locusts was established to coordinate policy and technical advisory on desert locust management which was tasked with activities like providing advisory to county administrations and any other stakeholders, planning the collection and collation of technical information and building capacity among stakeholders on integrated desert locust management (77). Surveys of terrain, state of habitat and locust population were performed so as to inform policy and decision making (83). Strengthening of existing systems to combat future outbreaks, for example the Locust Control Unit within the Plant Protection Service Division of Kenya's Ministry of Agriculture, was among the strategic aims of the funding from the World Bank (77). FAO encouraged country-level partners to record and transmit desert locust related surveillance data to ministerial organizations (like the Ministries of Agriculture) so that such essential information is included and utilized in FAO's Desert Locust Information Service (DLIS). In each country, a Locust Information Officer is responsible for collating, analyzing and transmitting this data to DLIS (84). In turn, the DLIS analyses the data and keeps countries informed of the current situation and expected developments by providing a forecast up to 6 weeks in advance (84). Data sharing for improved monitoring of desert locusts is also being boosted through mobile-phone based surveillance technology such as the eLocust3m and other platforms like the centralized Desert Hub platform (74). According to FAO, the primary method for controlling the 2020 desert locust swarm and hopper bands, is through organophosphate chemicals, delivered by vehicle mounted aerial sprayers and knapsack or hand-held sprayers (85). The main strategy involves targeting breeding grounds and controlling the hopper bands while still at the nymph stage-that is before they can fly (77). More recently, test drones equipped with mapping sensors and atomizers have been deployed to spray pesticides to tackle the desert locust swarms in East Africa. Governments and donor agencies (e.g., FAO and the World Bank) have ensured that there is disaster recovery relief provided, including inputs such as seed-fertilizer and pesticides to selected farmers faced with hardship and also provided fodder seed to affected communities to restore lost pastures, emergency food security mechanisms and direct cash transfers (76,83,86). In Uganda, for example, as part of World Bank's US$ 48 million loan, funds were set aside to boost existing savings and investment platforms/groups at village level through a Village Revolving Fund and seasonal income transfers to vulnerable (76). --- Building Expertise and Fully Staffed Workforce Capacity building of in-country human resource was conducted. For example, the FAO facilitated training of National Youth Service trainees as part of boosting the Government of Kenya ground surveillance for desert locusts (85). Governments also mobilized and trained communities to establish locust surveillance systems based at community, district and national levels so as to ensure the sustainability of mapping monitoring and surveillance (76). By the start of the desert locust disaster, FAO was already coordinating with over 100 NGOs in Ethiopia in using and building capacity for the use of eLocust3m (86). --- Accountability and Quality Improvement The FTS of the UNOCHA has played an integral role in enabling timely access to financial data on humanitarian funding flows on desert locust response across East Africa and the Horn of Africa (87). As part of quality improvement-in terms of the potential unintended negative consequences of pesticide use-initiatives were put in place to monitor and assess the environmental and human health risks attributable to their use (76,77). --- IMPLICATIONS FOR POLICY AND RESEARCH Evidence-informed policy and decision-making is crucial for ethical and sustainable response to PHEs. The generation and translation of evidence to inform policy and decisionmaking is often seen as a race against time-with the quality, depth and conciseness of available evidence directly affecting policy decision-making processes (88). Notable in this regard are rapid assessment tools, developed as a public health approach to speed up-and bring together-the processes of evidence-based decision-making during crisis management of PHEs (89,90). The present paper emphasizes the necessity of utilizing the rapid assessment approach in a more collaborative and engaging manner as a means of facilitating dialogue between decision-makers and other stakeholders (including scientists and communities) in relation to programme planning and interventions which, in turn, enable PHEP and responses "in-the-now" (91). Indeed, rapid assessment tools have been applied successfully around the world to generate evidence for decision-making in the management of a variety of PHEs, including HIV/AIDS (92), forced displacement due to conflict (93), natural disasters (90) and more recently COVID-19 (94). Such tools are vital for identifying and addressing context specific issues, in acting as a guide for resource allocation and providing key information in relation to response planning and implementation as evidenced during the 2013-2016 Ebola epidemic in West Africa (95). Each of the PHEs described in our paper has important lessons. Notably, comprehensive and reliable data generated through well-designed and well-executed research (e.g., realtime epidemic forecasting and disease surveillance through administrative data systems) will prove important in resolving important research questions and existing knowledge gaps. It follows that any research during PHEs should only be conducted if it has high social value (i.e., it provides information to support the immediate response, either through evidence to assist the decision-making process or targeted interventions aimed at minimizing the magnitude of the harm suffered by a population) (96). Importantly, the knowledge generated through research in anticipation of, during, and after a PHE is vital to building future capacity to better achieve the goals of preparedness and response: preventing illness, injury, disability, and death and supporting recovery (97). However, conducting research in PHE settings often presents a number of challenges, including an inability to access affected people, insecure settings and lack of research infrastructure (e.g., underdeveloped oversight and regulatory bodies of host countries) (98)(99)(100). In recent years digital health technologies have been harnessed as a means of data collection in a variety of research settings including PHEs (101). Specifically, these technologies improve the quality and efficiency of research studies via automated data capture and by improved data traceability, reliability and provenance (102). Additionally, digital technologies can improve study transparency, security, informed consent, handling of confidential patient information and data sharing (102). This stated, challenges exist in relation to data sharing mechanisms as well as the technical and legal ability to protect intellectual property (e.g., inventions/innovation and research publications), particularly in the context of LAMICs. Some solutions toward improving research in the PHE context include appointing a coordinator for scientific research-a role that involves coordinating the research process, identifying mechanisms and rapid funding schemes to support research, enlisting existing research networks in order to coordinate and accelerate research efforts (e.g., for data collection), and establishing a centralized institutional review board to provide timely reviews of multiagency studies (97). In line with the lessons learned from our paper, Khan and colleagues (103) identified a number of important considerations/factors for enhancing research, policy and practice related to preparedness and response to PHEs which include governance and leadership, community engagement, risk analysis, surveillance and monitoring, resources, investment in systems strengthening and capacity building, communication, research and learning, and evaluation (103). Each of these factors are described in further detail-in relation to their significance in policy decision-making processes-in Table 1. --- CONCLUSIONS Public health emergencies in the African region continue to exert an enormous toll on people's livelihoods-with some PHEs characterized by excessive mortality and morbidity rates-often testing collective resilience. Unfortunately, there are significant challenges that surround coordinated rapid response capacity, staffing and capacity building, and quality improvement with respect to most of the PHE response efforts on the African continent, which points to systemic fragility. Compounding matters further, PHEs and some of their secondary effects are bi-directional, and these secondary effects-such as distrust of health authorities and disruption to health services-are seen to make it harder to combat these humanitarian crises (104,105). Notwithstanding, there have been vital lessons learned from previous PHEs and African governments, institutions and partners have devised various initiatives toward more holistic PHEP. In light of the PHEs discussed in this paper, namely Ebola, COVID-19 and locust swarms, initiatives to strengthen preplanned and coordinated response have included: containment measures (e.g., social distancing and border restrictions); building local and international collaborations to leverage expertise, international aid, and other resources; scaling-up surveillance and monitoring activities; leveraging initiatives like COVAX to ensure vaccine roll-out and supply; management and treatment of survivors; social protection programs against shocks to livelihoods; and community engagement. To develop expertise and a well-staffed workforce, various training programs as part of capacity building (e.g., in surveillance, laboratory work, infection prevention and control, data management) coupled with mentorship and leadership training have been found to be beneficial. Foreign worker deployment has also been critical especially in relation to the Ebola virus outbreak. Lastly, as a means of improving accountability and quality improvement in PHEP, African governments and institutions have utilized the Financial Tracking Service of the UNOCHA to monitor humanitarian financial assistance and commitments. Dialogue between governments and institutions on innovative accountability mechanisms and crisis budget support operations has strengthened cross-learning on best practices for accountability. Some countries have also devised community monitoring approaches to improve trust and monitor quality of services during PHEs. Overall, adoption of system-wide approaches matched by scaling-up innovations to achieve impact may prove effective to better deal with the negative outcomes related to complex PHEs on the African continent. There will also be a need to refine policies on leadership relating to PHEP and response in conjunction with policies that focus on strengthening national and technical capacities that align with the IHR as a means of accelerating progress toward universal health coverage. --- DATA AVAILABILITY STATEMENT The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors. --- AUTHOR CONTRIBUTIONS All of the authors substantially contributed to the conception and drafting of this manuscript. --- Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Africa is home to 54 United Nation member states, each possessing a wealth of ethno-cultural, physiographic, and economic diversity. While Africa is credited as having the youngest population in the world, it also exhibits a unique set of "unfortunate realties" ranging from famine and poverty to volatile politics, conflicts, and diseases. These unfortunate realities all converge around social inequalities in health, that are compounded by fragile healthcare systems and a lack of political will by the continent's leaders to improve smart investment and infrastructure planning for the benefit of its people. Noteworthy are the disparities in responsive approaches to crises and emergencies that exist across African governments and institutions. In this context, the present article draws attention to 3 distinct public health emergencies (PHEs) that have occurred in Africa since 2010. We focus on the 2013-2016 Ebola outbreak in Western Africa, the ongoing COVID-19 pandemic which continues to spread throughout the continent, and the destructive locust swarms that ravaged crops across East Africa in 2020. Our aim is to provide an integrated perspective on how governments and institutions handled these PHEs and how scientific and technological innovation, along with educational response played a role in the decision-making process. We conclude by touching on public health policies and strategies to address the development of sustainable health care systems with the potential to improve the health and well-being of the African people.

Introduction Digital healthcare services (DHS) have become an integral part of the health services provided by healthcare organizations in Israel and in the world, and it has been called as "virtual service revolution". 1 It is part of the technological revolution developed in many fields and offers solutions to the growing burdens in the health systems. In addition, the DHS aims to decrease the gap between the requirements of available and accessible services and the lack of resources. 2 DHS can support health systems to deliver more health care, to promote health and prevent diseases. 3,4 Such services were also effective in hospitals, and they can reduce demand for (in-house) consultations, medical procedures, and unnecessary hospitalizations and improve postoperative monitoring of patients. 5,6 DHS can also be beneficial for individuals and patients with chronic diseases; it supports self-management and preventive behaviors related to chronic diseases. 3,7 It seems that in recent years the tendency to rely on technology in the field of healthcare services is expanding. 8,9 DHS includes a wide range of services like mobile applications of digital information technologies and more. In Israel, the four health maintenance organizations (HMOs) started developing such services many years ago, and it became essential after the outbreak of the corona virus. It includes websites, consultation with different specialized physicians, maternity care, service for receiving prescriptions and information about pharmacies, administrative services, and more. The utilization of DHS was less among certain groups based on various factors. These factors include advanced age, male gender, lower levels of education and income, and a disadvantaged socioeconomic background. [10][11][12][13] Obstacles to DHS utilization can also arise from the breakdown or interference of established resources or systems. 14 Huxley et al 15 in their review refer to barriers among marginalized groups (itinerant populations such as refugees, homeless people, unemployments) compared to the general population. The review revealed that marginalized groups reported access difficulties and stigmatizing reactions from health professionals and other patients. Previous qualitative, quantitative, and mixed articles analysis review 16 showed that eHealth can widen the gap between those at risk of social health inequalities and the rest of the population. Ethnicity and low income were the most commonly used characteristics to identify people at risk of social health inequality. Norman & Skinner 17 found that high levels of health literacy in general, and digital health literacy in particular, is needed for the utilization of DHS. Health literacy is defined as "the degree to which individuals can obtain, process, understand, and communicate about health-related information needed to make informed health decisions" p.16. 18 When digital health literacy, in addition, requires more skills to obtain online health information. 17,19 Previous research indicated that higher levels of digital health literacy related to better health, healthy behaviors, and increased knowledge regarding the management of chronic diseases. 20 Levels of digital health literacy were low among disadvantaged population groups with low socioeconomic status. A literature review and meta-analysis 21 found that accessibility to infrastructure and low levels of education were the main factors for this. Digital health literacy is a crucial means that now goes beyond restricted access to information to the denial of actual healthcare services. It is imperative to recognize that in the third millennium, digital literacy has evolved beyond mere technological expertise and has become a tool that empowers individuals to access various services, including healthcare, on an equal footing. Other studies also report that the most powerful predictors of not using information technology among older adults are cognitive decline associated with aging processes and attitudes such as anxiety about computer use and the perception that the technology was not useful for them. [22][23][24] Numerous models have been created to explore and comprehend the factors that influence the acceptance of computer technology. The technology acceptance model (TAM) proposed by Davis 25 was one of those theories. The theory proposes that user acceptance, which is affected by three elements, namely perceived usefulness, perceived ease of use, and attitudes towards usage of the system, can determine the effectiveness of a system. The theoretical frameworks utilized to examine user acceptance, adoption, and usage behavior. The Ministry of Health in Israel in 2017 2 declared a policy to encourage the use of digital services in order to improve the quality of care. They initiated a "National program for digital health". Even though surveys show significant disparities between groups in utilizing DHS. Low rates were found to be among Arab minorities in Israel. 2 Previous research from other countries also found low adherence rates of utilizing DHS among minority populations. 1,26 This research focused on the Arab minority living in Israel. They constitute about 21% of the population in Israel. 27 Almost 50% live in Northern region, 10% in the central region, 20% in Haifa, and 20% in the southern part of the country. The Arab community is characterized by low socioeconomic status and higher health disparities. 28,29 Little is known about utilizing DHS among the Arab community. Recent published data (in Hebrew) of Laron et al 30 found that more than 90% of Arabs use the internet and have smartphones. 60% of them reported that they use telehealth services just to set a doctor appointment. Two-thirds did use the health plan's application. The main barrier for using such services was a lack of awareness about using DHS, when previous acquaintances with the doctor and services in Arabic were facilitating factors. There was a significant correlation between education level and the utilization of telehealth for written communication with a known healthcare professional. They concluded that even though high percent of the Arabs have an access to the internet, the usage of the DHS is still limited. Thus, this research aims to deepen the knowledge about the other characteristics and barriers of DHS among the Arab community compared to the Jewish community in Israel and to examine the general model of literacy, including attitudes and usage, and to investigate the impact of ethnicity on individuals' patterns of use. Following, the research hypotheses are: 2. There will be a positive relationship between digital health literacy and the extent of use of DHS between the two ethnic groups. 3. There will be a positive relationship between attitudes towards DHS and the extent of use of DHS between the two ethnic groups. 4. Attitudes towards DHS will mediate the relationship between digital health literacy and the extent of use of DHS. --- Materials and Methods --- Study Design This study used a cross-sectional correlational design. An online survey was conducted during 13 September to 01 October 2022 using a closed digital questionnaire. The questionnaire was administered via a well-known survey institute to a panel of optional respondents sample of 609 Israeli citizens. The survey was conducted in Hebrew and Arabic. Participation in the survey was voluntary, and participants were not offered any compensation. --- Participants and Data Collection A representative sample of 609 subjects participated in the study: 165 from the central region, 123 from Tel Aviv, 88 from the northern region, 83 from Haifa, 80 from the southern region, 45 from Jerusalem, and 23 from the Judea and Samaria region. The inclusion criteria were Israeli adults Arabs and Jews. It should be noted that the Arab population in Israel is overrepresented in this study, and the sample represents the distribution of the insured in the various health maintenance organizations. Participants were first informed about the purpose of the study and maintained participant confidentiality. They were informed about the option to refuse to complete the questionnaire or stop filling it out at any time without any consequence to themselves. They then gave their informed consent. --- Variables and Measurements --- Demographic Characteristics The following demographic data were collected: gender, year of birth, place of residence, marital status, number of children, religion, level of religiousness, occupation, education, financial status, health status, and membership in a health fund. The Extent of Use of Digital Healthcare Services (DHS) Questionnaire The extent of use of DHS questionaire based on a questionnaire developed by Even-Zohar et al 31 and included eight digital healthcare services such as scheduling appointments and viewing test results. To validate the questionnaire and adapt it to the research purpose, it was forwarded to three experts who were asked about the degree of relevance of each of the items. In light of the experts' comments, one item was omitted from the questionnaire, and 3 new items were added. The final questionnaire included 10 items. The participants were asked to mark the frequency of use for each service on a scale between 6 levels: 0 -not familiar, 1 -familiar but never used, 2 -seldom, 3 -sometimes, 4 -in most cases, and 5 -whenever necessary. For data processing, one average was calculated for the scale of extent of use of DHS and a high score indicates a greater extent of use of DHS. The questionaire internal reliability (Cronbach's alpha) was <unk>=0.87. --- Attitudes Towards the Use of DHS Questionnaire The questionnaire was designed for the present study. The questionnaire consists of 6 positive attitudes towards DHSs, for example: "The digital services allow me to perform actions quickly" and 4 negative attitudes, for example, "It is difficult for me to use the digital services". The degree of agreement on each item is measured on a Likert scale between 1 -do not agree at all and 5 -agree to a very large extent. To validate the questionnaire, it was passed to three experts who were asked about the degree of relevance of each of the items. Considering the experts' comments, the wording of three of the items was corrected and 2 new (inverted) items were added. The final questionnaire included 12 items. For data processing, one average was calculated for the scale of attitudes towards using DHS (after reversing the negative items), and a high score indicates more positive attitudes toward DHS. The internal reliability (Cronbach's alpha) of the questioniare was <unk>=0.86. --- Digital Health Literacy Questionnaire The questionnaire is based on Norman's and Skinner's 17 research questionnaire. The questionnaire included eight items measuring knowledge and skill in locating, evaluating, and applying health information from digital sources. For example: "I know where to find effective health information on the Internet". For each item, there are five answer options: 1 -do not agree at all and up to 5 -agree to a large extent. The Arabic version of the questionnaire was translated and validated by W<unk>ngdahl et al 32 and high internal reliability was found (0.92). For data processing, one average was calculated for the scale of digital health literacy and a high score indicates a higher literacy towards DHS. In the present study, the internal reliability (Cronbach's alpha) was <unk>=0.91. --- Data Analysis Analyses were conducted using the IBM SPSS Statistics 25.0. The analysis was calculated on 606 responses. The missing values were less than 0.02% and were not replaced. Cronbach's <unk> coefficient was measured to verify the reliability of the measurement tools used in the study. Group comparisons were performed using the t-test for continuous variables and the <unk> 2 test for categorical variables. To compare the means of the research variables between the ethnic groups and gender we used 2<unk>2 ANOVA's. Correlations between the study variables were analyzed using Pearson correlations. We used Fisher r-to-z transformation to compare the correlations between the two ethnic groups. We conducted a hierarchical regression analysis to test all relationship variables contribution to predicting the extent of the use of DHS. Finally, we conducted an analysis using the PROCESS macro for SPSS (model 7) to examine the moderated mediation model for predicting the extent of use of DHS. 33 A 95% confidence interval (CI) was calculated for each regression coefficient included in the model. The moderated mediation approach utilizes a bootstrap test, for which we generated 5000 samples, to produce 95% confidence intervals, which indicate a significant indirect effect if they do not include 0. 34 --- Results --- Participant Demographic Characteristics The study included 606 participants from two ethnic groups: 445 Jews and 161 Arabs. Table 1 shows significant differences in gender, religiousness, income level, number of children and health condition between the two ethnic groups. Among the Jews, 55.3% were women, 60.9% were married, whose ages ranged from 20 to 84 years (M = 43.9, SD = 16.4), and parents to an average of 1.9 children (SD=1.8). Most of the Jews reported that they were secular or traditional (83.1%). Moreover, more than half of them have an academic education (55.5%), and most are salaried employees (70.9%) and almost half with lower than average level of income (41.6%). Most of the Jews participants reported being in good or very good health (90.3%), while only 28.8% have a chronic disease. Among the Arabs, 38.5% were women, 72.0% were married, whose ages ranged from 21 to 69 years (M = 41.7, SD = 11.9), and parents to an average of 2.5 children (SD=2.1). Most of the Arabs reported that they are secular or traditional (62.1%). Additionally, half of them hold an academic education (50.3%), and most are salaried employees (71.9%) and with lower-than-average levels of income (82.6%). Finally, most of the Arabs reported being in good or very good health (74.5%), with only 28.0% have a chronic disease. --- Differences in the Research Variables Between the Ethnic Groups and Gender A 2<unk>2 ANOVA was performed in order to compare the means of the research variables between the Jews and Arabs and between woman and men. health literacy between Jews and Arabs. In addition, there was no significant main effect for gender in all research variables. Finally, there was no significant interaction effect between gender by ethnicity in any of the measures. Table 3 shows <unk> 2 tests to examine the differences between Jews and Arabs in each item of the extent of DHS utilization. The table shows that except of using online pharmacy (item 10) the Jews use all operations more frequently than Arabs do. --- Relationship Among the Research Variables in the Two Ethnic Groups Correlations between the research variables (extent of use, attitudes, and digital literacy) were explored and are reported in Table 4. As can be seen, in both groups there was a significant positive correlation between the extent of use of DHS and attitudes towards DHS (r=0.39 for Jews and r=0.34 for Arabs). In Fisher Z test no significant difference was found between those correlations (z=0.62, p>0.05). In addition, a significant positive correlation was found between the extent of the use of DHS and digital health literacy in both groups. Although the correlation is stronger among the Jews (r=0.41) than among the Arabs (r=0.28), in Fisher Z test no significant difference was found between those correlations (z=1.60, p>0.05). --- 3539 In contrast, the correlation between digital health literacy and attitudes towards DHS is significantly stronger among Arabs (r=0.53) than Jews (r=0.34). In Fisher Z test, significant differences were found between those correlations (z = -2.55, p <unk> 0.01). --- Hierarchical Linear Regression Analysis for Predicting Extent of Use of DHS Hierarchical linear regression was used to predict the extent of the use of DHS in the two ethnic groups. Step one included demographic variables: gender (1-woman, 0-man), religiousness (1-religious/very religious, 0-secular/traditional), income (1-average or above, 0-below average), health condition (1-good, 0-not good), and number of children. Step two included the following variables: attitudes toward DHS and digital health literacy. The results of these analyses are presented in Table 5. In step one, income was found to be a significant predictor of the extent of use of DHS among the Jews, whereby Jewish subjects with an average income and above use digital health services to a greater extent. Among the Arabs, the demographic variables do not significantly predict the use of DHS. Combined, the demographic variables explained 5% of the variance of the extent of use among the Jews and 3% among the Arabs. In step two, attitudes towards DHS and digital health literacy were significant predictors of the extent of use among the Jews. It was found that the more positive the attitudes and the higher the literacy, the greater the extent of use of DHS. These research variables added an additional 20% to the explained variance. Among the Arabs, only attitudes towards DHS were found to be significant predictor of the extent of use of DHS. Step 2 added an additional 14% to the explained variance. In total, our model explained 25% of the variance of the extent of use among the Jews and 17% among the Arabs. Moreover, the model is statistically significant in both groups [F(7,436)=20.93, p<unk>0.001 for Jewish and F(7,152)=4.50, p<unk>0.001 for Arabs]. To test the moderated mediation model for predicting the extent of use of DHS, linear regression was used using Model 7 in the macro-PROCESS. 33 The independent variable was digital health literacy, and the mediating variable was attitudes toward DHS. The Ethnicity was chosen as moderating the relationship between digital health literacy and the attitudes towards DHS since a significant difference was found in the correlation between these two variables between Jews and Arabs. The results of this analysis are presented in Figure 1. The bootstrap analysis (5000 samples) 34 was used in PROCESS 3.0 Model 7. 33 The results of the analysis are presented in Table 6. As can be seen in Table 6, the interaction between digital health literacy (X) and religion (W) is significant. The effect of digital literacy (X) on attitudes (M) was more substantial for Arabs (b=0.46) than for Jews (b=0.29). In addition, the index of moderated mediation (IMM) indicated that religion (W) moderated the indirect effect of digital health literacy (X) on extent of use (Y) through attitudes (M). The indirect effect of digital health literacy on the extent of use through attitudes was significant at the two groups of the moderator but is stronger among the Arab group. This finding suggests that the effect of digital health literacy via attitudes on the extent of use was moderated by ethnicity. --- Discussion DHS has the potential to improve healthcare access and outcomes for all populations, including minorities. However, minority communities may face barriers in utilizing digital health tools due to a lack of technological access, lower income, language barriers, and limited digital health literacy. The current research investigated the characteristics of DHS utilization and the barriers related to it and compared among the communities living in Israel Jews as a majority and Arabs as a minority. The results show that on average, the Jewish participants use more digital services and have more positive attitudes toward them than the Arabs have. Examining the differences in the extent of use of DHS between Jews and Arabs in each of the utilizing issues revealed that Jews use all operations more frequently than Arabs except for online pharmacy (both have low usage rates). Significant predicting factors related to DHS utilization among Jews include positive attitudes and high digital health literacy. Among the Arabs, only attitudes towards DHS were found to be a significant predictor of the extent of DHS use. Moreover, digital health literacy significantly affects the extent of use through attitudes in both groups but it is stronger among the Arab group. This finding suggests that the effect of digital health literacy via attitudes on the extent of use was moderated by ethnicity. The Arab minority do not have limited access to technology, as the latest unpublished data from 2021 shows that more than 90% of Arabs use the internet and have smartphones. 30 Another research held in 2019 also reported small differences between the Jewish and Arab communities in the daily rate of internet use (82% compared to 77%, respectively) and a considerable gap in the rate of computer use among the same groups (77% Jews compared to 46% among Arabs). 35 Despite these facts, the results of the current study still indicate disparities in utilizing DHS between Arabs and Jews. One of the factors related to these disparities is cultural barriers. There may be a distrust that leads to discomfort in using this technology and utilizing DHS among the Arab community, which may prevent them from using it. Lack of trust can emerge due to concerns about privacy, data security, and the accuracy of health information obtained through digital platforms. Previous research indicates trust as a necessary aspect of successfully using electronic health records and other electronically stored health information. 36 It is particularly important for low-income communities. 37 Mistrust of the Arab community on the health system raised on the recent years during and after the COVID 19 pandemic. Research conducted in 2021 and 2022 38,39 found moderate to low levels of trust in the health system among the Arab community. A related concern to safety use is the fact that private information is being recorded and stored. 40 Moreover, language can be another cultural barrier to using DHS as it affects a person's ability to understand and navigate the technology. If DHS is not offered in a person's native language or if the language used is not accessible or understandable, it can limit their ability to use these services and benefit from them effectively. This can also lead to mistrust, confusion, and, ultimately, decreased usage of DHS among certain cultural groups. Language barriers can also affect the accuracy of information exchanged and the quality of care received, leading to further disparities in healthcare access and outcomes. [41][42][43] The participants reported on language translation and cultural adaptation as a crucial factors enabling them to use the DHS. Therefore, digital health providers need to consider the linguistic diversity of their target population and offer services in multiple languages to ensure equitable access and utilization of these resources. The study found that digital literacy and attitudes toward using DHS are factors that explain the differences in the use of DHS among different groups. According to the Technology Acceptance Model (TAM), 25 a person will use digital services if they have both the perceived ability and the skills to do so and perceive the usefulness of it. The study found that Jews had more positive attitudes towards using DHS and higher digital literacy levels, which led to higher usage of digital health services. However, among Arabs, the low levels of general literacy, which affects digital literacy and brings them to have negative attitudes towards using digital services, resulted in lower usage of DHS. It was also found that people with low digital literacy would only use technology for simple tasks, such as playing games or browsing websites. 44 To conclude, the study showed that compared to Arabs, Jews have more positive attitudes toward the use of DHS and a higher digital health literacy, which perhaps leads to higher utilization rates. The use of DHS has already been proven to contribute to improving and maintaining health in most studies, and it can also contribute to closing the gaps in health between population groups of different socioeconomic status, such as the Jewish and Arab populations. Addressing these barriers and ensuring equal access to DHS is essential to reducing healthcare disparities among the Arab communities in Israel. It can be done through targeted digital literacy education, removing barriers to technology access, offering services in Arabic, designing and implementing reward strategies to motivate them to use DHS, and collaborating with community organizations to reach underserved populations. --- Limitations Although the sample in the study was a representative, the questionnaire was distributed online, through the Internet. This method can be convenient, on the one hand, as the researchers can reach a large number of participants. Yet, on the other hand, it can be a limitation. Since online surveys may only reach a limited sample, as not all people have access to or are comfortable using the internet or electronic devices. This can lead to the underrepresentation of certain groups and might limit the generalizability of the findings. Participants in online surveys self-select to participate, which can introduce bias into the sample. People who choose to participate may differ from those who do not in important ways, which can affect the accuracy of the results. While the research, conducted in the form of a cross-sectional survey and testing a moderated mediation model, indicates a possible causal link between variables, it's essential to emphasize that the analysis serves as a supplementary tool and cannot definitively establish causal relationships. Thus, the model in this research provides evidence for a possible explanation of the relationship between variables, but it does not prove that the relationship is causal. --- Ethics Considerations This study complies with the Declaration of Helsinki. Ethical approval was obtained from the Yezreel Valley College Ethics Committee before data collection (Approval No. YVC EMEK 2022-61). --- Disclosure The authors report no conflicts of interest in this work. --- Journal of Multidisciplinary Healthcare --- Dovepress --- Publish your work in this journal The Journal of Multidisciplinary Healthcare is an international, peer-reviewed open-access journal that aims to represent and publish research in healthcare areas delivered by practitioners of different disciplines. This includes studies and reviews conducted by multidisciplinary teams as well as research which evaluates the results or conduct of such teams or healthcare processes in general. The journal covers a very wide range of areas and welcomes submissions from practitioners at all levels, from all over the world. The manuscript management system is completely online and includes a very quick and fair peer-review system. Visit http://www.dovepress.com/testimonials.php to read real quotes from published authors.

The purpose of this study was to examine ethnic disparities in the utilization of digital healthcare services (DHS) in Israel and explore the characteristics and factors influencing DHS use among the Arab minority and Jewish majority populations. Methods: A cross-sectional correlational design was employed to collect data from 606 Israeli participants, 445 Jews, and 161 Arabs. Participants completed a digital questionnaire that assessed DHS utilization, digital health literacy, attitudes towards DHS, and demographic variables.The findings reveal significant disparities in DHS utilization and attitudes between these ethnic groups, with Jewish participants demonstrating higher rates of utilization and positive attitudes toward DHS. The study also explores the predictive role of digital health literacy and attitudes in DHS use while considering ethnicity as a potential moderator. Significant predicting factors related to DHS utilization among Jews include positive attitudes and high health literacy. Among the Arabs, only attitudes towards DHS significantly predict the extent of DHS use. Digital health literacy affects the extent of use through attitudes at the two groups of the moderator significantly, but it is stronger among the Arab group. Conclusion: To improve healthcare outcomes and reduce disparities, efforts should focus on ensuring equitable access to DHS for the Arab minority population. Targeted interventions, including digital literacy education, removing technology access barriers, offering services in Arabic, and collaborating with community organizations, can help bridge the gap and promote equal utilization of DHS.

INTRODUCTION The legal system in Indonesia ensures the protection of the rights of Indigenous people, including the Indigenous Dayak Ngaju community. 1 They 1 See Mohammad Jamin et al, "Legal Protection of Indigenous Community in Protected Forest Areas Based Forest City," Bestuur 10, (2022): 198-212; and Yanarita et al, "Development of the Dayak Ngaju Community Forest in the Forest and Peatland Area, Central Kalimantan, Indonesia," Environmental Science, Toxicology and Food Technology 8, no. 3 (2014): 40-47. Yanarita characterizes Dayak Ngaju as indigenous communities residing in the Central Kalimantan province. In both literature and everyday language, Dayak Ngaju are often referred to as an "indigenous community," "tribe," or "subtribe." are guaranteed full involvement in decision-making and sustainable management of the ecosystem in their territory. However, the lack of empowerment and ineffective implementation of these standards hinder the Indigenous communities ability to fully leverage their knowledge, values, and wisdom in peatland management. 2 It is crucial to highlight that the Dayak people are Indigenous people of Kalimantan island, where Indonesia's new capital is planned to be relocated. 3 With over 400 subtribes, each having its unique traditional system of land and natural resource management, the Dayak community possesses valuable insights and practices. 4 Dayak community is categorized or grouped based on their dwellings in watersheds, such as the Kapuas, Katingans, Seruyan, and Barito river. Dayak people who live in the upper river in Central Kalimantan Province is called Dayak Ngaju. 5 Dayak Ngaju heavily relies on the peatland ecosystem in which they live. 6 Traditionally, they have utilized the peat swamp forests for various small-scale activities such as timber harvesting, gathering food and medicinal plants, and obtaining clean water. As forest-dwelling people, the environment has played a significant role in shaping their culture and way of life, with the peatlands deeply intertwined with the customs and traditions of the Indigenous Dayak people. 7 Although Law No. 32 of 2009 concerning Environmental Protection and Management promotes participatory principles in environmental conservation and management, with an emphasis on the value of local Although these terms may hold distinct legal and anthropological connotations, this paper employs them interchangeably in a general sense to represent a community known as the Dayak Ngaju that inhabit an upper stream in Central Kalimantan Province, united by collective ancestral connections to the land and natural resources, and steadfastly maintain the traditions passed down through generations. 2 See Gusniarjo Mokodompit et al, "Ensuring the Rights of Indigenous Peoples: International Legal Standards and National Implementation," The East Journal Law and Human Rights 1, no. 3 (2023) knowledge, there is still a need for further action to actively involve Indigenous peoples, particularly the Dayak community, in the management of peatland ecosystem.8 This necessity arises from the historical context prior to the Constitutional Court Decision No. 35/PUU-X/2012, where the legal framework in Indonesia did not recognize community ownership of forested land as previously stated in the Law No. 41 of 1999 concerning Forestry (Forestry Law). 9 According to the Forestry Law, country's forests are categorized as state forest and private forest. The law declares that any forestlands in Indonesia without private entitlements are considered state forestlands. Unfortunately, the implementation of forest-protection actions in the past often excluded local communities, resulting in no tangible benefits for them. 10 This lack of recognition resulted in conflicts between governmentdesignated zones and areas acknowledged under Customary Law. Recognizing the rights of Indigenous peoples, including the Dayak community, to govern their natural resources has long been a demand. 11 The recognition of customary forests as distinct from state forests, as stated in the Constitutional Court ruling, marked a significant change. 12 However, several studies examining the limitations and challenges associated with the implementation of the Constitutional Court Decision No. 35/PUU-X/2012 have identified several important points. Firstly, little power has been transferred in favor of Indigenous people, as the state retains full authority in determining the procedure of customary forest recognition. Secondly, the recognition of Indigenous communities has been hindered by concerns of maintaining national integration and the complex articulation of Indigenous identity due to historical and post-colonial dynamics. Additionally, the absence of an Indigenous peoples' rights law has been a common reason for many local governments to avoid recognizing Indigenous territories. 13 The Mega Rice Project, initiated in 1995 by President Soeharto in Central Kalimantan, aimed to enhance food security but neglected the active involvement of Indigenous Dayak communities. 14 The project involved extensive drainage canal construction, deforestation, and the introduction of Javanese transmigrants using unfamiliar farming methods, causing irreparable damage to the peatland and local communities. Similarly, the Food Estate project, characterized by intensive modern farming, fails to consider the Indigenous Dayak Ngaju Wisdom in agricultural practices and community land management. 15 This constitutes a violation of Indigenous Dayak people's rights as clearly recognized in the Article 18B (2) of the 1945 Constitution of the Republic of Indonesia and Law No. 39 of 1999 concerning Human Rights. 16 The peatland land use change in Kalimantan, particularly for agricultural purposes such as monoculture production of palm oil and rice, has often disregarded the valuable knowledge and wisdom of Indigenous communities. 17 This disregard has led to significant consequences, including the degradation of ecosystems, socio-economic conflicts, and land disputes. The shift towards monoculture plantations, particularly palm oil and rice, has disrupted the delicate balance of peatland ecosystems, resulting in the loss of biodiversity and the destruction of natural habitats. 18 The traditional knowledge held by Indigenous peoples, which encompasses sustainable land management practices and a deep understanding of the ecological intricacies of the region, has been marginalized and ignored. 19 As a result, the degradation of peatlands has not only had ecological implications but has also caused socio-economic conflicts as indigenous communities, whose livelihoods and cultural heritage are closely tied to the land, are marginalized and deprived of their rights. Additionally, the neglect of indigenous land rights has fueled land conflicts, further exacerbating tensions and instability in the region. 20 Recognizing and incorporating the knowledge and rights of Indigenous peoples is crucial for addressing these issues, promoting sustainable land use practices, and fostering a harmonious coexistence between agriculture and the environment. 21 Understanding the significance of the peatland ecosystem to develop effective management strategies requires acknowledging people's values, perceptions, and traditional knowledge systems. 22 The Indigenous Dayak people possess valuable traditional knowledge that provides a holistic understanding of the peatland ecosystem. The profound connectedness of indigenous people with nature and their wisdom have been highlighted as a vital aspect of the sustainable management of socio-ecological systems. By involving and valuing the knowledge and perspectives of Indigenous communities, a more inclusive and participatory approach to peatland conservation can be developed. 23 The present paper builds upon previous studies exploring the topic of indigenous land management and environmental justice. Boag (2016) conducted a comparative study on Australia and Indonesia, providing insights into the benefits and limitations of different policy strategies for Indigenous peoples in the Asian-Pacific region. 24 Jamin et al (2022) analyzed the protection of customary law communities in urban forest-based protected forest areas designated as the National Capital, recommending the inclusion of legal protection provisions for Indigenous Peoples in the Law on the State Capital. 25 Furthermore, Belliera and Preaud (2011) examined the transformative effects of recognizing Indigenous peoples globally, exploring various local contexts and strategies and uncovering transnational links and differences. 26 These studies contribute valuable perspectives to the understanding of Indigenous land management and the pursuit of environmental justice for Indigenous communities worldwide. This research aims to explore the values, cultural significance, and traditional practices held by the Indigenous Dayak communities regarding the peatland ecosystem. By understanding their perspectives, this study seeks to develop a comprehensive legal framework that actively involves and appreciates the values of local communities in ecosystem management, fostering ownership, responsibility, and stewardship. Ultimately, the goal is to establish effective and sustainable conservation strategies aligned with the aspirations and priorities of the Indigenous Dayak people, ensuring the long-term protection and well-being of both the peatlands and the Indigenous communities. The research was conducted in Tumbang Nusa village and Pilang village, located in Pulang Pisau Regency, Central Kalimantan Province. The selection of these villages as study areas was based on the following factors: First, the majority of the residents in both Tumbang Nusa and Pilang villages, specifically over 90%, belong to the Indigenous Dayak Ngaju community. These communities actively preserve and practice the traditional customs and culture of the Indigenous Dayak Ngaju. Second, both villages possess peatland ecosystems, which are currently undergoing changes in land use patterns. Third, the proximity of the villages to the researchers and their ease of accessibility were additional factors considered during the selection process. To collect and collate the data for this study, a cross-sectional survey was used. A one-month in-depth interview was conducted in Tumbang Nusa village and Pilang village from mid-September 2021 to mid-October 2021 using purposive sampling with a total of eighty-eight respondents. The indepth interview was conducted with village officials, Dayak elders, Mantir (Indigenous leaders), a Non-Governmental Organization (NGO) representative, researchers, and the provincial forestry department. --- RESULT AND ANALYSIS 2.1. Indigenous Dayak Ngaju Peatland Management There are no rules specifically governing peat forests in the Dayak Ngaju Indigenous community. Peat forest management is based on the zonation of the land. Peat forests can be used as agricultural areas, settlements, or protected forests. Peatland zonation is not based on the depth of the peat soil but based on the vegetation that grows in the area which then becomes a guideline for the community to determine the use of the area. It covers three zonations for the purposes of settlement, farming, and secondary forest. Regarding the settlement zonation within the Dayak Ngaju Indigenous community, the organization of villages follows traditional rules that have been passed down through generations. In the villages of Tumbang Nusa and Pilang, most of the stilt houses are constructed on flat land near the river. Consequently, peat areas adjacent to rivers are commonly utilized as settlement sites. This strategic choice is motivated by the convenience of accessing water, fish, and transportation routes, which are readily available in close proximity to riverside settlements. 27 For the farming zonation, the selection of land for farming, known as "ladang," is guided by several considerations, which are the presence of a nearby river or creek, the abundance of fresh and green leaves on plants within the primary forest, and the presence of specific grasses and trees, such as taro, suna, bungur tree (Lagerstroemia) and Jajangkit tree. These criteria do not apply to areas that were previously cleared for agriculture. In such cases, the land can be reused if there are trees with a trunk diameter exceeding 15 cm. 28 If a member of the Dayak Ngaju community clears land (A) within the primary forest for cultivation and subsequently re-cultivates it after a few years, the fertile land is referred to as "balik uwak." The individual who cleared the land is rewarded with rights to the land, acknowledging their hard work. According to Dayak Customary Law, the responsibility for managing forests converted into agricultural land rests with the initial cultivator. This principle is enshrined in Article 39 of the Customary Law of the Dayak, known as "Singer nalinjam bahu himba balikuwak." If another individual (B) wishes to work on the previously cultivated land (A), they are obligated to compensate the previous cultivator (A) with voluntary offerings such as rice, white chicken, whetstone, machete iron, pickaxe, and manas lilis. The subsequent rights to the former field, after being cultivated by individual (B) for one or two years, will revert back to the ownership of individual (A). 29 The Dayak Ngaju people have developed a deep understanding of the peatland ecosystem through keen observation of the vegetation present in the area. By closely observing the growth of specific plants they can discern the ecological characteristics of the land. This Indigenous knowledge guides their decision-making in choosing the most suitable areas for farming. They avoid utilizing deep peat for agricultural purposes, as they are well aware that the soil in such areas tends to be acidic and more beneficial for conserving as forested areas. For the secondary forest zonation, when a piece of land previously used for farming or "ladang" displays a thriving growth of valuable timber and fruit-bearing trees, it is conserved as a wood, fruit, medicine, vegetable, fish, and purun (grass species used for weaving) source forest. This designation ensures the availability of wood and fruits for the community's needs, allowing for sustainable utilization of these resources. --- Dayak Ngaju Farming Practice The farming practices of the Dayak people integrate ecosystem management and Dayak traditions. The stages of farming identified in their study are as follows: inspecting the land, determining the land area, cleaning farming tools, slashing, cutting trees, burning the land, planting, weeding, harvesting, and performing a thanksgiving ceremony. These ten stages are universally followed by the Dayak people and must be completed. 31 Dayak farming is typically initiated in May, during the most favorable season, according to an indigenous elder from Pilang and Tumbang Nusa Village. For the Dayak people, farming holds a deeper meaning beyond occupation; it is a spiritual connection with all beings, especially the Almighty, the Creator of the universe. Before opening and clearing the forest, the Dayak people perform a ritual called "Mangariau" to offer prayers to the spirits of the forest guards, requesting them to relocate. Mangariau is performed on small arable land, while for larger fertile lands, a ritual called "Manyanggar" is conducted, involving the offering of pigs or cows. 32 Slashing and burning practices, although controversial due to their association with forest fires, play a crucial role in shifting cultivation by clearing land and enhancing soil fertility. 33 The Dayak people, however, have been practicing this tradition responsibly for centuries. They conduct controlled burns simultaneously, equipped with water and fire extinguishers, solely for agricultural purposes, and ensure no socioeconomic issues arise from these fires. 34 Cooperation is evident in Dayak farming, as men create holes in the soil through dibbling (Manugal), while women sow seeds in these holes. This collaborative planting process is accompanied by joyful interactions, jokes, and displays of various arts and cultures. The harvest marks the final stage of farming and brings great joy to the Dayak people. They express gratitude by performing the 'Pakanan Batu Ceremony' or "feeding the rock ritual," acknowledging the farming tools they used. These rituals exemplify the Dayak people's respect for nature and all of creation, maintaining a harmonious relationship with their environment. Forestry Law and Law No. 18 of 2004 concerning Plantation prohibit land clearing through burning, except for specific exemptions introduced after the 1997-1998 forest fires. 36 Article 69 (2) of Environmental Protection and Management Law allows the continuation of fire use in traditional agricultural techniques while considering regional customs. 37 This exemption recognizes ancient practices of slash-and-burn as local knowledge protected under the law. However, there is ongoing debate and uncertainty about the definition and application of local knowledge. 38 During the drafting of Indonesia's new Job Creation Law, discussions were held regarding the repeal of this exemption. However, the implementation of the exception remains complex, and instances have been reported where traditional farmers practicing their local knowledge were detained for using fire to clear land. The enforcement of the zero-burning policy has caused fear among Indigenous communities, leading to disconnection from their land and traditions. 39 Additionally, the Food Estate program, which promotes zero-burning farming, has not effectively integrated Dayak Ngaju traditional values and knowledge in peatland management. The program's introduction of chemical fertilizers and non-local rice seeds conflicts with the regenerative farming system of the Dayak people, which relies on local seeds and avoids chemical inputs. The implementation of similar program, Mega Rice project in 90s, had led to a significant change in the traditional Dayak farming system, resulting in the abandonment of land in Tumbang Nusa and Pilang villages. The 2015 forest fire incidents prompted the government to introduce new regulations regarding land clearing without burning, posing challenges for the Dayak community, who relied on farming as their livelihood. As a result, many community members shifted their occupations to rubber farming and fishing, leading to the abandonment of land in the villages of Tumbang Nusa and Pilang. 41 --- Pukung Pahewan as Conservation Area The Indigenous Dayak Ngaju has designated a primary forest in Pukung Pahewan as a reserve or protected forest, ensuring the tribe's future. This forest also serves as a sacred home for the "forest spirits" believed by the Dayak Ngaju people to coexist with the community. According to Article 87 of Dayak Customary Law, known as 'Singer Karusak pahewan, Karamat, rutas dan Tajahan,' anyone who mocks, burns, slashes, cuts down trees, or steals from the sacred area will face penalties. The punishment includes a demand for an inheritance penalty or compensation for the nearest village, ranging from 15 to 30 kati ramu. The offender must also conduct a small ceremony at the location, offering a pig sacrifice and covering the expenses of a mediator who communicates with the forest spirits as an act of apology. 42 The Indigenous Dayak Ngaju firmly believe in their responsibility to protect and preserve Pukung Pahewan, as it holds sacred and mystical messages within its traditions and rituals. Any disturbance, destruction, or hunting of animals or plants in the area, whether intentional or unintentional, is met with sanctions. The community fears that not only the violators but the entire village may be subjected to punishment by their ancestors and the forest spirits. 43 Pukung Pahewan represents a restricted space with specific constraints, where trees, stones, and other sacred elements must not be disturbed or harmed, including the surrounding area. It serves as a conservation methodology for the Dayak people to protect nature and symbolizes their willingness to coexist harmoniously with all organisms in nature, including animals, plants, and forest spirits. 44 Despite various policies in Indonesia regarding the conservation of peat forests, there is currently no policy that specifically addresses the and Tumbang Nusa village" interviewed by Sumarni, Central Kalimantan, 14-20 September 2021. 41 Indigenous Dayak community and village official, loc.cit. 42 --- Leveraging Indigenous Dayak Participation in Peatland Management through Customary Forest Practices The Indigenous Dayak Ngaju people have a deep understanding of the interconnectedness between their lives and the ecosystems they inhabit. Their social, economic, and cultural aspects are intricately linked to the natural environment, and there exists a reciprocal relationship between the people and the land, encompassing a concept known as the "duty of care". 45 The duty of care implies a responsibility to care for and protect the land, which is closely tied to cultural norms and values. While the community benefits from the ecosystem, they recognize their duty to ensure its wellbeing. This perspective acknowledges that any benefits derived from the environment should be balanced with the preservation of cultural heritage and ecological integrity. 46 In the Dayak Ngaju community, the economy is not viewed as separate from the ecosystems but rather as an integral part of them. The well-being of the community's economy is closely intertwined with the health of the surrounding ecosystems. As a result, there is a mutual exchange between the two, with the community relying on the resources and services provided by the ecosystem, while also recognizing the need to sustainably manage and conserve those resources. However, throughout the colonial and New Order eras, the Indigenous Dayak Ngaju people faced the unfortunate reality of their customary rights not being recognized, which deprived them of the authority to manage their natural resources and apply their local wisdom. 47 This lack of recognition became evident during the Soeharto era when the Indonesian government initiated the ill-fated Mega Rice Project in the peatlands of Central Kalimantan. Tragically, the project's improper irrigation methods and degradation of the peatlands resulted in catastrophic forest fires in 1997, engulfing extensive areas that included Pilang and Tumbang Nusa. 48 Without Free, Prior, and Informed Consent (FPIC) from the community, the government proceeded with the construction of thousands of canals, resulting in the cutting and destruction of many villagers' farm areas. This 45 See Bulkani,Ilham and Darlan,loc.cit.;and Indigenous Dayak mantir,loc.cit. 46 See Sara A. lack of consultation and consent has had significant negative impacts on the affected communities, disrupting their livelihoods and causing environmental damage. The consequences were two-fold for the communities: not only did they suffer the loss of their lands without compensation, but they also witnessed the severe environmental damage caused by the ill-conceived project. 49 The implementation of FPIC is essential to ensure that decisions made by the government respect the rights, interests, and well-being of local communities and enable more sustainable and inclusive development. 50 The experience of facing discrimination in agrarian conflicts within Pilang village heightened the Indigenous peoples' awareness regarding the significance of obtaining formal recognition for their ancestral territories. In response, they have embarked on a determined struggle to secure official acknowledgment of their customary forest, which would grant them the autonomy to independently manage their ecosystem. In 2019, their long and arduous struggle began to yield positive outcomes with the issuance of a decree by the Pulang Pisau Regent recognizing indigenous peoples. Additionally, the Ministry of Environment and Forestry issued a decree acknowledging their customary forest, as specified in Minister of Environment and Forestry Decree No. 5447/MENLHK-PSKL/PKTHA/KUM.1/6/2019. 51 Notably, the Barasak Island Customary Forest, covering 102 hectares and designated for protection, stands as Central Kalimantan's only customary forest established through a social forestry scheme. The swift issuance of the decree for Barasak Island, located within an area with a different designated use, sets it apart from customary forests in forested areas, which typically require recognition through regional regulations (peraturan daerah) as mandated by the law. 52 --- Customary Forest: Recognising Indigenous Dayak Ngaju Land Management The recognition of customary forests in Pilang Village provides greater space for Indigenous Dayak peoples to use their traditional knowledge and wisdom in managing forests for the greatest prosperity of their people. This also promotes the utilization of time and space perspectives of the Indigenous Dayak people, while emphasizing handep collaboration and equal partnership among stakeholders. 49 This framework centers around the concept of time and space as utilized by the Indigenous Dayak people. Time refers to the understanding and respect for the temporal aspects of ecological processes and the intergenerational perspective. It recognizes that sustainable management of peatlands requires long-term thinking and planning, considering the needs and well-being of future generations. The Indigenous Dayak people's knowledge of the land passed down through generations, holds insights into the temporal dynamics of the ecosystem. Space refers to the Indigenous Dayak people's intimate connection with the physical and cultural landscapes of the peatlands. It recognizes the significance of their traditional practices, cultural values, and customary land management systems. The framework promotes the preservation and revitalization of Indigenous practices and institutions related to peatland management. By valuing their knowledge and expertise, the framework seeks to incorporate Indigenous perspectives into decision-making processes. 53 The framework also emphasizes handep collaboration and equal partnership among stakeholders. It recognizes that effective peatland management requires the active involvement and meaningful participation of Indigenous communities, government agencies, non-governmental organizations, and other relevant stakeholders. Handep refers to a custom practiced by the Dayak people, where they come together to collectively clear agricultural land. 54 When one villager is clearing land, others join in to provide assistance, with relatives also contributing their labor as repayment for previously received services while working on their own fields. Those who are unable to participate may feel a psychological and customary burden, as reciprocity is valued within the Dayak community. This sense of obligation to help one another fosters a strong sense of community among the Dayak people. 55 By fostering handep collaboration, the framework aims to create a more inclusive and equitable approach to peatland management, where the voices and rights of Indigenous Dayak communities are valued and integrated into decision-making processes. This will also allow the community to run their initiative, such as making tree nurseries for native peatlands species that give them economic benefits or making bee keeping. To make this movement viable, it must be supported financially and technically by providing mentoring and tools. The Indigenous Dayak Ngaju people in Pilang village have faced a challenging struggle to obtain recognition for their customary land, primarily due to the extensive documentation required and the reluctance of some local governments to acknowledge indigenous territories. The community was struggling to navigate the intricate legal processes required for recognition, including meeting various administrative requirements and complying with governmental regulations. Limited access to legal support and information further compounds the difficulties faced by Pilang community. The procedure of obtaining legal recognition for customary forests still follows the procedure required by the Forestry Law. To be able to manage the forest, Indigenous communities should be recognized by district or provincial governments, as stated in Article 67 of Forestry Law. 56 If their territories fall within the administrative jurisdiction of a single district, recognition should come from the district government. For territories spanning across multiple districts, recognition must be obtained from the provincial government. However, the practical implementation of this provision often hampers the recognition of customary forests since many local governments are unwilling to acknowledge Indigenous territories. This reluctance from local authorities constraints Indigenous communities from gaining the legal recognition they need to govern their natural resources effectively. 57 However, their efforts have been bolstered by the invaluable support and assistance of a third party, USAID-Lestari. This external entity has played a crucial role in providing the necessary backing and resources to navigate the complex process of formal recognition. With the aid of USAID-Lestari, the Indigenous community in Pilang village has been able to overcome barriers and advance their cause, paving the way for the recognition and preservation of their ancestral lands.58 --- Barriers in Community-led Peatland Management Indigenous communities may face various challenges in governing their natural resouses. The limited availability of resources to maintain and manage their natural resources might be the biggest barrier for them. For instance, in the case of Pilang village, even after obtaining legal recognition of their customary forest, the community encountered difficulties in securing funding to support their initiatives. They faced challenges in finding financial resources for essential infrastructure development, such as custom buildings, necessary for the development of ecotourism and generating economic benefits for the community. The lack of adequate resources can hinder communities from implementing sustainable practices and maximizing the potential of their natural resources. Additionally, Indigenous communities may also encounter knowledge and managerial challenges, which pose significant barriers to effective governance. The progress in government-led peatland restoration has been constrained primarily by socio-economic challenges faced by communities. 59 To achieve successful intervention, it is essential to comprehend community concerns and develop optimal short and mediumterm income solutions that facilitate the transition to sustainable income generation. 60 By addressing these socio-economic aspects, peatland restoration efforts can become more effective and inclusive, benefiting both the environment and local communities. However, historical marginalization and limited access to education and training opportunities have left some Indigenous communities lacking the necessary knowledge and expertise. This knowledge gap inhibits their ability to manage their customary forests efficiently and fully benefit from them. Addressing the barriers to community sovereignty in governing their natural resources and obtaining legal recognition of their customary forests necessitates collaborative efforts involving government agencies, civil society organizations, and the Indigenous communities themselves. These efforts should focus on overcoming the challenges related to recognition by local governments, securing sufficient resources, bridging knowledge gaps, and simplifying the legal processes. By empowering Indigenous communities and supporting their rights and stewardship over their ancestral lands and resources, a more inclusive and sustainable approach to natural resource governance can be achieved. --- CONCLUSION Peatland serves as a valuable ecosystem for the Indigenous Dayak Ngaju community. Their adoption of a zonation system for peatland use, along with the integration of ecosystem management and cultural rituals into their farming practices, showcases their deep connection and harmonious relationship with nature. However, despite the laws in Indonesia that recognize the rights of Indigenous peoples and aim to 59 safeguard their customary lands, the implementation and enforcement of these laws have demonstrated weaknesses and inconsistencies. Furthermore, several programs and policies have failed to prioritize seeking the consent and opinions of the community, despite the evident impact on their traditional way of life. To address these issues, it is essential to urgently implement the principle of FPIC and enhance the implementation and enforcement of legal protections. FPIC ensures that decisions affecting Indigenous communities are made in consultation with them, respecting their rights and interests. Strengthening legal protections will further safeguard the rights and well-being of Indigenous communities, providing them with the necessary legal mechanisms to protect their customary lands and maintain their traditional way of life. To bridge the gap between government policies and indigenous knowledge, it is important to foster a legal framework that recognizes and integrates the utilization of the 'time and space' perspectives of the Indigenous Dayak people, while placing emphasis on the collaborative practice of handep among all stakeholders involved.

In Indonesia, the legal system heavily favors state ownership of land, leading to the marginalization of Indigenous peoples and their traditional land management practices. The prioritization of economic development over environmental and cultural conservation has resulted in a limited understanding of the value of the peatland ecosystem for Indigenous Dayak communities, leading to inappropriate and ineffective peatland management policies. To address these challenges, this research adopts a descriptive qualitative approach, utilizing a cross-sectional research design that includes in-depth interviews and literature study to gather and analyze data from Indigenous Dayak Ngaju communities in Tumbang Nusa and Pilang villages, Pulang Pisau regency, Central Kalimantan province. The study uncovers that the Indigenous Dayak Ngaju community has established a zonation system for peatland use, comprising separate areas for settlement, farming, and forest protection (Pukung Pahewan). The creation of specific policies for managing sacred areas is crucial to preserving Indigenous values and practices. Moreover, the absence of free, prior, and informed consent in certain policies and programs, such as the Mega Rice project, Food Estate program, and Zero-burning policy, has caused social conflicts within the Indigenous Dayak community, leading to the destruction of their livelihoods. Despite existing laws in Indonesia that acknowledge the rights of Indigenous peoples and safeguard their customary lands, the implementation and enforcement of these laws have proven weak and inconsistent.

Introduction Culture is an important element in human life because it provides the implied meaning of various aspects of society. The significance of culture is strongly related to values, beliefs, ways of thinking, ways of living, and world views adopted by community members at certain times (Eko & Putranto, 2019). Culture is not a thing that is only owned by a particular group of people but is owned by everyone and can be a unifier of the nation. Human beings and culture are never separable. In daily life, the human being is never detached from culture. As social beings, people interact with each other and follow habits that can become a culture (Mahdayeni et al., 2019). Pesurnay (2018) defines culture as an expression of the will of man in recognizable structures shared by those who inhabit the same world; therefore, the relationship between man and his cultural world is dynamic and dialectical. This concept informs a theory of social construction, including objectivity, internalization, and externalization. Pujiati (2017) argues that these three concepts outline a process of forming a tradition that goes hand in hand with that tradition itself and evolves continuously. The individual human being becomes an instrument involved in creating an objective social reality through a process of externalization, as the individual interprets influences through a process of internalization. Ngangi (2011) explains that this social construction can be dialectically illustrated as in Figure 1. Each process for the dialectical scheme of social construction is presented in Table 1. The human being is also seen as the creator of culture. Culture is closely related to tradition, which can be seen as formed through the community's continuous transmission of a culture. Tradition is a habit, behavior, or attitude of a society passed down from generation to generation and preserved by the local community as a reflection of that society with a distinctive culture. Tradition is a spirit of culture that strengthens a cultural system. Culture and all its products are the results of the process of human life (Suarmika, 2022). Local wisdom is a cultural product that includes philosophy, values, norms, ethics, rituals, beliefs, habits, customs, and so on (Uge et al.,2019). wisdom usually comes from ancestors, who are followed by community members from generation to generation (Gadeng et al., 2018;Atahau et al., 2020;Raj et al., 2022). This local wisdom accumulates the good habits of generations. --- Table 1. Theoretical Dialectical Scheme of Social Construction No. --- Dialectical Scheme Description 1 Externalization Human generosity is directed towards the world in mental and physical activities. It is sometimes seen as the essence of man himself, and it is an anthropological imperative that man always devotes himself to the world in which he exists. Humans cannot understand themselves as detached self-enclosed beings separated from the outside world. --- Objectivation The results have been achieved both mentally and physically through human externalization activities. The results confront the producer himself because they are outside and different from the humans who produce them. Through this process, the community becomes a sui generis reality. Objectivation can manifest as sharing opinions concerning a social product that emerges within a community through public discourse, even without direct, in-person interaction between individuals and the creators of said social product. 3 --- Internalization It can be argued that the human person is an instrument in the process of creating objective social reality through a process of externalization, as she or he influences it through a process of internalization which reflects subjective reality. Individuals become members of society through this process of internalization or socialization. Local wisdom passed down by tradition becomes the basis for someone in a particular tribe to communicate with other tribes. Habits and the local wisdom they form give rise to a tradition with its customs, norms, and other cultural forms (Pratamawaty, 2017). Local wisdom is a form of community culture in the form of knowledge, products, and activities used for survival adapted to where they are from generation to generation. Local wisdom has philosophical values believed to be guidelines (thoughts, attitudes, and behavior) in life activities to maintain personal and group survival (Suarmika, 2022). Each region has a different tradition of commemorating or celebrating important events such as births, weddings and deaths. One area of Indonesia with a unique tradition is the Lahat Regency in South Sumatra Province. One of the national tribes that inhabits this region is the Besemah Tribe which is different from the peoples inhabiting other areas in celebrating weddings. The Besemah Tribe community still upholds the tradition of pantauan bunting inherited from its ancestral beginnings, especially when wedding celebrations are held. This is evident from researchers' direct observations; many still carry on this tradition. Some research shows that the Pantauan Bunting Tradition involves a pair of brides who will be accompanied by a man and a woman who in the Besemah language are called 'bujang ngantat' and 'gadis ngantat'. A 'bujang ngantat' or 'gadis ngantat' must be unmarried. The task of the 'gadis ngantat' and 'bujang ngantat' is to join the bride in surrounding the house of the residents who have called the couple. They also accompany the bride and groom from the time of the marriage proposal until the wedding reception (Arios, 2019). The research results of (Sari et al., 2021) reveal several new things; the Pantauan Bunting tradition is still practiced by the Pasemah community, especially the Sukarami village community, in a series of marriage ceremony activities in the form of an invitation to eat from the local community to a newly married partner. The tradition of Pantauan Bunting is carried out to profit Muji Jurai or honor the descendants, as an act of gratitude and respect for their descendants because they are married, or it is also said to be a gift from the community to the bride. Zaman (2017) also reports on the uniqueness of marriage events in the community as exemplified in the performance of rituals such as the Rokat Tek-tek kemanten tradition which is imbued with institutionalized community values. These practices involve symbols that have sacred meanings in the Rokat tektek kemanten tradition. as it is an ancestral heritage of appreciating "bujuk nia" which determines social reality in the community as an institutional belief system. The process of social community formation occurs through simultaneous awareness and solidarity. --- Research Methods This study utilizes the qualitative method of ethnography to provide a detailed description of the social construction process involved in the Pantauan Bunting tradition within the Besemah Tribe Community. This research aims to shed light on how social constructs are formed by examining the stages of externalization, objectivation, and internalization. Additionally, the study will offer insights into the distribution, uniqueness, and prevalence of the Pantauan Bunting tradition specifically within Lahat Regency. Through this comprehensive analysis, a deeper understanding of the cultural significance and dynamics of the Pantauan Bunting tradition can be attained. --- Research Location This research study will be conducted in three zones selected by researchers based on the strength of the influence and distribution of the construction of the Pantauan Bunting tradition. In one zone, the tradition is still strong, in the second it is fading away, and in the third it no longer exists. Agung City Village is included in the zone that is still practicing the tradition. Pulau Pinang Village is included in the transition zone where the tradition is rarely practiced. Selawi Village is included in the zone where the Pantauan tradition is barely practiced anymore. The location of research is presented in Figure 2. --- Research Subject In qualitative research, the research subject, commonly referred to as 'the informant', is someone who provides information about the data to be studied. In this study, the subject of research is the Besemah tribe community in Lahat Regency. The Pasemah tribe, commonly called the Besemah Tribe, is one of the ethnic groups residing in the Province of South Sumatra, Indonesia. The majority of the people live in and around Mount Dempo, Pagaralam City, Lahat Regency, Empat Lawang Regency, and Muara Enim Regency. A small part of the tribe is spread across other districts. The subjects of this study were mainly people in Kota Agung Village, Pulau Pinang Village, and Selawi Village (Table 4). The persons selected by the researcher to be used as informants in this study include (1) traditional figures in Kota Agung Village, Pulau Pinang Village, and Selawi Village; (2) Native people who are tribal to the Besemah nation in Kota Agung Village, Pulau Pinang Village, and Selawi Village. This study used the snowball sampling model to gather information. The method is called'snowball' sampling because a researcher determines a person to be a sample based on the recommendations of people who have been a sample before (Vincent et al., 2022). The specifics of this research study's informants are presented in Table 2. --- Research Instrument The instrument used in this study is the researcher himself (Human Instrument) because this study uses a qualitative approach that must interact directly with the surrounding community and involve interview guidelines, observations, recording tools, and documentation tools (in the form of photographs taken at the time of the enactment of the Pantauan Bunting tradition). --- Data Analysis Data analysis is a crucial stage in the process of conducting scientific research because it allows the researcher to arrive at answers or to the problems which the research has. Qualitative data analysis involves the sorting, coding, and thematizing of data derived from the data collection process involving interviewing study participants, recording the interview and taking notes, and reviewing the literature. --- Results and Discussion --- Social Construction: the Process of Forming the Pantauan Bunting Tradition in the Besemah Tribe Community Regions in Indonesia have many local forms of wisdom, cultures, traditions, customs, languages, and rituals or ceremonies that differ from one area to another (Hilman et al., 2020). The cultural diversity of Indonesia consists of customs or traditions that develop in society into a distinctive view held by a particular community which is embodied in its acting as it behaves (Hasmika & Suhendro, 2021). In Indonesia, several traditions have been highlighted by research into traditions in particular locations. The Pararem custom in the mass marriage tradition found in Pengotan Village, Bangli Regency, is underpinned by the community's economic situation, cultural preservation, hereditary sustainability, and culturally-specific views of happiness (Gede et al., 2021). The tradition of Javanese customary marriage that prevails in Kalidadi Village is the wetonan custom, involving taboo understood as a form of caution, a way parents choose prospective partners for their children and protect their children's households from all the possible adverse effects that could befall them in the future (Ruslan et al., 2021). Susantin & Rijal (2021) highlight that the marriage tradition in Madura differs from that in Java. In Madura, the majority of the population adheres to the matrilineal tradition in which after marriage, husbands and wives are required to live in the wife's house. Before the wedding, the future husband carries BhenGibhen (cabinets, chairs, beds, and other household furniture) to the wife's house. The wife has a house to occupy. This contrasts with the ampa sabae tradition (request for marriage by women), which has developed over a long period in the Ambalawi community. The Ambalawi community understands this ancestral tradition as a solution to the problem which arises when women suffer detriment due to men's actions. It is a way for women to hold men accountable (Elpipit & Safitri, 2021). A unique case is found in the bakar batu (stone burning) tradition, a traditional ceremony enacted by the Dani tribe. It involves the event of cooking a dish made from several pigs. This dish is served as the main part of the meal. A stone that had been burned as a cooking medium. The Dani tribe continues this tradition of baking stones as a form of gratitude to God as an expression of joy or of sorrow. It is a regular part of big events such as celebrations of marriages or births or the final tribute to God on the occasion of someone's death or thanksgiving for the blessings of the harvest (Nipur et al., 2022). The tradition of marriage in the Bugis-Makassar tribal community has several long stages including determining the amount of panai money that the groom's family will hand over to the bride's family (Mustafa & Syahriani, 2020). In this tradition, it is not surprising to find marriage occurring at an early age as documented by (Yodi et al., 2020). These researchers' findings shed light on the meanings attributed to early marriage customs in Nagari Tapan, Basa Ampek Balai Tapan District. Parents see early marriage as: 1) a way of avoiding shame, 2) an economic matter, 3) a rescue effort. A similar diversity of marriage traditions occurs in the Besemah tribe community. The marriage contract is usually signed the day before the wedding celebration. However, some people do this after the wedding. This is in accordance with the agreement of the community with the bride and groom when they will hold the Pantauan wedding festivity. Several processes are involved in keeping the tradition. They are presented in Table 3. The tradition of Pantauan Bunting continues to be carried out by the community, whether it is the Besemah tribe or even people outside the tribe. There are no traditional sanctions or customs about what other customs must be followed if people do not carry out the Pantauan Bunting tradition. However, the Pantauan Bunting Tradition uses a system of reciprocity. If we hold the traditional Pantauan Bunting celebration, relatives are going to be married, then the relatives will reciprocate by doing pantauan for us if we one day get married. Likewise, if we do not hold aPantauan Bunting celebration, other people will do the same thing and not have the celebration. This principle of reciprocity is also important in customary law in the Basemah community itself. Wardani (2021) argues that customary law is a living law that manifests itself as a community habit. This marital custom law could thus be considered an inseparable part of the human body in the community. Anyone going to get married informs the community beforehand by inviting the residents in the surrounding areas by visiting their houses (in the Besemah language 'besuare'). This is usually done two to three weeks before the wedding. The bride and groom who will be married usually first inform their families or close relatives. Thus, The community knows there will be a wedding so that they can prepare everything needed for Pantauan, including obtaining food and other necessities from far away. At this time, the inviting family will also bring 'lemang' (lahat specialty food containing glutinous rice cooked in bamboo) to people in the community. The people who receive the 'lemang' understand that it signifies they are obliged to hold a Pantauan Bunting celebration. 2 Bemasak After being notified by the bride's family of the date of wedding, the community will prepare everything necessary for pantauan. The most prominent feature in the Pantauan Bunting Tradition is the presence of various dishes ranging from snacks such as cakes and fruits to more substantial food such as rice and side dishes. For this reason, the community will usually make or cook food to be served when the Pantauan Bunting celebration is held. Usually the day before the pantauan, the community will have prepared a dish that will be served during the pantauan. Some people even start gradually making food a week before the pantauan. There are no stipulations regarding what food should be available and served during the pantauan. However, some dishes are almost always part of the Pantauan Bunting tradition, lemang, dodol and pepes ikan, a typical food from the Besemah tribe. In addition to these two types of food, there is also always meat, but it is only served by close family members. This symbolizes that the family conducting the Pantauan Bunting celebration still has a blood relationship with the bride. --- Mantau Bunting The day before the wedding celebration, Mantau Bunting or calling the bride is done. After the marriage contract is signed, the community usually calls the bride and invite the bride to visit their homes. The Pantauan Bunting feast can also be held after the wedding celebration is over because there will not be enough time for the bride and groom to visit the homes of the residents of the surrounding areas in a single day. For this reason, before the pantauan begins, the community will agree when they will hold the traditional Pantauan Bunting celebration. Source: Field Research Results (2022) In Social Construction Theory, something can be formed due to the dialectic of externalization (the adjustment of individuals to their environment). This involves objectivity (individuals are aware that they are part of society) and internalization (individuals create a social reality in everyday life) (Susanto et al., 2020). The following is an explanation of the three social processes associated with this Pantauan Bunting Tradition. Based on this explanation of the dialectical process of Pantauan Bunting, the researcher sees it as having component parts as shown in Figure 3. --- Distribution of Pantauan Bunting Traditions in the Besemah Tribe Community Pasemah Tribe, commonly called the Besemah Tribe: The origins of the name 'Besemah' for this community are believed to be derived from the name of a fish that was formerly found in the Pagaralam area of South Sumatra Province. The Semah fish is a type of goldfish that lives in murky streams among rocks that are overgrown with moss and shaded by trees. However, many people call the Besemah Tribe the Pasemah Tribe. The name 'Pasemah' stopped being used because Dutch colonists found it difficult to pronounce the phoneme "pa" and pronounce it "be", so the name "Pasemah" became "besemah" (Asrin et al., 2016;Refisrul, 2019). For the purposes of government authority and administration, the Besemah cultural area includes Pagaralam City, Lahat Regency, Empat Lawang Regency, Muara Enim Regency, and South Ogan Komering Ulus Regency in South Sumatra Province. Lahat Regency, specifically, includes the area of Jarai District, Tanjung Sakti District, and the area around Kota Agung District. The Besemah cultural area in Bengkulu Province includes Kaur Regency, Seluma Regency, and South Bengkulu Regency. Specifically, Kaur Regency includes Padang Guci Hulu District and Padang Guci Hilir District. Besemah culture is also found in Lampung Province, namely in South Lampung Regency. The spread of Besemah culture to various areas outside Pagaralam City was followed by changes and the formation of new cultural identities in these different regions but they all still recognize that their origin lies in Pagaralam (Arios, 2019;Asrin et al., 2016). The Besemah Tribe is one of the tribes that inhabit Lahat Regency and its surroundings. The Besemah people are scattered in all areas both inside and outside the Province of South Sumatra. Thus, the Pantauan Bunting Tradition is not only to be found in one place in South Sumatra, but elsewhere too. In this study, researchers selected three different villages in Lahat Regency as a research location: Kota Agung Village, Pulau Pinang Village, and Selawi Village. Based on the results of the study, the researchers can say that Pantauan Bunting Tradition is still alive in these three areas. However, the numbers of members of the Besemah tribe differ in these areas. In Kota Agung Village, the Besemah tribe community is still widespread, and it can even be said that the majority of people in this village are from the Besemah tribe community. In Penang Village, by contrast, it is fairly uncommon to find people from the Besemah tribe because the majority of people in this village are from the Gumai Lembak tribe. Finally, in Selawi Village it is very difficult to find Besemah people because Selawi Village is located very close to the city center where there has been singificant mixing of tribes and cultures. The results of this research study also show that several factors cause the distribution of the Besemah tribe in Lahat Regency to vary from area to area. The first factor is physical factors that include location and distance. Kota Agung Village is quite far from the city center makes it difficult for people to move to other inhabited places. Therefore, the people of the Besemah Tribe tend to live in the village of Kota Agung. On the other hand, in Pulau Pinang Village, members of the Besemah community are rare. This is because Penang Village is a transition area between the village and the city center, so it is quite easy for people to move from one place to another. The concentration of the Besemah tribe in Selawi Village is almost the same as that in Penang Village; it is quite difficult to find people from the Besemah tribe because it is located very close to the city center. Social factors also affect differences in the distribution of the Besemah tribe in Lahat Regency. Social factors come from society itself. As explained above, in Kota Agung Village, the majority of the people are from the Besemah tribe. The people in Kota Agung Village are still classified as a traditional community. People still adhere to the existing traditions. The closed attitude of the community makes it difficult for this village to accept anything new that comes from outside and it tends to maintain its original culture. Therefore, if we visited Kota Agung Village, the Pantauan Bunting Tradition would still be very easy to find. The next research area is Pulau Pinang Village. The distribution of members of the Besemah tribe in Pulau Pinang Village is sparse because Pulau Pinang Village is a transition area between the countryside and the city center so there has been a lot of cultural mixing. The indigenous people in Penang are not the Besemah tribe, but the Gumai Lembak tribe; the ancestors who lived in this village came from the Gumai Lembak tribe. Most people now in Penang Village are not from the Besemah tribe, but the Gumai Lembak tribe. Researchers also find it difficult to interview informants about the Pantauan Bunting tradition. The Besemah tribe is only a minority and its members are immigrants. However, we can still see the Pantauan Bunting tradition in the Pulau Pinang Village. The Besemah people in this village still continue the tradition even though they are only few in number and are not the village's original inhabitants. In Selawi Village, members of the Besemah community are few and difficult to find. This is because this village has seen significant cultural mixing. This phenomenon cannot be avoided because Selawi Village is located very close to the city center, so many people from different ethnic groups migrate to this village on account of work, education, or marriage. In Selawi Village many people have moved with the times and have weddings with a modern vibe. This is also what makes the existence of the Pantauan Bunting tradition in this village difficult to find. The community has left many traditions behind that have been considered ancient. Social factors also affect the distribution of the Pantauan Bunting Tradition. One is amalgamation, the marriage of members of different tribes. Amalgamation can lead to assimilation and acculturation. Assimilation is a meeting between two cultures that brings about a new culture and replaces and erases the old culture. Acculturation, conversely, is the meeting of two different cultures that creates a new culture but does not eliminate or abandon the old one. This phenomenon does not only occur in the Basemah tribe. Several studies show the same process in different locations and involving different populations such as the amalgamation of Chinese and Madura ethnic groups in Bangkalan Madura Regency (Rahmatina & Hidayat, 2021). Another example is the marriage amalgamation of Batak and Malay ethnic groups in Pangkalpinang City (Siagian et al., 2021). Ethnic differences between Flores and Chinese in Trubus Village, Central Bangka Regency have likewise been amalgamated through marriage (Aprilia, 2021). A final example is that of Chinese amalgamated through marriage with members of indigenous peoples in Java (Winarni, 2017). It depends on the community itself to assimilate or acculturate in a particular cultural environment. This is also the case with the Pantauan Bunting tradition. Society chooses to maintain or abandon this tradition in daily life. Every region in Java has an amalgamated community, with the exception of the three villages that the researchers chose, Kota Agung Village, Penang Village, and Selawi Village. Based on the results of field observations, the researchers can say that although there have been many amalgamations in Kota Agung and Pulau Pinang village, the community still maintains the traditional culture. Even if there is an amalgamation, other tribes follow the customs or traditions of the Besemah tribe. On the other hand, in Selawi Village, because there have been many assimilations and influences due to the location of the village very close to the city center, most people abandon existing traditions and keep up with the times slowly. The distribution of the Pantauan Bunting Tradition among the three villages exhibits variations, with some villages still having easy access to it, while others find it more challenging or even extremely rare to come across. These disparities can be attributed to the factors mentioned earlier. Despite these differences, the Besemah people strive to uphold and sustain this tradition. They view the Pantauan Bunting as an integral part of their lives that holds significant cultural value, hence the community's commitment to its preservation and continuation. --- The Existence of the Pantauan Bunting Tradition in the Besemah Tribal Community in Lahat Regency The Pantauan Bunting tradition is one of the customs which is distinctively that of the Besemah tribe. People will invite a bride to come to their house. The community will entertain the bride and groom, offering them various dishes ranging from snacks to heavy food. Pantauan Bunting is a tradition constructed by the Besemah people in ancient times and still exists today even though it is not as strong as it once was. The tradition is a special form of human interaction response with its living environment. The formation of an environment is determined by several factors, one of which is the local community's culture (Astuti et al., 2021;Fitriana, 2018;Musafiri et al., 2016). The tradition of Pantauan Bunting is still found throughout all areas of Lahat District. Although in the modern era people compete to have wedding celebrations r with a modern feel and leave behind traditions considered old-fashioned, the Besemah community still continues the Pantauan Bunting tradition. These forms of kinship solidarity include the role of parents, routine social activities, feelings of energetic life, and traditional Javanese customs and rules of thumb that still apply to everyday life. The existence of Pantauan Bunting Tradition has many benefits for the community. In this study, the researchers selected 3 villages: Kota Agung Village, Pulau Pinang Village, and Selawi Village. The three villages were selected based on the difference in distance from the city. One village is far from the city center, namely Kota Agung Village, one village is in the transition area between the village and the city center, namely Pulau Pinang Village, and one village is close to the city center, Selawi Village. With these differences in distance from the city center, the researchers can compare the villages to see whether villages that are located within a close proximity to the city center and villages that are located long distances from the city center exhibit differences in their levels and methods of maintaining the existence of the Pantauan Bunting tradition. The numbers of people living in each area of Lahat District (2022) number are shown in Table 4. The Besemah tribe still maintains the Pantauan Bunting tradition. In Kota Agung, a majority the people are from the Besemah tribe. They still maintain the practices of the Pantauan Bunting tradition. By contrast, in Penang Village, the majority of the people are from the Gumai Lembak tribe, while people from the Besemah tribe form a minority. Nevertheless, the Besemah community in this village still carries on with the Pantauan Bunting tradition. In Selawi Village, located very close to the city center, significant cultural mixing has caused people to abandon their culture and traditions. The community has begun to be diverse and it is very rare to find people from the Besemah tribe who still maintain the Pantauan Bunting tradition. We therefore know that the villages' different locations and distances between the three villages and the city center impacts, but never quite eliminates, the Pantauan Bunting tradition. Many still hold to their traditions firmly and ensure that this tradition does not fade or disappear even in an advanced modern era. --- Conclusion The Pantauan Bunting tradition originates in the Besemah tribal community. It involves occasions when a couple in the community is to be marred. The bride-to-be is invited to come to the community members' houses in the village where the community will provide various dishes ranging from snacks to heavy food. This tradition is usually done the day before the wedding celebration is held. In doing Pantauan Bunting tradition, a bride-to-be will be accompanied by a single man and woman (bujang ngantat and gadis ngantat). The Pantauan Bunting Tradition is widespread in various areas across Lahat Regency. Several factors cause Pantauan Bunting Tradition in Lahat Regency to differ in each area where it is kept. Physical factors include the village's location and its distance from a city center. Social factors come from the community itself. In this study, we find that Kota Agung Village is a village where we can easily see the Pantauan Bunting tradition. This is because the majority of Kota Agung village community members are still classified as traditional and closed to something new. In addition, the people in this village have not intermingled much from other ethnic groups so the purity of the indigenous Besemah tribe in this village is still maintained. Whereas in Pulau Pinang Village and Selawi Village, the tradition has been difficult to find because this area has become a transition area between village and city where there have been many cultural confluences caused by amalgamation. Despite the advancing and modern times, the existence of Pantauan Bunting tradition continues to be preserved by the Besemah people in Lahat Regency. --- Author Contributions

Indonesia is a multi-cultural country characterised by hereditary traditions passed down by ancestors. Strands of this traditional culture are often specific to particular communities, for example the Pantuan Bunting tradition is expressed and passed down in the customs of the Besemah community of Lahat Regency, South Sumatra. The purpose of this study is to analyze the social construction of the process of the formation of Pantauan Bunting tradition, the distribution of Pantauan Bunting tradition in different regions, and the existence of Pantauan Bunting tradition in Besemah Tribe community in Lahat Regency. The method used in this study was qualitative with ethnographic, historiographic, and spatial approaches. This research was conducted in three locations, namely in Kota Agung Village, Penang Village, and Selawi Village where the research subjects consisted of traditional leaders, religious leaders, community leaders, and the Besemah community. The results of this study showed that (1) the Besemah community constructs Pantauan Bunting tradition since the time of its earliest ancestors and this process of transmission still continues. This tradition is characterised by a public invitation to prospective brides to come to their prospective bridegrooms' homes. (2) Pantauan Bunting Tradition has spread to various areas in Lahat Regency, such as Kota Agung Village, Pulau Pinang Village, and Selawi Village. (3) in the modern era, the existence of Pantauan Bunting tradition is maintained by the Besemah community, and we can still find it in various areas in Lahat Regency. The Pantauan Bunting tradition, practised in various parts of Lahat Regency since ancient times and firmly ingrained in the Besemah community, highlights the community's resiliency and commitment to preserving its cultural legacy.

Introduction Migration is a global phenomenon with 272 million international migrants in 2019; more than half are migrant workers and most are from low-and middle-income countries (LMICs) (International Organization for Migration[IOM], 2020). With over half of the households having a current or returnee migrant, migration is common in Nepal (International Organization for Migration [IOM], 2018) and it is increasing trend (Government of Nepal [GoN], 2020). Labor migrants from Nepal are predominantly men (GoN, 2020;Sharma et al., 2014) with most migrating to work in India, Malaysia, and the six countries of Gulf Co-operation Council (GCC) (GoN, 2020). Due to lack of money (Telve, 2019), stringent migration laws in host countries (Acedera & Yeoh, 2020) and the need to leave someone in charge of assets at home (Lu, 2012), most labor migrants from LMICs, emigrate to destination countries alone. Leaving spouses/family behind, which has resulted in an increasing number of transnational families globally (Démurger, 2015;Lu, 2012) and in Nepal (Lokshin & Glinskaya, 2009). There is growing literature on the sexual harassment of female migrants, for example in China (Zong et al., 2017), but left-behind family members of labor migrants (Acedera & Yeoh, 2020;Shattuck et al., 2019), non-migrating spouses remain relatively understudied (Archambault, 2020;Fernández-Sánchez et al., 2020) particularly in Nepal (Maharjan et al., 2012). Prolonged absence of their migrant partner affects spouses in several ways. While leftbehind men with migrant wives as breadwinners of the family are also affected (Acedera & Yeoh, 2019;Elizabeth et al., 2020;Hoang & Yeoh, 2011) in patriarchal settings, left-behind women are more affected since women generally have a lower status than men and they are considered to be dependent on men (Fernández-Sánchez et al., 2020). While non-migrating women receive financial support in the form of remittances, their husbands' absence often creates both social and personal problems (Kunwar, 2015). The societal perception is that women in the absence of their partners are vulnerable to sexual violence/abuse/harassment (Krug et al., 2002). The World Health Organization (WHO) identifies sexual violence as a serious problem with short-and long-term consequences on women's physical, mental, and sexual and reproductive health (WHO, 2021). The WHO defines sexual violence as: "any sexual act, attempt to obtain a sexual act, unwanted sexual comments or advances, or acts to traffic, or otherwise directed, against a person's sexuality using coercion, by any person regardless of their relationship to the victim, in any setting, including but not limited to home and work" (Krug et al., 2002). Studies have looked at sexual harassment experienced by female migrant workers in the carpet industry in Kathmandu Valley (Puri & Cleland, 2007), or internal migrants in Chine (Hu et al., 2022). Other studies reported that many women experienced sexual abuse/harassment in the absence of their migrating husbands (Ahmed, 2020;Kamal, 2019). This study aimed to explore the experiences of non-migrating female spouses of Nepali international migrant workers on sexual harassment/abuse and its impact on their mental health. --- Methods and Materials This study employed a general qualitative approach using in-depth interviews. Women were recruited using snowball sampling (Green & Thorogood, 2018). The initial participant, whose husband was working abroad, was chosen for an interview. She then introduced additional participants, and subsequent selections were made based on the subsequent participants. Owing to the ongoing coronavirus pandemic, 14 in-depth interviews (Van Teijlingen & Forrest, 2004) were conducted online with wives of international migrants residing within the Kathmandu Valley via Facebook messenger and Viber by a female interviewer. A semi-structured interview guide was developed, participants were informed about the study objectives, confidentiality and anonymity, and voluntary participation and verbal consent was taken from all prior to the interview. Interviews were conducted in Nepali language. Since data were collected virtually, participants requested not to be recorded. Therefore, the first author, who conducted all the interviews, took notes and after each interview checked these with the participants. All interviews were then transcribed and translated into English language for analysis. After translation, the first author manually coded the transcripts using thematic analysis (Green & Thorogood, 2018). Participants' verbatim quotes formed a major part of the findings. The names of participants have been changed to maintain anonymity. Ethical approval was sought from the Nepal Health Research Council (Ref: 163/2019) and informed consent was obtained from all participants prior to the interviews. --- Results --- Participants' Characteristics Most women (86%) were identifying themselves as Hindus. With 64% having higher secondary education or above, most participants were better educated than the average women in Nepal. Most were engaged in some form of employment, such as government or privatesector jobs (57%) or business (14%), while 21% identified themselves as housewives. Twothirds of participants (36%) were under 30 years old, 43% were between 30-40 years old and 21% were over 40 years old (Table 1). Sexual Harassment Among Nepali Non-Migrating Female Partners --- Participants' Circumstantial Characteristics Most (57%) reported their husbands were working in GCC (Gulf Corporation Council) countries and participants were living in joint family and 57% resided in joint families. Most lived in rented accommodations (Annex 1). Most participants living in joint families shared that the relationship with their in-laws was unpleasant which caused them stress, for example: My relatives told negative things about my character to my mother-in-law. Things like I am very beautiful, and I might have extramarital affairs with my colleagues and asked her to take care of me. After that my mother-in-law started to doubt my character and threatened and scolded me. My job is to collect money from the market so sometimes I am late to come home... [Pooja (pseudo name), 27 years] Another woman explained why she had to stop living with her husband's family: I lost my baby girl three days after birth. After that, my father-in-law and mother-in-law started to behave very badly with me. They used to put cold water and excess salt in my dal (lentil soup). They started to scold me and tortured me so much... I used to get to college in tears... They took my husband's income; we always provided them money... It's been a few months I started living separately, otherwise I would have died. [Radha, 44 years] Thus, many migrant wives lived in challenging circumstances in their husbands' absence. Participants elaborated that their husbands had to work abroad because they were unable to get good jobs in Nepal. However, participants reported that their husbands loved them, send them remittances and communicated with them regularly. Owing to increased mobile ownership and the Internet availability in Nepal, women communicated with their husbands regularly using Viber and Messenger and phone calls. --- Experiences of Harassment Most of the participants reported experiencing some form of harassment at some point in their lives, five had experienced sexual harassment by men who knew their husbands were abroad. They reported that physical harassment in the form of touching/pinching; verbal abuse wherein the perpetrator used vulgar words and repeated propositions for dating. Participants shared that as women they faced similar but less severe forms of harassment commonly in their lives. Two participants were sexually harassed by their teachers....while writing my master's thesis, I was sexually harassed by a teacher... He was the head of that department and he asked me to date him and have sexual relations with him. He told me to visit him at a hotel in Nagarkot where he would check my thesis. Before this, he used to touch my body parts as if unknowingly and he told me that if I needed anything, he would fulfill my demands. I used to ignore all this... then he created problems in my thesis correction and viva process and stopped everything (stopped the natural progression of thesis correction and finalization). After some time, his head of department post was terminated, only then I could submit my thesis for correction and viva... through the new head of department. [(Bindu,27 years] Another woman reported similar abuse from her academic tutor: Touching/pinching, using vulgar words... these are very common in my life which I have been facing in the streets and public buses. But I had such bitter experience of sexual harassment from my teacher, I cannot forget. I was a science student. I passed my master's degree with a first division. However, I was sexually harassed severely and abused by my thesis guide (supervisor). He asked me to date him and meet him in different locations and used to touch my private parts. Because of his behavior, I felt scared, and depressed as well. I had decided to give up my master's degree. But, due to my sister's encouragement, I could continue my study. I changed my thesis guide and finally I completed my thesis. [(Sushila, 27 years)] Two women reported being harassed by work colleagues. Nepal's workplace follows a strict hierarchical structure wherein women facing harassment may find it difficult to challenge their seniors, for example: In our organization, there is a hard and strong chain of command and generally, people with a lower profile cannot oppose their seniors. In this case, most harassment cases might be overlooked or kept secret. In real working life, I have also faced some bitter experiences of sexual harassment in my career from our seniors and male friends like touching, body brushing, use of vulgar words, sexual messages via social media, and propositions for dating also. Some lady friends have shared to me their bitter experiences about sexual harassment from seniors like the request for sexual relations too. [(Suman, 28 years)] One participant, 34-year-old Sita (pseudo name), was sexually harassed by her brother-inlaw: My husband has been in Dubai for the past 10 years...My sister-in-law and brother-in-law also live with us. I feel unsafe from my brother-in-law... I am experiencing domestic violence from my mother-in-law and sister-in-law, and sexual harassment like verbal harassment... and touching and brushing body parts (from brother-in-law)... The provided text highlights the disturbing reality faced by women living alone while their husbands work abroad. These women encounter various forms of sexual harassment from multiple sources, such as their teachers, office colleagues, and even their own family members. Many migrant wives, particularly the ones who experienced harassment, shared two main reasons why they thought migrant wives faced harassment. First, men's mindset that migrant wives in the absence of their husbands are easy targets as they are alone and may need sexual partners. Second, the patriarchal societal norms where victim-blaming and shaming women as being immoral instead of accusing men who make the advances. It is very common in our life, people (men) think that married women need sexual pleasure, therefore, wives of migrant men who are alone in Nepal might need a sexual partner. With such a wrong ideology, men are motivated to sexually harass wives of migrant men. Such sexual harassment might be less severe type (touching, teasing) to extremely severe type (force dating and abuse)... because of my beauty, I have faced different types of sexual harassment like propositions for dating and sexual relationship too. Once my senior in the office proposed me to marry him. [(Pooja, 27 years)] Many participants mentioned that weak laws on sexual harassment and poor implementation of existing laws, as offenders do not get punished, encourage perpetrators of harassment, and discourage victims to speak up. Thus, our participants who had experienced harassment, did/could not report what was happening to them, despite being relatively empowered in terms of education and employment. --- Impact of Harassment Participants reported facing several negative effects due to sexual harassment including irritation, fear, frustration, humiliation, and depression. They reported that in addition to their mental health, their academic/professional life was also affected as they could not perform their jobs well. I felt mental and professional stress due to sexual harassment. I was afraid, stressed, and depressed while my thesis supervisor was torturing and harassing me sexually. I was not able to complete my thesis. At that point, I was thinking of leaving my studies due to the mental pressure.... At that time, I was very stressed, depressed, and humiliated, because of my thesis guide... [Bindu,27 years] It also negatively affected how they saw their employing organization: Yes, I feel humiliated, tortured, flustered, due to the sexual harassment at work from friends and seniors. Simple types of sexually harassing behaviors are common and we also take it lightly but some cases of sexual harassment are hard to forget and makes us nervous always (for fear it will happen again). It makes me lose respect for my organization and causes mental stress. [Suman,28 years] Depending on available support network and life circumstances, sexual harassment may add to already complicated life for some migrant wives. Such was the case for 31-year-old Dhanmati: After the death of my husband, people falsely blamed me for it. Family members as well as the society, and colleagues tried to harass me sexually... that made me depressed and scared... I thought of committing suicide, but at the same time, I thought about my child and my responsibilities toward him... to protect me from such a bad environment, I took support from a kind male friend (her current boyfriend). Now he is helping me to solve my problem as a good friend and I am feeling quite safe now. [Dhanmati,31 years] Even in such challenging circumstances, Dhanmati was still justifying entering another relationship. This reflects the strict moral values that women are subjected to in patriarchal societies: My family members and my late husband used to blame me as characterless woman, I ignored it because I was not like what they thought of me. After my husband, I have to depend on some other person for my safety and needs. He is just a good friend. But in society sometimes I tell other people that he is my husband to protect me from other sexually deprived persons (sexual predators). (Dhanmati, 31 years) Therefore, the issue of sexual harassment has led to challenging circumstances for the wives of immigrant male spouses, prompting them to employ various strategies and approaches to shield themselves from such negative behaviors. --- Suggestions to Mitigate Sexual Harassment When asked what could be done to prevent such incidents, most migrant wives' suggestions were these. First, that the society should take responsibility and create a safe space for migrant women in the absence of their husbands. Instead of blaming and shaming migrant wives for being harassed, and subjecting those to doubts of immoral behaviors and affairs, a culture of respect should be developed. Second, the government and judiciary should formulate and implement strong laws against sexual harassment. Also, the judiciary should ensure the legal procedures are victim-friendly. Finally, awareness should be created regarding sexual harassment prevention and reporting to create a safe environment for women in society. --- Discussion In this study, we explored the experiences of sexual harassment of non-migrating wives of international labor migrants from Nepal and participants were found suffered from several types of harassment such as touching/pinching, verbal abuse and repeated requests for dating. Sexual harassment is particularly common in patriarchal societies (Berman et al., 2000;Foulis & McCabe, 1997). Participants also noted that they were vulnerable to sexual abuse/harassment from men because of the patriarchal society in Nepal which places women in a subordinate position. Less severe forms of sexual harassment such as name-calling, wolfwhistling, and other forms of public sexual harassment are extremely common in Nepal (Kunwar et al., 2014). All five women suffered from sexual/abuse harassment at the hands of people they knew and who knew their husbands were away. Some studies also found migrant wives being sexually harassed by people they knew and who knew their husbands were away including male relatives and friends of their husbands (Ahmed, 2020;Kamal, 2019). Thus, we argue that the societal, particularly the perpetrators', mindset is such that non-migrating wives can be harassed as her husband is not there to protect her, and also because she may be willing to engage in an affair. Sexual harassment in academic institutions (Dunne et al., 2004;United Nations Children Fund[UNICEF], 2016) and workplaces (Kunwar et al., 2014) is common in Nepal and that in many cases perpetrators are people the women know (Puri & Cleland, 2007;Thapalia et al., 2020). Although most of our participants were highly educated and employed, they did not mention reporting or taking an active approach to dealing with harassment. They reported being very disturbed by it, so much so that their studies/work was suffering. This is consistent with existing literature where women tend not to report sexual violence (Krug et al., 2002;Puri et al., 2011) to authorities because they feel ashamed and afraid of being blamed or mistreated (Krug et al., 2002). Many participants mentioned that their in-laws, neighbors, and people in their community generally doubted their loyalty towards their husbands and in some cases even accused them of infidelity. Similar to our finding, a qualitative study among migrant wives in rural Bangladesh (Kamal, 2020) found that some participants experienced sexual harassment from the men in the family or community, but did not report or take any action for fear that if they revealed being harassed, the society would blame them for not being'respectable' (Kamal, 2020). Participants reported facing several negative effects due to sexual harassment including irritation, fear, frustration, humiliation, shame, and depression. They reported that in addition to their mental health, their academic/professional life was also affected as they could not perform their jobs well. According to McLaughlin et al. (2017), women suffering from sexual harassment are more likely to quit jobs leading to financial stress. Already, these women were vulnerable as they were struggling with finances, missing their husbands, and adjusting with their husbands' family and life without the intimacy and support of their husbands. The additional stress of being sexually harassed further complicated their personal and even academic/professional lives. Further, except for a few women who sought help from their close family members/friends, women were unable to seek help. Berman et al. (2000) observed that girls and women who experienced sexual harassment reported feeling afraid, intimidated, belittled and a decreased sense of confidence. With no studies to date focusing on sexual abuse/harassment experiences of non-migrating wives, our study fills an important gap in literature. It highlights the urgency to focus on the protection and the needs of the non-migrating wives. However, our study's qualitative nature is a potential weakness, because we only interviewed women residing in the Kathmandu Valley who were recruiting using snowball sampling, hence the generalizability of our findings is very low. We suggest further qualitative exploratory work on non-migrating wives in different parts of Nepal including in rural/urban and mountain/hill/terai (plain) areas. We also recommend quantitative surveys to produce more generalizable findings. --- Conclusion The findings of this study show that sexual abuse/harassment is common among the nonmigrating wives and that the perpetrators are usually men they know and men who know their husbands are away. None of the women experiencing harassment reported/took any action against the perpetrator due to fear of self-blaming and hopeless to get justice in society. Patriarchal social norms where family and community view non-migrating wives dubiously and question their loyalty to their husbands, and who perceive women as weak in the absence of their husbands encourage perpetrators to commit such acts and discourage victims to report such acts. Our recommendations include that we need to develop a culture of respect. Secondly, the government should initiate support groups for migrant wives locally where they can discuss their problems and support each other. Thirdly, the government should encourage women suffering from harassment to report the act(s) and the judiciary should formulate and implement strong laws against sexual harassment. Again, the judiciary should ensure the legal procedures are victim-friendly. Fourthly, psychosocial counselling should be made available at the local level wherein women seeking help could go. Finally, awareness should be created in the society regarding sexual harassment prevention and reporting to create a safe environment for women in the society. --- Role of Authors PS and EvT: Designed the study. KG: Conducted, collected, and analyzed the data. KG: Wrote the first draft. KG and SM: Wrote the second draft with advice from PS, EvT and RCS. All authors reviewed the manuscript and provided critical feedback and suggestions. --- Conflict of Interest The authors declare that they have no competing interests.

International migration shows an increasing trend around the world. The majority of labor migrants, particularly low/semi-skilled migrants from low-and middle-income countries, immigrate to destination countries leaving their family members behind, leading to an increasing number of transnational families. While non-migrating spouses often receive financial support in the form of remittances, their husbands' migration also creates numerous social and personal problems. This general qualitative study aimed to explore non-migrating spouses' experience of sexual harassment/abuse and its impact on their mental health. Fourteen in-depth interviews were conducted to collect data. Participants reported experiencing harassment by men they knew, including their teachers and colleagues, who knew their husbands were abroad. None of the women reported taking any action against the perpetrators. Policy level changes to spread awareness on sexual harassment, encouraging victims to report such acts, and establishing and implementing appropriate laws are essential to mitigate this serious problem.

Introduction As internet penetration becomes more extensive, the range of purposes for which it has been employed has equally increased. Some of these purposes contain the potential to educate and improve citizen well-being in ways that were previously not possible. This is particularly evident in the area of health. For example, online health communities (OHCs) enable individuals to interact with others who share similar health concerns in order to learn from their experiences and gain useful advice (Eysenbach et al., 2004;Hajli et al., 2014) and to reciprocate by sharing health information that is frequently based on personal experience (Ziebland & Wyke, 2012). Patients and those who support their care can use these networks to expand their understanding of diseases, treatments, or recommended healthy practices (Goonawardene & Tan, 2013;Ram et al., 2008;Rupert et al., 2016). They can source information about many aspects of medical conditions or concerns, making the issue seem less complex and more manageable. These networks also enable them to receive much-needed psychological support (Yan & Tan, 2014). This is particularly salient since a lack of informational and psychological support is consistently highlighted by those with serious illness and their caregivers (Luszczynska et al., 2013). OHCs can also increase inclusion by providing a supportive environment for those who may not be able to access health information easily due to location or socially stigmatized conditions and associated privacy sensitivities (Still, 2008), enabling them to overcome spatial or temporal limitations (Fan et al., 2014). For these reasons, OHCs are a valuable resource for expanding the understanding of medical conditions, treatments, or recommended healthy practices (Goonawardene & Tan, 2013;Ram et al., 2008), empowering patients to become more informed about how to self-manage their conditions and take an active role in their treatment, thereby improving clinical outcomes. In this way, these communities also have the potential to contribute to preventive healthcare (Goh et al., 2016), something that in a context of changing demographics and strained healthcare systems (England & Azzopardi-Muscat, 2017) has assumed greater social and economic significance. Notwithstanding the potential value of OHCs, research has shown that engagement in online health communities is highly variable-in some cases, as few as 1% of members contribute up to 75% of information (Carron-Arthur, 2014, Van Mierlo, 2014). The underpinning reasons for this appear to be trust related. For example, a recent survey found that only 4% of those surveyed said that they trust the health and medical information available on social media, 5% reported believing what they read on discussion forums, and only 15% stated that they trust information available on health websites (IPSOS MRBI and MSD, 2019). That deficit of trust is critical, as it limits individuals' engagement with OHCs and the positive potential contained therein. While research has begun to identify the factors that may increase an individual's trust in an OHC (Fan & Lederman, 2018;Fan et al., 2014;Fan et al., 2010), research on how those same factors influence OHC engagement remains limited. This gap in understanding has important implications as it limits the potential of these communities to support health self-management and improved health outcomes. This research addresses this deficiency in a number of distinctive ways. First, it advances our contextual understanding of trust generation in OHCs, illustrating how trust antecedents influence engagement in OHCs and, through engagement, influence knowledge contribution and knowledge adoption. We theorize that three trust-related antecedents influence member engagement and behavioral trust (knowledge adoption and knowledge contribution). By examining this relationship and its formation pathways, our findings yield important implications for research and practice, providing insight into how more engaged membership of these communities and their associated positive outcomes can be supported and maintained. The fact that the majority of online community users are lurkers who do not participate by contributing or adopting knowledge (Amichai-Hamburger et al., 2016;Rafaeli et al., 2004;Sun et al., 2014) has amplified the need to understand how the trustworthiness of online health environments can be more effectively developed in order to increase the active participation of their members and accelerate the realization of these communities' empowering benefits. This study advances this understanding and is therefore not just interesting but important (Tihanyi, 2020). Second, our focus on OHCs complements the existing literature. For example, much attention has been paid to trust in online transactional contexts (e.g., Connolly & Bannister, 2007;Fang et al., 2014;Gefen et al., 2003;Lee et al., 2011;Pavlou & Gefen, 2004) and, to a lesser extent, to trust in general online social networks (e.g., Grabner-Kräuter & Bitter, 2015;Matook et al., 2015). While this is valuable, the findings of these studies are bounded to those contexts and research focusing on trust in an OHC context remains limited. Moreover, what does exist varies considerably in focus, ranging from examinations of cognitive and affective trust development mechanisms (Fan & Lederman, 2018;Tacco et al., 2018) to trust stage progression (Fan et al., 2014), language, and similarity cues that indicate member trustworthiness (Sillence, 2013) and the consideration of trust dimensions in tandem with several other constructs in the context of value co-creation (Zhao et al., 2013b;Zhao et al., 2015). Furthermore, no empirical research has examined the relationship between trust antecedents and engagement, or the consequents of that relationship, in an OHC context, despite the fact that the empowerment of OHC members has been shown to relate directly to their level of engagement with the community (Oh & Lee, 2012), manifested through information disclosure (Petri<unk> & Petrov<unk>i<unk>, 2014) and knowledge adoption (Johnston et al., 2013), both of which are trust behaviors. However, this research does just that, answering repeated calls to address this absence of research on engagement in OHCs-answering, in particular, the call to investigate whether the outcomes of such research are similar to those obtained in other contexts (Hur et al., 2019), as well as the call (Demiris, 2006) to clarify how engagement in OHCs might empower members to make healthcare decisions. Our research yields insights that contribute to the small but growing body of knowledge on engagement in the OHC context, empirically illustrating the role of engagement as a mediator between trust antecedents and behavioral trust responses in the unique context of OHCs. Additionally, this research answers calls from the IS field for trust research that focuses on trust targets other than technology (Söllner et al., 2016) by focusing on OHC members and their responses to nontechnical trust antecedents. Finally, our findings advance an understanding useful to community hosts. Engagement is critical in determining the sustainability of social networks (Thielst, 2011), and researchers (Wang et al., 2017) have shown that those who contribute informational support in an OHC context remain members of those communities for longer periods of time than those who simply seek and receive informational support. The findings of our study yield important insights into how trust can be more effectively generated in an OHC context to support increased member engagement, thereby providing valuable guidance for those seeking to promote the sustainability of these platforms. This study is structured as follows. First, we outline the theoretical background of this examination of trust in the OHC context. This includes a review of the relevant literature and the study hypotheses. Then, we describe the methodology employed to test the research model. Finally, we discuss the study findings and their implications for theory and practice. The paper concludes with an outline of study limitations and potential directions for future studies in this area. --- Theoretical Background The objective of this paper is to examine trust formation in OHCs. We examine the relationship between trust and engagement, both in terms of the specific trust antecedents that predict engagement and in terms of the trust responses related to engagement. To that end, we draw on social capital theory and social exchange theory. Social capital is a term used to describe the "norms and networks that facilitate collective actions for mutual benefits" (Woolcock, 1998, p. 155). It has been described (Beaudoin & Tao, 2007) as the actual or potential resources that result from social connections and senses of reciprocity and trust, which can bring about outcomes at the individual and collective levels. It has been argued (Nahapiet & Ghoshal, 1998) that social capital encompasses distinct structural, relational, and cognitive dimensions. In the OHC context, the structural dimension is represented by social interaction links and ties between members of the community, as manifested in network density, interaction frequency, duration, and depth. These structural links are conduits for resources, such as credible information and experiential knowledge. The relational dimension encompasses relationship connections between community members and trust and identification with other members, as evidenced in the perceived support and perceived responsiveness of the online community. The cognitive dimension is represented by the shared understanding, values, and normative expectations of the community, all of which bind a community together and facilitate the achievement of its objectives. In the context of this study, it is proposed that the presence of these three dimensions is likely to influence a trust response and the intent to engage with the community. However, the unique nature of OHCs means that they are characterized by particular vulnerabilitiesspecifically, the adoption of incorrect health advice may result in significant consequences for the individual; similarly, the disclosure of personal health information represents privacy loss. As a consequence, both of the trust responses examined in this study involve a risk-benefit calculus with more impactful outcomes than would be the case in many other online contexts. In the case of knowledge contribution, it contains the elements of social exchange, one that takes place under a condition of risk, which in this case is loss of privacy. We thus employ social exchange theory (SET) as one of the theoretical frameworks guiding this study because the transfer of personal information is an exchange between social actors that involves awareness of the risks associated with the disclosure of this information (Youn & Hall, 2008). SET bridges disciplines, including anthropology, social psychology, and sociology, and conceptualizes social behavior as an exchange process in which individuals evaluate relationships in terms of their benefits and risks. It therefore emphasizes behavior as a process of resource exchange (Emerson, 1976) where one person evaluates the cost associated with exchanging a resource (such as health information) with someone else in order to receive a specific benefit (such as advice). The explanatory power of this theory has been applied to examine issues as diverse as psychological contracts (Rousseau, 1995), employee responses (Jones, 2010), trust generation, and privacy concerns (Luo, 2002). In an online context, it has been used (Tsai & Kang, 2019) to examine reciprocal intention in knowledge seeking, online repurchase intentions (Chou & Hsu, 2016), and knowledge sharing in OHCs (Yan et al., 2016). The literature has made it clear that trust is only required in conditions of uncertainty and risk and is necessary for exchange relationships to succeed. This applies to disclosure relationships, as without some form of trust among online community members, most individuals would be reticent to disclose personal information, particularly to online community members with whom they are unfamiliar. SET is also relevant to knowledge contribution from a benefit-evaluation perspective, as it emphasizes the intrinsic rewards that accrue from information sharing, which include feelings of belonging, network ties, trust, and community commitment-all of which are rewards that strengthen the further development of social capital. The unit of exchange (in this case personal health information) may also contain intrinsic socioemotional value for the recipient of that information, motivating their desire to reciprocate (Cropanzano & Mitchell, 2005). That value relates to the fact that disclosure of such information demonstrates trust, respect, and appreciation of their expertise. For example, researchers (Foa & Foa, 1980, 1974) have long contended that units of exchange (including information) may provide symbolic benefit to the recipient, a benefit that conveys a meaning that transcends objective worth to the individual and enables these units of value to be exchanged in a more open-ended manner. This is particularly true in the context of an online health community, where the disclosure of personal health information and the request for guidance regarding the management of one's health conveys a message that the recipients' expertise is trusted, respected, and needed (Shore, Tetrick & Barksdale, 2001). In this way, the provision of personal health information may be evaluated by the recipient as having intrinsic social value (Redmond, 2015), which facilitates their participation in an altruistically motivated interpersonal exchange, motivating their contribution to the development of an online community that they value. The SET framework therefore provides an empirically tested scaffolding for exploring the normative aspects of exchange that affect online information sharing choices, specifically trust in the online community. In the context of the current study, it indicates that when OHC members evaluate the informational and socioemotional supports that the community is providing as trustworthy and aligned with their needs, they are more likely to actively engage through applying that information, contributing health advice, and disclosing their own experiences. Our decision to integrate social exchange with social capital is consistent with an increasing body of work that has recognized the value of this integrated approach in examining trust or trust-related factors in the online community context (Ho & Lin, 2016;Jin et al., 2015;Munzel & Kunz, 2014;Wang & Liu, 2019). --- Trust Trust is a construct of enduring interest whose value and contribution to interpersonal, interorganizational, and transactional relationships is widely acknowledged by researchers and practitioners. The former seek to understand the antecedents of trust, whereas the latter seek to use those insights to reduce risk and improve interaction outcomes in situations of uncertainty. Golembiewski and McConkie (1976, p. 131) remark that there is "no single variable which so thoroughly influences interpersonal and group behavior as does trust." Notwithstanding significant interest in the construct by the academic community, there are numerous conceptualizations of trust. The multiplicity of definitions and the conceptual diversity that surrounds the construct results from the different disciplines of researchers and their different research foci and emphases (McKnight et al., 1998). Nonetheless, some points of commonality are evident in the literature, with trust frequently defined in terms of optimistic expectations or confidence. For example, McAllister (1995) perceives trust in terms of positive expectations regarding consequent behavior, while Jarvenpaa and Leidner (1999) define trust as the optimistic expectation that the trusted person will act ethically and morally, even without being monitored (Jarvenpaa & Leidner, 1999;Moorman et al., 1992). While Hosmer (1995) describes trust as a positive expectation that the other party will not exploit or take advantage of a situation through opportunistic behavior, a slightly more nuanced approach is provided by Golembiewski and McConkie (1976), who view trust in terms of confidence in an event, person or process based upon personal perceptions and experiences. Interestingly, they also view trust as a dynamic phenomenon, one that can evolve over time and can be influenced by positive experiences. Trust definitions frequently reference issues such as the potential for exploitation or perceived risk, thereby pointing to the fact that trust is critical for the success of all social interactions that involve uncertainty and dependency. In fact, Mayer et al. (1995, p. 711) note that the need for trust only arises in a situation of risk. It has also been asserted that "willingness to take risks may be one of the few characteristics common to all trust situations" (Johnson-George & Swap, 1982, p. 1306). Engaging with an OHC, either through disclosing personal information or acting on health advice, places participants in a position of vulnerability and risk. Because it involves a significant dependency on that community (for the provision of trustworthy health advice) in order to ensure a positive outcome, the potential vulnerability and risk from opportunistic behaviors are correspondingly greater. Since this study focuses on the OHC context, it incorporates these perspectives and draws on Corritore et al. (2003) to define trust as "an attitude of confident expectation in an online health community context that one's vulnerabilities will not be exploited." Trust research typically examines the relationship between trust antecedents, cognitive and affective trust (e.g., Fan & Lederman, 2018;Johnson & Grayson, 2005;Kanawattanachai & Yoo, 2002), and outcomes such as the intention to adopt technology. In our study, we advance knowledge in two ways: First, we advance knowledge in our choice of outcomes, we focus on knowledge adoption and knowledge contribution, which aligns with our OHC context. Our outcome variables also focus on behaviors rather than intention. Second, we advance knowledge by focusing on engagement as the mediating mechanism between trust antecedents and outcomes. There are two main reasons for this: First, the trust antecedents employed in this study conflate with cognitive and affective trust because they are all measured perceptually. For example, the literature has repeatedly confirmed that evaluations of information credibility reflect a cognitive trust judgment, and evaluations of community support and community responsiveness influence both cognitive and affective trust perceptions. Second, we believe that engagement is a more important mediating mechanism to examine than cognitive and affective trust because its strong association with behavior makes it particularly salient to the stability and continuation of online communities (Algesheimer & Dholakia, 2005). Thus, our model comprises an examination of cognitive and affective trust antecedents, engagement, and behavioral trusting responses, providing new insight into which type of trust antecedent and which pathway is most effective in influencing engagement and behavioral trust responses in OHCs. We turn now to an examination of engagement in the literature. --- Engagement Engagement-conceptualized in this study as a state of involvement and connection between the individual and community that creates value for the individual, as manifested by behavioral outcomes-has its roots in the marketing literature. Engagement has been found to be related to consumption and purchase behaviors (van Doorn et al., 2010), online brand community engagement (Wirtz et al., 2013), and online engagement and advertising effectiveness (Calder et al., 2009). Researchers have criticized the definitional confusion associated with engagement (e.g., Ray et al., 2014;Suh et al., 2017) and Cheung et al. (2011) observed that the definition, dimensionality, and consequent operationalization of customer engagement in many marketing studies is inconsistent and mixed. Our definition of engagement builds on existing research (Webster & Ahuja, 2006;Webster & Ho, 1997) that defines engagement in a system as something that "holds [users'] attention and they are attracted to it for intrinsic rewards" (Jacques et al., 1995, p. 58). This definition is also consistent with Higgins's (2006, p. 442) description of engagement as being involved, occupied, and interested in something, and Calder and Malthouse's (2008) view of engagement as a state of involvement and connectedness between the user and the object of engagement that can motivate behavioral outcomes. This involvement and holding of attention is not temporally bounded to a specific instance of information exchange (Eldor, 2021;Eldor & Harpaz, 2015;Brodie et al., 2011) because a responsive and supportive community can provide an intrinsic socioemotional reward that equally interests and captures the attention of community members. A small but steadily increasing body of work has started to examine engagement in more diverse and non-product-specific online community contexts where the focus is on interaction and value co-creation (Hollebeek et al., 2017). These include online magazine communities (Heinonen, 2018), social media platforms (Di Gangi & Wasko, 2016), online travel communities (Fang et al., 2018), online learning communities (Ryle & Cumming, 2007), online gaming communities (Chuang, 2020), and OHCs (Hur et al., 2019;Litchman et al., 2018), among others. Within these examinations, the locus of attention varies considerably, ranging from usage metrics, antecedents of online engagement, and the consequents of that engagement to motivations and valence, social identity, and telepresence-variations that have, at times, bounded the dimensionality and generalizability of these examinations. However, this focal diversity is accompanied by valuable conceptual work, including literature reviews that have provided much-needed structure and guidance regarding the construct (Suh & Cheung, 2019;Unal et al., 2017). One point on which most researchers agree is that engagement is context dependent (Brodie & Hollebeek, 2011;Brodie et al., 2011;Brodie et al., 2013;de Oliveira et al., 2016); as a consequence, there is a need for further research within more diverse social and cultural contexts in order to progress our understanding of the predictors and consequents of the construct and to increase its conceptual clarity (Cheung et al., 2011;Dessart et al., 2015;Suh et al., 2017). Engagement and trust are related in that they share cognitive and affective elements; nonetheless, they remain distinctive constructs, as is evident in their conceptual composition and expression. For example, trust is frequently defined in terms of beliefs and an attitude of confident expectation that vulnerabilities will not be exploited, in contrast to engagement, which is typically conceptualized as a state of involvement and connection that creates value for the individual. Engagement has a stronger association with behavior and has been found to influence trust responses in the online community context (Islam & Rahman, 2016;Kang et al., 2016). For example, Ray et al. (2014) demonstrated the relationship between online engagement and the trust-related outcomes of satisfaction and knowledge contribution, while Rich et al. (2010) found that engagement mediates behavior. However, the majority of studies examining the relationship between engagement and trust focus on online brand communities or discussion communities; as a consequence, whether that relationship extends to the specific OHC context remains undetermined. Both cognitive (rational evaluation) and emotional (indicating an affective perception) factors enable the expression of engagement (Kahn, 1990). By examining information credibility, community support, and community responsiveness as trustrelated determinants of engagement, this study explores the cognitive and emotional components of engagement that trigger the behavioral activation component, demonstrating that engagement can effectively be measured through specific interactions (and that trust-related components can influence this). In doing so, it advances understanding of the nature of the trust-engagement relationship in an online health community context, answering calls (e.g., Ray et al., 2014) to better understand engagement through expanded frameworks that incorporate related constructs. --- Model Development The research model for this study is shown in Figure 1. It proposes that OHC engagement is influenced by information credibility, community support, community responsiveness, and the propensity to trust. Community support is conceptualized as a reflective second-order construct, with four dimensions corresponding to the four facets of community support (Chiu et al., 2015). The model also shows that engagement influences knowledge adoption and knowledge contribution behaviors, both of which are also influenced by the propensity to trust. Information Credibility In an OHC, participants seek credible information to help them cope with the uncertainty associated with the illness they are trying to overcome. This is a significant challenge, as much of the communication in online groups is subjective, discursive, experiential, and frequently anonymous (Fan et al., 2010). Moreover, it is a challenge with potentially serious consequences (Hilligoss & Rieh, 2008), as acting on incorrect information regarding aspects of a disease or its management could negatively impact health outcomes (Hajli, 2014;Hajli et al., 2014;Lober & Flowers, 2011;Maloney-Krichmar & Preece, 2005). We argue that information credibility influences engagement, knowledge adoption, and knowledge contribution in the online health community context. As online communities are characterized by the lack of face-to-face interaction and the inability to verify expertise, this amplifies perceived and behavioral uncertainty regarding the credibility of information provided by other members. As a result, members of these communities place greater reliance on signals of information trustworthiness, such as member feedback (Pavlou & Dimoka, 2006) expressed in comments or posts, treating these as important indicators of information approval (Fan et al., 2014;Flanagin & Metzger, 2013). The presence of such signals of information credibility has been shown to stimulate members' participation in general online communities (Benlian & Hess, 2011) and is likely to be equally relevant to OHC contexts. Based on the above discussion, we propose: H1a: In the OHC context, information credibility is positively related to engagement. Research by Fan and Lederman (2018) on patient OHCs found that perceived information credibility influences knowledge adoption. Similarly, a reexamination of trust antecedents in internet-based health information (Sillence et al., 2019) confirms the predictive importance of information credibility on the intention to act on that information. Although in this case, the focus was health websites rather than online communities, it is likely that the same outcome may extend to OHCs. We therefore propose: H1b: In the OHC context, information credibility is positively related to knowledge adoption. Moreover, because credible information benefits the recipient by enhancing their knowledge, this results in an increase of social capital and their desire to reciprocate through information contribution. Empirical support for this is found in prior work (Benlian & Hess, 2011) showing that quality-assured content shapes the trust perceptions of online community users, thereby increasing their participation behavior. Researchers such as Chan and Li (2010) have demonstrated that interactivity or engagement in a virtual context can be developed via structural or experiential routes, both of which influence reciprocity. The structural route comprises community features that provide credible information resources to users (with the experiential route comprising social bonds and enjoyment that provide socioemotional resources to users). These authors have shown that both routes to interactivity influence the norm of reciprocity and voluntary co-creation behaviors, which in the case of this study is expressed through the contribution of knowledge. The current study is therefore consistent with extant research in proposing that the provision of information which is perceived as credible strengthens the structural bonds that stimulate community involvement and connection, motivating reciprocal engagement as expressed through knowledge contribution. Based on the above discussion, we propose: H1c: In the OHC context, information credibility is positively related to knowledge contribution. --- Online Community Support Supportive interactions among individuals in a traditional healthcare environment can play a protective role in countering the health-related effects and life-stressing consequences of a disease situation, thus contributing to participants' well-being (Cobb, 1976;Schaefer et al., 1981). OHCs can also perform this protective role by promoting social interaction. Further, participants benefit from learning from the experience of others, resulting in improved health outcomes and greater engagement in the selfmanagement of disease (Yan & Tan, 2014). Community support is a multidimensional construct comprising facets such as emotional support, informational support, tangible support, network support, and esteem support (Mattson & Hall, 2011;Schaefer et al., 1981). While tangible support does not apply in the context of online communities, the other support categories do apply and serve as manifestations of social support within online communities. Such support provides an intrinsic socioemotional reward that equally interests and captures the attention of community members. We therefore propose: H2a: OHC support is positively related to engagement. Once an individual has been diagnosed with a disease, it is understandable that they would search for health information and advice regarding how best to proceed in treating their illness (Yan & Tan, 2014;Schaefer et al., 1981). When members of an OHC perceive that they are receiving informational support, through salient information, valuable advice, and informed guidance on specific issues, this is likely to engender beliefs regarding the competency of other community members. In this way, informational support aligns with the ability dimension of trust, contributing to the decision to engage in trusting behavior. Emotional support, reflecting the demonstration of concern and care, fills the affective needs of the individual. Such concern and care has been described as empathy and sympathy (Yoo et al., 2014), encouragement and security, and care and affection. It helps to engender a sense that the community is positively intentioned and genuinely supportive of the individual and their wellbeing (Schueller, 2009) and thus aligns with the concept of benevolence (Mayer et al., 1995). Esteem support can be expressed through online interactions that reinforce the individual's self-esteem and their belief in their capacity to cope with the situation by moving through the stages of their health condition (Mattson & Hall, 2011). Because of their positive intention, such interactions are also analogous to the trust concept of benevolence. Finally, network support demonstrates that the individual is a member of a support network that is available to assist others, thereby providing the participant with a sense of belonging to the community and the ability to share their experiences (Yan & Tan, 2014;Schaefer et al., 1981;van Uden-Kraan et al., 2008). Research (Tsai & Hung, 2019) has shown that a sense of belonging or identification with an online community influences both cognitive and affective trust formation, which, in turn, predict continuous use intentions. In the literature, explicit support is provided for the predictive influence of social support on engagement and trust-related behavioral outcomes. For example, recent work by Mirsaei and Esmaeilzadeh (2021) in the U.S. found that perceived social support (as an indicator of channel richness) influences engagement in OHCs, as well as patient participation in care management. The work of Wang et al. (2021) demonstrated that social support is a key predictor of a new user's continued engagement in an OHC. Similarly, Yang et al. (2017) revealed the relationship between perceived social support, trust in health information, and engagement in health informationseeking actions, while an earlier study by Jin et al. (2016) confirmed the influence of emotional support on healthcare knowledge adoption behavior within an online community context. We therefore propose: H2b: OHC support is positively related to knowledge adoption. As previously noted, Chan and Li (2010) confirmed that interactivity or engagement in a virtual context can be developed via experiential routes that include the provision of socioemotional resources to users and that interactivity developed in this way influences the norm of reciprocity and voluntary co-creation behaviors, which in the case of this study is expressed through the contribution of knowledge. More recently, Abidin et al. (2020) demonstrated the relationship between social support and trust formation within an OHC, showing its influence on knowledge sharing and community promotion. Based on this discussion, we propose: H2c: OHC support is positively related to knowledge contribution. --- Online Community Responsiveness Many individuals join an OHC to increase their knowledge regarding a specific health concern and prefer to receive answers to their questions from others who have either experienced or are familiar with their health issue and can therefore provide informed insights. Consequently, an OHC that is perceived as being responsive to information requests by providing timely responses to posts, is likely to result in more satisfied members and higher levels of member participation and is more likely to be evaluated as trustworthy (Zhao et al., 2013a). In this study, we argue that community responsiveness positively influences engagement, knowledge adoption, and knowledge contribution in the online health community context. If other community members respond speedily to member requests, this indicates that they have the competence to provide informed guidance, are willing to do so, and are interested in the needs of the community, demonstrating their ability, integrity, and benevolence. It also builds confidence in the community as a valuable source of socioemotional support for guiding decisions. As a result, individuals are more likely to increase their participation in the community over time by reciprocating through responding to other members' posts. Lin and Lee's (2006) examination of the determinants of success for online communities confirmed the importance of perceived responsiveness to behavioral intentions, which in turn increases member loyalty to the community, as indicated by participation in the community. Later, work by Singh (2012) also showed that responsiveness can strongly influence the participation of new members of communities. Similarly, Casaló et al. (2013) found that response speed, value, and frequency influence online community members' satisfaction and their participation intentions, while Sheng (2019) empirically demonstrated that perceived responsiveness is a motivational driver of customer engagement. Although the context of these studies was general, technical, travel, and review online communities, it is a reasonable expectation that these relationships would equally extend to the OHC context. Based on this discussion, we propose: H3a: OHC responsiveness is positively related to engagement. Because a responsive online community provides a range of informed and supportive perspectives, this increases trust in the perceived competence, integrity, and benevolence of the community (Zhao et al., 2013a) and correspondingly reduces the perception of risk associated with acting on information provided by community members. This is consistent with Bagozzi and Lee's (2002) view that social processes are important determinants of decision-making. We therefore propose: H3b: OHC responsiveness is positively related to knowledge adoption. and Chen (2005) in the context of an online breast cancer discussion board, which demonstrated that the orientation of members who make frequent posts tends to change over time from an emphasis on seeking information to one of supporting other members through the provision of information. Thus, we propose: H3c: OHC responsiveness is positively related to knowledge contribution. --- Engagement and its Relationship to Knowledge Adoption and Contribution Our final hypotheses focus on the relationship between engagement and the outcomes of knowledge adoption and knowledge contribution. We reason that the more a person actively participates in an OHC-for example, by posting questions or requesting advicethe broader the range of information they will accumulate from other members, which they can then evaluate and use to guide their behavior. In addition, the psychosocial and relational benefits that result from participation will increase their confidence in member benevolence and reassure them that they are making informed and correct decisions. Support for this position is provided by Jin et al. (2016), who found that the level of involvement positively affects online community members' adoption of healthcare information. Similarly, Zhou (2020) found that informational support and emotional support, through their effect on social capital, influence Chinese online community users' participation, as expressed through health knowledge acquisition and contribution. This is also consistent with the work of Liao andChou (2012, 2017), which showed that prior positive exchanges with an online health community engender the trust necessary for leveraging a contributor's social capital for the purpose of information adoption. As a result, we propose: H4a: OHC engagement positively influences knowledge adoption. Because members of OHCs seek to protect their privacy (and avoid negative repercussions such as being trolled if dealing with a stigmatized health condition), a significant perceived risk is associated with the selfdisclosure of personal health information. We contend that active participation in an OHC generates the trust necessary to overcome that perception of risk, reasoning that observing and learning from the posts of others and their responses generates user confidence in the expertise, integrity, and benevolence

Online health communities (OHCs) represent a popular and valuable resource for those seeking health information, support, or advice. They have the potential to reduce dependency on traditional health information channels, increase health literacy and empower a broader range of individuals in relation to their health management decisions. Successful communities are characterized by high levels of trust in user-generated contributions, which is reflected in increased engagement and expressed through knowledge adoption and knowledge contribution. However, research shows that the majority of OHCs are composed of passive participants who do not contribute via posts, thereby threatening the sustainability of many communities and their potential for empowerment. Despite this fact, the relationship between trust and engagement, specifically the trust antecedents that influence engagement in the OHC community context has not been adequately explained in past research. In this study, we leverage social capital behavior and social exchange theory frameworks in order to provide a more granular trust-based elucidation of the factors that influence individuals' engagement in OHCs. We collected data from 410 Brazilian participants of Facebook OHCs and tested the research model using partial least squares. The results confirm two new constructs-online community responsiveness and community support-as trust antecedents that influence engagement in OHCs, resulting in knowledge adoption and knowledge contribution responses. These findings contribute to the trust and engagement literatures and to social media research knowledge. From a practitioner perspective, the study findings can serve as an important guide for moderators and managers seeking to develop trusted and impactful OHCs.

decisions. Support for this position is provided by Jin et al. (2016), who found that the level of involvement positively affects online community members' adoption of healthcare information. Similarly, Zhou (2020) found that informational support and emotional support, through their effect on social capital, influence Chinese online community users' participation, as expressed through health knowledge acquisition and contribution. This is also consistent with the work of Liao andChou (2012, 2017), which showed that prior positive exchanges with an online health community engender the trust necessary for leveraging a contributor's social capital for the purpose of information adoption. As a result, we propose: H4a: OHC engagement positively influences knowledge adoption. Because members of OHCs seek to protect their privacy (and avoid negative repercussions such as being trolled if dealing with a stigmatized health condition), a significant perceived risk is associated with the selfdisclosure of personal health information. We contend that active participation in an OHC generates the trust necessary to overcome that perception of risk, reasoning that observing and learning from the posts of others and their responses generates user confidence in the expertise, integrity, and benevolence of other members. It also reinforces knowledge efficacy. Over time, we predict the social capital that this generates increases the desire to reciprocate and contribute to the community through the provision of information. Support for this is provided by Kuem et al. (2020), who found that Instagram community engagement positively influences active contribution behaviors. Cheung et al. (2015) found that the posts and member recommendations in online social shopping communities influence subsequent customer information contribution behavior, with the latter exerting the stronger effect. This confirms that positive feedback and the advice of other online community members reinforce learning and drive information contribution behaviors. Similarly, Chan and Li (2010) showed that interactivity in a virtual community stimulates the norm of reciprocity and voluntary behaviors. This is consistent with work (Rodgers & Chen, 2005) showing that OHC member orientation tends to progress over time from an informationseeking orientation to one that supports other members through providing information. We thus propose: H4b: OHC engagement strongly influences knowledge contribution. --- The Propensity to Trust Researchers such as Rotter (1967) have conceptualized trust as a personality characteristic that influences an individual's likelihood of trusting others. This has alternately been described as a trust propensity (Mayer et al., 1995) or as dispositional trust (Kramer, 1999) and indicates a general willingness to trust others across a broad range of trust situations and trust targets (McKnight et al., 1998). The propensity to trust influences the amount and level of trust that a person has for another party in the absence of available or experiential information on which to base a judgment (Rotter, 1971(Rotter,, 1980)). Because of this, the propensity to trust is particularly important in the early stages of relationships involving interpersonal interactions with unfamiliar actors when there are insufficient situational cues or information about the trustee available (Bigley & Pearce, 1998;Colquitt et al., 2014;McKnight et al., 1998). Moreover, the propensity to trust retains its impact and can continue to influence trusting beliefs even after information about the trustee becomes available because it serves as filter or lens through which the behavior of others is then viewed (Colquitt et al., 2007). Research has shown that dispositional trust influences trust beliefs in relation to web vendors (e.g., Chen et al., 2015;Gefen, 2000;Kim et al., 2009). Similar outcomes are evident in nontransactional contexts as well. For example, Tait and Jeske (2015) found that the propensity to trust predicts the disclosure of potentially sensitive and identifying information in an online information-sharing context. The propensity to trust also exerts a significant influence on risk-related beliefs and the intention to adopt health information from online health infomediaries (Song & Zahedi, 2007). Similarly, Heldman and Enste (2018) found that dispositional trust determines the level of trust placed in the recipient of private data, especially when the person is unfamiliar with this recipient. Thus, we propose: --- H5a: The propensity to trust is positively related to OHC knowledge adoption. --- H5b: The propensity to trust is positively related to OHC engagement. --- H5c: The propensity to trust is positively related to OHC knowledge contribution. --- Methodology This study is aligned with the pragmatic philosophical paradigm, which encourages practical and applied action (Teddlie & Tashakkori, 2008). In order to provide an in-depth examination of the relationships between different constructs, specifically the relationship between trust antecedents, engagement and trust responses, the most appropriate method was determined to be a quantitative survey. --- Data Collection To sample from the target population of participants of OHCs, we surveyed members of OHCs on Facebook Brazil. We identified six OHC types: pregnancy/breastfeeding/motherhood (PBM), nutrition/alimentation/ dietary (NAD), beauty/esthetics (BES), disease treatment (DTR), fitness (FIT), and animal care (ANC). Respondents were asked to identify one of these online communities in which they participated and then to answer the remaining questions with regard to that community. Invitations were published in 10 OHCs, with a total of 813,223 registered members, after securing authorization from group managers. To encourage respondent participation, nine raffles of USD 20 were announced. Each raffle targeted different groups such as moderators, managers, and participants. We received 602 responses. After eliminating those with high levels of missing data, we were left with 410 valid responses. As a preliminary validity test, we checked for alterations in the mean of the responses possibly associated with the time spent answering the questionnaire, without finding relevant differences. The majority of respondents were female 93.2% (n = 382); almost half were between 26-35 years old (46.6%); 58.0% were married and 56.8% were educated to at least college level. The majority (74.2%) had more than one year of experience using online groups on Facebook and were active participants. Most (81.2%) visited online health groups at least once a day, participating in online health groups with similar themes (79%). Almost half (48.0%) considered themselves to be active participants, regularly contributing through posting questions, responding to questions, and "liking" others' posts. In order to ensure that participants had enough knowledge and experience using OHCs to be able to assess information credibility, community support, and community responsiveness, respondents with low participation frequency were excluded from the study. This was achieved by retaining only respondents who self-reported accessing the online health group "once a day" or "once a week" (n = 27 respondents were removed). Also, only communities with a sample size of at least 30 were retained for the sake of statistical representativeness. As a consequence, we additionally removed "fitness" (n = 13), and "animal care" (n = 12), resulting in a final sample of 358 responses across the four communities. Table 1 shows the sample distribution based on the community type, frequency of use, and user experience. The nature of our data collection, where no real information about the population is available, precludes a full assessment of nonresponse bias. However, we followed the procedures recommended by Armstrong and Overton (1977) to assess the likelihood of nonresponse bias. We compared the earliest and latest responses received, based on the assumption that those who respond less readily are likely to be more similar to nonrespondents than those who respond immediately. We assessed the differences in the means of each of the 40 items that make up our measures between the first and last 10% of responses received and observed only two significant differences (Knowledge Contribution Item 2 and Esteem Support 2). The limited observed differences suggest sufficient similarity between early and late responders, thus diminishing the risk of nonresponse bias as an alternative explanation for our findings. --- Measures We measured information credibility using items from Lederman et al. (2014). The community support construct combined items from previous research, which measured four dimensions: emotional, informational, esteem, and network support (Chiu et al., 2015;Schaefer et al., 1981). Although these researchers had proposed tangible support as an additional dimension, it was not considered relevant in this study, as our virtual context provides no physical interaction between participants. Community responsiveness items were drawn from Wagner et al. (2014). Knowledge adoption was measured using items adopted from Chou et al. (2015), while knowledge contribution was measured using items from Meng and Agarwal (2007) and Zhao et al. (2013a). Engagement was assessed by adapting items from Webster and Ahuja (2006). We dropped one item (ENG 5) because it originally referred to "how fun" the respondent feels the experience of using the system is, which we found inappropriate for the context of online health communities. The measures were translated into Portuguese and two pretests were conducted in order to retain meaning and idiomatic equivalence (Cha et al., 2007). In the first pretest, expert researchers in the field were invited to respond to the questionnaire and provide feedback to improve the items. For the second pretest, the process was repeated with the moderators and group managers of each OHC. During this process, we examined the validity of the scales based on statistical procedures proposed by MacKenzie et al. (2011). The research participants were asked to answer the questions using a 5-point Likert scale ranging from 1 = "strongly disagree" to 5 = "strongly agree." Overall, no significant changes to items were required, but some were slightly adjusted in order to maintain the meaning and to ensure compliance with Portuguese grammatical requirements. Appendix 1 shows the items used. --- Data Analysis The model was tested using partial least squares (PLS) structural equation modeling (SEM) as implemented in SmartPLS (Ringle et al., 2015). PLS-SEM is appropriate when the objective is to identify key driver constructs in a relatively complex model that deals with multiple latent variables and relationships, without being subject to rigorous distributional assumptions (Hair et al., 2017b). Power analysis using GPower (Buchner et al., 2014), indicated that our sample was more than sufficient to detect a medium effect size of f 2 = 0.15 (Cohen, 1988) with 90% power. --- Results --- Measurement Model Table 2 shows the individual items and cross-loadings. All but one load at greater than 0.71 on their intended construct, meaning that the item loading accounts for more than 50% of the overlapping variance, which is considered excellent (MacKenzie et al., 2011). We considered the marginal value of ENG2 (<unk> = 0.67) to be acceptable, as it does not pose any threat to the other measures of reliability and validity of the construct. Following the rule of thumb in Tabachnick and Fidell (2014) and Comrey and Lee (2016), we found that the majority of the cross-loadings are below the value of 0.32 (10% of overlapping variance), while scattered occurrences are under 0.45 (20% of overlapping variance), a threshold that is considered fair enough to show any notable interconstruct confounding effects. We also focused on the few occurrences where values were above 0.45 but below 0.55 (30% of overlapping variance), such as in relation to the information support latent variable (items IS01, IS02, and IS03). We then performed a post hoc analysis with the structural model by eliminating the entire construct to evaluate the potential effect on the stability of the structural model results (which will be presented in the subsequent sections) and found no substantial differences. However, we decided to maintain the construct since it is conceptually linked to the community support construct. The average variance extracted (AVE) of all constructs (Table 3) are above the threshold of 0.5 (Fornell & Lacker, 1981). Cronbach's alpha (CA) and composite reliability (CR) values were all above 0.79, indicating satisfactory reliability. Finally, the square root of the AVE for each construct is higher than the correlations with the other constructs, thus providing evidence of discriminant validity. Variance inflation factor (VIF) values were all below 2 (the highest was 1.68), indicating that multicollinearity did not exert a biasing influence on the results (Hair et al., 2017b). In any data collection with a single instrument at a single period in time, common method bias (CMB) is a potential alternative explanation for the results. To mitigate this risk, we first undertook procedural remedies (Podsakoff et al., 2003) through careful construction of the survey to deal with ambiguity, conciseness, uniqueness of content, and lack of focus. We then empirically assessed the potential concern of CMB using two procedures. First, we performed the Harman's single-factor test (Podsakoff et al., 2003). The unrotated factor solution did not converge on a single factor and the largest covariance explained by any factor was 19.5%. Second, as suggested by Kock (2015), we assessed CMB in our structural model using lateral multicollinearity assessment (Kock & Lynn, 2012). All the variance inflation factors (VIF) were below the recommended threshold of 3.3, with the highest being 1.21. Given our procedural remedies and the lack of evidence in the empirical assessments, we do not consider CMB to be a significant threat. --- Hypothesis Testing The results provide partial support for our hypotheses. Consistent with H1a, information credibility showed a positive influence on engagement (<unk> = 0.15, p <unk> 0.001). Similarly, the data showed a positive relationship between information credibility and knowledge adoption (<unk> = 0.22, p <unk> 0.001), thus confirming H1b. However, the relationship between information credibility and knowledge contribution was not significant (<unk> = -0.07, ns), rejecting H1c. With regard to the influence of engagement, the data show that engagement influences knowledge adoption (<unk> = 0.24, p <unk> 0.001) but does not exert any significant influence on knowledge contribution (<unk> = 0.05, ns), thus supporting H4a but not H4b. The propensity to trust was found to exert a positive influence on engagement (<unk> = 0.10, p <unk> 0.001) and knowledge adoption (<unk> = 0.15, p <unk> 0.001), offering support to H5a and H5b. However, no effect was observed on knowledge contribution (<unk> = 0.04, ns); thus H5c is not supported. In summary, our findings indicate that knowledge adoption is influenced by information credibility, community responsiveness, engagement, and the propensity to trust, each of which exerts a similar effect. Community support influences knowledge adoption indirectly through its effect on engagement. In the case of knowledge contribution, the source of influence is more bounded with community support exerting a strong influence on this behavioral trust response. As shown in Figure 2, these findings support many of the proposed relationships and explain 41% of the variance in engagement in OHCs, 45.2% of the variance in knowledge adoption, and 23.8% of the variance in knowledge contribution within this context. --- Mediation Analysis of the Role of Engagement We examined the importance of engagement as a mediator variable in the model following the procedures of Hair et al. (2017b). After confirming that our measurement model is reliable and valid, a crucial prerequisite to determining mediation effects, we estimated the direct and indirect effects by bootstrapping with 5000 subsamples the complete model. This technique implements the method of Preacher and Hayes (2004) and others (Hayes, 2013;Zhao et al., 2010) in the context of PLS-SEM. Following Zhao et al. (2010) and Hair et al. (2017b) we subsequently calculated the mean and standard errors of the paths in the model and determined the multiple mediation roles of engagement revealed by the significance of the corresponding direct and indirect effects paths, as shown in Table 4. The results demonstrate that, in relation to knowledge adoption, engagement partially mediates information credibility and community responsiveness, while fully mediating community support. These results additionally clarify the importance of user engagement in OHCs. The indirect effect of the community responsiveness on knowledge adoption represents about 20% of the total effect. We further explore the mediating role of engagement in our post hoc tests below. --- Post Hoc Tests Following our assessment of the formal hypotheses we conducted post hoc tests to explore whether community type or user experience influenced the relationships in our model. We conducted a multigroup analysis (MGA) in PLS. We assessed measurement equivalence using the measurement invariance of composite models (MICOM) procedure (Henseler et al., 2016), which assessed configurational and compositional invariances across the groups. --- Community Type Our analysis of community type was restricted to the two largest communities to ensure adequate sample sizes. Table 5 shows the MGA results for the nutrition/ alimentation/dietary (NAD) (n = 127) and pregnancy/ breast-feeding/motherhood (PBM) (n = 131) groups. We found partial measurement invariance between the two groups for all latent variables in the model, which allowed for path coefficients comparisons by means of a multigroup analysis. The specific differences between groups paths (<unk>) are discussed below. The results provide interesting insights regarding commonalities and distinctions. First, community support influences engagement in both types of communities, with the effect on PBM (<unk> = 0.47, p <unk> 0.001) being stronger (<unk>PBM, <unk>NAD = 0.27, p <unk> 0.03) than NAD (<unk> = 0.20, p <unk> 0.001). Other factors contribute to engagement in both cases. In NAD, engagement also depends on information credibility (<unk> = 0.34, p <unk> 0.001) and community responsiveness (<unk> = 0.20, p <unk> 0.001), while in PBM it depends on the propensity to trust (<unk> = 0.18, p <unk> 0.05). For both communities, engagement influences knowledge adoption (<unk>NAD = 0.23, p <unk> 0.05; <unk>PBM = 0.30, p <unk> 0.001, but does not exert a significant influence on knowledge contribution. In addition to the influence of engagement, information credibility also influences knowledge adoption to a similar degree for both types of community (<unk>NAD = 0.24, p <unk> 0.001; <unk>PBM = 0.26, p <unk> 0.001). However, interesting distinctions emerge because knowledge adoption in NAD strongly depends on community responsiveness (<unk>NAD = 0.32, p <unk> 0.001; <unk>NAD, <unk>PBM = 0.23, p <unk>0.06), while in PBM it is dependent on community support (<unk>PBM = 0.23, p <unk> 0.05; <unk>NAD, <unk>PBM = -0.21, p <unk>0.13). Moreover, although knowledge contribution is dependent on the effect of community support for both community types, in the case of NAD it is marginally important (<unk> = 0.20, p <unk> 0.10), whereas for PBM it exerts a much stronger effect (<unk> = 0.78, p <unk> 0.001; <unk>NAD, <unk>PBM = -0.57, p <unk>0.00). Finally, the propensity to trust plays a marginal role in influencing knowledge contribution in both groups (<unk> = 0.14) with no statistical difference observed between the two groups (p <unk>0.99). Further results are shown in Appendix 2. --- Mediating Role of Engagement by Community Type Examining the mediating role of engagement by community type (Table 6) also highlights the important role of context, showing that the effect of engagement differs according to community type, particularly in relation to the outcome of knowledge adoption. For example, engagement partly mediates the effect of information credibility on knowledge adoption for both NAD and PBM communities. However, while it fully mediates the effect of community support on knowledge adoption in the case of PBM, it has no effect in the case of NAD. In addition, when the mediating effect of community responsiveness is examined, the opposite outcome applies, with engagement partly mediating in the case of NAD but not in the case of the PBM community. This important distinction in outcomes confirms that the mediating effect of engagement on knowledge adoption differs according to the nature of community type. However, engagement does not mediate the effect of community support on knowledge contribution in either community. Experience We assessed differences in three levels of online community experience (1-6 months, 6 months to 2 years, and more than 2 years). To avoid confounding with community type, before performing group analyses, we compared the two types of OHCs-NAD (nutrition/alimentation/dietary) and PBM (pregnancy/ breast-feeding/motherhood) and found that no differences existed in distributions of users based on experience (<unk>2= 2.88, df = 2, p-value = 0.236), indicating no significant cross-effects between experience and community type. Following this, we performed all the steps of multigroup invariance analysis to ensure that we could compare the structural paths of the three experience levels. Since we were interested in comparing three user-experience levels, we employed the sequence of three pairwise comparisons with the Bonferroni correction to avoid Type I error inflation (Hair et al., 2018). We found configurational and compositional invariance across the groups. Additionally, we tested for the equality of the composite mean values and variances and found no statistical evidence of differences. The analysis of user experience (Table 7) provides interesting insights regarding the mechanics of engagement, knowledge adoption, and knowledge contribution in OHCs. The findings show that engagement is influenced by community support (1-6M: <unk>1-6M = 0.46, p <unk> 0.001; 6M-2Y: <unk>6M-2Y = 0.38, p <unk> 0.001 and 2+Y (<unk>2+Y = 0.34, p <unk> 0.001), irrespective of the length of user experience in online health communities (higher intergroup path coefficient difference: <unk>1-6M,<unk>2+Y = 0.11, p > 0.46). However, they also speak to the changing nature of the trust development process. For example, although significant only at p <unk>0.10, the results show that in the early stages of experience, the propensity to trust exerts an influence on engagement (<unk>1-6M = 0.18; p <unk> 0.10) that lessens as the user gains experience. This indicates an experience-dependent repertoire of factors that illustrates the progressive nature of engagement in online communities. --- Mediating Role of Engagement by User Experience Further assessment of the mediating role of engagement according to the user's level of experience with OHCs supports the overall finding that engagement is particularly relevant in the initial and medium stages of experience, particularly in relation to the outcome of knowledge adoption. The effect of community support on both knowledge adoption and knowledge contribution is partly mediated by engagement during the early (1-6 months) stage of experience. As experience increases (6 months to 2 years), the effect of community support on knowledge adoption is fully mediated by engagement, while the effect on knowledge contribution is not. For users with the greatest amount of experience, engagement no longer mediates the effect of community support. No mediation is present for information credibility and community responsiveness. These variations in effect are important, as they reveal that the level of users' experience is an important consideration when seeking to understand the mediating role of engagement on the influence of community support in the OHC context. --- Discussion This study examines the factors that influence trust and engagement in OHCs. It does so by leveraging social capital theory and social exchange theory to examine the relationship between trust and engagement, as reflected in trust antecedents that predict engagement and trust responses that result from that engagement. This study extends theory in a number of important ways, contributing significantly to the IS literature by providing a more complete understanding of the relationship between trust and engagement in the OHC context, as well as illustrating the need for incorporating contextual influence when examining this relationship. --- Contributions First, it shows that the key trust responses (knowledge adoption and knowledge contribution) are influenced by different community attributes. Knowledge contribution in OHCs is directly influenced by perceived community support, a factor that relates to whether community members offer information and advice that helps individuals cope with their health situation and healthrelated decision-making. Previous research has pointed to a diverse range of possible factors that can influence knowledge contribution in online social communities, ranging from IT-based features and identity verification (Ma & Agarwal, 2007); performance expectancy, selfefficacy and professional experience (Tseng et al., 2014;Wang & Lai, 2006); and the influence of selfpresentation, peer recognition, and social learning (Jin et al., 2015) to the rewards associated with altruism and fulfillment (Lin & Huang, 2013), social presence and identification (Shen et al., 2010), and even egoistic motives (Yu et al., 2011). Moreover, previous research has conceptualized knowledge contribution as being dependent on interconnected prior variables, including member satisfaction (Chou, 2020;Ma & Agarwal, 2007). In contrast, our research shows that community support is the dominant/singular driver of knowledge contribution in the OHC context and that its influence is direct and independent of other variables. Adding to the richness of the contribution is the fact that because we conceptualized community support as a second-order construct comprising four support subdimensionsemotional, esteem, information, and network support, our findings also clarify the exact nature of that support. This advances insight into how support can be implemented in an online health context, something that is of particular importance to the sustainability of these communities. On the other hand, our findings show that knowledge adoption in OHCs is influenced directly by information credibility, and community responsiveness and indirectly by community support. This extends the Fan and Lederman (2018), which focuses on the influence of information credibility (and contributor attributes) on knowledge adoption in OHCs by showing that community responsiveness and support are equally important considerations for understanding the formation of this trust outcome. A second contribution relates to the centrality of engagement to knowledge adoption as part of the trust formation pathway. Our findings show that engagement in OHCs is driven by information credibility, community support and community responsiveness. However, although our findings show that engagement influences knowledge adoption, it does not influence knowledge contribution behavior. This may indicate that the privacy concerns of OHC members are distinctively stronger than would be the case for members of more general virtual communities and that additional trust generation mechanisms are required to ensure that increased engagement translates into knowledge contribution. The fact that community support is the only attribute that influenced knowledge contribution points to the likely nature of such mechanisms. This contrasting finding places a cautionary pause on the assumption that increased engagement in virtual communities will automatically motivate member cooperation (Porter et al., 2011). In our OHC context, it did not. An associated contribution relates to the direction of the trust-engagement relationship, an issue that has long been a matter of contention in the academic community and the focus of calls (Islam & Rahman, 2016) for empirical work to determine whether trust is an antecedent or consequent of engagement. In conceptualizing trust in terms of both distinct trust antecedents and also the trust responses that arise from engagement, this study progresses beyond the limited binary perspectives that tend to characterize such discussions, affording much-needed insight into the cyclical nature of that relationship in the OHC context. The findings confirm that trust antecedents influence engagement and that a positive and direct relationship exists between engagement and one specific trust response, that of knowledge adoption. Our findings build on the work of Kang et al. (2016) which indicated a positive relationship between engagement and trust in a general online community, but we deepen that insight by showing the precise pathway and behavioral expression of that trust response, as well as the limits of this relationship in the OHC context. A third contribution relates to the importance of context. The study sample was predominantly composed of women respondents, a reflection of the fact that participants of online health support communities are more likely to be women (Ginossar, 2008); further, women are the population of interest in the context of the specific online communities examined in this study. For this reason, our finding of the importance of community support as a driver of knowledge contribution should be evaluated in relation to the study context (OHCs) and the nature of the respondent sample, both of which are interconnected. For example, research shows that women place particular value on community support in the virtual community context (Klemm et al., 1999;Sun et al., 2020). This may be due to gender-based socialization (Meyers-Levy & Loken, 2015;Reevy & Maslach, 2001), the greater emphasis that women have been shown to place on cues (Porter et al., 2012;Riedl et al., 2010;Rowley et al., 2017), and/or the fact that women's perception of risk and severity of consequences is stronger than that of men (Garbarino & Strahilevitz, 2004). Since community support is a multidimensional construct that is strongly aligned to trust components of perceived ability, benevolence, and integrity, all of which reduce perceived risk, the fact that it should emerge as the predominant influence on knowledge contribution (a risk behavior) for the study sample is therefore not entirely surprising. In light of this fact, the study findings progress the understanding of the factors that influence trust formation, engagement, and trust outcomes in OHCs that are particularly relevant for women. Nonetheless, it is interesting that engagement did not produce a stronger effect in relation to knowledge contribution for this sample. The explanation may lie in the fact that other factors specific to this sample may be inhibiting knowledge contribution. For example, Amichai-Hamburger et al. ( 2016) identified a number of psychological factors that may potentially influence individuals' lack of participation in online community discussions. These include individual differences, such as the need for gratification, personality dispositions, lack of time available, and self-efficacy, in addition to social group processes and technological issues. Additional issues such as introversion and social inhibition have also been shown to inhibit knowledge contribution. For example, Nonnecke and Preece (2001) found that nearly 30% of respondents were shy about posting and Rafaeli et al. (2004) found that those with high introversion scores tend not to actively engage in online groups. Confidence in having valuable information to contribute may also explain this outcome; Ray et al. (2014) found that the contributions of the most knowledgeable online community members do not derive purely from engagement but also from a competing sense of knowledge self-efficacy. Similarly, Preece et al. (2004) found that nearly one quarter of respondents explained their lack of participation in the online community in terms of having no knowledge to offer. Our post hoc assessments of differences across community types further reinforce the sensitivity of trust responses to context, with different antecedents showing greater importance in the different types of communities. For example, in this study, we compared two types of communities. The nutrition/ alimentation/ dietary (NAD) community places particular value on structured, precise, and timely information, while the pregnancy/breast-feeding/motherhood (PBM) community values experiential knowledge. We thus consider the former a more transactional type of community and the latter more relational. The study findings show that information credibility and engagement influence knowledge adoption for both types of communities, but they also show that in the case of NAD, community responsiveness directly affects knowledge adoption. In the case of this community type, structured, transactional aspects of community evaluation, such as information credibility and community responsiveness, influence the engagement decision. On the other hand, in the case of PBM, knowledge adoption is directly influenced by community support. Similarly, for this latter type of community, the decision to engage is influenced by less structured but more relational assessments, such as the evaluation of the community support level. Although the findings show that the strength of community support influences knowledge contribution outcomes for both communities, the behavioral response is stronger in the relational community than in the transactional community. These findings provide a particularly important contribution to the body of knowledge because they show that user engagement and active participation in OHCs, as manifested through the adoption or contribution of knowledge, are influenced by an assessment of the adequacy of specific types (transactional or relational) of information, which vary according to different types of health communities. A related contextual issue is the influence of user experience on engagement in OHCs. The findings of this study show this to be an evolving and phased dynamic with engagement, knowledge contribution, and adoption outcomes shifting according to increased user experience levels. For example, the findings show that in the initial phase of exposure to the OHC, the user's propensity to trust influences their decision to engage with the community, as manifested through knowledge contribution and adoption responses. However, as the user's experience with the community increases, that influence diminishes while the effect of community responsiveness on engagement grows. As the user's experience further increases, there is a shift toward a more informational, transactional perspective. In this more mature phase, it is utilitarian evaluations of community knowledge, such as information credibility and community responsiveness, that primarily sustain knowledge adoption. The motivation for contributing knowledge also changes in line with increasing levels of experience, becoming entirely sustained by community support. Our analysis of the mediating role of engagement by community type and user experience further reinforces the importance of context in understanding trust and engagement in the context of OHCs. The different forms of mediation between the PBM and NAD communities and across levels of user experience show the complexities of the influence of context. In doing so, we highlight the need for other scholars interested in understanding engagement in online communities to further theorize community types and experience levels in order to provide more granular insight into how the characteristics of their context influence user engagement and, in turn, shape behavioral outcomes. --- Implications for Practitioners The insights from our research provide practical guidance for social media practitioners interested in increasing participation and engagement in online communities, particularly communities that provide information or advice on sensitive issues, such as health information. First, the results clearly suggest that online community administrators should employ organizational mechanisms to increase user trust in the information provided by participants. This can be achieved through the inclusion of design features that allow participants to rate answers in terms of their helpfulness, thereby guiding users of the community to information that has been deemed credible by and useful to other users. Second, helpful answers should be made easily accessible to users through the provision of search options and Q&A design features. Utilizing design features that increase the speed of access to relevant, and helpful answers will in turn increase the perception of community support and responsiveness for users. The resultant increased engagement will strengthen the likelihood of users not only using that information but also contributing their own experiences, strengthening the norm of reciprocity that will increase the perception of community support and responsiveness. Similarly, the provision of design features that enable users to interact with community members who share similar backgrounds and experiences will influence their readiness to use the information provided and share information with others. The implications of the study findings have the potential to improve user engagement and result in more trusted and successful OHCs. As the mechanisms by which users adopt knowledge vary according to community type, moderators should tailor how knowledge is structured in a way that reflects the needs of their end users. In OHCs where the availability of precise, structured, and timely information is of the highest importance, this could be achieved through online community designers providing easily accessible drop-down search lists based on frequent word tags, which also show the date of provision of the response. However, in communities where social relationships are valued as much if not more than just factual information, website designers should provide links to "my experience testimonials" that are accessible on the basis of the type of information required. --- Limitations and Future Research This study provides insights that increase our understanding of the relationship between trust and engagement in OHCs, but as is the case with all studies, it also contains limitations. First, our results are based on a sample of respondents who are users of OHC websites in Brazil. Previous research has called for greater attention to the need for research in countries other than the US, UK, and Australia (Fan & Lederman, 2018).

Online health communities (OHCs) represent a popular and valuable resource for those seeking health information, support, or advice. They have the potential to reduce dependency on traditional health information channels, increase health literacy and empower a broader range of individuals in relation to their health management decisions. Successful communities are characterized by high levels of trust in user-generated contributions, which is reflected in increased engagement and expressed through knowledge adoption and knowledge contribution. However, research shows that the majority of OHCs are composed of passive participants who do not contribute via posts, thereby threatening the sustainability of many communities and their potential for empowerment. Despite this fact, the relationship between trust and engagement, specifically the trust antecedents that influence engagement in the OHC community context has not been adequately explained in past research. In this study, we leverage social capital behavior and social exchange theory frameworks in order to provide a more granular trust-based elucidation of the factors that influence individuals' engagement in OHCs. We collected data from 410 Brazilian participants of Facebook OHCs and tested the research model using partial least squares. The results confirm two new constructs-online community responsiveness and community support-as trust antecedents that influence engagement in OHCs, resulting in knowledge adoption and knowledge contribution responses. These findings contribute to the trust and engagement literatures and to social media research knowledge. From a practitioner perspective, the study findings can serve as an important guide for moderators and managers seeking to develop trusted and impactful OHCs.

the study findings have the potential to improve user engagement and result in more trusted and successful OHCs. As the mechanisms by which users adopt knowledge vary according to community type, moderators should tailor how knowledge is structured in a way that reflects the needs of their end users. In OHCs where the availability of precise, structured, and timely information is of the highest importance, this could be achieved through online community designers providing easily accessible drop-down search lists based on frequent word tags, which also show the date of provision of the response. However, in communities where social relationships are valued as much if not more than just factual information, website designers should provide links to "my experience testimonials" that are accessible on the basis of the type of information required. --- Limitations and Future Research This study provides insights that increase our understanding of the relationship between trust and engagement in OHCs, but as is the case with all studies, it also contains limitations. First, our results are based on a sample of respondents who are users of OHC websites in Brazil. Previous research has called for greater attention to the need for research in countries other than the US, UK, and Australia (Fan & Lederman, 2018). Our work thus addresses an important gap in the literature. Nonetheless, our sample also bounds the findings to some extent. While it is unlikely that national culture would fundamentally alter the dynamics that underpin the trust and engagement relationship, it is possible that culture may influence some aspects of trust formation. For example, a comparative analysis of the trust-based drivers of health disclosure (Lin et al., 2016) found evidence of different cultural emphases, and previous research by Gefen and Heart (2006) showed differences in trust formation and trust outcomes in individualist and collectivist cultures, albeit in an online transaction context. Consequently, it is possible that perceived information credibility may exert a higher trust formative influence on people from individualist cultures, whereas people from collectivist cultures may place greater weight on community responsiveness and knowledge contribution. Future research to test the generalizability of the study results by applying this framework to other national cultures can determine whether that is in fact the case. A second point worth noting is that our sample was predominantly comprised of women. Gender-related behavior is contextually influenced (Deaux & Major, 1987), and the OHCs (breastfeeding/pregnancy/ motherhood; beauty/aesthetics; and nutrition/diet) that form the contextual backdrop to this study are normatively skewed toward women, thus making women the predominant population of interest. Because these types of health information are typically of greatest interest to women, our sample is relevant for the context of our study and provides important insight into the specific factors influencing trust formation, engagement, and trust outcomes in OHCs, which are particularly important for those respondents. It does, however, bound the research findings, and future studies using more normatively neutral community types would enable greater opportunity for gender-based comparison. Similarly, while in this study we measure gender as a biological construct, future studies that include the effect of social, psychological, or cultural constructs of genderorientation could improve the understanding of gender differences in relation to online trust formation and engagement. For example, Hupfer and Detlor (2006) demonstrated the value of measuring specific self-concept traits that are associated with gender identity in relation to predicting web shopping site design preferences, rather than assuming their existence as a consequence of biological sex. Third, we focused on OHCs, which are characterized by the need for timely and accurate advice and where inaccurate information can result in very serious consequences for community members. In such a significant environment, the emotional, information, and network support provided by an online community may explain the strength of influence on knowledge adoption and knowledge contribution behaviors. Future research conducted in different (nonhealth) contexts would be beneficial in determining whether the strength of the relationships between trust antecedents, engagement, and trust outcomes remains the same, regardless of context type. Finally, in light of the finding that the mechanisms by which users adopt knowledge vary according to community type, future research could focus on chronic and acute health conditions to determine the role of medical conditions on knowledge adoption and knowledge contribution outcomes. Examinations of responsiveness that include an explicit recognition of different valences and measure their influence on engagement in OHCs also represent a valuable avenue for future research. --- Conclusion OHCs have the potential to positively impact healthcare outcomes through user value co-creation, but the way in which that value is achieved has received limited attention to date. This study empirically examines the factors that influence how individuals engage and cocreate value in OHCs. It extends existing theory through the inclusion and empirical testing of new variables that have received little attention as antecedents of trust in the OHC context: online community support and online community responsiveness. It also extends insight into trust formation by examining the predictive influence of these constructs on different trust responses as evidenced through engagement, knowledge adoption, and knowledge contribution. In doing so, it illustrates that different community attributes drive the formation of knowledge adoption and knowledge contribution responses in OHCs, and also reveals the different influence of engagement as a formation pathway for both of those responses. Finally, conceptualizing trust in terms of distinct trust antecedents and trust outcomes provides more granular insight into the cyclical relationship between trust and engagement. Our findings contribute both to the trust and engagement literatures and to social media research knowledge. From a practitioner perspective, the study findings can serve as a guide for moderators and managers seeking to develop trusted and impactful OHCs. --- When trust responses are examined in the context of the spectrum of user experience (Table 8), it is evident that knowledge adoption is consistently influenced by information credibility (<unk>1-6M = 0.26, p <unk> 0.05; <unk>6M-2Y = 0.38, p <unk> 0.001; <unk>2+Y = 0.29, p <unk> 0.05). However, the relationship between other factors and knowledge adoption is more variable as the user acquires greater experience of OHCs. For example, engagement shows an influence on knowledge adoption (<unk>1-6M = 0.22; p <unk> 0.10 and <unk>2+Y = 0.35; p <unk> 0.001) in the early stages of user experience, as does community support (<unk>1-6M,<unk>2+Y = 0.36, p <unk> 0.04). However, after this initial period, community responsiveness emerges as the dominant factor influencing knowledge adoption (<unk>2+Y, <unk>1-6M = 0.36, p <unk> 0.02). A similar change in influence applies to knowledge contribution. In a context of limited experience, it is initially influenced by engagement (<unk>1-6M, <unk>2+Y = 0.30, p <unk> 0.06) and the propensity to trust (<unk>1-6M, <unk>2+Y = 0.38, p <unk> 0.06). However, as the user's experience increases, the influence of community support also increases until it becomes the most influential factor (<unk>1-6M, <unk>2+Y = 0.45, p <unk> 0.07). Figures 3,4, and 5 illustrate how the evolution of user experience influences engagement, knowledge adoption, and knowledge contribution respectively. Copyright <unk> 2023 by the Association for Information Systems. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and full citation on the first page. Copyright for components of this work owned by others than the Association for Information Systems must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists requires prior specific permission and/or fee. Request permission to publish from: AIS Administrative Office, P.O. Box 2712 Atlanta, GA, 30301-2712 Attn: Reprints, or via email from [email protected].

Online health communities (OHCs) represent a popular and valuable resource for those seeking health information, support, or advice. They have the potential to reduce dependency on traditional health information channels, increase health literacy and empower a broader range of individuals in relation to their health management decisions. Successful communities are characterized by high levels of trust in user-generated contributions, which is reflected in increased engagement and expressed through knowledge adoption and knowledge contribution. However, research shows that the majority of OHCs are composed of passive participants who do not contribute via posts, thereby threatening the sustainability of many communities and their potential for empowerment. Despite this fact, the relationship between trust and engagement, specifically the trust antecedents that influence engagement in the OHC community context has not been adequately explained in past research. In this study, we leverage social capital behavior and social exchange theory frameworks in order to provide a more granular trust-based elucidation of the factors that influence individuals' engagement in OHCs. We collected data from 410 Brazilian participants of Facebook OHCs and tested the research model using partial least squares. The results confirm two new constructs-online community responsiveness and community support-as trust antecedents that influence engagement in OHCs, resulting in knowledge adoption and knowledge contribution responses. These findings contribute to the trust and engagement literatures and to social media research knowledge. From a practitioner perspective, the study findings can serve as an important guide for moderators and managers seeking to develop trusted and impactful OHCs.

INTRODUCTION Several contagious diseases infected a lot of people in past and continuing their effect in future such as AIDS (Acquired Immune Deficiency Syndrome), plague, influenza, cholera, Ebola and SARS (Severe Acute Respiratory Syndrome) (Cutsem et al., 2016;Kupferschmidt, 2016). Among these knowing diseases, in the late 2019 another disease was introduced named as, COVID-19. According to (WHO, 2020) Coronavirus was early named as, <unk>2019 novel coronavirus<unk> now officially named COVID-19. The virus is just like SARS (severe acute respiratory syndrome) in which the respiratory infection takes the form of pneumonia but it is actually COVID-19. Investigated to be an infectious disease that transfers through saliva excretions or droplets of sneezing or coughing from the affected to the normal people. Until now, no devised vaccine or specified medicine investigated to prevent its dispersion more effectively among the public. However, the vaccine helps to lessen the effect of the virus if an individual is infected (ElBagoury et al., 2021). Thereby, according to the situation, less physical contact as well as to keep oneself as clean as possible is also a remedy to avoid its widespread. During these circumstances, it is thereby challenging for people to work while being at their workplace (Haleem et al., 2020). Organizations cannot stop working while waiting for the unknown period of the continuity of this disease. They need to manage their work from a place with less physical interaction just for the sake of everyone's safety. Therefore, the possibility to work from outside the work territory entailed the significance of work from home. Although work from home is an appropriate facilitation to employees in their workplace due to some of the circumstances. In this arrangement, employees can dedicate their duty hours outside the boundaries of the organization. It provides a balance between employee work and family life (Christensen et al., 2013). Employees can work in any mobile place, either home or can be any place (Venkataraman et al., 2018). It is for the well-being of the employee and therefore, work from home facility helps to retain the employees (Rodwell & Martin, 2013). Moreover, it is beneficial for organizations to strategize their policies more effectively. Therefore, efficiency is also increased (Jansen & Hlongwane, 2019). Working in a safe and sound environment enables the employees to augment their productivity Christensen et al., (2013) and this enhancement plays significance in employee personal and professional life. According to Taamneh et al. (2018), internal satisfaction of employees from work leads them towards organization well-being. The debate about those practices which are employee friendly and that help to balance employee work and family life (Shakir, 2019). With these facilitations, employees found themselves more empowered to work, hence these relaxations at work increase their loyalty. Sometimes negating the idea to facilitate the employee to work out of the boundaries of work results in lack of loyalty (Kim & Lee, 2019). On the other hand, increase in efficiency and effectiveness of employees are also identified (Jansen & Hlongwane, 2019). This study involved a situation in which the virus is infectious and influencing many of the population worldwide (Haleem et al., 2020). While, giving importance to the individuals who need to go out either to work or for pray, is one of the main points of discussion nowadays. With the help of game theory, it has been made easy and understandable that to avoid circumstances, we need to strategize our plans and routines (Begley et al., 2020). Such as it is a zero-sum theory for rational decision making in which every participant of the game or the situation has the same benefit and loss as compared to the others. Referring to the game theory, individuals change their behaviours according to situations, to avoid the worst circumstances. For example, in this COVID-19, the only solution is social distancing to avoid oneself in contact with the infected ones. Some of them, who are not understanding the situation same time, will soon change their behaviour after observing those who are taking precautionary measures and are staying safe from the pandemic. Social distancing is one of the solutions to save oneself in a situation like infectious or viral diseases. Therefore, from the above discussion, the following points and questions are raised that are entailed to be discussed in this study: 1. Effects of COVID-19 in the organizational context on how it is affecting the employers and employees. 2. Does social distancing is effective to avoid the infectious disease? 3. What is the Islamic perspective behind this? --- LITERATURE REVIEW --- Effect of Coronavirus on the Population COVID-19, a pandemic disease, escalated like fire in the jungle. Its effectiveness is increasing day by day (Asrani et al., 2021). According to WHO (2020), the pandemic started on 31st December 2019 and it is continued with its second wave (Xu & Li, 2020). The pandemic with a sudden push infected around 428 million population worldwide and the latest number of deaths reached up to 5.19 million WHO (2021). The affected countries from different territories across the border were infected by this disease. This coronavirus named COVID-19 caused disruptions in the social life of the individuals living in that specified affected territories. With the effectiveness of the virus, it caused an effect on 1/3 of the world population (Begley et al., 2020). Asserting the influence on the population, quarantine was suggested that is, a state of being isolated at a place due to hygienic reasons and to avoid contact to several epidemics and infectious diseases (Altinoz et al., 2012). The prevention to spread the disease is the best remedy for infectious and viral diseases (WHO, 2020). Thereby, to provide safety from the affected ones to the unaffected, self-isolation is the best remedy. Among the seven continents of the world, Asia is the largest continent with a huge figure of the population (Moen et al., 2017). Discovered as an initiator of the COVID-19 due to the number of population and tourism. North America, although ranked fourth in the populated country's list, recent surveys found the maximum number of cases in the United States (WHO, 2020). Europe being on the third level in the most populated continent (Moen et al., 2017) and is the majority of countries facing the maximum number of cases (WHO, 2020). The United States to date is the most effected territory with the highest number of cases 104,671, while Italy is on the second number of positive cases 86,498 has the highest number of fatalities 9,134. China, who initiated the virus now is at the highest recovery rate of 74,971 patients. Other countries like, Spain, Germany, France are also with the high rate of positive cases. Iran found to be with the highest rate of fatalities and continuing with the positive cases among the Muslim countries. --- Social Distancing Distancing is the isolation and hostility, and social distancing is to make a specified distance in view to disassociate oneself with others. Individual's level of closeness within the same or different groups in their social association (Arenas et al., 2004) due to differences in age, race, ethnicity, religion, culture or maybe gender-based (Elaine, 2008). The tenderness and warmth felt by them within a similar group or with those who are the intruders (Bogardus, 1947). According to Robert E. Park (1924), social distancing is the diminution in the understanding, involvement or intimacy that is present in an individual or collective relationship. --- Globally Infected Muslims Among all, Muslim ummah is also suffering from this chronic disease. Several countries having a huge spread of the disease, with the lack of knowledge, precautions, and resources. A major spread was on the Friday prayer, where most the Muslim were gathered, took the hype of the disease, Countries included who were affected after those gatherings are, Malaysia, Lebanon, Turkey, Egypt, Iraq, Iran, Jordan, Sudan, and Saudi Arabia (Dwyer, 2020). Iran, found to be the highest and most effected country among Muslims, with 21638 to date cases and several deaths (WHO, 2020). Other countries with Muslim majority affected by viral diseases such as, Pakistan, Qatar, Saudi Arabia, Bahrain, Egypt, Iraq, Lebanon, Kuwait and also the United Arab Emirates. The widespread of this among Muslims was due to religious gatherings (Emont & Shah, 2020). These gatherings caused a severe dispersion of the virus and the migrants of the other countries spread the virus to those host countries. Workplaces, offices, shops and every kind of business were suspended, even the air service was also disconnected by the affected countries to avoid the circumstances from getting worse. Globally, organizations are pursuing to work from home facilitations to their employees just to avoid the country level situation COVID-19. People were required to stay at home and to do their work from the home (Lufkin, 2020). Even though the major worship holly places like, the Grand Mosque of Mecca and the Prophet's (SAW) Mosque of Medina were closed for the public to minimize their contact (AFP, 2020). --- Theory of Maqasid Al-Syariah This study relates to the theory in the perspective of taking care of oneself and others by preventing the disease through getting vaccine and other precautionary measures. Theory of Maqasid Al-Syariah focuses on the betterment of society, it explains that the purpose of Shariah is happiness and security (Bahri et al., 2019). It shows the Maslahat of humanity where only the social benefit is important. It ascertains the individuals gain economic and social benefit with no harm to others, thus creates a balanced environment (Amaroh & Masturin, 2018). The theory also ensures that the followers are protected from each side, i.e. their health, money, property and community are protected to ensure welfare of the society (Nordin et al., 2017). --- Islamic Perspective on Health Safety There are some verses from the Quran teaching the sense of self-caring and safety. <unk> "Everything good that happens to you (O Man) is from God, everything bad that happens to you is from your own actions". (Quran 4:79). It is an assessment from Allah, so people who will save themselves, take precautionary measures will be saved from the COVID-19 disease. Those who ignore the safety measures and keep socializing themselves will be the suffered ones without any doubt. <unk> "O mankind: Eat of what is lawful and good on earth" (Quran 2: 168). This verse of the Quran defends the prohibition of eating non-halal food, which can be dangerous for health. The non-halal food impedes several kinds of germs which are dangerous and somehow poisonous to the human body. COVID-19 found to be originated in the non-halal food market (Jewell, 2020) and propagated to others who are not in consumption to that. --- METHODOLOGY Methodology of the paper explains the summary of research design. Data has been gathered by collecting the information from the recently published research articles. Researchers collected those specific articles with the content related to global pandemic and its effect on the community. Along with that, the online published articles were considered with up-to-date effects of pandemic COVID-globally. The researchers searched the secondary data form the different journals with the help of google scholar, that helped to find the most recent and relevant articles from Elsevier, science direct, MDPI, WHO and research gate. The researchers also gathered the information for recent statistics from world health organization and bureau of statistics. --- RESEARCH IMPLICATIONS --- Benefits of Work from Home Thus, work from home incorporates benefits to both parties i.e., employers and employees. --- Benefits to Employers There are several studies which have discussed the work from home facility to the employees (Agus & Selvaraj, 2020;Beauregard, 2011;Fuller & Hirsh, 2019;Germeys et al., 2019;Hytter, 2007;Judge & Ilies, 2004;Ke & Deng, 2018;Mauno & Ruokolainen, 2017;Mayo et al., 2016;McNall et al., 2009;Russell et al., 2009). Researchers identified some of the benefits to the organizations by the practice work from home. a. Operating Cost. Offices, normally catering to several people working as employees. They used to work normally for 9 hours, accommodating themselves under these working hours. Utilization of office stationery, lots of documents, electricity and other utilities somehow make a huge expense towards the organization. Moreover, the rental incurred by the office place. Working from home, helps to reduce the cost of this work operating cost. b. Productivity. Increase in productivity when employees feel more relaxed while working because of a sense of autonomy and change in the working environment. Employees feel more comfortable and work more proficiently, hence, lessening their work anxiety and stress. Thereby, increasing workability and productivity. c. Turnover. Reduced turnover from several studies was investigated if the employees are satisfied with their jobs. This satisfaction leads them towards staying with their jobs. Flexibility at the workplace helps to intensify their mode of happiness and joy, removing stress, lessen the number of leaves. Turnover is itself a huge cost for the employers, if employees are leaving so frequently it will cause, bad word-of-mouth for the prospective employees, also the time and money cost associated with hiring, the overall recruitment activity will be incurred by the employers. d. Behavioural Positivity. Leading towards the feeling of freedom of workplace, employee freedom of workplace, employers take benefit in the form of a positive attitude of employees. e. Impact on the Environment. Reducing transport to travel, reduces the fuel cost and the associated environmental cost therewith. Communicating via electronic media, the employees need no traveling, hence, the environmental benefit is also on the employer's side. --- Benefits to Employees According to previous studies on work from home (Allen et al., 2016;Amorim & Santos, 2017;Fuller & Hirsh, 2019;Mansour & Tremblay, 2018;Sharma & Yadav, 2019), employees also benefit if they are allowed to work from their place such as the following: a. Work-life Balance. It creates a balance between employee work and family life. When employees are provided with deadlines, they schedule their work timing accordingly to complete their tasks on the deadline. It makes an easy for them to take care of their dependents as well as to do their work. b. Productivity. Increases as the employee are working under less pressure of the workplace environment. c. Stress. Lessening of stress by the variance of tasks associated with work and life. Employees as family members take care of the family, as well as working for the job. This balance of life makes him or her mentally relaxed and therefore, reduces stress. d. Less Cost. Staying at home, not traveling every day to the workplace, sometimes more far from home, reduces the time and money costs associated with traveling. e. Satisfaction. More if an individual is working in a family environment and flexible hours. This job autonomy gives the person more satisfaction with the job and family life. f. Less Health Hazard. If someone at the workplace is suffering from an infectious disease, it will be transmitted to the persons working in the surroundings. Work from home can reduce this effect. --- Limitations of Work from Home Work from home seems like a perfect and ideal situation (Madell, 2019) in which an employee will not be having any trouble getting ready, work in an office environment, working under supervision, and fuel-saving (Bussing, 2019). But, along with the benefits associated with both employers and employees, it also has some limitations as described below. a. Lack of Productivity. Without supervision, sometimes it becomes difficult to manage workrelated tasks (Michelle Kiss, 2019). It is thereby, investigated that while sitting out of the boundaries of the workplace, it is difficult for the employees to work efficiently and effectively. b. Absence of concentration. A balance between work and family life disturbs if an employee is somehow not very compatible to create a balance between both. This time mismanagement may thereby trigger a lack of concentration (Madell, 2019). c. Telecom cost. Being online and connected with the colleagues and subordinates may also cause the telecom cost to the employees and the employer. This stays connected practice can cause a huge amount of bills to the parties involved (Michael Hurd). d. Less incorporation. Coordination can also be subordinated if the employees do not face to face connected. Sometimes some issues cannot be handed over without the presence of the relevant staff, causes disruptions in the services (Michael Hurd). --- CONCLUSION COVID-19, a global catastrophe, calamity started from a wet market of Wuhan, China (Citroner, 2020) affected up to 100 countries worldwide (WHO, 2020). Including among the Muslim community, which are those infected individuals who had direct contact with the infected ones in religious gatherings. Muslim worship places with restricted time before but now suspended their operations until the situation gets better (Soumaré & Crétois, 2020). Thousands of Muslim migrants who were there for worship in Iran were infected by this virus (Emont & Shah, 2020). Regarding the situation, there are many controversies by religious bodies. Such as, the worships especially, the Mosque of Makkah and the Prophet's (SAW) Mosque of Madinah should be always operational until the day of judgment. Taking the point towards practicality and avoiding the current scenario to get it to the worst situation, it became mandatory to avoid those places in which people are gathered and are in direct contact with each other. Work from home, practice intended for employees to work from home or remote place. This facilitation is nowadays implemented due to chronic disease and infectious nature triggered many casualties. To avoid the circumstances to get more miserable, it is thereby executed by most of the companies worldwide, Microsoft, Google, Twitter, Spotify, Hitachi, Apple, Amazon and Chevron (Lufkin, 2020). Thereby, less probability or occurrence of getting infected. It can be elaborated as if an employee suffering, others can also be transmitted by casual contact while working together. Covid-19, also specified as a precaution to stay at home, intensifying the possibility of work by staying at home. Considering all the facts, COVID-19 propagating its effects day by day, from one to several individuals who are in contact. It can only be avoided by making the distance in a social aspect. Islam also taught to save oneself and others from loss of health. Theory of Maqasid Al-Syariah also intends towards the safety of health and life of individuals. This can only be possible when they avoid the circumstances to get infected by the virus. In the social circumstance, to get vaccinated, maintaining social distancing, wearing masks, washing hands are some of the ways to avoid the situation to get worse. In the organization perspective, flexible work practices are the other way to avoid the spread of this virus. Flexible work practices facilitate people working in the organizations by providing flexibilities in working hours, working day etc. Hence, there is a possibility to work and to pray from home to avoid and control the situation. Although, these are not permanent solutions but also the situation will not sustain for a prolonged time. So, for the betterment of the individuals is to keep distance, keep connected, keep working and keep praying at home.

Contemporary worldwide disaster COVID-19 encountered by people seemed to be widely spread like a fire in the jungle. This surge impacted the lives of many, even caused fatalities. Staying at home and not to go outside is one of the hot topics of the moment. The population around the world found a major lapse when it was noticed that the epidemic COVID-19 is infecting both Muslims and non-Muslims. Meanwhile, due to this worldwide disaster, countries facing the issue altered their work status into work from home. This study was conducted by collecting the readily available secondary data from previous studies and recently published articles on the several worldwide pandemics. This study discussed the effects of COVID-19 on the population especially, organizations with the practice of work from home. It also discussed the effectiveness of social distancing to avoid the spread of this virus. Moreover, this study discussed the benefits of work from home along with enlightening the Islamic perspectives to strengthen the study.

Social networks are critical sources of informal support, especially for older adults. Informal social support is important for coping with a range of social issues including physical and mental health problems (Cohen, Brittney, & Gottlieb, 2000) and daily life stressors (Benin & Keith, 1995). For instance, social support is linked to higher levels of overall well-being (Nguyen, Chatters, Taylor, & Mouzon, 2015;Smith, Cichy, & Montoro-Rodriguez, 2015) and lower rates of serious psychological distress (Chatters, Taylor, Woodward, & Nicklett, 2015;Gonzalez & Barnett, 2014), depression (Fagan, 2009), and social anxiety disorder (Levine, Taylor, Nguyen, Chatters, & Himle, 2015). Social networks also provide instrumental assistance such as financial assistance, household work and transportation (Sarkisian & Gerstel, 2004). Research universally indicates that belonging to a supportive social network is critical for the healthy functioning for older adults (Cacioppo & Cacioppo, 2014). The present study examines social network types and aims to determine (a) whether the probability of belonging to specific social network typologies varies by race and ethnicity among African American, Caribbean black, and non-Hispanic white older adults and (b) whether race/ethnicity interacts with sociodemographic characteristics, such as age and education, to influence the probability of belonging to specific network types. Although no study to date has systematically investigated racial and ethnic differences in social network types among older individuals in the United States, prior research verifies racial differences in key social network characteristics. The following sections of the literature review discuss prior research on social network typologies, along with background research examining race and social networks among older adults. This is followed by a section describing research on the social networks of Caribbean blacks. The literature review concludes with a discussion of the focus of the present investigation and study hypotheses. --- Social Network Types Social network types are a growing area of research on social relationships that uses an innovative, person-centered approach to examine varied configurations of social network characteristics (Fiori, Smith, & Antonucci, 2007;Li & Zhang, 2015;Litwin, 2001;S. Park, Smith, & Dunkle, 2014;Wenger, 1996). Information on network characteristics (e.g., network size and composition, frequency of contact) is aggregated to identify distinct typologies or profiles of social networks. In synthesizing findings from this research area, a number of distinct network types have been identified. Studies of social networks typologies (e.g., Fiori, Antonucci, & Cortina, 2006;Litwin, 2001;Wenger, 1996) have consistently identified four general, archetypal network typologies: diverse, family-focused, nonkin-focused, and restricted. Network types that belong to the diverse type are characterized by high levels of social integration and different network role composition. In contrast, network types belonging to the archetypal restricted network typology are characterized by high levels of social isolation. Network types belonging to the family-focused and nonkin-focused archetypal typologies are characterized by high levels of integration within almost exclusively family and nonkin networks (e.g., friends, congregants, neighbors), respectively. Diverse and nonkin-focused are the most prevalent typologies identified among general samples of the American population, while the family-focused type is the least prevalent (Fiori, Antonucci, & Akiyama, 2008;Fiori et al., 2006;Shiovitz-Ezra & Litwin, 2012). Network types identified in international samples are also closely aligned with these four general archetypal typologies (Burholt & Dobbs, 2014;Cheng, Lee, Chan, Leung, & Lee, 2009;Doubova, Pérez-Cuevas, Espinosa-Alarcón, & Flores-Hernández, 2010;Fiori et al., 2008;Li & Zhang, 2015;Litwin, 2001;N. S. Park et al., 2013N. S. Park et al.,, 2014)). Overall, studies of non-US samples indicate that the most prevalent network types were the family-focused and diverse types, while the least prevalent types were the nonkin-focused and restricted types. The prevalence and distribution of these specific network types is consistent with an emphasis on strong family orientation and values (i.e., familism, filial piety) that is characteristic of several of these cultures (Baca Zinn, 2000;Ikels, 2004;Lin, 2013;Mucchi-Faina, Pacilli, & Verma, 2010;I. H. Park & Cho, 1995). The body of evidence of cultural differences in derived network typologies suggests that the distribution of social network types within the US population may also differ by race and ethnicity. Although race/ethnic differences in network types among older Americans have yet to be explored, there is a body of research examining race differences in social networks. --- Negative Interaction With Network Members Research on support typologies generally does not include measures of negative interaction such as criticisms and conflict. The inclusion of measures of negative interaction in this study represents an important innovation that more accurately reflects interactions within support networks. Negative interactions are a natural feature of social life and are a fairly common occurrence among family members (Rook & Ituarte, 1999), as well as church members (Krause & Batisda, 2011). Further, distinctive support network characteristics are associated with negative interactions. For example, more frequent interaction with support network members (Lincoln et al., 2013), as well as circumstances in which extensive support is provided (Newsom & Schulz, 1998) are both associated with negative interactions. Overall, negative interactions do not occur as frequently as positive emotional exchanges. Nonetheless, they are associated with emotional distress including clinicallyrelevant mood and anxiety psychiatric disorders (Lincoln et al., 2010). Consequently, we would expect that as a common and central characteristic of social life, negative interactions will emerge as an important feature of network typologies. --- Race Differences in Social Networks Over the past 50 years there has been considerable debate (both theoretical and empirical) as to whether African American families can be characterized as being stable, disorganized or reflective of alternative patterns of family (Allen, 1978;Sarkisian & Gerstel, 2004). Early theories in this area characterized African American families in a largely negative manner as being deficient and dysfunctional or they have idealized the family networks of typically poor African Americans. Further, this area of research has, either explicitly or implicitly, used non-Hispanic white families as the comparison group for African American families. Researchers, for the most part, now generally accept the view that African American and white families are different and these differences are not solely a reflection of social class (Gerstel, 2011;Sarkisian & Gerstel, 2004;Taylor et al., 2015). In particular, Allen (1978) uses the term culturally variant to explain these differences in family structure and function. He argues that, unlike a deficit perspective on African American families, a culturally variant perspective does not view differences in African American families as indicators of pathology or deficiencies. Instead a cultural variant perspective acknowledges that African American and white families exist in different social and cultural environments and, as such, these differences are manifested in a variety of family indicators. Research and theories on racial differences in support networks further argue that, in order to fully appreciate support network structure and processes, it is important to extend our conception of families beyond nuclear to extended families (Gerstel, 2011), as well as to investigate both kin and nonkin as members of social networks (Taylor et al., 2015). Despite the importance of this issue, there is a paucity of research on racial differences in social networks. Further, the available research on racial differences in social network characteristics among older African Americans and whites is equivocal in relation to network composition. Some studies indicate that older African Americans have smaller networks than older whites (Antonucci, Ajrouch, & Birditt, 2006;Barnes, Mendes de Leon, Bienias, & Evans, 2004;Magai et al., 2001), whereas others indicate no racial differences in network size (Mendes de Leon, Gold, Glass, Kaplan, & George, 2001). Research on social involvement (e.g., contact, supportive exchanges) with network members is also mixed, with some studies reporting higher involvement among older African Americans, especially within the extended family network (Antonucci et al., 2006;Johnson & Barer, 1995;Peek, Coward, & Peek, 2000), while others indicate higher involvement among older whites (Mendes de Leon et al., 2001). Finally, two recent studies indicate that older African Americans are more likely to live in extended family households (US Census Bureau, 2014) and are more likely to be involved in church support networks (Krause & Batisda, 2011;Taylor et al., 2015). There is a limited amount of research on racial differences in the quality of relationships with in support networks. African Americans have more interaction with congregation members and more negative interaction with church members (Krause & Batisda, 2011;Taylor et al., 2013). Taylor and colleagues (2013) did not find any differences in subjective closeness to friends or subjective closeness to family between Caribbean blacks, African Americans, and non-Hispanic Whies. Sarkisian and Gerstel (2004) argue that in terms of understanding kinship networks, demographic diversity within African American and non-Hispanic white populations (e.g., age) may be more important than racial differences. Therefore, it is important to examine whether race and various sociodemographic characteristics (e.g., age, education) in combination have an interactive influence on social network characteristics. For example, Ajrouch, Antonucci, and Janevic (2001) found that African Americans reported more kin in their networks than whites, but this difference was attenuated among older persons. Antonucci and colleagues (2006) found that African Americans with higher levels of education had more extended family members in their networks, while whites with higher levels of education had fewer extended family members in their networks. For our present investigation, these studies suggest that race may interact with particular sociodemographic characteristics in shaping the network typologies of older adults. --- Caribbean Blacks in the United States Despite the rapid expansion of older minority populations in the United States, most gerontology research focuses on the general, non-Hispanic white population. Further, research comparing the social networks of African Americans and Caribbean blacks and Caribbean blacks and non-Hispanic whites is extremely limited. The current investigation seeks to address this major gap in knowledge and provide a better understanding of race and ethnicity influences on social networks by including Caribbean black older adults in this study. Both family and nonkin are important sources of support for Caribbean blacks, particularly during the migration process (Taylor, Forsythe-Brown, Lincoln, & Chatters, 2015;Taylor et al., 2013). Upon arrival in the United States, social networks provide an array of support to Caribbean blacks including assistance with housing, employment, and legal documentation (Basch, 2001;Bashi, 2007). In preparing for migration, individuals rely on their extended families to fund their migration. In the postmigration period, extended family members in the home country provide care (i.e., child fostering) for children who stay behind (Bashi, 2007;Waters, 1999). Religious institutions are important community and cultural resources that assist Caribbean blacks in their migration to and settlement in the United States. Immigrant churches provide tangible social support and fellowship, serve as a cultural repository and broker (Taylor, Chatters, & Jackson, 2007), and provide access to services from clergy to manage life problems (Taylor, Woodward, Chatters, Mattis, & Jackson, 2011). In sum, Caribbean blacks rely on assistance from a diverse group of support resources comprised of kin and nonkin, both in the United States and in their home countries. There are substantial differences between Caribbean blacks and African Americans in life circumstances (e.g., family structure, immigration status) and culture (Taylor et al., 2013). However, these two groups are rarely identified as representing distinct ethnic groups, but rather are seen collectively as black American. Unfortunately, only a few race comparative studies of social networks recognize these ethnic variations within the black population and their relevance for support network structure and functioning. The broad characterization of African Americans and Caribbean blacks as constituting one undifferentiated group (i.e., black Americans), is both inaccurate and problematic for developing a well-defined understanding of potential ethnic differences in social network characteristics. Research focusing on social network typologies among racially and ethnically diverse older Americans are necessary for the development of social network interventions that are culturally sensitive and relevant. --- Focus of the Present Study The present study addresses a critical gap in knowledge on racial/ethnic differences in social network types among older adults in the United States by assessing whether the probability of belonging to specific social network types varied by race among older African Americans, Caribbean blacks, and non-Hispanic whites. I expect that the four general archetypal network typologies that have been identified in previous studies discussed in the literature review-diverse, family-focused, nonkin-focused, and restricted-will be identified (Hypothesis 1). This analysis includes measures of negative interaction as social network typology indicators. Accordingly, two subtypes of the diverse and nonkinfocused network types (the most prevalent types identified in previous work using US samples) are also expected (Hypothesis 2). Specifically, the analysis will yield a positive diverse subtype (i.e., with low levels of negative interaction) and a negative diverse subtype (i.e., with high levels of negative interaction). Similarly, a positive nonkin-focused subtype as well as a negative nonkin-focused subtype will be identified. Moreover, the probability of being in a particular social network type will vary by race/ethnicity (Hypothesis 3). Specifically, Caribbean blacks, who are often characterized as having transnational family ties (e.g., family geographic dispersion), will rely heavily on extended family, friends, and congregants for assistance. Thus, Caribbean blacks will be more likely than whites to belong to the diverse type. Moreover, it is anticipated that given the centrality of the extended family among African Americans, they will be more likely than whites to belong to the familyfocused type. Additionally, this study examines how age and education interacts with race to influence the probability of belonging to certain network types. Prior literature indicates that the effect of race (i.e., comparing African Americans and whites) on network composition varies by educational attainment and age (Ajrouch et al., 2001;Antonucci et al., 2006), specifically in relation to the presence of kin in the network. Accordingly, the associations between education and social network types and age and social network types are expected to vary by race and ethnicity (Hypothesis 4). --- Method --- Sample The study sample was drawn from the National Survey of American Life (NSAL) conducted by the Program for Research on Black Americans at the University of Michigan's Institute for Social Research. The African American sample is the core sample of the NSAL. The core sample consists of 64 primary sampling units, of which 56 overlap substantially with established national sampling areas. The remaining eight primary areas are located in the South, ensuring the sample represents the national distribution of African Americans. The African American sample is a nationally representative sample of households located in the 48 coterminous states with at least one black adult aged 18 or older who did not report ancestral ties in the Caribbean. The NSAL also included the first major probability sample of Caribbean blacks in the United States. This study used a subsample of respondents aged 55 and older, featuring 837 African Americans, 298 non-Hispanic whites, and 304 blacks of Caribbean descent. --- Measures Race and Sociodemographic Variables Race/ethnicity was coded as African American, Caribbean black, or non-Hispanic whites. For the purpose of this study, Caribbean blacks were defined as individuals who trace their ethnic heritage to a Caribbean country but now reside in the United States, are racially classified as black, and speak English (but may also speak another language). Control variables included gender, family income, marital status, parental status, and living arrangement. Gender, parental status, and living arrangement were dummy coded (0 = male, 1 = female; 0 = nonparent, 1 = parent; 0 = does not live alone, 1 = lives alone). Age, education, and family income were scored continuously; age and education were assessed in years. Family income was coded in dollars and log transformation was used to minimize variance and account for its skewed distribution. Missing data for income and education were imputed using an iterative regression-based multiple imputation approach incorporating information about age, sex, region, race, employment status, marital status, home ownership, and nativity of household residents. Marital status was coded to differentiate respondents who were married or partnered, separated, divorced, widowed, and never married. --- Social Network Type Indicators Indicators for family, friendship, and church networks were based on respondents' perceptions of their relationships and were used to identify network types. Frequency of contact with family was measured by asking: "How often do you see, write or talk on the telephone with family or relatives who do not live with you?" Possible responses ranged from 1 (never) to 7 (nearly every day). Subjective closeness to family was assessed by asking: "How close do you feel towards your family members?" Response categories ranged from 1 (not close at all) to 4 (very close). Emotional support from family was measured by asking: "Other than your (spouse/partner), how often do your family members: (a) make you feel loved and cared for, (b) listen to you talk about your private problems and concerns, (c) express interest and concern in your well-being?" Negative interaction with family members was assessed by three questions: "Other than your (spouse/partner) how often do your family members: (a) make too many demands on you, (b) criticize you and the things you do, and (c) try to take advantage of you?" Response categories for emotional support and negative interaction questions ranged from 1 (never) to 4 (very often). Similar questions and response options were used to measure frequency of contact with church members and friends, subjective closeness to church members and friends, emotional support from church members, and negative interactions with church members. To facilitate analysis and interpretation of results, all indicators were dichotomized using median split. --- Analysis Strategy Latent class analysis (LCA) was used to identify network types. LCA uses a person-centered approach to classify respondents into subgroups (i.e., latent classes) based on response patterns across dichotomous class indicators. Latent class multinomial logistic regression analysis, in which class probabilities are regressed on sociodemographic variables, was used to determine correlates of network types. This was conducted using the three-step LCA approach to avoid the inclusion of sociodemographic variables in the class extraction process. To determine whether the effects of race on network type varies by education and age, two interaction terms (Race <unk> Education and Race <unk> Age) were constructed and tested in latent class multinomial logistic regression models. All analyses used sampling weights and accounted for the complex multistage clustered design of the NSAL sample, unequal probabilities of selection, nonresponse, and poststratification to calculate weighted, nationally representative population estimates and standard errors. --- Results --- Social Network Types (Hypothesis 1 and 2) LCA indicated that the best-fitting model featured four classes. Goodness of fit was determined using the AIC and sample-size-adjusted BIC. The four identified network types were: optimal, family-centered, strained, and ambivalent (Supplementary Figure 1). The optimal type, which was most prevalent (30.36% of the sample), had high levels of subjective closeness, contact, and emotional support involving family and church members, and low levels of negative family and church interactions. Moreover, these respondents reported high levels of subjective closeness and contact with friends. The ambivalent type, the least prevalent (19.09% of the sample), was similar to the optimal type, with the exception that respondents in this network typology reported high levels of negative family and church interactions. The family-centered network typology (30.15% of the sample) featured high levels of subjective closeness, contact, and emotional support involving family members and low levels of negative family interactions. Additionally, members of this class reported low levels of subjective closeness, contact, emotional support, and negative interactions involving church members and low levels of subjective closeness and contact involving friends. Finally the strained type (20.39% of the sample) featured low levels of subjective closeness and contact with family, church members, and friends, coupled with low levels of emotional support from family and church members. Further, respondents in the strained type indicated moderate levels of negative family interactions and low levels of negative church interactions. Race/Ethnicity and Social Network Types (Hypothesis 3 and 4) The distribution of network types across the three racial/ ethnic groups (Table 1) indicated that for African Americans and Caribbean blacks, the optimal type was most prevalent (35.12% of the African American sample; 33.12% of the Caribbean black sample), and the strained type was the least prevalent (18.24% of the African American sample; 17.65% of the Caribbean black sample). In contrast, among whites, the family-centered type was the most prevalent (35.62% of the subsample) and the ambivalent type was the least prevalent (15.5% of the subsample). Results from the latent class multinomial logistic regression analysis (using optimal type as the reference category) did not yield a significant association between race/ethnicity and network type (Table 2). However, interactions between race/ethnicity and education and age were statistically significant. The interaction between education and race/ethnicity indicated that among respondents with lower levels of education, Caribbean blacks had a substantially greater probability of belonging to the ambivalent type (high integration and negative interaction) compared to whites (Figure 1). For whites, as education increased, the probability of belonging to the ambivalent type marginally increased. However, for Caribbean blacks, an increase in education was associated with a substantial decrease in the probability of belong to the ambivalent type. Thus, at the highest education level, Caribbean blacks and whites had similar probabilities of belonging to the ambivalent type. Figure 2 depicts an interaction between race/ethnicity and education and the likelihood of membership in the strained type (low integration, high negative interaction). This interaction revealed that at the lowest educational attainment level, African Americans had a higher probability of belonging to the strained type compared to whites. For white respondents, the probability of belonging to the strained type increased with education level, whereas African American respondents' probability remained stable across education levels. Consequently, among the most educated respondents, whites had a substantially higher probability of belonging to the strained type than African Americans. Two significant interactions between race/ethnicity and age (Figure 3) indicated that for younger respondents in this sample of persons 55 years and over, whites and African Americans had the same probability of being in the ambivalent type. As age increased, this probability decreased for both groups. However, the decline was more precipitous for whites, such that among the oldest respondents, whites were less likely to belong to the ambivalent type than African Americans. Additionally, among younger respondents, Caribbean blacks had a lower probability of belonging to the ambivalent type than their white counterparts. While the probability of belonging to the ambivalent type decreased with age for whites, it increased with age for Caribbean blacks. As a result, older Caribbean blacks had a notably greater probability of being in the ambivalent type than older whites. --- Discussion The present analysis is the first to investigate racial and ethnic differences in the likelihood of being in particular social network types among older African Americans, Caribbean blacks, and non-Hispanic whites in the United States using a national probability sample. This is an important contribution to the study of social networks, as few studies have examined network types among racial/ ethnic minorities and none have examined racial and ethnic differences among older Americans. Collectively, the findings underscore within-group heterogeneity in social relationships in the black population (i.e., African Americans and Caribbean blacks) and differences between black and white populations. Although race and ethnicity alone did not influence membership in specific social network types, it was relevant when examined jointly with education and age, underscoring the complex nature of interactions involving race/ethnicity and sociodemographic factors. Four distinct network types-optimal, ambivalent, family-centered, strained-were derived in the present investigation from a nationally representative sample of older African Americans, Caribbean blacks, and non-Hispanic whites. The study findings partially supported Hypotheses 1 and 2. These four network typologies are representative of the archetypal diverse, family-focused, and restricted network types identified in the synthesis of the literature on social network types. The optimal and ambivalent types identified in this current analysis are characteristically similar to the archetypal diverse network type in their high levels of social integration. The family-centered type, which was characterized by high levels of social integration within primarily the extended family network, is similar to the family-focused type previously identified in the literature. The strained type most closely reflects the archetypal restricted network type in its low levels of social integration. However, a nonkin-focused network type was not identified in this analysis. Further, both positive and negative diverse network subtypes (i.e., the optimal and ambivalent network types) were confirmed, but only a single restricted network typology characterized by high levels of negative interaction (i.e., the strained type) was identified. Turning to findings for race/ethnic differences with respect to typology membership, the analysis did not support Hypothesis 3; Caribbean black and African American respondents did not differ from white respondents in the probability of belonging to the optimal and family-centered types. The data did support Hypothesis 4 regarding interactive effects; race/ethnicity significantly interacted with education and age in predicting network types. Education was negatively associated with the probability of being in the ambivalent type for Caribbean blacks, but the opposite was true for whites. This difference may be linked to unique life circumstances of Caribbean blacks, who may experience more obligations to provide support to their extended families (e.g., sending remittances to family), particularly family members in their home countries. Moreover, recent migrants to the United States with lower levels of education may be more limited in their socioeconomic resources and thus burdened by support exchanges. In fact, research indicates that socioeconomic status is positively correlated with sending remittances to family members residing abroad (Menjivar, DaVanzo, Greenwell, & Valdez, 1998). Thus, Caribbean blacks with less education may find it more difficult to meet the needs of their extended family, generating negative interactions based on mismatched expectations for assistance and/or unmet needs. This combination of high negative interaction coupled with high positive social involvement leads to ambivalent ties for Caribbean blacks with less education. Furthermore, a number of studies have indicated that education is positively associated with acculturation (Romano, Tippetts, Blackman, & Voas, 2005;Shen & Takeuchi, 2001) and that lower levels of acculturation is associated with increased relational conflict (Chung, 2001;Farver, Narang, & Bhadha, 2002). Given this, an alternative explanation for this interaction could be that Caribbean blacks with less education are more likely to be less acculturated. Thus, in addition to reporting high levels of positive social involvement with their networks, they nonetheless experienced more relational conflict, which is associated with a greater likelihood of being in the ambivalent network type. A second interaction indicated that higher levels of education increased the probability of being in the strained type for whites only. In contrast, the probability of being in the strained type was virtually the same across all education levels for African Americans. This pattern is consistent with research indicating that socially disadvantaged groups are more integrated within their social networks and tend to rely more heavily on them for informal support (Gerstel, 2011;Sarkisian & Gerstel, 2004). For African Americans, however, educational attainment was not associated with the probability of belonging to the strained type, which is consistent with research indicating that, at all socioeconomic levels, informal social networks, particularly extended family networks, are important for African Americans (Gerstel, 2011;O'Brien, 2012). Finally, interactions between race/ethnicity and age involving all three racial/ethnic groups were noted. For both white and African American respondents, the probability of belonging to the ambivalent type decreased as age increased. However, this decrease was smaller for African Americans. Potential qualitative differences in social relationship dynamics for African American and white older adults may contribute to more negative interactions among African Americans. For example, older African Americans are more likely than older whites to be in poorer health, have fewer financial resources, and reside with extended family (Williams & Wilson, 2001). The effects of these factors intensify with advanced age, increasing both reliance on connections to social networks and the stressors and strains that accompany these circumstances and potentially contribute to ambivalent ties. Differences in objective life circumstances (e.g., health, income) and the cultural relevance of informal social networks may account for the differential impact of age on ambivalent ties for the two groups of older adults. In contrast, the relationship between age and membership in the ambivalent type was reversed for Caribbean blacks, whose probability of belonging to the ambivalent type increased with age. This may be a function of expectations of support from network members and diminished ability to provide support among older Caribbean blacks. Caribbean black culture has been described as a "culture of reciprocity" that underscores the importance of equity in supportive exchanges (Bashi, 2007). Due to financial and physical limitations, older adults are likely to have limited means of providing support. Thus, uneven support exchanges (i.e., receiving more than providing) between older Caribbean blacks and their social networks may contribute to ambivalence with network members, especially when support reciprocity is an expected cultural norm. --- Study Limitations and Conclusions Several limitations of the current analysis should be noted. All social relationship measures in this study were self-reported and subject to recall and social desirability biases. Given the cross-sectional nature of the data, causal relationships between sociodemographic factors and network types could not be assessed. Future studies should use prospective data to investigate the causal relationships between sociodemographic factors and network types. Additionally, because the NSAL did not include negative interaction with friends, this issue could not be addressed in this study. Without information on negative friendship interactions, this study was unable to determine whether there are network types that are delineated specifically by this characteristic. Consequently, the network types identified in this analysis may be incomplete and provisional. Another limitation of this study is the lack of differentiation between ethnic groups within the non-Hispanic white sample, which the NSAL did not assess, and the sociodemographic diversity within non-Hispanic white ethnic groups. An important contribution of this study is the use of diverse sources of informal support and relational strains. Examining multiple sources of support coupled with negative interaction provides a more complete understanding of the idiosyncratic contributions of different network members to older adults' social environments. Negative interactions, while relatively common, are an often overlooked feature of social relationships. The inclusion of both positive and negative social relationships in this analysis contributes to the literature because it identifies network types that reflect an enhanced and more realistic representation of older adults' relationships. Although previous studies have examined ambivalent relationship types (Connidis & McMullin, 2002;Fingerman et al., 2004;Rook et al., 2012;Uchino et al., 2012), these studies examined ambivalent types by constructing them based on a priori assumptions of relational ambivalence. This study builds on these prior studies and extends the literature on ambivalent relationship types by using latent variable modeling to identify an ambivalent network type. Thus, the identification of an ambivalent network type in this study confirms the existence of an ambivalent relationship type that researchers have long proposed. This is an important finding that requires further investigation because despite the presence of positive relationship qualities, ambivalent ties are associated with poor mental and physical health outcomes, such as depression, inflammation, high blood pressure, and functional health limitations (Holt-Lunstad, Uchino, Smith, & Hicks, 2007;Kiecolt, Blieszner, & Savla, 2011;Rook, Luong, Sorkin, Newsom, & Krause, 2012;Uchino et al., 2013). In fact, ambivalent ties are associated with worse physical health status than exclusively negative ties, such as relationships within the strained type (Rook et al., 2012). Moreover, this is the first analysis, to my knowledge, of social network typologies among a racially/ethnically diverse population of older Americans. Another contribution of this study is the use of multiple indicators of church-based relationships (e.g., frequency of contact with congregants, subjective closeness to congregants, emotional support from congregants, and negative interaction with congregants). Prior studies of social network typologies typically used religious service attendance as the sole indicator of church-based social networks, which captures only a single facet of these networks. Additionally, the current analysis used LCA, an innovative analytical methodology based on a person-centered data analysis approach that addresses several limitations of cluster analysis, which is the statistical analysis typically used in social network typology research. In sum, the present analysis extends previous work by examining social network types for the first time in a nationally representative sample of older African Americans, Caribbean blacks, and non-Hispanic whites. This innovation contributes to work on social network types by providing a more complete understanding of race/ethnicity in relation to network typologies and the interactive effects of race/ethnicity, age, and education on social network typologies in two traditionally under-researched populations. This investigation represents a preliminary effort to understand the role of race and ethnicity in social relationships as manifested in network typologies. Future studies should explore the implications of these differences in relation to mental health and subjective well-being, which have been linked to network types. --- Supplementary Material Supplementary data are available at The Journals of Gerontology, Series B: Psychological and Social Sciences online.

Objectives: This study examined race differences in the probability of belonging to a specific social network typology of family, friends, and church members. Method: Samples of African Americans, Caribbean blacks, and non-Hispanic whites aged 55+ were drawn from the National Survey of American Life. Typology indicators related to social integration and negative interactions with family, friendship, and church networks were used. Latent class analysis was used to identify typologies, and latent class multinomial logistic regression was used to assess the influence of race, and interactions between race and age, and race and education on typology membership. Results: Four network typologies were identified: optimal (high social integration, low negative interaction), family-centered (high social integration within primarily the extended family network, low negative interaction), strained (low social integration, high negative interaction), and ambivalent (high social integration and high negative interaction). Findings for race and age and race and education interactions indicated that the effects of education and age on typology membership varied by race. Discussion: Overall, the findings demonstrate how race interacts with age and education to influence the probability of belonging to particular network types. A better understanding of the influence of race, education, and age on social network typologies will inform future research and theoretical developments in this area.

thnicity in relation to network typologies and the interactive effects of race/ethnicity, age, and education on social network typologies in two traditionally under-researched populations. This investigation represents a preliminary effort to understand the role of race and ethnicity in social relationships as manifested in network typologies. Future studies should explore the implications of these differences in relation to mental health and subjective well-being, which have been linked to network types. --- Supplementary Material Supplementary data are available at The Journals of Gerontology, Series B: Psychological and Social Sciences online.

Objectives: This study examined race differences in the probability of belonging to a specific social network typology of family, friends, and church members. Method: Samples of African Americans, Caribbean blacks, and non-Hispanic whites aged 55+ were drawn from the National Survey of American Life. Typology indicators related to social integration and negative interactions with family, friendship, and church networks were used. Latent class analysis was used to identify typologies, and latent class multinomial logistic regression was used to assess the influence of race, and interactions between race and age, and race and education on typology membership. Results: Four network typologies were identified: optimal (high social integration, low negative interaction), family-centered (high social integration within primarily the extended family network, low negative interaction), strained (low social integration, high negative interaction), and ambivalent (high social integration and high negative interaction). Findings for race and age and race and education interactions indicated that the effects of education and age on typology membership varied by race. Discussion: Overall, the findings demonstrate how race interacts with age and education to influence the probability of belonging to particular network types. A better understanding of the influence of race, education, and age on social network typologies will inform future research and theoretical developments in this area.

INTRODUCTION Goals aimed at adapting to climate change in sustainable and just ways are embedded in global agreements such as the Paris Agreement, Sustainable Development Goals and the New Urban Agenda. These agreements seek to move environmental and climate concerns into the urban policy action arena by developing strategies for risk management. Ideally, these strategies would be supported by the three pillars of sustainability (economy, equality, and environment), while increasing cities' resilience to chronic and acute physical, social, and economic stressors and hazards (Zeemering, 2009;Campbell, 2013;Romero-Lankao et al., 2016a;Simon et al., 2016). However, in practice, tradeoffs are often present that shrink the size one pillar and augment another. In the last decade, scholars and decisionmakers have shown increased interest in the mechanisms by which urbanization and climate change are coevolving to compound the unequal risk of floods, wildfires, and other hazards to urban populations and their supporting food, energy, and water (FEW) systems. However, actions to improve equality on the ground have been less evident (Revi et al., 2014;Romero-Lankao et al., 2017c). Incorporation of equality into urban adaptation plans is important because the most vulnerable communities within cities, most often are more exposed, have lower socio-economic status, make lower contribution to GHG emissions, and have lower levels of access to FEW systems, and livelihood options to mitigate risk and adapt (Boone, 2010;Hughes, 2013;Agyeman et al., 2016;Romero-Lankao et al., 2016a;Shi et al., 2016;Reckien and Lwasa, 2017). It is widely accepted, in the literature of social vulnerably, that social inequality shapes differences in climate risk and vulnerability and in capacity to mitigate and adapt to these hazards (Ribot, 2010;Romero-Lankao et al., 2016a). However, largely unexamined, are the ways in which different narrative understandings relate to suggested actions in existing adaptation plans. In this paper, we examine whether and how adaptation plans from 43 C40 cities address inequality in risk, by planning ways to reduce inequality in hazard exposure or tackling the drivers of social vulnerability (Reckien and Lwasa, 2017). We apply a mixed methods approach, including a discourse analysis and meta-analysis of adaptation plans for 43 C40 cities (Figure 1 and Supplemental Table 1A). In this approach, the discourse analysis helps unpack framings of urban equality issues as they relate to policy actions, and the meta-analysis seeks to quantitatively investigate patterns of framing and policy across adaptation plans. --- TRACING EXISTING SCHOLARSHIP Three areas of scholarship, relevant to this paper, include urban adaptation, and governance, inequality in climate risk, and the food, energy, and water (FEW) nexus (Leck et al., 2015;Araos et al., 2016;Shi et al., 2016;Romero-Lankao et al., 2017c;Wiegleb and Bruns, 2018;Heikkinen et al., 2019). We use findings in these areas as a basis to suggest a conceptual framework (section Conceptual Framework), which will be used to map attention given, in urban adaptation plans, to FEW interactions with inequality, and thereby gain knowledge of how far these considerations have penetrated urban adaptation planning. --- Urban Adaptation and Climate Governance Having proven to be important agents of change globally, cities, and transnational networks occupy a central role in the global governance of climate change because of many reasons (Bulkeley and Betsill, 2013;Romero-Lankao et al., 2018). There is a wide acknowledgment among scholars of the incapacity of national actors alone to produce policy actions that can address the complex dynamics of climatic risk (Gordon and Johnson, 2017). Attention has shifted to the array of governance initiatives undertaken outside of interstate climate negotiations and policies. These initiatives, taken by state, municipal, market, and civil society actors operating at multiple local to global levels, are seen as key to creating the kinds of innovations necessary to address environmental change and climate risk (Acuto, 2013;Shi et al., 2015;Gordon and Johnson, 2017). In recent years, in what has been termed the second wave of urban climate governance (Bulkeley, 2010), cities have moved beyond symbolic commitment to climate change action, to its integration into their planning and development policies (Aylett, 2014). For many cities, part of this movement has included participation in local and city-networks such as ICLEI, the World Association of Major Metropolises (Metropolis) and the C40 Cities Climate Leadership Group (C40) (Bouteligier, 2013;Gordon and Johnson, 2017). C40 is a network of 94 of the world's cities concentrating more than 650 million people and one quarter of the global economy. This peer network of cities seeks to address climate change through the design and implementation of policies seeking to mitigate greenhouse gas (GHG) emissions and climate risks (https://www.c40.org, February 28th,2019). A body of literature has examined different aspects of the C40 global and city governance influence. For instance, some portray the C40 as an orchestrator of global urban climate governance steering member cities toward particular climate actions (Gordon and Johnson, 2017), or creating new inequalities and sometimes even intensifying existing ones (Bouteligier, 2013). Others analyze whether the kind of change the network promotes is incremental, reformistic, or transformational (Heikkinen et al., 2019). In this study, we start from the assumption that member city agendas may differ from that of the C40 network (Heikkinen et al., 2019), and examine how, in their adaptation plans, city officials understand and manage inequality in climate risk to urban populations and FEW-systems. --- Risk and the FEW-nexus Studies on FEW nexus have grown recently (Endo et al., 2015). As it pertains to human food, energy, and water systems, the term nexus refers to the relationships, as defined by linkages and interdependencies, between two or more FEW resources and systems, including trade-offs and feedbacks between them (Leck et al., 2015;Romero-Lankao et al., 2017c). FEW-nexus scholarship has grown in recent years, but differences in motivation, purpose, and scope pervade the field (Stringer et al.,FIGURE 1 | Cities covered in the analysis of adaptation plans. Based on World Bank income category as of 1 July 2015, at the country-level. Low-income economies are those with a GNI per capita, of $1,045 or less in 2014; middle-income economies are those with a GNI per capita of more than $1,045 but <unk>$12,736; high-income economies are those with a GNI per capita of $12,736 or more. Lower-middle-income and upper-middle-income economies are separated at a GNI per capita of $4,125. --- 2018). A FEW-nexus approach can be used to analytically examine links and interdependencies between FEW-systems, but it also functions as a boundary object that engages decision makers and academics across a science-policy interface aimed at understanding and managing FEW-system links and interdependencies (Wiegleb and Bruns, 2018). In governance, its concepts are sometimes used to achieve integrated management across FEW sectors and jurisdictions (Bizikova et al., 2013). Here we will examine how linkages and interdependencies between FEW-systems are acknowledged and prioritized at the city level and whether integrated FEW-management is a goal of adaptation plans. Or if, as suggested by existing scholarship, bringing together diverse policy domains creates its own set of challenges. The most important is given by the difficulties involved in moving decision makers beyond their accustomed ways of understandings and action precisely because this involves a collective engagement of disparate sectors, ways of knowing, levels of government, power, and values (Romero-Lankao et al., 2017c). FEW-nexus studies tend to be motivated either by the scarcity of FEW resources or by threats to FEW-resource security due to development and environmental pressures (Galaitsi et al., 2018). We will focus on the latter, which tends to be framed using either a security or a risk approach (Corry, 2012). In the security approach, the focus is on an existing threat such as an ongoing drought or disruption of energy or food supplies (Comfort, 2005). In the risk approach, however, the emphasis is on how human development and environmental dynamics are interplaying (or might interplay) to create the potential for harmful events (Trombetta, 2008). While security thinking leads decision making to look for the current, direct causes of harm to urban populations and FEW-systems, risk analysis analyzes the potential causes of harm, current or future. We use a risk approach here, because it fits better with both climate change scholarship, ours included, and the framing used in 87% of the adaptation plans (Field et al., 2014;Romero-Lankao et al., 2017a) (Figure 2). Within our sample, we look at how adaptation plans address inequality in risk. Following the IPCC, we define risk as the potential for adverse effects on lives, livelihoods, health, and assets (Field et al., 2014). Risk may spring from exposure to floods, sea level rise, and other threats and vulnerability of people and the FEW-systems that support them. Such vulnerability, or the propensity to be negatively affected by events or impacts, results from the multiscale interplay of factors in five domains: Socio-demographic, Economic, Technoinfrastructural, Environmental, and Governance (SETEG), which have been used by Arup and by us in prior work (Arup, 2014;Romero-Lankao and Gnatz, 2016). While people can be susceptible to hazards, they also have capacity and agency to modify their circumstances and behavior to mitigate risks or adapt. Capacity is the unequally distributed pool of resources, assets, and options governmental, private, and non-governmental actors can draw on to mitigate and adapt FIGURE 2 | Framing the adaptation vision. After reading and summarizing each adaptation plan, four notions capturing cities' broader frame or vision were identified. See Supplemental Table 1B. to risks, while pursuing their development goals and values (Vincent, 2007). To understand how policymakers are prioritizing these issues, we examine how in their adaptation plans, city officials attribute climate risk to a series of locational and SETEG factors, and what policy actions they suggest to manage these (section Study Design). --- Urban Adaptation, Inequality, and Equality For centuries, the notions of inequality, equality, and justice have been the subject of compelling philosophical, conceptual, and ethical debates, with persistent disagreements in definition, scope and policy implications whose discussion is beyond the scope of this paper (Ikeme, 2003;Agyeman et al., 2016). The concepts of fairness and justice can be related to discussions of the differences in definitions of equal and equitable. The word justice comes from the Latin jus, meaning right or law, and refers to either an actual or ideal situation in which: (a) benefits and burdens in society are distributed according to a set of allocation principles where the basic rights and needs of individuals and groups are considered and respected (distributive element); (b) the rules and regulations that govern decision making preserve basic rights, liberties, and entitlements of individuals, groups, or communities (procedural element); and human and other biological beings are treated with respect and dignity by all parties involved (interactional element) (Jost and Kay, 2010). Likewise, equality, which we use here in its opposite, conveys an ideal state of perfectly balanced or even distribution of goods and services across populations, while equitable can allow an element of self-determination. In a neo-liberal conception, as long as each member or group has an equal chance to obtain access to resources and options, a distribution can be termed equitable because it is self-determined on an equal playing field. Such equitable distributions are seen in this conception as fair or just because no one has had an advantage in gaining access to resources and options (Ikeme, 2003;Hughes, 2013). However, this conception ignores the power of assets and options, once attained by some individuals and groups, to create or compound differential access a to assets and options for others thus creating social inequality (Agyeman et al., 2016). Social inequality thus creates self-feeding systems that are not fair or equitable because they deny, to marginalized people and groups, access to assets and options necessary to avoid risks at the same they deny access to police systems and institutional features that could help them gain access those assets and options. Inequality determines differential location and access to places, water, food, energy resources, and decision-making options in a city where resources are distributed unevenly across populations (Reckien and Lwasa, 2017). Typically such uneven distributions result from markets, power, other institutional mechanisms and risk mitigation and adaptation policies that engender or perpetuate socially defined categories of wealthy or poor, or of included and excluded populations (Stein, 2011;Romero-Lankao et al., 2016b) based on class, caste, gender, profession, race, ethnicity, age, and ability (real or perceived). Undergirding our analysis in this paper is an assumption that, in the context of city climate action, an understanding of how inequality creates differences in exposure and vulnerability is fundamental to creating fair and effective risk mitigation and adaptation. Policies aimed at creating risk-equality should contain mechanisms to ensure the fair distribution of risks of negative impacts and of benefits (assets and options) to undertake climate action across city populations (distributive justice). Creating equality also means generating equal opportunities for participation and recognition for all, including underrepresented groups (procedural justice) (Bulkeley et al., 2013;Hughes, 2013;Reckien and Lwasa, 2017). Among the resources and options that vary with inequality to create differential urban vulnerability, access to food, energy, and water are so basic and primary that they can be used as bellwethers of an uneven distribution of many other resources conditioning vulnerability Biggs et al., 2015;Romero-Lankao et al., 2016b. When considering the fair distribution of resources, assets and services related to distributive justice, it is important to recognize that differences in gender, race, socioeconomic status, and culture are part of procedural barriers that condition participation in policies affecting distribution. Thus, a cultural value can inhibit poor and marginalized populations from effectively participating in decisions (e.g., where to locate infrastructural investments in water and electricity) that affect their wellbeing, property, resources, climate risks, and capacities to adapt and mitigate. --- CONCEPTUAL FRAMEWORK Using discourse analysis, we qualitatively unpack how, in their adaptation plans, city officials' frame inequality in urban climate risk. We then combine discourse analysis and adaptation analysis to examine some of the issues addressed by the adaptation actions suggested in the plans. Lastly, we use a meta-analysis approach to quantitatively investigate patterns of framing and adaptation action across cities. We will map narrative understandings in the adaptation plans of how inequality creates differences in exposure and vulnerability. We will also examine if and how adaptation actions contain mechanisms to ensure the fair distribution of assets and options to manage climate risks (distributive justice), and generate equal opportunities for participation and recognition for all, including underrepresented groups (procedural justice). --- Discourse Analysis Various strands of social science scholarship have used discourse analysis to examine texts, images, papers, books, and reports to define the ideas and concepts-which we will call understandings-through which actors understand and act upon the world (Foucault, 1972;Sharp and Richardson, 2001;Hajer, 2004;Keller, 2011;Wiegleb and Bruns, 2018). Rather than being neutral, these narrative understandings privilege some socio-environmental facts and may suggest some policy actions over others (Sharp and Richardson, 2001;Hajer, 2004;O'Brien et al., 2007;Trombetta, 2008). We draw on section Conceptual Framework and on the Sociology of Knowledge Approach to Discourse to map the discourse of 43 adaptation plans (Keller, 2011). The sociology of knowledge analysis of discourse includes three components: knowledge structuring, discourse production, and power effects. Here we will only focus on the first and the third. We excluded the second, which entails an examination of the influence of sociopolitical context on framing and action (Keller, 2011), because our study focuses on discourse as it crystallized in the plans, and not on the influence of each city's sociopolitical context on framing and action. To help us determine knowledge structuring, we mapped, through their references to issues of concern, the general interpretative frame city officials use to make sense of a climate change issue in their adaptation plans. For instance, do city officials frame climate adaptation as a problem of risk, or of resilience? However, setting issues such as those related to inequality in climatic risk on the adaptation agenda also relates to the way in which city officials determine what kind of problem climate change is. What causal SETEG factors are involved in the creation of climate change impacts? Are these impacts only the result of location and geography, or exposure? Or are they also the result of prior policies and unequal patterns of development determining differences in the vulnerability of people and FEWsystems within cities? Drawing on the discussion of existing literature (section Conceptual Framework), we will map how adaptation plans address inequality in hazard exposure and in the following multiscale (SETEG) factors determining vulnerability (Arup, 2014;Romero-Lankao and Gnatz, 2016). -Locational (exposure) factors conditioned by the presence of populations and critical FEW infrastructures in places that could be adversely affected by floods, heatwaves, and other climate hazards (Nicholls et al., 2008). -Socio-demographic factors consist of age, gender, and demographic structure of a city or the behavior of individuals and groups (Donner and Rodr<unk>guez, 2008). -Economic factors relate to uneven economic growth, urbanization, income, and affordability of food, energy, water, and other resources (Uejio et al., 2011). -Techno-infrastructural and built environmental factors include land use change and the distribution, quality, and robustness of water, sanitation, electricity and related, FEW critical infrastructures, and systems. Critical FEW infrastructures include electric power, natural gas and oil, water supply, and food distribution systems, but because we acknowledge the role of transportation, telecommunications, health, emergency and other services, we also included these as critical urban FEW infrastructural systems (Rinaldi et al., 2001). -Environmental factors such as the biophysical and climatic characteristics affecting an urban area's predisposition to hazards relate to exposure. For instance, coastal cities are prone to sea level rise, storm surge and coastal flooding, saltwater intrusion and tropical storms. -Governance factors consist of the fit between areas of concern and authority, cooperation, and cohesiveness among governing bodies and levels of government, policies and actions, and the legacies of actions and policies around-land use planning; and through investments, location and climate proofing of FEW infrastructure and service networks, which shape the geography of urban risk (Aylett, 2014). Power effects relate to the intended or unintended consequences emerging from the discourse. Elements of the power effects include the dispositifs, a French word describing the institutional, organizational and infrastructural elements, which we define here following Foucault and Keller as the suggested apparatuses of adaptation action, such as a) Personnel and organizations charged with undertaking adaptation policies; b) Institutional and organizational processes seeking to evaluate, monitor and understand the climate change problem, or to foster awareness among city actors, decision makers, and populations. We will include these under institutionalbehavioral adaptation actions (note that (a) and (b) seek to address the sociodemographic and governance factors within our SETEG framework); c) Investments in and climate proofing of critical FEW infrastructure (artifacts), which we will include under techno-infrastructural actions. (These address the technoinfrastructural factors within our SETEG framework); and d) Other discursive or non-discursive adaptation actions, such as environmental and economic adaptation actions (which address respective factors within our SETEG framework). Such "dispositifs" are shown in the literature to hold the potential to address climate risk to people and FEW-systems in cities. In our analysis we sort "dispositifs" among techno-infrastructural, institutional-behavioral, economic, and environmental action categories (Romero-Lankao et al., 2017b). --- Adaptation Analysis We also include insights from the climate adaptation literature to add accuracy to our discourse analysis. In the climate adaptation literature, institutional-behavioral actions include changes in the procedures, incentives, or practices of city actors, and often work through existing urban competencies and hybrid actor arrangements in sectors, such as urban planning, health, water, energy, and disaster risk management (Fisher, 2013;Romero-Lankao et al., 2017b). Institutional behavioral actions entail the creation of organizations charged with mainstreaming adaptation into other sectoral and developmental policies such as urban planning, transportation, and disaster management; with evaluating, monitoring and understanding the climate change problem; and with fostering awareness among city decision makers and populations. In the environmental justice literature, these actions are fundamental to procedural justice by broadening participation in, recognition, and commitment to adaptation across governmental, private, civil society, and community actors (Bulkeley et al., 2013;Shi et al., 2016;Reckien and Lwasa, 2017). Techno-infrastructural actions are critical in the creation of artifacts, such as energy, water and sanitation. They are often framed in the climate adaptation literature, as efforts to discourage growth in risk-prone areas and to protect critical urban infrastructural systems through investments in climate proofing, and changes to design, operational, and maintenance practices (Romero-Lankao et al., 2017b). Other adaptation actions include economic and environmental policies. The former aim at creating enabling conditions for autonomous action by governmental and nongovernmental actors, and to support broader development goals. Funding programs from public and private sectors are fundamental. By strategically allocating funding (whose amount and sources vary widely across cities), local governments can effectively respond to climatic risks (Aylett, 2014). Environmental actions seek to manage the biophysical, climatic, and hydrological factors affecting an area's predisposition to hazards (Brink et al., 2016;Kabisch et al., 2016). Environmental actions take into account and manage the role of biodiversity, greenspaces, and other ecosystem services in mitigating hazard risk and reducing the vulnerability of urban populations and FEW systems to climate change (Levy et al., 2014). --- STUDY DESIGN Meta-analysis is often applied to find commonalities within a variety of research papers and methods (Littell et al., 2008). It involves the pooling of data that quantitatively examine whether causal relations described in individual papers (e.g., drivers of climate risk, determinants of vulnerability of food, energy, and water insecurity) hold across a broader body of scholarship (Misselhorn, 2005;Romero-Lankao et al., 2012). While meta-analysis is frequently combined with systematic literature reviews to synthesize the results of previous research, in our approach, we combine meta-analysis with discourse analysis to systematically investigate patterns on the framing of inequality in risks within a selection of 43 adaptation plans. --- Selection and Analysis of the Adaptation Plans This study resulted from a prior report commissioned by the C40. Although the C40 has 94 affiliated cities, we only got access to 60 adaptation plans for analysis. Of these, we selected 43 plans, 4 of which are from cities located in lower-income, 12 in middleincome and 27 in upper-income countries. As can be seen in Figure 1, our selected sample also has a good representation of C-40 cities from Latin America, Europe, North America, Africa, and South-East Asia. We built on our prior work on FEW nexus, climate adaptation and inequality cited in section Conceptual Framework, and on the review of the adaptation plans, to map how city officials prioritize policy actions to manage inequality in risk. Although we couldn't analyze how individual city officials actually understand the climate change adaptation and FEW issues we studied, we did analyze the understandings of these issues conveyed in the plan. We will refer to these understandings, conveyed in the plans, as narrative understandings. Our data extraction and synthesis followed an examination of discourses and a meta-analysis approach (Littell et al., 2008;Keller, 2011;Romero-Lankao et al., 2012;Wiegleb and Bruns, 2018). Our conceptual framework functioned as a starting point to design and test a review template and to agree on our own definition of terms and fields (available upon readers' request). We then used this template to extract data from each of the 43 adaptation plans. First, each selected plan was carefully reviewed by at least two members of our research team to ensure systematic and consistent data extraction. Factors influencing risk to people and FEW-systems were identified and coded into the five SETEG domains (i.e., sociodemographic, economic, technoinfrastructural, environmental, and governance). Adaptation actions were classified into institutional-behavioral, technoinfrastructural, economic, and environmental. We further subdivided these categories of SETEG factors and adaptation actions into terms, as described in the second column of Supplemental Tables 1A,B, 2A-E, 3A-D). After summarizing each adaptation plan, mention counters were developed, based on mention of the terms, to capture overall narrative understanding (Supplemental Tables 1A,B, 2A-E, 3A-D). Once a term was found, the counter maxed at "1" for that particular topic to avoid duplicate counting. Limiting mention counts to one per plan is the most effective way to avoid bias in answering The first gives a view of the relative importance, attributed by urban policymakers, to particular issues within plans compared with all plans. The second gives a view of the relative importance, attributed by urban policy makers, to particular issues compared to all issues within a given category (e.g., techno-infrastructural vs. institutional-behavioral actions). Together, these measures give a two-scoped view of the relative priorities given by urban policymakers to the issues addressed in the plans. Although we feel this study offers many relevant insights, it was faced with some constraints that may affect its outcomes. While we included 43 cities from low-, middle-, and high-income countries, these were not selected using a sampling approach. Due to our determination to have at least two members review each plan, and our group's language limitations, we could only review plans written in English and Spanish. This meant we were not able to analyze the discourse in many plans that might have offered additional insights. Readers of this paper should, therefore, keep in mind that while the combination of discourse analysis with meta-analysis to identify patterns in understanding and action is innovative, our study is exploratory in nature. Furthermore, while our use of a discourse analysis to examine the framings of inequality in risks exposed some of the narrative understandings conditioning policy actions, it did not include an examination of why and how the socio-political and geographical contexts in which city officials operate shape their interpretations and planned actions. Lastly, since we studied plans and not implementations we could not determine how (or if) the suggested adaptation actions were implemented. While ethical questions regarding this study might be raised around the fact that it was commissioned by the C40 to study the adaptation plans of C40 cities, giving rise to concerns about scientific objectivity, we feel that our analysis of these plans was objective and sound for two reasons: (1) We studied the adaptation plans as independent documents and not as they pertain to the C40 or its mission; and (2) The methods used in the study were evenly applied across city adaptation plans without regard to any city's membership, income level or status in the C40. --- NARRATIVE UNDERSTANDINGS AND POLICIES IN THE ADAPTATION PLANS This section is organized around three topics. The first and second include a mapping of the narrative understandings-or knowledge structuring-crystallized in the adaptation plans. This not only in terms of what interpretative frame is used but also in terms of what locational and SETEG factors are identified as key determinants of climate risk, and whether inequality is considered in this conveyed understanding. The third topic refers to the power effects in the form of adaptation actions suggested in the adaptation plans to address inequality in risk to people and FEW-systems. --- Interpretative Frames We found that the urban adaptation plans analyzed here embed adaptation in a larger vision for the city, often with a multiplicity of coexisting frames. Many of these interpretive frames are not only full of symbolism, as in the resilience framing we will describe later in this section, they also feature keyand sometimes, contradictory-organizing principles of policy action (Figure 2). Rather than converge toward an integrated understanding, these concepts often coexist in a tension of incongruent and unbalanced sets of principles and related actions. In this disharmony, economic and investment concerns and interests (e.g., infrastructural and economic investments) tend to take precedence over concerns and interests for the environment and the marginalized (see next subsection). Frequently cities appear in the adaptation plan narratives as leaders, development hubs or engines of innovation and investment, key to growth and stability nationally, and internationally. Adaptation in this context forms part of a broader sustainability vision present in many cases for the creation of a vibrant, economically prosperous, and socially just cities, or cities that are habitable, secure, resourceefficient, socially and economically inclusive, and competitive internationally (Seattle, Tshwani). In many adaptation plans, city officials frequently see climate change as posing risks, but also offering opportunities. These include opportunities to attract investment, generate high-value jobs, strengthen research and development, or foster circular or green economies. For instance, the Singapore plan states that the city is poised to tap economic opportunities offered by global warming, such as investments in new growth areas, the creation of high-value jobs, the promotion of green growth, and of R&D capabilities. Interestingly, 87% or 37 of the reports apply a risk approach to frame climate change issues (Figure 2). Risk is often framed in the adaptation plans as the probability of occurrence of a hazard, such as sea level rise, multiplied by a consequence such as property damage. While differences in emphasis exist, a dominant narrative emerges, underlying the risk approaches in these plans. Common to this narrative is the idea that strategies for the protection of urban areas from the risks and FEW constraints associated with climate change require a scientifically grounded technical assessment of how changes in temperature, precipitation, and sea level are likely to affect critical infrastructures, resources and economic activities in the cities. Adaptation plans reviewed in this study illustrate that resilience is, increasingly, becoming embedded in the discourses of urban decision-makers. Resilience is not only seen in the plans as an ecological principle, but also, frequently, as an opportunity. Such opportunities, when coupled with appropriate actions, can increase a city's economic, energy, environmental, and food security, in addition to protecting the quality of life and safeguarding property (e.g., Durban). It is, therefore, common for the adaptation plans to frame the hazards and disruptions brought about by climate change as somewhat of a blessing in disguise. In this discursive thread, cities may even view themselves as symbolically endowed with a power of resilience like "the mythic phoenix, " able to take advantage of disruptive events and carry on through challenges over the years. In such cases cities become a phoenix aware of how the threats cities face-and their responses to these threats-expose several interdependencies that city officials must better comprehend (San Francisco). An almost mythic idea of its own resilience can also be found, for instance, in the New Orleans plan, which describes a city certain that the creativity and resilience of its people and places have been key in its capacity to bounce forward, after being faced with a decade of hurricanes, oil spills, and the Great Recession. --- Inequality in Climate Risk We compared levels of attention paid to climate risk associated with five selected SETEG factors, and examined whether the plans mentioned inequality in reference to these factors (inequality within each domain, Figure 3). This comparison revealed that because city officials are, by necessity, generalists, adaptation plans deal with many climate change issues at a time, from those related to economic development and land tenure to those associated with health, disaster management, housing and critical FEW infrastructures (Supplemental Tables 2A-D). Evidence from the narrative understandings conveyed by the plans suggests that FEW-nexus thinking is not yet embedded in city officials' priorities, or that such considerations create a conundrum that officials are reluctant to tackle. Of the total of risk factors, those related to food, energy and water systems were mentioned in 6, 14, and 20 reports, respectively (Figure 3). Where they did appear, food, energy, or water systems are treated separately, in most cases, without consideration of how their interdependencies can amplify or mitigate risk. The influence and vulnerability of FEW-systems was often framed in terms of techno-infrastructural issues associated with age, design or capacity characteristics (Blue bars, Figure 3). For example, the plans mention that FEW-systems and infrastructures are vulnerable because they are old, designed without consideration of the new (and unstable) normal that climate change will bring, and in need of retrofitting and climate-proofing actions. Buildings are also vulnerable because of poor quality design and construction, age, and lack of maintenance (Figure 3; Supplemental Table 2C). Inequality also tends to be given a lower priority and appears mainly in relation to other factors and very rarely in relation to FEW systems. Inequality considerations were included in 24 plans and represented 26 percent of the total mentions of technoinfrastructural risk factors. However, scant consideration was given to how techno-infrastructural and built environment factors condition unequal risk through such distributive mechanisms as differential access to water or sanitation, or differences in the provision and placement of infrastructures and services such as electricity, waste disposal, tree shading, parks, hurricane shelters, and evacuation routes. Locational (exposure) factors were mentioned in 32 plans (green bars, Figure 3) as related to differential exposure of populations and FEW-systems to climate hazards. Adaptation plans in Peru, Mexico City, and Cape Town point to how the poor are priced out of desirable

Goals aimed at adapting to climate change in sustainable and just ways are embedded in global agreements such as the Sustainable Development Goals and the New Urban Agenda. However, largely unexamined, are the ways that narrative understandings conveyed in adaptation plans consider and attempt to address inequality in climate risk to urban populations and FEW-systems. In this paper, we examine whether and how adaptation plans from C40 member cities address inequality in risk, by planning actions to reduce hazard exposure or tackling the drivers of social vulnerability. C40 is a network of 94 of the world's cities fostering policies to address climate change. We apply a mixed methods approach, including a discourse analysis and meta-analysis of adaptation plans. The discourse analysis helps to unpack framings of urban equity issues as they relate to policy actions, and the meta-analysis seeks to quantitatively investigate patterns of framing and policy across adaptation plans. Our findings suggest that FEW-nexus thinking is not yet embedded in narrative understandings of risk and planned adaptation actions, within the adaptation plans we studied. In the city adaptation plans we analyzed, we found multiple frames coexisting behind the broader adaptation visions (e.g., risk and resilience). Rather than converging, issues, and principles such as those of equality, coexist with economic issues in an imbalance of incongruent political movements and priorities. Techno-infrastructural and economic investments and concerns tend to take precedence over concerns and interests for inequality in climate risks. We discuss some of the institutional factors explaining this. Knowledge integration, for instance, is constrained by the existence of a plurality of sectors, levels of government, power, values, and ways of understanding and managing climate risk. We also suggest that the relatively low importance of equality considerations in the adaptation plans will likely limit the capacity of cities to support broader goals such as those of the New Urban Agenda and the Sustainable Development Goals.

are old, designed without consideration of the new (and unstable) normal that climate change will bring, and in need of retrofitting and climate-proofing actions. Buildings are also vulnerable because of poor quality design and construction, age, and lack of maintenance (Figure 3; Supplemental Table 2C). Inequality also tends to be given a lower priority and appears mainly in relation to other factors and very rarely in relation to FEW systems. Inequality considerations were included in 24 plans and represented 26 percent of the total mentions of technoinfrastructural risk factors. However, scant consideration was given to how techno-infrastructural and built environment factors condition unequal risk through such distributive mechanisms as differential access to water or sanitation, or differences in the provision and placement of infrastructures and services such as electricity, waste disposal, tree shading, parks, hurricane shelters, and evacuation routes. Locational (exposure) factors were mentioned in 32 plans (green bars, Figure 3) as related to differential exposure of populations and FEW-systems to climate hazards. Adaptation plans in Peru, Mexico City, and Cape Town point to how the poor are priced out of desirable neighborhoods and are often forced to live in hazardous areas. In Seattle, San Francisco, and New Orleans, adaptation plans show concerns for how inequality makes poorer populations more likely to occupy low-lying areas, prone to flooding or more likely to experience heat island effects because these areas are more affordable. Related to location, environmental risk factors were mentioned in 12 plans (green bars, Figure 3). Some of these mention that many informal settlements locate on areas, where the high-water table and inadequate infrastructure make them particularly vulnerable to flooding (e.g., Cape Town, Buenos Aires, Tshwane, Mexico City, and Lima). Cities from the Global North also offer examples of how low-income communities living in brownfields or in flood risk areas face higher levels of exposures not only to sea level rise, floods and heatwaves but also to contaminated land (e.g., New York, and New Orleans). Regarding economic factors, twenty-seven adaptation plans (67%) refer to economic development as a key determinant of risk, and twenty-three (53%) of all plans mention urbanization as a broader driver of risk (yellow bars, Figure 3). Interestingly, 27 or 62% of the adaptation plans referred to unequal economic growth conditioning access to determinants of a population's capacity to mitigate risks and to adapt. Such determinants include location, and access to secure land, affordable, accessible, and good quality housing, energy, water, food, and transportation (yellow bar, Figure 3). In the adaptation plans of Lima, Mexico City and Cape Town, the narratives acknowledge deep inequalities and high poverty rates that relate to the existence of informal, unplanned settlements whose populations have precarious housing without adequate FEW resources necessary to protect themselves against hazards. Recognition of such conditions is rare in the adaptation plans of the global north. New York is one of the handful of such cities indicating that nearly half of its people live in or near poverty, and lack access to good quality housing and other resources needed to adapt. While 17 adaptation plans refer to socio-demographic factors such as population size and growth, age, gender, and pre-existing medical conditions as determinants of vulnerability, 20 plans convey an understanding of governance as a determinant of risk and vulnerability (purple bars, Figure 3). Such governance conditioned risks operate through investments and the location of FEW infrastructures and service networks, and through the legacies of actions and policies around-land use planning or its lack though this is not generally acknowledged in the plans (orange bars, Figure 3). As for inequality, socio-demographic and governance factors, creating social exclusion by class, gender, race, migration, and minority status were mentioned in 13 and 5 plans, respectively, (orange and purple bars, Figure 3). Adaptation plans from cities in middle-and low-income countries tended to mention the influence of social exclusion on inequality in access to affordable energy, water, food, and sanitation, and reliable transportation systems more often than plans from high-income countries. Race, however, appears in the adaptation plans of the US cities of New York, New Orleans and San Francisco as a predictor of risk. These plans indicate that people of color are more likely to live in areas more at risk of flooding and subsidence, to live in poverty, to be unemployed and to have pre-existing health conditions associated with higher hazard risks. These plans also recognize that their marginalized populations have lower capacities to mitigate and adapt (Supplemental Tables 3A-D). --- Policy Actions to Address Inequality in Risk and FEW-nexus In our mapping of the power effects emerging from adaptation discourse among policymakers, we examined whether planned adaptation actions aimed at either reducing hazard exposure or tackling the drivers of social vulnerability considered inequality. The adaptation actions identified were organized into "dispositifs" as defined in section Tracing Existing Scholarship. We sorted "dispositifs" among techno-infrastructural, institutional-behavioral, economic, and environmental action categories. Our findings suggest that, while proposed adaptation actions tend to target many issues at a time, they also tend to prioritize infrastructural and economic issues, and that inequality is a secondary concern. Furthermore, city officials tend not to address the links and feedbacks between critical FEW infrastructural systems but rather to suggest actions to manage each infrastructural system at a time. Technological-infrastructural actions, which can be a means of fostering distributive justice, received the highest number of mentions (with 124, or 41%, blue bars, Figure 4). However, by and large distributive justice was not considered. Instead, actions were presented in the plans as a means to protect buildings and infrastructure through changes to design. Similar to what we found in our examination of narrative understanding, suggested policy action did not address the links and interdependencies among critical FEW-systems but rather focused on one sector at a time. Examples of planned infrastructural adaptation actions included: • Improving energy redundancy and reliability of (e.g., distributed power), flood fitting the design of surfaces, and increasing the extent of cooler, green surroundings (Changwon, Chicago, Karachi, New Orleans, Paris, Seattle). • Introducing low-carbon or renewable energy sources, reducing coal usage for electricity generation, promoting energy-efficient and resilient technologies, appliances, and designs in buildings and developments-e.g., cooling systems, LED and fluorescent lighting (Amsterdam, Quito). • Adapting water infrastructures to withstand heavy rain events, drought, and heat. Climate-proofing water systems and implementing a water sensitive approach to urban design and flood mitigation through blue and green infrastructures (Copenhagen, New York, Rotterdam, San Francisco). Techno-infrastructural actions were most frequently organized around resilience, low-carbon utilities and buildings, promoting a circular economy, and risk as a source of investment opportunity (Supplemental Table 3A). For instance, Amsterdam and Boston suggested fostering a circular economy to reduce waste and increase recycling throughout economic activities and districts. Other cities, such as Copenhagen, suggested basing adaptation on a risk and resilience approach aimed at improving infrastructure adaptability to new or unexpected conditions by achieving a city-wide, multiple-purposed, and longer-term risk mitigation vision. There were a few exceptions were plans used technoinfrastructural actions aimed at addressing inequalities in risk. For instance, the following actions were suggested: • Reducing intra-urban differences in water scarcity, access and use; increasing water coverage to poor and informal populations without regular, safe, and continuous water service (Cape Town, Durban, Johannesburg, Kolkata, and Mexico City); and providing access to weatherization of homes to low income families (Seattle). • Scaling up development tied to renewable energy services to accomplish a lower energy impact while achieving reduced poverty and promoting economic development (Durban, Tshwane). • Fostering structural investments that consider the consequences from interrupted energy supply during and after extreme events, and target those that are more affected (Durban, Tshwane). • Renovating slums, informal, or poor settlements (Addis Ababa, Buenos Aires, Cape Town, Durban, Kolkata, Mexico City, and Tshwane). Institutional-behavioral actions were second in the number of mentions (118 or 39% of the total). The focus in order of importance was on knowledge and awareness, monitoring, urban planning, disaster risk management, and institution building (orange bars, Figure 4). Awareness and knowledge, and monitoring were addressed in 31 and 29 of the plans, respectively. These plans suggest a suite of strategies to systematically evaluate, assess, understand, and monitor the kinds of climate risks and vulnerabilities they face (Supplemental Table 3B). They also suggest using scientific and technical expertise as a vital source of knowledge. For instance, Amsterdam suggests improving the city's knowledge and understanding of data to become active partners, steering events toward sustainability based on a knowledge of interconnections between systems such as energy and water. Two crucial adaptation instruments received attention in 22 adaptation plans each: disaster risk reduction (DRR) and urban planning. Elements of DRR included early warning systems, cooling centers for poorer populations, and climate-sensitive management protocols (e.g., Bogota, Kolkata, Mexico City, San Francisco, Quito, Rio De Janeiro, and Sydney). Urban planning was mentioned as a fundamental tool for anticipating climate change impacts, fostering early action and even preventing risks (orange bars, Figure 4). Some plans (e.g., Lima and Tshwane) acknowledged institutional barriers to effective implementation, such as weak law enforcement. Others pointed to gaps in the levels of authority and autonomy to control the investments and decisions that are fundamental not only for effective urban planning but also for managing the drivers of climate risk in the city. FEW thinking with relation to equality received scant attention within planned institutional-behavioral actions. We found only the following few examples of strategies to enhance equality within each sector: • Community based adaptation actions such as upgrading informal settlements, building flood-water drainage, and sewer systems in poor areas (Mexico City and Tshwane), and training poor communities for the management and attention of disasters (Bogota). • Increasing the share of renewable energy per capita through demand management actions, such as agreements with a number of utilities, incentives that support energy efficient practices, and reduced electricity consumption during peak hours (Amsterdam, Durban). • Inducing water conservation through water restriction, tariffs, and reduction of leaks (Cape Town). • Enforcing polices and by-laws that make healthy food accessible to all (Boston) and reserve space for local decentralized food hubs that can supply small traders while reducing ecological impact, through the support of small scale, sustainable farming practices (Durban). Within the economic instruments suggested in 38 adaptation plans, equality considerations were, likewise, virtually absent. While many of the plans seek to create enabling environments for independent action by both governmental and nongovernmental actors, for example through infrastructural investments, they largely aim at enhancing their economies without regard for structural inequality or uneven distribution. Through these actions, the plans also aim to support broader goals such as the Sustainable Development Goals. Indeed, the governments that produced many of the adaptation plans we analyzed are driving investments in major flood defenses, and in the transportation, water, and sanitary services sectors, but generally steer away from equality considerations in these investments and are more concerned with how they will fund them. Some cities, particularly from highincome countries, are explicitly and actively partnering with the private sector (Amsterdam, Copenhagen). One of these plans acknowledges that society at large will pay a large dividend to have infrastructures privately constructed and operated (Copenhagen). Environmental actions were considered in 40% of the plans, and many of these contain actions primarily focused on increasing or protecting biodiversity (e.g., Karachi, Montreal, Seoul, and Los Angeles), and on strategies for managing ecosystem services (green bar, Figure 4 and Supplemental Table 3D). For instance, the plans suggest actions to green the cities' streets, parks, and open spaces in order to serve multiple risk mitigation purposes. Other planned actions include efforts to increase biodiversity and reduce the urban heat island effects (e.g., Sydney, Vancouver, Melbourne), to increase urban agriculture (Seoul), and to better manage such hazards as runoff or fires (e.g., Rotterdam, Melbourne, Rio de Janeiro, and Portland). Nature-or ecosystem-based adaptation actions are also suggested to increase the resilience of vegetation to climatic and ecological impacts (such as erosion, Montreal), or to establish temporary rainwater catchment systems (Mexico City). Some cities also suggest conservation or rehabilitation of degraded ecosystems (Tshwane, Quito, and Mexico City) and protecting or restoring natural protections in coastal areas (New Orleans). --- ADAPTATION PLANS AND RISK INEQUALITY In this study, we examined evidence from 43 adaptation plans to determine whether and how they considered the factors driving inequality in exposure and vulnerability of people and the FEW systems that support them. To do this, we combined a discourse analysis with a meta-analysis of adaptation plans for 43 C40 cities. We are not the first scholars to conduct metanalysis. Examples of existing literature include (Misselhorn, 2005;Romero-Lankao et al., 2012;Endo et al., 2015). Nor are we the first to examine environmental discourse, even with regard to FEW systems. For instance, existing discourse scholarship has shown that a risk approach is prevalent among FEW nexus scholars (Wiegleb and Bruns, 2018). Because risks lack immediacy-says the analysisdiscourse around FEW risks entails connecting a future scenario to a policy, "presented as a way of preventing that risk from materializing into real harm" (Corry, 2012. p. 244). Our methodological innovation lies, rather, in our combination of discourse analysis with meta-analysis. We used this combination to examine narrative understanding and planned adaptation actions in 43 city adaptation plans. We integrated several theoretical strands of scholarship, such as FEW-nexus thinking, adaptation, and inequality, climate change risk, and adaptation and discourse analysis. Nevertheless, we did not examine why and how the socio-political and geographical contexts, in which city officials operate shape their interpretations and planned actions. Nor were we able to determine how or if the suggested adaptation actions were implemented. These represent the short-comings and limitations of our study that make it largely exploratory in nature. Notwithstanding these limitations, however, some clear patterns emerged that can help guide future research and policy. We found that FEW-nexus thinking is not yet embedded in city officials' narrative understandings of risk and planned adaptation actions, even when unpacking interdependencies among food, energy, and water systems may help cities tackle some of the root causes of vulnerability and risk (Romero-Lankao and Norton, 2018). Other scholars have already pointed to the fact that, while promising, FEW-nexus thinking faces many practical challenges. For instance, knowledge integration is constrained by the existence of a plurality of sectors, levels of government, power, values and ways of understanding and managing climate risk (Leck et al., 2015;Romero-Lankao et al., 2017c). Scholars also suggest that local governments lack the institutional and organizational capacities needed to appropriately manage the complexity and uncertainty associated with climate risks, let alone inequalities in the vulnerability of people, or how that vulnerability interplays with FEW systems. Officials within sectors involved in managing climate risk, such as food, energy, water, disaster risk management, and urban planning hold diverse organizational and cultural values. They lack the incentives, rights, financial resources, and responsibilities needed to work across sectors and jurisdictions (Scott et al., 2015). Additionally, decision makers involved in DRR and adaptation policies lack interaction and coordination because of differences in language and political culture (Schipper, 2009). An examination of these factors is an essential first step to develop the skill sets, tools, funding, and incentives needed to foster nexus thinking in risk mitigation and adaptation practice. In the city adaptation plans we analyzed, we found multiple frames coexisting behind the broader adaptation visions conveyed in their narratives. Rather than converging, issues and principles such as those of equality, coexist with economic issues in an imbalance of incongruent political movements and priorities (Anguelovski and Carmin, 2011;Campbell, 2013). In this disharmony, techno-infrastructural and economic investments and concerns tend to take precedence over concerns and interests for inequality or the environment in climate risks. Clearly, challenges exist with under-investments, backlogs and deferred maintenance of infrastructure. Urban infrastructures in many developed countries are deteriorating, and in developing countries infrastructure construction and maintenance have often failed to keep pace with the dynamics of urbanization (Kraas et al., 2016). Adaptation plans recognize that by working as a risk amplifier, climate change is projected to intensify these challenges, through at least two mechanisms: long-term, slow impacts such as constant deterioration of storm water system due to floods mentioned in the adaptation plans of 27 cities, or extreme events such as hurricanes (mentioned by 10 cities) and damaging critical FEW infrastructural systems. Still, with a few exceptions, equality concerns were not the priority. In the adaptation plans, narrative understanding and policies to address techno-infrastructural challenges were frequently organized around resilience, low-carbon utilities and buildings, promoting a circular economy, and risk as a source of investment opportunity. All these strategic decisions advance cities as centers of economic and infrastructural growth. However, they run the danger of fostering inequality in access, related to distributional justice, by creating climate proof places that become more exclusive and expensive, pricing out marginalized populations who end up living in less desirable areas and lacking access to critical FEW infrastructures (Coutard, 2008;Zérah, 2008). In their adaptation plans, cities of high-income countries are seeking to explicitly and actively partner with the private sector (Amsterdam, Copenhagen). Policy-makers in these cities reason that moving infrastructural development and operation to the private sector can be a way of diverting development costs away from government and reducing the need for politically unpopular taxes. However, this hasn't often shown itself to be a good strategy, as private interests must inevitably draw profits from their projects, leaving less for the public good. Ultimately, this will have implications for inequality in risk, as the poor communities, those most in need of investments in climate proofing, are more likely to be excluded not only from decisions (procedural justice) but also from reaping the benefits of techno-infrastructural interventions (distributional justice) (Coutard, 2008;Zérah, 2008;Revi et al., 2014). Socio-institutional actions relate to the distributive and procedural aspects of equality in different ways (Reckien and Lwasa, 2017). For instance, by involving vulnerable populations in decisions on land use and location of infrastructural investments, in the generation of knowledge, or in the monitoring of climate risks (Moser, 1998;Moser and Satterthwaite, 2010;Bouzarovski, 2014). Nonetheless, rather than using participatory instruments such as community based adaptation (Ebi and Semenza, 2008;Dodman and Mitlin, 2013), the plans mostly suggest using scientific and technical expertise as a vital source of knowledge. There are reasons for this. Climate change adaptation is highly data-dependent, demanding that city officials engage in new ways of gathering data, collaborating with scientists, using scientific information, and dealing with uncertainty (Hughes and Romero-Lankao, 2014). Yet, the focus on technical knowledge is a key element of prevalent cultural values that inhibit poor and marginalized populations from effectively participating in decisions on where to locate FEW critical infrastructural investments that affect their well-being, property, resources, climate risks, and capacities to adapt and mitigate. Although our current study, based purely on textual analysis, did not attempt to examine socio-political context (knowledge production), our conclusions do suggest that sociopolitical context was at play in the creation of the plans. Even beyond that, they suggest that common elements in socio-political context may be drawing cities away from actions based on effectively addressing such complex concerns as vulnerability and inequality toward those least conflicting with economic priorities. The relatively low importance of equality considerations in the adaptation plans will likely limit the capacity of cities to support broader goals such as the Sustainable Development Goals, Sendai Protocol for Disaster Risk Reduction and New Urban Agenda (Simon et al., 2016). The purposefully inclusive scope of the New Urban Agenda and of the targets and indicators in the urban SDG (Goal 11) provide a unique opportunity to include equality considerations in adaptation (Romero-Lankao et al., 2018). Prospects for progressing and mainstreaming climate change agendas, therefore, depend on demonstrating that climate agendas do not always and irreconcilably conflict with development priorities, such as those related to equality. From a longer-term perspective, they are essential and complementary to them. Atmospheric Research, sponsored by the National Science Foundation. Open access publication of this article was funded by the National Renewable Energy Laboratory, operated by Alliance for Sustainable Energy, LLC, for the U.S. Department of Energy (DOE) under contract DE-AC36-08GO28308. We want to thank Dakota Smith, Adelmut X. Duffing Romero, and Olivia Pearman for their support reading and analyzing the plans. We also want to thank our C40 partners Neuni Farhad, Caterina Sarfatti, Snigdha Garg, and Amanda Ikert for their keen insights in the reviews of the report that inspired this paper. --- AUTHOR CONTRIBUTIONS PR-L led the design, gathering, analysis and interpretation of data for the work. She also drafted and revised the work critically for important intellectual content. DG contributed to the design, analysis and interpretation of data for the work. He also drafted and revised the work critically for important intellectual content. --- SUPPLEMENTARY MATERIAL The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fsoc.2019. 00031/full#supplementary-material --- Conflict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Goals aimed at adapting to climate change in sustainable and just ways are embedded in global agreements such as the Sustainable Development Goals and the New Urban Agenda. However, largely unexamined, are the ways that narrative understandings conveyed in adaptation plans consider and attempt to address inequality in climate risk to urban populations and FEW-systems. In this paper, we examine whether and how adaptation plans from C40 member cities address inequality in risk, by planning actions to reduce hazard exposure or tackling the drivers of social vulnerability. C40 is a network of 94 of the world's cities fostering policies to address climate change. We apply a mixed methods approach, including a discourse analysis and meta-analysis of adaptation plans. The discourse analysis helps to unpack framings of urban equity issues as they relate to policy actions, and the meta-analysis seeks to quantitatively investigate patterns of framing and policy across adaptation plans. Our findings suggest that FEW-nexus thinking is not yet embedded in narrative understandings of risk and planned adaptation actions, within the adaptation plans we studied. In the city adaptation plans we analyzed, we found multiple frames coexisting behind the broader adaptation visions (e.g., risk and resilience). Rather than converging, issues, and principles such as those of equality, coexist with economic issues in an imbalance of incongruent political movements and priorities. Techno-infrastructural and economic investments and concerns tend to take precedence over concerns and interests for inequality in climate risks. We discuss some of the institutional factors explaining this. Knowledge integration, for instance, is constrained by the existence of a plurality of sectors, levels of government, power, values, and ways of understanding and managing climate risk. We also suggest that the relatively low importance of equality considerations in the adaptation plans will likely limit the capacity of cities to support broader goals such as those of the New Urban Agenda and the Sustainable Development Goals.

INTRODUCTION The 145 million smokers in the Region of the Americas account for 12% of the more than 1 billion smokers in the world [1]. The region lies in the fourth place among the six regions of the World Health Organization (WHO) with a 22% smoking rate among the adult population [1]. Tobacco is a major preventable risk factor for major non-communicable chronic diseases (NCDs), which are currently responsible for almost two-thirds of deaths worldwide. In the Region of the Americas, NCDs are responsible for 77% of all deaths: among these, tobacco is responsible for 15% of deaths from cardiovascular diseases, 26% of deaths from cancer, and 51% of deaths from respiratory diseases [1]. According to the WHO, tobacco use and exposure to secondhand smoke kill about 1 million people annually in the Americas [2]. In Argentina, tobacco is responsible for 14% of all NCDs, compared with 8% of all communicable diseases [2]. Despite some decrease observed during recent years, high smoking prevalence and related harm still remains a significant public health concern in Argentina and Uruguay [1,3]. Apart from preventing smoking tobacco among young people, encouraging cessation is essential to ending the tobacco epidemic. Cessation of tobacco use has the potential to provide the most immediate benefits of tobacco control and maximize the benefits in terms of preventable disease morbidity and mortality [4]. However, achieving substantial improvement will depend on successful implementation of the relevant tobacco control measures that can increase the smoking cessation rate at the population level in Argentina, Uruguay and other Latin America countries. In general, smoking prevalence and tobacco consumption is much higher in certain social groups [5]. Correspondingly, an increased susceptibility to tobacco related illnesses was found in low income groups, especially in all-cause mortality, lung diseases and low birth weight [5]. Likewise, several studies have indicated the social gradient in tobacco use in Argentina as well as in Uruguay. Fleischer et al. showed that better socio-economic status, measured through education, was related to less smoking and higher odds for recent quitting [6]. The most recent study by De Maio et al. revealed social gradients in tobacco use, exposure to secondhand smoke and cessation attempts among Argentinians and Uruguayans [3]. Therefore, social context cannot be overlooked when discussing applicable strategies to improve the design and implementation of appropriate tobacco policies and cessation programs in both countries. Data on the factors associated with successful smoking cessation that can be analyzed by socioeconomic factors beyond age and gender are crucial for the development of potential, high-impact population smoking cessation strategy [7]. In view of that, the purpose of our study was to examine the socio-economic dimensions associated with successful smoking cessation among adults in Argentina and Uruguay. --- MATERIALS AND METHOD The data source was the Global Adult Tobacco Survey [9]. GATS is a nationally representative household survey designed to monitor key tobacco control indicators. The target population of GATS includes all non-institutionalized men and women 15 years of age or older. The study protocol and questionnaire is based on standard methodology with some country-specific adaptations. Detailed methodology of the survey has been described elsewhere [8,9,10]. A multi-stage, geographically-clustered sample design was used to produce nationally representative data. The GATS questionnaires were administrated by trained survey staff during in-person interviews. There was a total of 6,645 and 5,581 completed individual interviews with an overall response rate of 74.3% in Argentina and 95.6% in Uruguay. The missing data were excluded from the analysis. After exclusion of respondents younger than 25 years, the final sample used in this study consisted of 5,383 Argentineans and 4,833 Uruguayans. --- Study variables. The main outcome variable was successful smoking cessation among adults in Argentina. Previous studies on quitting smoking are not homogenous in defining successful quitting, and many different measures of success have been suggested [11,12,13,14]. Some studies have shown that the risk of relapse is relatively high for people who abstain from smoking for short periods, and are at the early stages of smoking cessation. About 65% -75% of these groups at risk would relapse within a year [11,12,15,16]. In the presented study, successful quitting is defined as having abstained from smoking for a year or more [17]. A sustained quitter was defined as a former daily smoker who had been smoking for at least 1 year or longer, and had stopped smoking for 12 months or more prior to the interview. Those subjects who had given up smoking in more recent periods were considered recent quitters. A continuous smoker was defined as a current daily smoker who had smoked more than an average of one cigarette per day on a regular basis for at least one year. The ever smokers group compriseds all the above-mentioned categories, including respondents who were current, former smokers and recent quitters. Overall lifetime cessation rates or 'quit rates' were calculated, as the number of former smokers divided by the number of ever smokers and multiplied by 100% [18]. The independent variables applied for determining associations of successful cessation were demographics: gender (male, female) and age of the respondents. Age was studied in five groups: 25-29, 30-39, 40-49, 50-59, and <unk>60 years old. Age at smoking onset -the age at which respondents started to smoke tobacco on a regular basiswas also considered (<unk>17, 18-20, 21 years or over). Moreover, socio-economic status, including education, economic activity, monthly household income and ownership of different household items were evaluated. Educational attainment was regarded as: primary or less, secondary, and higher education. Economic activity differentiated subjects who were currently employed, self-employed, homemakers, unemployed. The variable called 'asset index' was created, based on a summative score of possession of the following assets: functioning electricity, flush toilet, fixed telephone, cell telephone, television, radio, refrigerator, car, washing machine, computer, internet access. The summative score was then divided into, high, medium low. Analogous methodology has been implemented elsewhere [19]. Additionally, awareness of the negative health consequences of smoking was assessed. Respondents were categorized as aware (those who answered 'yes' to the question: 'Do you think that tobacco smoking causes serious diseases?'), and not aware (those who answered 'no' and 'do not know'). Similarly, awareness of the adverse health consequences of environmental tobacco smoke (ETS) exposure was determined, and respondents were characterized as aware and not aware. Cohabitation with a smoker (yes, no) was also taken into account. --- Analysis and statistics. The STATISTICA Windows XP version 8.0 programme was used to carry out the statistical analysis. All analyses were performed separately for men and women. Firstly, a descriptive analysis for all variables involved in the analysis was completed. Categorical variables were studied by chi-square test. Univariable and multivariable logistic regression analyses, with results being presented as odds ratios (OR) with 95% confidence intervals, was applied to study differences among those respondents who sustained smoking abstinence for one year or longer with those who continued smoking. In multivariable analyses, all statistically significant socio-economic variables were simultaneously included in the model. Significance level for relevant calculations was set at 0.05. --- RESULTS The characteristics of the respondents are described in Table 1. In Argentina and Uruguay, there are more ever male smokers than female smokers. Argentina recorded 40.7% male smokers vs. 25.8% female smokers, and Uruguay had 60.4% male vs. 36.1% female smokers (p<unk>0.001). Similarly in both countries, more men started smoking before women, before or by the age of 17, while more women started smoking before men by or after the age of 21. Before or by the age of 17, 58.0% men and 43.0% women started smoking in Argentina, and 58.6% men and 47.5% women in Uruguay (p<unk>0.001). On the other hand, 23.4% women vs. 11.6% men in Argentina and 24.9% women vs. 9.8% men in Uruguay started smoking later by or after the age of 21 (p<unk>0.001). Smokers in both countries differed by economic activity. Male smokers and quitters in Uruguay tended to be older than their counterparts in Argentina, while the women were quite similar in age. The average age of male ever smokers in Argentina was 47.8<unk>15.3 In the same vein, current male smokers in Argentina were 43.1<unk>13.2 years vs. 47.0<unk>13.8 years in Uruguay, while the female smokers were 44.6<unk>13.7 years and 44.9<unk>13.6 years in Argentina and Uruguay, respectively (p > 0.05). At the mean, former smokers were a bit older in both countries; 54.3<unk>15.8 years and 58.6<unk>15.4 years for men and 51.8<unk> 16.1 years and 51.2<unk> 16.1 years for women in Argentina and Uruguay, respectively (p <unk> 0.04). Recent quitters were 39.7<unk>125 vs. 48.0<unk>14.4 for men and 39.2<unk>13.5 vs. 48.0<unk>14.4 years for women in both Argentina and Uruguay, in that order (p > 0.05). Following the same trend, women started smoking later than men in both countries (data not shown). Former and current male smokers started smoking by 17.3<unk>5.1 and 17.1<unk>4.5 years in Argentina vs. 16.7<unk>4.6 and 17.0<unk>5.0 years in Uruguay, respectively. Also, female former and current smokers in Argentina started at 19.6<unk>6.5 and 19.4<unk>7.5 years vs. 19.8<unk>7.8 and 19.2<unk>6.7 years in Uruguay, respectively (men vs. women p<unk>0.001). Alternatively, a higher quit rate was observed among women relative to men in Argentina; 39.6% for women compared to 38% for men, and a lower quit rate in Uruguay; 50.1% for women compared to 52.7% for men (p>0.05). Interestingly, women who successfully quit, did so at a slightly younger age than men. The mean age of quitting for male and female former smokers was 38.9<unk>13.2 and 37.7<unk>13.9 years, respectively, in Argentina, and 41.7<unk>14.3 and 38.3<unk>14.3 years, respectively in Uruguay (p > 0.05). --- Univariate regression. In both countries, men older than 45 years were more likely to be long-term quitters relative to those aged 35-44, but those over the age of 65 had the highest likelihood to maintain cessation; Argentina (OR=7.61; 95% CI 4.76 -12.16) and Uruguay (OR=4.70; 95% CI 3.29 -6.73; p<unk>0.001). Similar results were obtained among women (Tab. 3). In Argentina, men with complete or incomplete secondary education had a lower likelihood to be long-term quitters (OR=0.62; 95% CI 0.42 -0.92; p<unk>0.05) relative to those in the tertiary level (Tab. 2). Results for women in Argentina were not statistically significant. In Uruguay, education did not produce statistical significant results for either men or women. Retired men in Argentina had higher odds of quitting smoking for the long-term then employed men (OR=5.47; 95% CI 3.77 -7.94; p<unk>0.001). Results were statistically insignificant among Uruguayan men. Similarly among women, retired respondents showed better prospects to be long-term quitters in Argentina (OR=3.58; 95% CI 2.36 -5.44; p<unk>0.001) and Uruguay (OR=4.70; 95% CI 3.29 -6.73; p<unk>0.001). Asset index was also a significant predictor of long- The evidence showed no statistically significant association between economic activity and being a long-term quitter among men in both countries, and women in Argentina. On the other hand, retired women in Uruguay (OR=1.33; 95% CI 1.09 -2.24; p<unk>0.05) were more likely to be long-term quitters relative to those currently employed. Similar to the univariate section, men and women with a high asset index had an increased likelihood of maintaining their status as long-term quitters. --- DISCUSSION Understanding potential social gradients in the population and its relation to quitting have significant implications for the development of a future population strategy for smoking cessation. The majority of studies on smoking cessation are derived from a Western context; it was therefore uncertain whether these findings would apply to two neighbouring Latin American countries -Argentina and Uruguay. Firstly, in Argentina, a lower lifetime quit rate was noticed compared to Uruguay (39.2% vs. 51.7%). This data coincides with trends observed in recent years showing greater progress in Uruguay than in Argentina, as it relates to many areas of tobacco control. Recent trends also showed intensified tobacco industry endeavours to postpone or undermine tobacco control legislation and policy in Argentina [20,21,22]. However, in both countries, quit rates were higher compared to middle-income European countries like Romania and Poland where one third of the people who have ever smoked gave up smoking [14,23]. Conversely, quit rates in Argentina and Uruguay are lower when compared to more developed countries, for example, Canada, where the quit rate reaches 60% [24]. This suggests that huge gaps still exist among countries in terms of the implementation, enforcement, and comprehensiveness of tobacco control efforts to curb the tobacco epidemic, including cessation measures. While the majority of studies on socio-economic inequalities in smoking have focused on education and used smoking prevalence as the outcome interest, the presented analysis focused on more than one particular dimension and being a successful quitter [16]. Although there is some variability in the findings, socio-economic conditions have been identified as a predictor of quit attempts and quitting success in a number of studies [25,26,27]. De Maio et al. found a reverse gradient, based on the GATS data, although lacking statistical significance, in smoking cessation attempts which were reported more frequently in the recent year by Argentineans and Uruguayans with lower levels of education [3]. However, when analyzing education and cessation success, in the current study it was found that men in Argentina with lower education attainment also had reduced odds to achieve tobacco abstinence for a year or more. Lower education results from the regression analysis for women in Argentina and respondents from Uruguay did not produce any statistically significant results. In general, this may suggest that male Argentineans with lower educational background are more likely to attempt to quit, but they are less likely to sustain abstinence compared to those with higher education. This is in line with the findings of Kotz et al. who indicated that smokers in more deprived socioeconomic groups are just as likely as those in higher groups to attempt stopping smoking. However, there is a strong gradient of success across socio-economic groups, with those in the lowest group being half as likely to succeed compared with the highest [28]. On the other hand, some studies have not found a relationship between socio-economic factors and quitting, particularly in multivariate analyses which also include other important characteristics [14]. The figures of the International Tobacco Control Four Country Survey showed that education was not generally associated with cessation success, although a few particular levels in certain countries were significantly associated with quitting success [29]. Furthermore, Siahpush et al. in a study of a national sample of Australians confirmed that while education had the strongest relationship with smoking cessation, of all the factors controlled, the relationship between higher education and increased odds of cessation no longer existed when other environmental and individual variables were included in the model [30]. Moreover, in the presented study it was noticed that retired women from Uruguay had higher odds of successfully quitting. This success can be linked with the fact that this group covers older people who are more likely to quit mostly due to health reasons, as previously discussed. In Argentina and Uruguay, unemployed respondents had decreased odds for successfully quitting, but the results were not statistically significant compared to each other. Figures from other GATS-based studies brought mixed results in this area. Being economically active was associated with long-term quitting among men in Romania [14]. In GATS Poland, employed males also had more than twice the probability of giving up smoking compared with the unemployed [23]. The association with employment status among women has not been observed in either country. However, GATS revealed that long-term smoking cessation was harder for men from disadvantaged groups with low asset indices from Argentina and Uruguay. Lower socio-economic groups are generally less likely to be successful quitters, although there is some variation [6,27,31,32,33]. These findings are mostly based on education and/or income data and cannot be compared with GATS results directly considering asset index. Further studies of the expected social gradients in quitting and asset index are needed. Study strengths and limitations. The data derived from GATS is the most recent, nationally representative data based on a high number of respondents. It considers various potential cessation predictors which may also contain some limitations. For the purpose of this study, subjects were selected who were aged 25 years or older at the time of the survey. The analysis was restricted to individuals aged 25 and above because they might still be engaged in the process of smoking uptake [34]. Moreover, subjects under 25 might not have completed the maximum level of education [35]. In addition, continuous abstinence for twelve months or longer was assessed by self-reporting and not validated. Selfreport methods are the most convenient and cheapest way to collect data on smoking tobacco from a large number of respondents in a short time. However, the possible limitation in obtaining answers about smoking may be recall bias, which might lead to underestimation of tobacco consumption. Nonetheless, self-report techniques are stated to be a valid tool for population studies, as addressed in previous papers [36]. Although the GATS questionnaire included questions on duration of tobacco smoking and age of smoking onset, the nicotine dependence or heaviness of smoking that are considered important determinants of, were not obtained for former smokers who maintained tobacco abstinence over one year in this data. There was no information on successful quitting for sustained quitters, such as number of quit attempts, duration per quit attempt, or details on assisted or unassisted quitting. Due to the unavailability of data, it was also not possible to compare some other information from Argentina and Uruguay with other countries; quitting motivations, impact of previous tobacco control measures, including tobacco tax increase. Another limitation is the inability to draw conclusions in causality or directionality of some results based on the cross-sectional study design. Nevertheless, in contrast to studies evaluating the efficacy of smoking cessation treatment programmes, or cessation in high risk groups of heart disease patients, the presented study population should be more representative of the great majority of quitters who quit on their own [18] --- CONCLUSIONS The GATS study revealed that a social gradient in tobacco quitting exists in Argentina and Uruguay. It also identified characteristics associated with long-term sustained tobacco abstinence in both countries. This study provided an insight in specific categories beyond age and gender that were not broadly studied previously, such as asset index. The current study also highlighted the need to encourage tobacco measures that focus on the population that have a harder time quitting smoking. These include younger people, and special attention should be paid to young groups aged 25-34, particularly men in Uruguay and women in Argentina, low educated people and those with lower economic position characterized by asset index. A number of evidence-based individual or community-based policies delivered according to the social context that successfully work in other countries and targeted socially disadvantaged groups, could be adopted in Argentina and Uruguay [37,38,39]. This may facilitate the reduction of inequalities in tobacco-related harm within populations. This is because if tobacco consumption is to be addressed across all social groups, without the distribution of impacts, the improvement will not be experienced equally everywhere, or by everyone [7]. Finally, further systematic research is needed to understand factors that are driving differences in quitting tobacco smoking between diverse social groups in Latin America countries, to ensure tobacco control policies work effectively for all population groups.

Introduction. Cessation of tobacco use has the potential to provide the greatest immediate benefits for tobacco control. Understanding the social determinants of smoking cessation is an essential requirement for increasing smoking cessation at the population level. The purpose of this study was to analyze the socio-economic dimensions associated with cessation success among adults in Argentina and Uruguay. Materials and methods. Data from the Global Adult Tobacco Survey (GATS), a cross-sectional, population-based, nationally representative survey conducted in Argentina (n=5,383) and Uruguay (n=4,833) was utilized. Univariable and multivariable logistic regression analyses with results being presented as odds ratios (OR) with 95% confidence intervals were applied to study differences among those respondents who sustained smoking abstinence (≥1 year) and those who continued smoking. Results. The GATS study revealed that social gradients in tobacco quitting exist in Argentina and Uruguay. Being aged 25-34, particularly men in Uruguay, women in Argentina, low educated men in Argentina and having a lower asset index were associated with reduced odds for quitting.Factors that are driving differences in smoking cessation between diverse social groups in Latin America countries need to be considered when implementing relevant interventions to ensure tobacco control strategies work effectively for all population segments.

BACKGROUND The utilization of basic public health services is an important aspect of the access migrants have to healthcare in the form of screening, preventive services, general practitioners, specialists, emergency rooms, and hospitals (1). In China, basic public health services are mainly provided by community health service centers (stations) in the community, with a large number of medical needs coming from large aging populations. One of the proposed solutions was to establish the elderly support systems in community health service centers (stations) through primary healthcare (2). The core functions of basic public health services in China that included prevention, case detection and management, gatekeeping, referral, care coordination, and so on were to be provided by community health service centers in urban communities (3,4). This was expected to be an effective way to relieve the congestion of superior hospitals; however, unlike the hierarchical diagnosis and treatment system in developed countries, patients could choose different types of hospitals at will in China (5). Studies in developed countries have shown that immigrants have lower rates of health insurance and use less healthcare than local populations (2,6). Specific determinants of health service utilization by immigrants were also inconsistent (7). Studies conducted in the Americas indicated that associations have been found between the length of stay and healthcare utilization of immigrants, and these findings suggested that acculturation or assimilation strongly correlated with the length of stay in the host society, which could be an important determinant of health status (8). However, anthropological studies thought that cultural differences did not necessarily fade with time (9). Along with the increase in urbanization and development of the economy, internal migrants who move between regions within the country have gradually become an integral part of migration in the context of national labor shortages in China (10). The number of internal migrants reached 241 million in 2018, of which 18 million were more than 60 years of age according to the dynamic monitoring survey for internal migrants. Furthermore, the elderly migrant population over 60 years old was growing each year (11). As is known, household registration policy, commonly known as "hukou, " which was classified by origin into urban or rural "hukou, " was implemented in 1958 by Chinese authorities (12). "Hukou" gave households access to social benefits in their registration area but limited access to those outside their registration area (12,13). Only migrants who started working for the government or were highly educated could change their "hukou" registration (14). In contemporary society, an increasing number of families are migrating with their elderly members to look after children, find jobs, or access better healthcare services in China. In addition, most of these elderly members are more than 60 years old (15). Strengthening the utilization of primary healthcare facilities is considered an effective approach to providing affordable, equitable access to quality basic health care for all Chinese citizens by 2020, as was pledged by China (13,16). As the number of internal elderly migrants increases, their use of basic public health services should also improve to contribute to health equity. For example, as one of the basic public health services, the national policy proposed free health checkups in community health service centers (stations) for the elderly aged 65 each year, where these internal elderly migrants are not subject to household registration restrictions (17). However, many studies have indicated that <unk>40% of elderly migrants have participated in the free community health checkups in the past year. In addition, <unk>40% of internal elderly migrants follow up on chronic diseases, and the level of other behaviors, such as establishing health records and seeking medical attention, is also low (15,18). This suggests that there are deficiencies in the health management of elderly migrants in China. In addition, this forms an institutional challenge for basic public health service utilization among elderly migrants. Therefore, it is necessary to explore how to improve the utilization of basic public health services among internal elderly migrants. Studies have shown that the factors affecting the utilization of basic public health services for the elderly include demographic characteristics and social factors (2,19,20). Some studies believe that the migration characteristics of elderly migrants directly affect the utilization of health services (21), but other studies have suggested that migration characteristics are associated with basic public health service utilization through social adaption, acculturation, and other social factors (22). For example, migrating characteristics such as years of residence are one factor of the social support network and acculturation for migrants that contribute to basic public health service utilization (23). In general, it is evident that migrating characteristics and social factors, such as community engagement, social mobilization, ability to communicate, reason for migration, and length of stay in host countries, were related to health service delivery (24,25). However, little work has been done to explore the possible pathways between social factors, migration characteristics, and the utilization of basic public health services, especially for internal elderly people (26)(27)(28)(29). Therefore, our study explored the potential pathway of the impact of migration characteristics on the utilization of basic public health services for internal elderly migrants in China. In our study, migration characteristics were represented by two dimensions: years of residence and reasons for migration, which were consistent with the literature (25). Social factors were represented by social contact, which plays an important role in determining individual health behaviors as a key dimension of poverty and well-being. However, the challenge of measuring social contact was daunting. A unified definition and measurement were not given by the vast and diverse conceptual literature on social contact (30). Empirical studies have explored different aspects of social contact, including physical isolation and access to social resources (31), such as "people feel that their communities" as a proxy for physical isolation and "ties with other people" as a proxy for access to social resources. However, these methods draw attention away from simply counting numbers of social contacts (32). For internal elderly people in China, due to retirement and migration, interactions with friends are the main social ties, and exercise in the community is the main way of feeling their communities. Referring to the literature and considering the actual situations of the population in question, the number of local friends and exercise time per day in the community were chosen as indicators of measuring two aspects of social contacts (33). In this study, we presented the situation of the utilization of basic public health services for internal elderly migrants and explored the potential pathway through which migration characteristics impact the utilization of basic public health services. The objective of this study was to complement the existing literature by providing further insights into the pathway that might influence basic public health service utilization for internal elderly migrants. The results might help policymakers design appropriate social policies to promote the utilization of basic public health services in this disadvantaged population. --- METHODS --- Data and Sampling Data were derived from the dynamic monitoring survey for internal migrants-a special survey on internal elderly migrants of the National Health Commission of the People's Republic of China in 2015. The survey was organized by the National Health Commission in 2015 (formerly the National Health and Family Planning Commission). It was part of the regular data collection of the government, and face-to-face home-based interviews were conducted by the unified training investigators. The response rates for the survey were not announced by the governments. Our local institutional review board (IRB) exempted the analysis of the public-access data because it involved analyzing existing data that had been de-identified; ethical approval was not required for secondary data. Stratified, multistage sampling based on a probability proportionate to size (PPS) sampling method was adopted. The basic sampling goal was all migrant households that did not have "hukou" (registered resident certificate) in the local area and had been living there for more than a month as reported by each village or neighborhood. Townships were randomly selected, followed by villages or neighborhoods. In each village or neighborhood, households of migrants were selected. Eight pilot cities (Beijing, Shanghai, Dalian, Wuxi, Hangzhou, Hefei, Guangzhou, and Guiyang) were chosen as sampling cities. Regarding location: Beijing, Shanghai, Hangzhou, Guangzhou, and Wuxi are located in the east and are more economically developed, Guiyang belongs to the western region, Hefei is in the central region, and Dalian is in the northeast region. Finally, 16,960 migrant households, which included 1,544 (5.5%) internal elderly migrants aged over 60, participated in the survey. In view of the bias caused by different regions, regional classification was incorporated into the model as fixed effects to disaggregate the influence of different areas. --- Measures Based on the existing data, considering that internal elderly migrants who moved between regions within the country and were more than 60 years old had reached the statutory retirement age, the informal networks of these migrants were their social ties. Since a unified definition and measurement of social contact was not given by literature, we chose the number of local friends and exercise time per day as the variables of social contacts following relevant researches (5,7,32). The respondents were asked how many friends they possessed in the local city; for comparative analysis, we adjusted the number of friends to a categorical variable with 7 categories: 0, 1-2, 3-4, 5-6, 7-8, 9-10, and 10 and above. The average daily exercise time was the time spent on physical exercise every day, such as walking more than 40 min, running, playing ball, aerobics, and swimming. If the respondents did not exercise every day, the weekly exercise time was divided by 7 to obtain the average daily exercise time. Similarly, the average daily exercise time was converted to 6 categories of categorical variables (0, within 30, 31-60, 61-90, 91-120 min, and over 120 min) for comparison in this study. Furthermore, migration characteristics were represented by years of residence in the local city and reasons for migration (25). Demographic characteristics included age, gender, education, marital status, medical insurance, average monthly household income, and self-reported health (Table 1). --- Utilization of Basic Public Health Services Health management services, which include lifestyle and health status assessment, medical check-ups, auxiliary examinations, and health guidance, were clearly proposed in the basic public health services launched in 2009 in China (20). This project is considered to be of great significance in controlling chronic diseases of the elderly and improving the health of the elderly. The "National Basic Public Health Service Standards (2011 Edition)" retains and improves the above-mentioned content of the "Elderly Health Management Service Standards." Furthermore, the document requires community health service centers (stations) to provide health management services for the elderly once a year as one of the basic public health services provided by community health service agencies (2). Free health checkups are still the core content of health management services for the elderly, and it is a prerequisite that the elderly enjoy follow-up basic public health services. Free health checkups include routine blood, blood pressure, blood lipid, fasting blood glucose, and blood uric acid tests, ECG examination, digital chest radiography (liver, ultrasonography of gallbladder and kidney, etc.), and screening for some malignant tumors. In recent years, the classification standard for elderly individuals who enjoy free health checkups is 60 years old and above in most areas, and they are not restricted by their place of household registration. In general, community health service centers (stations) usually provide free health checkup services once a year for elderly people over 60 years old in the community within a fixed period each year. The potential participants are informed by issuing leaflets, posting posters, calling through the telephone, and using other methods within the community. Recently, results showed that the percentage of migrant older adults receiving free medical checkups was 36.2% (15). Therefore, our study selected free health checkups as the measurement of basic public health services. Whether elderly migrants participated in free health checkups within the past year provided by community health service agencies was adopted as an indicator of their utilization of basic public health services in the context of free medical check-up services provided to elderly people over 60 years old. If the respondent participated in free health checkups provided by community health service agencies within the past year, in other words, his answer was "Yes, " we thought basic public health services utilization occurred and vice versa. --- Statistical Analyses Demographics, migrating characteristics, and social contacts were presented by descriptive analyses. We calculated the average and standard deviations in age, average monthly household income, and number of local friends. Furthermore, we conducted a classification process and calculated the frequency of each categorical variable. Chi-squared analysis was used to test the relevance of each categorical variable to the utilization of basic public health services. Multivariate binary logistic regression was developed to understand the associations of the variables with the likelihood of using basic public health services through modeling odds ratios. To explore the pathway of migrating characteristics and social contacts for basic public health services, three models were employed. First, the demographic variables were entered (model I), then the variables of migration characteristics were added (model II), and finally, the variables of social contacts were added (model III). In models II and III, we adjusted for demographic variables as fixed effects, and other independent variables were subjected to multifactor analysis using forward stepwise regression (forward: LR) based on the maximum likelihood estimation with a significance level of.05. --- RESULTS --- Demographic Characteristics A total of 1,544 internal elderly migrants were included across 4 districts. A total of 55.6% of internal elderly migrants were men, the mean age was 66.34 years (SD, 5.94), and 50.2% of the internal elderly migrants were aged 60-64 years old. On average, we observed a low level of education: 88.6% of the individuals had received high school education or less. A total of 78.2% were married, and 94.7% rated their health as healthy or basically healthy. Most of the respondents (74.7%) were in the eastern region. The 25-75% interquartile range of the average monthly household income of the migrants was 5,000-12,000 RMB (or US$784-1,881). More than half of them (52.5%) had New Rural Cooperative Medical Care Insurance (NCMS). This meant that the majority of internal elderly migrants came from rural areas, which was the result of the household registration policy ("hukou") that people with rural household registration (rural "hukou") can only participate in the NCMS in their hometown ("hukou" location). It was also found that 33% of internal elderly migrants had participated in a free medical checkup in community health service centers (stations) over the past year (Table 2). --- Migration Characteristics and Social Contacts The median migration time was 5 years, the interquartile range (IQR) was 2-10 years, migrants with <unk>1 year of residence accounted for 7.2% of the respondents, and 59.9% of migrants had a residence time of over 10 years. The main reasons for migrating were taking care of grandchildren (31.1%) and spending their remaining life with children (27.2%). Regarding local friends, the average number of local friends was 8.29 (SD, 11.90), but 12.9% had no local friends. With respect to exercise time, 5.2% did not exercise, and most (62.2%) of their exercise time was 60 min or less per day (Table 3). Significant differences were observed in migration characteristics and social contacts. --- Pathways of the Impacts of Migration Characteristics on the Utilization of Basic Public Health Services In the model with demographic variables (model I), the result of the Hosmer and Lemeshow test was 0.778 (>0.05), which meant that the information in the current data had been completely extracted. The classification accuracy was 62.3%, suggesting that the mean regression model could correctly classify 62.3% of the observations. Age, region, and medical insurance were significant predictors of basic public health service use. Then, after adding the variables of migration characteristics (model II), the classification accuracy increased to 69.4%. We found that age was no longer significant, and years of residence and region were significant influencing factors (p <unk> 0.05). Finally, social contacts were added to the model (model III), demographic variables were controlled, and the remaining variables were gradually entered into the model by forward: LR. The classification accuracy became 80.7%, which was much higher than those of model I (62.3%) and model II (69.4%). In the final model, owing to variable filtering, migration reasons were removed from the model. Furthermore, age, region, and medical insurance showed significant differences among confounders. It is worth noting that years of residence were no longer significant compared with model II, and it was replaced by the number of social contacts variables (number of local friends, exercise time per day). Comparing the results of model II and model III, we believe that, for internal elderly migrants, migration characteristics were complete mediators between social contacts and their utilization of basic public health services, which might have an effect due to the social opportunities provided for them by years of residence (Table 4). Evidence from model III suggested a significant variation in the utilization of basic public health services across regions. Respondents in different regions had different probabilities of using basic public health services. Internal elderly migrants in the western region even had 4-fold higher odds than those in the eastern region (OR = 4.661, 95% CI: 3.196-6.796) where relatively developed cities such as Beijing, Shanghai, Guangzhou, Hangzhou, etc. are located. The reason would be that the income in the eastern region would have been greater than that in the other areas, and perhaps the individuals who earned more had less reason to access the free clinic; therefore, the utilization of paid health services in this cohort could be explored. The probability that respondents who were more than 75 years old used basic public health services was more than twice as high as that for 60-64 year-olds (OR = 2.032, 95% CI: 1.304-3.168; OR = 2.136, 95% CI: 1.081-4.224). Similarly, respondents who had social medical insurance care had a higher likelihood of utilizing basic public health services than those who did not, especially for people who had urban and rural residents cooperative medical insurance (OR = 2.338, 95% CI: 1.190-4.529). Regardless of the demographic variables, the associations tended to be stronger for the number of local friends than for the other factors (p <unk> 0.001). Respondents who reported having local friends had almost 3-fold (OR = 2.988, 95% CI: 1.745-5.118) to more than 4-fold (OR = 4.350, 95% CI: 2.593-7.452) higher odds of utilizing basic public health services than those without local friends. In short, the more local friends a respondent possessed, the more likely the respondent was to use basic public health services. Furthermore, respondents who exercised 61-90 min per day had more than three times higher odds of utilizing basic public health services (OR = 3.459, 95% CI: 1.511-7.919) than those who did not exercise as much. In other words, internal elderly migrants who had many local friends and engaged in 61-90 min of exercise time were more inclined to use basic public health services (Table 4 and Figure 1). --- DISCUSSION As one would expect, when excluding the influences of demographic characteristics, migration characteristics affected the use of public health through social factors, which was similar to some other studies. In fact, we disproved the direct effect of migrating characteristics on the utilization of public health services. Social factors caused the migrating characteristics to lose their significant influence. This study supported the potential pathway of migration characteristics influencing the utilization of public health services through social factors (6)(7)(8)11). In this study, the number of local friends and exercise time per day were significantly associated with the utilization of basic public health services rather than migration characteristics. We found that the more local friends an elderly migrant possessed, the more likely they were to use community health services, which might be due to the information and support their friends provide. Furthermore, we noted that exercise should be encouraged, as exercise time between 60 and 90 min per day was more beneficial for promoting the utilization of basic public health services. The descriptive statistics revealed that most of the internal elderly migrants were aged 60-64. There were more men than women, and all tended to have a lower level of education, which was consistent with the characteristics of internal migrants in general (34,35). Self-reported health was also observed. In this research, the finding regarding self-reported health was similar to previous research that most healthy people were inclined to choose to move (8). However, we did not find a correlation between health and the utilization of basic public health services, which was contradictory to other studies on migrants in general (9). This might account for the elderly individuals who were able to move away from their hometowns being more physically fit, and those who were not inclined to seek or used to seeking health services from health institutions. Another inconsistency with previous studies was the finding that marital status was irrelevant to the utilization of basic public health services. Other studies showed that persons who were 65 years of age or older and living with others were less likely to see a doctor than persons living alone (15). First, this might be due to immigration health effects; second, the companionship coming from family might have replaced the role of marriage (regarding reasons for migration, the results showed that 73.1% of the respondents in the study were accompanied by their families). Furthermore, we found that region was related to the utilization of basic public health services. This was confirmed by earlier findings of other studies that persons who lived in larger communities had a lower rate of general practitioner visits than those who lived in smaller communities (10). As mentioned above, the reason would be that the individuals with higher incomes or who were busier in metropolitan areas had less reason to access a free clinic. Additionally, a total of 8.9% of elderly internal migrants did not have medical insurance, which means they were in an inferior position compared to the 95% coverage of the entire population by three public insurance plans [NCMS, Urban Resident Basic Medical Insurance (URBMI), and Urban Employee Basic Medical Insurance (UEBMI)] (11). As some studies showed, family migration has become a trend in the migration process in China in recent years (13), and the migration characteristics of our study supplied evidence for this trend. In addition, 12.9% of the elderly we studied had no local friends, probably because they were far from their hometowns and old friendship circles and their social circles of local friends needed to be rebuilt. In terms of demographic characteristics, the significant predictors were age and medical insurance. We did not find that gender, marriage, education, economic income, or self-reported health had significant impacts on the utilization of basic public health services, as other studies discovered (18,20,36). In this case, we believe that the discrepancy may be attributed to our target research groups and the type of public health services. First, our objects were internal elderly migrants who tended to live with their families because most elderly migrants moved with their children to take care of their grandchildren; therefore, the impacts on the children might be greater than the impacts on the elderly themselves. Second, we focused on free medical check-up services in the community, so it was reasonable to suggest that, in this case, economic income was not significant. Our data coming from secondary data were the major limitation of this study, which limited the variables that we could employ, especially in the variables of social factors. The study only contained two social contact variables. We believe that it was necessary to further explore the influence of other social factors, such as social support and social integration, which would be our next research direction. The prevalence of chronic diseases and even acute diseases were also not considered in the study. Second, this study was a cross-sectional study, so causality could not be inferred. Last, the sample size was small; furthermore, the number of internal elderly migrants in different sampling areas varied widely, which affected the representativeness of the study. --- CONCLUSIONS To conclude, this study provided a more in-depth examination of the relationship between the studied variables and confirmed the mediating effect of social factors between migration characteristics and the utilization of basic public health services. Because only one-third of the respondents used basic public health services, some other obstacles did exist (18,24). The findings supported the need to increase the opportunities for social contacts between local elderly individuals and internal elderly migrants (31,37). --- DATA AVAILABILITY STATEMENT The datasets presented in this article are not readily available because the data used in this paper were provided by the National Health Commission of the People's Republic of China and we have signed a legally binding agreement with the Commission that we will not share any original data to any third parties. Requests to access the datasets should be directed to [email protected]. --- AUTHOR CONTRIBUTIONS YL: conceived, designed, and performed the study and wrote the paper. YL, TW, and TZ: analyzed the data. All authors contributed to the article and approved the submitted version. --- Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

The literature shows that migration characteristics are a potential pathway through which migration can influence basic healthcare service utilization. The goal of the study was to explore the effect of migration characteristics on the utilization of basic public health services for internal elderly migrants in China and to identify the pathways that might promote their utilization of basic public health services.We studied 1,544 internal elderly migrants. The utilization of basic public health services was determined through participation in free health checkups organized by community health service institutions in the past year. Migration characteristics were represented by years of residence and reasons for migration. Other variables included demographic characteristics and social factors, e.g., the number of local friends and exercise time per day were measured to represent social contacts. Multivariate binary logistic regression was employed to explore the association of the variables with the likelihood of using community health services. Results: A total of 55.6% of respondents were men, and the mean age was 66.34 years (SD 5.94). A low level of education was observed. A total of 59.9% of migrants had been residents for over 10 years, and the main reason for migrating was related to family. Of these migrants, 12.9% had no local friends. Furthermore, 5.2% did not exercise every day. Social contacts were complete mediators of the impact of migration characteristics on the utilization of primary healthcare.Our study highlighted the mediating role of social factors in the relationship between migration characteristics and the utilization of basic public health services among Chinese internal elderly migrants. The findings supported the need to increase the opportunities for social contacts between local elderly individuals and internal elderly migrants.

Introduction Nowadays, with the existence of the internet in life, people began creating communities online. Social media networks have become an essential part of the Internet for billions of people all over the world. These networks are equally important for both people and organizations over the globe. The information and photo sharing, spreading ideas, exchange experience and recommendations are attractive properties of social networks. They enable clients to express their opinions, show loyalty and start conversations with their favorite companies. On the other hand, companies use social media networks to collect information about markets, clients, and competitors in an easy way. Furthermore, companies can communicate with clients to enhance their image and meet their expectations. Based on this view, it is critical to build a trust in using the social media for both people and companies. In fact, building trust in social websites plays a critical role in enhancing the quality of social media networks and implementing security for them. One of the key aspects when dealing with social networks is to try to minimize the effect of hate speech which may lead to violence, terrorism, and security problem to the political systems [1]. One of the most popular methods that help in detecting certain types of semantics in a data is Sentiment Analysis (SA) [2]. SA is a way for text mining, computational treatment of opinions, sentiments and subjectivity of the text [3]. SA is a field of study that measures people's opinions, sentiments through natural language processing (NLP), computational linguistics and text analysis that are used to extract and analyze subjective information from the web, mostly social media and similar sources [4]. In this study, we are concerned with the activity of Jordanian people in the major social networks. According to Arabic social media report, there were 156 million Facebook users in the Arab world of whom 5 million in Jordan and 11.1 million Twitter users of whom 200 thousand in Jordan in March 2017 [5]. In Facebook and Twitter, people share their likes, dislikes, beliefs, political and sports opinions. Among these opinions there are a significant percentage of violence and abuse statements. Sites such as Facebook and Twitter have become a priority to actively combat hate speech which leads to violence [6]. Some of these sites such as Facebook added options to report violent in the feedback options for any post. The importance of detecting hate speech is clear from the strong relation between hate speech and actual vio-lence. Early detecting hate speech could enable outreach programs that attempt to prevent an escalation from speech to action. In this research, we built a model for violence detecting by using people (tweets, posts) written in Jordan Arabic dialect through popular social media networks sites (Twitter, Facebook). Despite the increasingly massive number of Arabic users on the Internet [5], Arabic language is considered amidst top six major languages of the world. The number of native speakers exceeds 200 million and it is the formal language used in over twenty countries [7], there is a weakness for building a strong corporation to exploit it in different applications. There are three different forms of Arabic language [2]: Modern Standard Arabic (MSA), Dialectal Arabic (DA) and Classical Arabic. This study deals with DA, but still, there are many challenges when using Arabic language as listed below: 1. Colloquial Arabic parsers: Many people used their Dialectal language instead of MSA on social media, parsing MSA is an already complex task towards which many efforts have been directed. However, colloquial Arabic different from MSA phonological, morphological such as ("walad", which means "a boy") ("waldan", which means "two boys"), and ("awlad", which means "more than two boys" [7], and lexical and does not have a standard orthography which complicates the task of building morphological analyzers and part of speech taggers [8]. 2. Pronunciation: Some pronunciations do not exist in English such as "Gh" as in "Gharb" [7]. 3. Diacritics: such as, "teacher" <unk>"<unk> <unk> <unk> <unk> <unk>"<unk> and "school" <unk>"<unk> <unk> <unk> <unk> <unk> <unk>"<unk> [9]. 4. Poor of sentiment lexicons: Sentiment lexicons contain opinions with their polarity, they are an important part of any sentiment analysis. There are currently a few number of publicly available colloquial Arabic sentiment lexicons, so building a colloquial Arabic polarity lexicon is still an open research area. 5. Named entity recognition: Named entity recognition becomes an important part of sentiment analysis when identifying the polarity of the opinion. Person name recognition becomes a requirement even for the task of determining semantic orientation, such as <unk>"<unk> <unk>,<unk> ". 6. Phrases and idioms: phrases and idioms are very commonly used by Arabic speakers in social media to express their opinions and feelings in a sarcastic way, old wisdom and idioms, consequently, we shall exclude them because it is hard to deal with them. 7. Negation: Negation can be an important concern in opinion and sentiment analysis. For example, " <unk> <unk> <unk> <unk> " " <unk> <unk> <unk> <unk> <unk> ". There is a list of negation terms in Arabic language which can change the sentiment polarity of terms from negative to positive and vice versa. 8. Emoticons: Arabic smiles and sad emoticons are often mistakenly interchanged; so many tweets have words and emoticons that are contradictory in sentiment mainly due to mixing the text orientation while typing emoticons. As mentioned above, DA is the target of our study so we have collected the dataset manually and it has been annotated into (Violence, Normal) by Arabic Jordan native speakers. In this study, we have built our own dataset that consists of set of words that were extracted from Facebook posts and twitter tweets in the Jordanian dialect. The collected dataset is then handled by applying set of pre-processing techniques to enhance the generalization performance of violence detection. This manipulated dataset and the lexicon are used as a source for the extracted features. In order to detect violence using the extracted features, three well-known classification algorithms are used (Na<unk>ve Bayes, Support Vector Machines and K-Nearest Neighbour) Many experiments have been conducted to study the effect of applying various preprocessing techniques and different classifiers. The results show that SVM classifier is doing better than KNN and NB when used with the collected dataset This paper is organized as follows. There are five more sections. In section 2, Related work presents an overview of related work. In section 3, Proposed method will be discussed Under this section data collection and pre-processing, building lexicon, features extraction, machine learning approach will be discussed. Section 4 discusses Experimental Results. Finally, in section 5, conclusion and future work. --- Related Work Millions of users share opinions on different aspects of life every day. Therefore, social media web-sites like (Facebook, Twitter) are rich sources of data for opinion mining and sentiment analysis [24]. One of the application of this mining and analysis process is the detection of violent and hate speech. Different approaches were proposed in the literature to tackle the problem of sentiment analysis. Hammer (2014) [10] presented a method of using machine learning to detect threats of violence from a data set of YouTube comments written in English language. The method described in the paper uses logistic LASSO regression analysis on bigrams of important words to classify sentences as violence or not. The dataset contains 24840 sentences from YouTube and was manually annotated as violent threat or not. The features are bigrams of two of these important words observed in the same sentence. The paper did not describe properly how these important words were selected, report only that words were chosen that were correlated with the response (violent/non-violent). However, it appears likely that the words were arrived at using LASSO regression. The obtained result shows Accuracy of 0.9466. The shortcoming of this study was the using of the logistic LASSO regression analysis, which has the limitation of performing an implicit feature selection while estimating the model. In another work, Djuric et al. [6] proposed a method to detect hate speech comments in English language by using two steps. First, they used paragraph2vec to convert a generic block of text into a vector using continuous bag of words (CBOW) which used for predicting the word given its context. Second, they used logistic regression classifier. The proposed model compared to TF and TF-IDF using area under curve (AUC) metric. The results show that Paragraph2vec outperforms TF and TFidf. In their paper, Yadav and Manwatkar (2015) [11] developed a social media networks prototype that aims to automatic filtering of offensive content in social media networks before sharing it. They applied AHO-Corasick string pattern matching algorithm. The idea of the proposed algorithm is based on matching the pattern (i.e. offensive keywords) from the input text with database that have offensive keywords collected from different datasets. After the word is detected they simply replaced it with some special character to prevent it from share. They used breadth first search to find the offensive keywords. The prototype shows good results even when using slang language. Gitari et al. (2015) [12] created a model classifier that uses sentiment analysis techniques and in particular a lexicon-based approach to automatically detect hate speech in online forums, blogs and comments in news reviews using semantic and subjectivity features. They collected blogs from Raymond Franklin sites that are considered to be generally offensive, they refer to this as first corpus. Second corpus consists of largely paragraph related to the Israel-Palestinian conflict. They concentrated classifying hate speech detection into three key target groups of race, nationality and religion. The achieved results show that precision, recall, and F-score are the best when using semantic orientation, hate verbs and theme-based as feature sets in first corpus and using subjective sentences only, while the results much less when used same feature sets but without using subjective sentences. From my opinion they could increase precision and recall scores if they applied machine learning. Waseem and Hovey (2016) [13] presented a method for hate speech detection on Twitter that is written in English. The method is divided into three parts. First, they collected 16K tweets that contain racist and sexist hate speech, then they proposed a list to identify hate speech which helps the annotators to reliably identify hate speech. Second, they examined the impact of different extra-linguistic features in coupling with character n-grams for hate speech detection instead of using word n-gram due to character n-gram being far less sparse than the word n-gram. In third part, they used a grid search in order to select the most suitable features, after that they evaluated the selection features by using logistic regression classifier and 10-fold cross validation. The obtained results show that using character n-grams of length up to 4 with gender as an additional feature have got 73.93 F1-score, which it is the best compared with location and word n-grams. The limitation in this article centred in extracted location feature which it need to consider more than just the tags Twitter provides, which affected on the final result. Alhelbawy et al. ( 2016) [14] presented a new corpus of violent tweets event; the dataset was manually labeled for seven classes of violence (Human rights abuse, Political opinion, Accident, Crime, Conflict, Crisis, Other) and annotated using popular crowdsourcing platform crowdflower. The work was targeting the Arabic tweets that are related to violence. Two filters are then applied to filter (Redundant tweets, emotional tweets, delete short tweets and sexual adverts), after that they used the confidence score to evaluate different sub-sets for each tweets. The obtained results show that crowd classification is overall reliable and can be used for further research on violence on social media. Mubarak et al. (2017) [15] presented an automated method to create and expand a list of obscene words that help to detect abusive language on Arabic social media. They classify Twitter users based on the using of abusive words or not. By using Twitter API with language filter set to Arabic they collected tweets patterns that are usually used in offensive communications. After that, they manually considered if the words that appeared in collected tweets are obscene or not. Magu et al. ( 2017) [1] introduced a mechanism to identify hate coded content on social media written in English language. For example, user have used words skype, google to represent Jew, black respectively. This hate coded leads to violence over social media. They extracted 1999 tweets separated to 1048 labeled hateful and the rest were labeled non-hateful. They used dataset to calculate the Pearson correlation coefficient appearance of every term exist in the dataset and the class label, this method helps them to extract the most correlated terms for identifying hate tweets. After that, they ran the Apriori algorithm to extract frequent item set. Lastly, to train classifier they applied a bag of words model to represent the most popular word in the corpora of training dataset with a Boolean feature vector, then they collected 23,401 tweets to build model using support vector machine with linear kernel to be able separate the hateful tweets and using 10-fold cross validation. They achieved an accuracy of 0.794 with precision of 0.794 and recall of 0.795. Abdelfatah et al. (2017) [16] proposed a new framework aiming at separating violent and non-violent Arabic tweets over Twitter by using sparse Gaussian process latent variable model (SGPLVM) followed by k-means. Experiment started by collecting 16234 Arabic tweets then pre-processing tweets by removing stop word and web links, after that they annotated manually by at least five different annotators and only tweets have a confidence score more than 0.7 are applied, then two set of experiments have been carried out. The first one is reduced dimensions with PCA and then apply K-means. Second experiment is using SGPLVM to reduce dimension then apply K-means, after that they compare between two results. The obtained results show that using SGPLVM with K-means it better than PCA with K-means and using unsupervised techniques to detect violent tweets in low dimensional representation be better than applying clustering on the original data. --- Proposed Method This section describes in details the methodology adapted in this research. The main processes in the research are: Data collection, Data preprocessing, Arabic lexicon, Feature extraction, Data classification and Model evaluation. Figure 1 illustrates the various processing steps of the proposed method. --- Data collection Data collection is the process of gathering large amount of data, in an established systematic fashion that enables one to prove research questions, test hypotheses, and evaluate results. In order to test our model, we need to use some datasets that contain data from Arabic audience. There are few datasets that contain this information but they are not useful in our research since they are collected in Arabic delicts that are different from our target delict which is the Jordanian Arabic delicts. In order to overcome this situation, we have built dataset from scratch that contain Jordanian dialect opinions about subjects related to politics and sports. The created dataset contains set of comments collected in various times from two popular social networks -Facebook and twitter [25]. These comments are political and sport comments that might lead to violence. Table 1 shows example of the collected data divided into violence and normal data. --- Data collection methods In this study, two methods were used to collect data from Facebook and Twitter: automatic and manual collection. In the first method, we used scripts to collect data automatically from the tweets of twitter. The number of tweets collected is almost 10,330 tweets. The number of useful tweets after excluding redundant and advertisement tweets is 385, 335 among them are violence and the rest is regular. The automatic collection of data was not enough, so we used manual collection. The main reasons for adopting this method despite the time and effort needed are: • Most of tweets returned from automatic collection are violence and it is hard to find keyword not including violence content. • There is no automatic data collection for Facebook. • Huge number of returned tweets are advertisement, links and comments on picture. Three persons that are aware of Jordanian dialect, politics and sport performed the data collection. They revised around 60000 tweets and comments. The final number of collected tweets and comments is 2057. Table 2. shows source, the domain and the number of followers of the groups that were used to collect data from. The next step after data collection is data annotation. Each tweet or comment should be classified as violence or normal. Table 3 summarizes the results of data annotations. --- Data preprocessing In the pre-processing stage, various Natural Language Processing (NLP) techniques are applied [9]. Several preprocessing strategies can be applied on sentiment analysis that affect the accuracy when applied to Arabic text. The pre-processing is performed on different stages: Tokenization, normalization, stop-word removal, and stemming. In this research, we applied different pre-processing techniques. At the beginning, tokenization is used to break up the text into tokens. Next, normalization is applied to convert all various forms of a word to a common form. It is the process of transform-ing text into a single canonical form since input is guaranteed to be consistent before operations are performed on it [17]. In this study, normalizer performs this specific task according to the rules listed below: • Replacing: in this stage we want to replace any words contain characters " <unk> " replace by " <unk> ", <unk>"<unk>"<unk> replace by <unk>"<unk>"<unk>. • Removing the "tatweel" character " _ ". • Removing the diacritics: " <unk> <unk> <unk> <unk> <unk> <unk> <unk> <unk> <unk> " will be <unk>"<unk> <unk>."<unk> • Removing punctuation and special characters. • Remove English letters and numbers. • Remove hyperlink: some of collected data contain link such as for continue reading the news, in this situation we deleted the link and keep the news text. • Remove duplicate letters such as <unk>"<unk>"<unk> which should be <unk>."<unk>"<unk> The normalized text is then handled in order to remove stop words. These words are removed because they don't add any new information to the text. A list of stop words such as" <unk> <unk>,<unk> <unk>,<unk> " is prepared and applied to the text. The last process applied is stemming. In stemming, all affixes of the words are removed. For instance, the word <unk>"<unk>"<unk> is stemmed into <unk>."<unk>"<unk> Light stemming python library used for this purpose. It provides a configurable stemmer and segmented for Arabic text (Zerrouki,2012). --- Normalizing the dataset In this phase, after extracting the features, the dataset is normalized. Normalization is generally performed during the data preprocessing. Since the scaling of values of dataset attributes range varies widely so it helps to fit into specific range. There are different normalization types such as Z-score and Min-Max normalization [18]. In this research Min-Max normalization type is used to scale data into range [0,1] because the dataset features have different range such as the range of feature the total frequency of violence words is [0,6]. The following equation is the formula of Min-Max normalization to scale the data in range [0,1]. --- v′ = v-min A max A -min A (1) Where v refers to original value and v′ refers to new value. --- Arabic lexicon Many researches work in the problem of sentiment analysis use sentiment lexicons also known as senti-lexicons. In this research we relied on Arabic senti-lexicon (ASL) that proposed by [19] where each word is assigned a polarity (positive, negative). The sentiment lexicon is considered as the most crucial resource for features extraction that help us to make model more accurate. ASL includes 13,760 positive and negative words, 3880 synset terms which have different words and same meaning that collected manually and dialects synst(D-synset) which are words used in the different Arabic dialects and have the same meaning. Based on ASL, we built our own lexicon to detect violence. Firstly, ASL was modified to become appropriate for violence detection where negative is changed to normal and positive is changed to violence. Secondly, 304 violence words and 225 normal words were added manually to ASL. Finally, we deleted D-synset because we apply Jordanian Dialect and we deleted the score field. In addition, we added set of part of speech (POS) and synset and inflection manually. Our new lexicon includes a list of 10,443 violence and normal words and 2450 violence and normal sentiment. An example from our new lexicon is shown in table 5. --- Features extraction Feature extraction is an important task in the sentiment analysis and more generally in text categorization. Since the text is unstructured, we need to convert original documents into feature vectors which is the main step in any supervised learning approach attached to sentiment analysis to select the right features that determine the overall performance of sentiment classification. Consequently, this work studies the effect of applying various pre-processing techniques on the extracted set of features and its impact on the overall performance. In this research, we define a group of features. These features can be grouped into four main feature groups: 1. Feature based on sentiment word of (Violence, Normal), presence and frequency. In this group, we defined six features: The presence of violence words, the total frequency of violence words, the presence of normal words, the total frequency of normal words, the ratio of the presence of violence to normal words, and the ratio of the total frequency of violence to normal words. 2. Bag-Of-Words (BOW) [20]: According to the BOW model, the document is represented as a vector of words in Euclidian space where each word is independent from others [21]. We used the feature n-grams and its Term Frequency-Inverse Document Frequency (TF/IDF). The n-grams of texts are used in text mining and natural language processing tasks. N-grams are simply all combinations of adjacent words or letters of length N that you can find in text. An n-gram of (N= 1) is referred to "unigram"; (N=2) is a "bigram": N= 3 is a "trigram". In this research we used the unigram, bigram and trigram. --- Features based on POS: The process of POS tagging allows to tag each word of text in terms of which part of speech it belongs to: noun adjective, verb, etc. The goal is to extract patterns in text based on analysis of frequency distributions of these part-of-speech. There is no common opinion about whether POS tagging improves the results of sentiment classification. Barbosa and Feng (2010) [22] reported positive results using POS tagging, while Kouloumpis et al. ( 2011) [23] reported a decrease in performance, so it depends on domain that you work in. We defined nine features: total number of violence adjective, total number of violence verb, total number of violence noun, total number of normal adjective, total number of normal verb, total number of normal noun, the ratio of the total frequency of violence adjective to normal words, the ratio of the total frequency of violence verb to normal words, and the ratio of the total frequency of violence noun to normal words. 4. Other features: We defined two features: the presence of negation word in sentence and the presence of violence and normal emoticons in sentence. Table 6 represents the summary of features extracted from (Tweet, Post). Another result from the models to measure is the AUC (area under curve), to calculate this measure values using Python Sklearn matrix, the model should be first trained and tested using the test data, then the evaluation methods can be applied to calculate the final results (which are presented in the experiment results section). These metrics of AUC and ACC are commonly used in the evaluation of the link prediction problems. After the model building and evaluation phases, the main results from this research have to be evaluated by comparing the results from the different models. There are two ways to compare these results. The first traditional one is by having that comparison manually using the resulted values from the described evaluation matrices. The second approach is more scientific method to compare the whole models' accuracy using the statistical significance tests and then apply the AUC after having confidence in the differences between models. --- Experiments In this research, three classifier methods were used to detect violence in Arabic language using sentiment analysis; SVM, NB and KNN. These methods used due to their effectiveness, simplicity and accurateness. --- Parameters of classification algorithms Generally, many machine learning algorithms need set of parameters to be assigned. Table 7 lists the algorithms used, the parameters of classification algorithms as well as the selected values. --- Table 7. parameters of classification algorithms --- Model evaluation To evaluate the quality and usefulness of the model, several experiments were conducted on our dataset. All algorithms were evaluated using 10-fold cross validation. The measures used to evaluate the models are based on the confusion matrix depicted on Table 8. In most of sentiment analysis problems, three measures are used to evaluate the model: Accuracy, precision and recall, in addition we used f-measure to measure the accuracy of test data as it considers both precision and recall. The following describe these measures: --- Table 8. Confusion Matrix • Accuracy: Tthe proportion of the total number of predictions that were correct classified. Accuracy = TP+TN TP+TN+FP+FN(3) • Recall (R): Tthe proportion of the number of correct positive predictions divided by the total number of positives. Recall = TP TP+FN(4) • Precision (P): the proportion of the number of correct positive predictions divided by the total number of positive predictions. Precision = TP TP+FP(5) • F-Measure: is the weighted harmonic mean of precision and recall. F-Measure = 2 * R * P (R+P)(6) In order to evaluate the robustness of the classifier, the standard deviation (Stdv) of the 10-folds is calculated and reported. Classifiers with lower Stdv show more robustness. The formula of Stdv is listed below: (7) where n is the number of data points,x <unk> is the mean of X i and X i is each of the values of the data. --- Experimental results This section provides the detailed experiment results. The purpose of this experiment is to evaluate the violence detection over social media by developing language resources for Arabic sentiment analysis. It lists the results of features that extracted from sentence without using Arabic lexicon such as n-gram and features extracted using Arabic lexicon such as number of violence words in sentence, then a compari-son of the results for various preprocessing techniques that were used with the features extracted. All experiments have been performed using Python 2.7, Anaconda Spyder 3.2.3. The experiments have been conducted in a computer with Intel(R) Core(TM) i7-6700HQ CPU @ 2.60GHz 2.60GHz running Windows 10 64-bits. The computer contains 32 GB RAM. --- Results based on n-gram feature with various preprocessing files This section represents the comparison results between the different types of ngram features (unigram, bigram and trigram) on various preprocessing techniques by using three classifiers, SVM, NB and KNN. In order to evaluate n-gram feature, four measures used: recall, precision, accuracy and f-measure. Table 9 shows the results of n-gram feature. In first column is the file that result after applying the following preprocessing techniques: 1. Normalization file: This file contains data that has been applied to normalization without stemming and remove stop words. 2. Stop words file: This file contains data that has been applied to normalization and remove stop words. 3. ISRI stemming file: This file contains data that has been applied to normalization and ISRI stemming. 4. Light stemming file: This file contains data that has been applied to normalization and light stemming. 5. ISRI stemming and stop words file: This file contains data that has been applied to normalization, remove stop words and ISRI stemming. 6. Light stemming and stop words file: This file contains data that has been applied to normalization, remove stop words and light stemming. We can notice that, NB performed better than the others on normalization file with bigram, stop words with unigram, light stemming with bigram. On the other hand, SVM surpass others on ISRI stemming file unigram feature and ISRI stemming with stop words file bigram feature. Overall, the best results recorded using SVM on ISRI stemming with stop words file with bigram feature. The least result was when using KNN classifier with trigram feature on ISRI stemming and stop word file. --- Results based on sentiment words of presence and frequency features This section presents the comparison results between different classifiers when we used Sentiment words of presence and frequency features (presence of violence word, the total frequency of violence words, presence of normal words, total frequency of normal words, ratio of the presence of violence to normal word and ratio of the total frequency of violence to normal words) that extracted from Arabic lexicon. In this case, there is no need to compare between preprocess files. Table 10 reveals that the best measured results recorded by SVM, followed by KNN and the worst performance is reported by NB. Moreover, classifies show stable models and the differences are insignificant. --- Results based on Part of Speech (POS) features This section represents the comparison results between different classifiers when we used part of speech features (total number of violence adjective, total number of violence verb, total number of violence noun, total number of normal adjective, total number of normal verb, total number of normal noun, the ratio of the total frequency of violence adjective to normal words, the ratio of the total frequency of violence verb to normal words, the ratio of the total frequency of violence noun to normal words) that extracted from Arabic lexicon so there is no need to compare between preprocess files. Table 11 lists the results Based on POS features. NB followed by SVM and finally KNN records the best results in terms of accuracy, precision, recall and Fmeasure. --- Results based on other features This section represents the comparison results between different classifiers when we used other features (presence of negation word in sentence, presence of violence emoticons and presence of normal emoticons in sentence). Table 12 depicts the results based on other features. It shows that SVM performs better than NB and KNN and produces more stable models. --- Results based on all features without n-gram This section represents the comparison results between different classifiers when we used all features except n-gram feature. Table 13 represents the results Based on all features except n-gram. As shown in table 13, it reveals the best measured results when using NB classifier. --- Results based on all features with n-gram feature on various preprocessing files This section represents the comparison results between different classifiers when we used all features with the n-gram feature on various preprocessing techniques as presented in the Table 14. when using POS features that are extracted from lexicon, NB outperformed the other classifiers. NB also outperformed other methods in the experiment that uses all feature extracted from lexicon with n-gram. Furthermore, the experiment of using all features that extracted from lexicon with N-gram feature on various preprocessing files shows that SVM outperformed the other methods in ISRI stemming and stop word file among the six files when used with unigrams feature. SVM again outperformed other methods when applied on ISRI stemming with stop words file with unigram and all features. --- Conclusion and Future Work This study uses sentiment analysis to detect violence over social media in Arabic language. Set of steps have been applied systematically to address this problem. First, data was collected from two of the popular social media web sites (Facebook and Twitter), then this data was annotated as (Violence or Normal). After collecting the data, set of preprocessing techniques applied to normalize the data. Then, ASL was adapted and modified to capture the objective of this study and features extraction process was conducted to distinguish the sentences. These sentences were classified using (SVM, NB and KNN). Finally, the model was evaluated using (recall, precision, accuracy, F-measure and study). In the light of this methodology, research questions have been addressed in details, and it has been found that using sentiment analysis can significantly detect the violence in Arabic language (Jordanian dialect). also, it has been concluded that the using of the Arabic lexicon with modification would help to extract features for detecting violence, what distinguishes this lexicon is the ability to determine whether the sentences are normal or violence as well as it can determine POS in the words. Moreover, this research shows that combining features extracted from lexicon with features extracted from sentence such as n-gram generates models that are more accurate. In addition, the study discusses the effect of various preprocessing techniques on the performance of the generated models. The experiments proved that using SVM classifier on ISRI stemming with stop words file with unigram and all features gave best results compared with other experiments. In the future, the method can be extended to cover other Arabic dialects and the dataset can be expanded. --- Table 14 shows that NB performs better than SVM and KNN on the normalization files regardless of the n-gram selected. It gives the same results using all features when n-gram is unigram, bigram and trigram. In the case of stop word file NB recorded better results too in all features with all n-gram variations. On the other hand, SVM is performing better than the NB and KNN in the other files (ISRI stemming, ISRI stemming with stop words, light stemming and light stemming with stop). We can notice that the maximum accuracy scored was by SVM classifier on ISRI stemming with stop words file with unigram and all features. To sum up, in the experiment of N-gram feature with various preprocessing files, the optimized SVM outperformed other classification algorithms in ISRI stemming and stop word file out of six files when used with bigram feature. The experiment of using Sentiment words of presence and frequency features that extracted from lexicon on dataset shows that SVM outperformed other classification algorithms. Moreover, Authors Monther Khalafat is a Jordanian computer scientist and the main contributor to this work, and was a student at the University of Jordan, King Abdullah II School for Information Technology, Department of Information Technology, Amman (Jordan). Email: [email protected] Dr. Ja'far Alqatawna is a Jordanian Dr. of Business Security at the University of Jordan, King Abdullah II School for Information Technology, Department of Information Technology, Amman (Jordan). Currently, he is on sabbatical leave at the Higher Colleges of Technology, Faculty of Computer Information Systems, Dubai, UAE. E-mails: [email protected], [email protected] Prof. Rizik Al-Sayyed is a Jordanian Prof. of Networks, Databases, and Data Science at the University of Jordan, King Abdullah II School for Information Technology, Department of Information Technology, Amman (Jordan). E-mail: [email protected] Dr. Mohammad Eshtay is a Jordanian Dr. of Machine Learning at LTUC, from the University of Jordan, King Abdullah II School of Information Technology, Amman, Jordan. E-mail: [email protected] Dr. Thaeer Kobbaey is

Today, the influence of the social media on different aspects of our lives is increasing, many scholars from various disciplines and majors looking at the social media networks as the ongoing revolution. In Social media networks, many bonds and connections can be established whether being direct or indirect ties. In fact, Social networks are used not only by people but also by companies. People usually create their own profiles and join communities to discuss different common issues that they have interest in. On the other hand, companies also can create their virtual presence on the social media networks to benefit from this media to understand the customers and gather richer information about them. With all of the benefits and advantages of social media networks, they should not always be seen as a safe place for communicating, sharing information and ideas, and establishing virtual communities. These information and ideas could carry with them hatred speeches that must be detected to avoid raising violence. Therefore, web content mining can be used to handle this issue. Web content mining is gaining more concern because of its importance for many businesses and institutions. Sentiment Analysis (SA) is an important sub-area of web content mining. The purpose of SA is to determine the overall sentiment attitude of writer towards a specific entity and classify these opinions automatically. There are two main approaches to build systems of sentiment analysis: the machine learning approach and the lexicon-based approach. This research presents the design and implementation for violence detection over social media using machine learning approach. Our system works on Jordanian Arabic dialect instead of Modern Standard Arabic (MSA). The data was collected from two popular social media websites (Facebook, Twitter) and has used native speakers to annotate the data. Moreover, different preprocessing techniques have been used to show their effect on our model accuracy. The Arabic lexicon was used for generating feature vectors and separate them to features set. Here, we have three well known machine learning algorithms: Support Vector Machine (SVM), Naive Bayes (NB) and k-Nearest Neighbors Paper-Violence Detection over Online Social Networks: An Arabic Sentiment Analysis Approach (KNN). Building on this view, Information Science Research Institute's (ISRI) stemming and stop word file as a result of preprocessing were used to extract the features. Indeed, several features have been extracted; however, using the SVM classifier reveals that unigram and features extracted from lexicon are characterized by the highest accuracy to detect violence.

of Business Security at the University of Jordan, King Abdullah II School for Information Technology, Department of Information Technology, Amman (Jordan). Currently, he is on sabbatical leave at the Higher Colleges of Technology, Faculty of Computer Information Systems, Dubai, UAE. E-mails: [email protected], [email protected] Prof. Rizik Al-Sayyed is a Jordanian Prof. of Networks, Databases, and Data Science at the University of Jordan, King Abdullah II School for Information Technology, Department of Information Technology, Amman (Jordan). E-mail: [email protected] Dr. Mohammad Eshtay is a Jordanian Dr. of Machine Learning at LTUC, from the University of Jordan, King Abdullah II School of Information Technology, Amman, Jordan. E-mail: [email protected] Dr. Thaeer Kobbaey is a Jordanian Dr. of Data Mining and Big Data at The Higher Colleges of Technology, Dubai (UAE). E-mail: [email protected]

Today, the influence of the social media on different aspects of our lives is increasing, many scholars from various disciplines and majors looking at the social media networks as the ongoing revolution. In Social media networks, many bonds and connections can be established whether being direct or indirect ties. In fact, Social networks are used not only by people but also by companies. People usually create their own profiles and join communities to discuss different common issues that they have interest in. On the other hand, companies also can create their virtual presence on the social media networks to benefit from this media to understand the customers and gather richer information about them. With all of the benefits and advantages of social media networks, they should not always be seen as a safe place for communicating, sharing information and ideas, and establishing virtual communities. These information and ideas could carry with them hatred speeches that must be detected to avoid raising violence. Therefore, web content mining can be used to handle this issue. Web content mining is gaining more concern because of its importance for many businesses and institutions. Sentiment Analysis (SA) is an important sub-area of web content mining. The purpose of SA is to determine the overall sentiment attitude of writer towards a specific entity and classify these opinions automatically. There are two main approaches to build systems of sentiment analysis: the machine learning approach and the lexicon-based approach. This research presents the design and implementation for violence detection over social media using machine learning approach. Our system works on Jordanian Arabic dialect instead of Modern Standard Arabic (MSA). The data was collected from two popular social media websites (Facebook, Twitter) and has used native speakers to annotate the data. Moreover, different preprocessing techniques have been used to show their effect on our model accuracy. The Arabic lexicon was used for generating feature vectors and separate them to features set. Here, we have three well known machine learning algorithms: Support Vector Machine (SVM), Naive Bayes (NB) and k-Nearest Neighbors Paper-Violence Detection over Online Social Networks: An Arabic Sentiment Analysis Approach (KNN). Building on this view, Information Science Research Institute's (ISRI) stemming and stop word file as a result of preprocessing were used to extract the features. Indeed, several features have been extracted; however, using the SVM classifier reveals that unigram and features extracted from lexicon are characterized by the highest accuracy to detect violence.

Introduction There is growing evidence that neighbourhood green space is beneficial for mental health (Alcock et al. 2014;Di Nardo et al. 2012;Hartig et al. 2014;Van den Berg et al. 2015). The neighbourhood social environment has been suggested to be one of the mechanisms. The presence of green, such as trees or vegetation increases the attractiveness of common spaces in the neighbourhood, thereby potentially increasing their use (Coley et al. 1997;Kuo et al. 1998), and facilitating informal social contacts between community members (Hartig et al. 2014;Kuo et al. 1998). Social contacts are health promoting; for instance through the social support they can offer (Cohen 2004). By facilitating social contacts, neighbourhood green can contribute to the development of neighbourhood social cohesion, i.e. the connectedness and solidarity among community members, which has proven to benefit people's health (Di Nardo et al. 2012;Kawachi and Berkman 2000). Furthermore, having green areas in the neighbourhood increases the attractiveness of the living environment, thereby enhancing people's attachment to the physical neighbourhood environment (Di Nardo et al. 2012). Place attachment helps to create group identity, which translates into a general sense of well-being (Brown et al. 2003) and has been associated with reduced loneliness and better mental health (Hagerty and Williams 1999;Pretty et al. 1994). The neighbourhood social environment as a mechanism for the impact of neighbourhood green space on mental health has received some research attention in the past years. Some studies found that social cohesion mediated the relation between green space and mental health (de Vries et al. 2013;Sugiyama et al. 2008), while others did not (Triguero-Mas et al. 2015). Lack of social support and feelings of loneliness were reported to mediate the relationship between green space and mental health (Maas et al. 2009), but not social contacts (Maas et al. 2009;Sugiyama et al. 2008). Inconsistencies between studies might be explained by different operationalisations of the social environment (e.g. social cohesion, individual social contacts, loneliness). It is also possible that the relationship between neighbourhood green, social environment and mental health differs across cultures (Hartig et al. 2014). For instance, in more individual oriented cultures, green space might be more important for the facilitation of social interactions than in more collectivist cultures where communal life is already more common. In the current study, we investigate the relationship between neighbourhood green space, neighbourhood social environment, and mental health in four European cities to examine if the social environment might be one of the mechanisms between neighbourhood green and mental health. The following research questions are addressed: is neighbourhood green space related to the neighbourhood social environment in four European cities? Are the neighbourhood social environment and neighbourhood green space related to mental health in these cities? This study uses a range of social environment measures (social cohesion, neighbourhood attachment, and individual social contacts) to examine if the associations depend on the operationalisation of social environment. Our green measures comprise both the amount and quality of neighbourhood green, to accommodate the increasing evidence stressing the importance of quality of green space and its impact on health (Francis et al. 2012;Hartig et al. 2014;Van Dillen et al. 2012). Furthermore, objective audit and subjective green measures are used as they may capture different aspects of greenness i.e. more emotional aspects with subjective measures and more tangible aspects with objective measures (Francis et al. 2012). These aspects may relate to the social environment characteristics and mental health differently (Leslie et al. 2010). --- Methods --- Study background This EU-funded PHENOTYPE study examined the health effects of the natural environment and its underlying mechanisms. A cross-sectional survey was carried out from May to October 2013 in four cities across Europe: Stoke-on-Trent (United Kingdom), Doetinchem (Netherlands), Barcelona (Spain), and Kaunas (Lithuania) (Nieuwenhuijsen et al. 2014). --- Study population and data collection In each city, 30 neighbourhoods varying in neighbourhood green space and socioeconomic status (SES) were selected (see Table 1 for a description of the neighbourhoods). Survey data were collected using face-to-face interviews, with the exception of Lithuania, where data were collected with a postal questionnaire. Around 1000 adults aged 18-75 years, were interviewed per city (n = 3947, overall response rate 20%) across 124 neighbourhoods. For further details on the data collection see Online Resource 1. We selected respondents with complete data for the indicators of interest, providing a sample of 3771 respondents in 124 neighbourhoods (96% of the study population). Additionally, in each neighbourhood an audit was carried out to assess the amount and quality of green space. For each neighbourhood a purposeful sample of streets was selected, ensuring that rare, but important, features of the neighborhood were included (e.g. parks). To do so, we divided each neighbourhood into more or less homogeneous sub-areas by means of land use maps in combination with local knowledge of the areas. Per sub-area, several streets were selected and combined into a route that was inspected by two trained auditors (in a small number of cases by one auditor) in a systematic way, using a form containing closed questions. --- Measures --- Mental health Mental health was measured using the mental health inventory (MHI-5) (Ware Jr and Sherbourne 1992). MHI-5 assesses nervousness and feelings of depression in the past month, with answers ranging from 'all the time' to 'never' on a six-point scale. Sum scores of the five answers were transformed into a scale from 0 to 100 (Ware Jr et al. 1995), with higher scores reflecting better mental health. The scale has proven to be of good validity and reliability (Ware Jr 2000). --- Neighbourhood green space Audit amount and quality of neighbourhood green space Amount of neighbourhood green space was based on six items containing information about the fraction of visible gardens, garden size, the arrangement of the gardens, number of trees, size of public green spaces, and size of public blue spaces (Cronbach's alpha 0.66). Quality of neighbourhood green space was derived from one question, answered by the auditors: 'what is your general impression of the quality of the green space in this neighbourhood'? Answers ranged from 1 (very negative) to 5 (very positive). Indicators were standardised using Z-scores, calculated for each city separately. This way, neighbourhood green was compared between the neighbourhoods within one city and not across all cities, allowing the examination of the relative effect of green space on mental health. Subjective amount and quality of neighbourhood green space Subjective amount of neighbourhood green space was measured by asking the respondents: 'How would you describe your neighbourhood in terms of green space', with answers on a five-point Likert scale from 'not green at all' (1) to'very green' (5). Subjective quality of neighbourhood green space was measured by asking: 'Overall, in your neighbourhood, how satisfied are you with the quality of the green/blue environment?' Answers ranged on a five-point Likert scale, with a higher score meaning more satisfaction with the quality. We conducted ecometric analyses to calculate the average perception of neighbourhood green space (see Online Resource 2 for a description of the ecometric analysis) (Raudenbush and Sampson 1999). This way, we can include subjective assessments of neighbourhood green space, while avoiding'same-source bias' (also measured at the same-time) (de Jong et al. 2011;Wheaton et al. 2015). Ecometric average scores were calculated (stratified by city) and standardised into country-specific Z-scores. We use the term neighbourhood green space for our natural environment measures, because the audit showed that the neighbourhood natural environment consisted foremost of green elements and because mainly green space is relevant for the social interaction mechanism. --- Social environment We measured three aspects of the social environment. Social cohesion constructed by summing the answers to five statements from the social cohesion and trust scale (Sampson et al. 1997): 'People are willing to help their neighbours', 'This is a close-knit neighbourhood', 'People in this neighbourhood can be trusted', 'People in this neighbourhood generally don't get along with each other' (reversed), and 'People in this neighbourhood do not share the same values' (reversed). Using a 5-point Likert scale, answers ranged on from 'totally disagree' to 'totally agree'. Negatively stated items were recoded so that a higher score reflected higher levels of social cohesion (Cronbach's alpha 0.76). Neighbourhood attachment measured by summing the answers to three statements: 'I feel attached to this neighbourhood', 'I feel at home in this neighbourhood', and 'I live in a nice neighbourhood were people have a sense of belonging', using a 5-point Likert scale, answers ranged on from 'totally disagree' to totally agree. A higher score reflected stronger neighbourhood attachment (Cronbach's alpha 0.80). Social contacts respondents were asked how often they had contact with their neighbours. Answers were: 'daily', 'at least once a week', '1-3 times per month', 'less than once a month', and'seldom or never'. Social contacts was dichotomised into 'at least once a week' versus 'less often' for the analyses with social contacts as outcome measure. Similar to the subjective green measures, ecometric analyses were conducted to calculate the neighbourhood average scores of social cohesion and neighbourhood attachment (see Online Resource 2). Social contacts were included at the individual level. The correlations between the neighbourhood characteristics (Online Resource 3) show that the audit and perceived green measures were moderately related, suggesting that these indicators measured different aspects of neighbourhood green space. --- Confounders Individual control variables in all analyses were sex, age (in years), highest achieved educational level (primary school /no education; secondary school/further education; university degree or higher), nationality (country nationality; other), employment status (fulltime employed; other), household composition (with children under 12 years; other), and homeownership (yes; no). Neighbourhood socioeconomic status (SES) (low; intermediate; high; based on country-specific data, see Online Resource 1) was included as a neighbourhood level confounder. See Table 2 for the descriptive statistics. --- Analyses Multilevel linear and logistic regression analyses were performed, with individuals at level one, neighbourhoods at level two, and city at level three. City was included as level to adjust for systematic differences in the intercept between the four cities, i.e. city differences caused by, for instance, policy differences. The green variables were allowed to have a different effect (slope) on social environment and health for every city, by creating a separate green indicator variable for every city (green indicator X city-dummy (1 = belongs to this city, 0 = does not belong to this city)). All four city green variables are added to the model (Weisberg 2005). First, multilevel models assessed the association between neighbourhood green space and individual level social contacts in the four cities. Ecological models at the neighbourhood level assessed the associations between neighbourhood green, social cohesion, and neighbourhood attachment, respectively. Next, we examined the associations between social cohesion, neighbourhood attachment, social contacts and mental health in the four cities, while adjusting for green space. Finally, we examined the associations between green space at the neighbourhood level and mental health in the four cities. The analyses with the subjective neighbourhood level green measures were also adjusted for the individual perception of neighbourhood green space, to distinguish the contextual health effect of green space from the individual level effect. Analyses were conducted using SAS 9.3. --- Results --- Neighbourhood green space and the social environment More cohesive neighbourhoods were greener and had better quality green space in Doetinchem (perceived and audit) and in Stoke-on-Trent (perceived amount; perceived and audit quality) (Table 3). In Barcelona and Kaunas, neighbourhood-level green space was not related to neighbourhood social cohesion. Stronger neighbourhoods attachment was found in greener neighbourhoods (perceived) and neighbourhoods with better quality green space (audit and perceived) in Doetinchem (Table 3). Better perceived quality of neighbourhood green was associated with stronger neighbourhood attachment in Barcelona and Stoke-on-Trent as well. Neighbourhood green space was not associated with social contacts in any of the cities. --- Social environment and mental health Residents living in neighbourhoods with more social cohesion or with stronger neighbourhood attachment reported better mental health only in Stoke-on-Trent, not in the other cities (Table 4). Having more frequent social contacts was associated with better mental health consistently in all four cities. --- Neighbourhood green space and mental health In Barcelona, a higher amount of neighbourhood green (audit) was associated with better mental health 4). In the other three cities, neighbourhood green space was not associated with mental health. --- The social environment as possible mechanism In Barcelona, we found no associations between neighbourhood green space and (one aspect of) the social environment (Table 3) and between the (same aspect of the) social environment and mental health (Table 4). In the other cities, we found no associations between neighbourhood green space and mental health (Table 4). Therefore, we found no indications that the social environment could be an underlying mechanism between neighbourhood green space and mental health. --- Discussion Greener neighbourhoods and neighbourhoods with better quality green space were more cohesive and had higher levels of neighbourhood attachment in Doetinchem and Stoke-on-Trent. More neighbourhood cohesion and stronger neighbourhood attachment were associated with better mental health in Stoke-on-Trent only. Only in Barcelona, however, the neighbourhood green space was associated with better mental health, but there, we found no indications that the social environment could be the underlying mechanism. --- Study limitations The cross-sectional design of this study prevents conclusions about the causality of the relationships (Galster 2008). We therefore did not implement statistical tests for mediation, as mediation implies causal processes. Another limitation is the low response rate (see Online Resource 1), resulting in an underrepresentation of low educated people in all four cities. It is suggested that people with a low socioeconomic status (SES) may benefit more from neighbourhood green space than those with a high SES (Mitchell and Popham 2008). The underrepresentation of low educated people may therefore have resulted in an underestimation of the relationship between green space and mental health. Third, in Kaunas, there was no variation between neighbourhoods in neighbourhood attachment and, as pointed out by the low reliability scores of green space and social cohesion in Table 2, only little neighbourhood variation in case of the other neighbourhood measures (Hox 2010). Because of the low reliability scores, we excluded results from Kaunas based on the perception measures in the discussion of the implications. Finally, the neighbourhoods in Barcelona were considerably smaller in size compared to the other cities. This could have increased the chance that the Spanish neighbourhoods were more homogeneous in terms of the amount and quality of neighbourhood green space, which could have resulted in more precise audit assessment of the neighbourhood green space in Barcelona. We cannot rule out completely that a more precise audit assessment of the green space in Barcelona resulted in finding a relation between audit amount of green space and mental health there. Neighbourhood green space and the social environment Our study showed that green space at the neighbourhood level was related to the neighbourhood social environment. Our findings that related social cohesion consistently to neighbourhood green space in Doetinchem and Stoke-on-Trent strengthens the evidence on the influence of green space on the development of social cohesion. Furthermore, in line with Arnberger and Eder (2012) we report neighbourhood attachment to be consistently associated with neighbourhood green space in Doetinchem, as well as the subjective quality of neighbourhood green in Barcelona and Stoke-on-Trent. We found no evidence that neighbourhood green space is related to more contacts between neighbours, in line with Maas et al. (2009) Our findings corroborate the argument by Hartig et al. (2014) that physical neighbourhood characteristics, such as green space, influence other area characteristics, e.g. social cohesion, more easily than individual characteristics, e.g. individual social contacts. --- Green space, social environment and the relation with mental health Our finding that individual social contacts were associated consistently with better mental health, while social cohesion and neighbourhood attachment were related to better mental health in Stoke-on-Trent, UK exclusively, underlines the fact that the neighbourhood environment is in general less important for individual health than individual characteristics (Pickett and Pearl 2001). Despite of that, studying neighbourhood characteristics such as neighbourhood green is relevant as it can influence the health of many people, therewith contributing substantially to the health of the population. We found only weak evidence for a relationship between neighbourhood green space and mental health. A study that used similar green data, i.e. audit information, reported no relation between the presence of green and general health (Dunstan et al. 2013), though another study reported that the amount of green was related to mental health (Van Dillen et al. 2012). We could only replicate this association between the amount of green space and mental health in Barcelona. The Barcelona neighbourhoods were considerably less green than the neighbourhoods in other cities (see Table 1). Possibly living in greener neighbourhoods in Barcelona is more strongly related to mental health than in other cities, because of the scarcity of green space in general. Another explanation for finding an association between green space and mental health in Barcelona only, is that especially nearby green space seems important for mental health (Kaplan 2001;Triguero-Mas et al. 2015;Van Dillen et al. 2012), as the Barcelona neighbourhoods were by far the smallest in this study. When we conducted post-hoc analysis using individual perception of neighbourhood green, assuming that the individual perception is based on nearby green space more than the neighbourhood average perception of green, we indeed found associations between green space and mental health in Doetinchem as well. In our study, quality of neighbourhood-level green was not associated with mental health, which is in contrast with previous studies (Francis et al. 2012;Van Dillen et al. 2012). We used a crude measure for quality of green space. Possibly this measure was not specific enough to detect a relationship with mental health. We found no indications that the neighbourhood social environment serves as a possible mechanism between neighbourhood green space and mental health. We either failed to find a relation between neighbourhood green space and mental health (i.e. Kaunas, Doetinchem, Stoke-on-Trent), or did not find associations between neighbourhood green space and (one aspect of) the social environment and between the (same one aspect of the) social environment and mental health (i.e. in Barcelona). In Barcelona, a highly urbanized city, restoration from daily stress might be a more relevant mechanism underlying the association between green space and mental health than the social environment. Unfortunately, we were unable to examine this hypothesis with the available PHENOTYPE dataset. --- Comparison of the cities There were marked differences between the cities with regard to the relevance of the neighbourhood environment for mental health. The Intra-Class Correlations of the cities (ICC), which estimates the proportion of variation in mental health between residents that is related to neighbourhood characteristics, reflects these differences. For example, in Doetinchem, the ICC was very low (0.51%) and both green space and the social neighbourhood characteristics were unrelated to mental health, in contrast with Stoke-on-Trent and Barcelona with ICCs of 8.51 and 6.71%, respectively. In Barcelona, this ICC reflected the relation between neighbourhood green space and mental health and in Stokeon-Trent the neighbourhood social environment was related to mental health. The different findings across the cities might reflect geographical and cultural differences (Hartig et al. 2014). The differences could also reflect that, despite the use of identical measurements, data might still not be comparable due to cultural differences in the interpretation of survey questions and audit. The use of more objective measures, such as GIS data, could improve the comparability of the findings, but this might at the same time not be the environmental characteristics that have the biggest impact on mental health. Furthermore, more objective data on the quality of neighbourhood green or the social neighbourhood characteristics will be much more difficult to achieve. Future comparative studies should make efforts to also incorporate objective data to allow even better comparison between European settings. --- Conclusion Neighbourhood green and the neighbourhood social environment were related to one another in two cities, but did not translate into better mental health there. Neighbourhood green was related to mental health only in Barcelona, but there we found no indication that the neighbourhood social environment could be the underlying mechanism. Our study found no indications that improving neighbourhood green space could be a relevant public health policy, nor were there indications that health benefits of green space would occur through the improvement of the neighbourhood social environment. Future studies should use longitudinal data to further investigate the possibility of this mechanism. To improve the comparison between European settings, studies should try to incorporate objective measures of both green and the social environment. --- Key points • The neighbourhood social environment has been suggested to be one of the mechanisms responsible for the beneficial effect of green space on health. • This study examines the relationship between neighbourhood green space, social cohesion, neighbourhood attachment, social contacts, and mental health in four European cities. • We find no evidence that the neighbourhood social environment could be the underlying mechanism between neighbourhood green space and mental health. • The relevance of this mechanism needs further investigation with longitudinal data and with more objective data to improve the comparison between European settings. --- Ethical approval This article does not contain any studies with animals performed by any of the authors. All procedures performed in studies involving human participants were in accordance with the ethi-cal standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Informed consent Informed consent was obtained from all individual participants included in the study. --- Conflict of interest The authors declare that they have no conflict of interest.

Publisher's PDF, also known as Version of record document license Article 25fa Dutch Copyright Act Link to publication in VU Research Portal citation for published version (APA)

Background In recent years, increased attention has been focused on concepts supporting the pivotal role of the individual's perceptions in healthcare [1][2][3]. This focus has been accompanied and reinforced by the marketization of Western society and the prominence given to concepts of service management [4,5]. In the 1980s, experts in the field of service quality argued for quality as a subjective perception, which varies from customer to customer [6]. The next decade saw an increased focus on the concept of value [7] as an individualized and even unique perception of the customer [8,9]. Individualization in healthcare is often proclaimed in order to enhance the healthcare user's position by becoming more participative and well-informed [10,11], ranging from co-developing treatment plans to choosing healthcare providers [12,13]. For multiple reasons, not all patients or inhabitants have the possibilities or prerequisites to be participative or well-informed [14,15]. Barriers may be constituted due to language skills [16,17], one's economic situation [18], or long travelling distances [19]. Possibilities to be participative or well-informed may also be constrained by the provider's normative or stereotypical expectations and perceptions [18]. For instance, Hedegaard et al. [10] found physicians to unconsciously be more amenable toward native Swedish-speaking than non-native speaking patients despite the latter group communicating more in align with "patient-centered care" (e.g. being wellinformed, and actively asking questions). Stereotypes have also been reported in the setting of the current paper, the Swedish child health centers, in which "family" was heteronormatively assumed to consist of child, mother and father in information given to parents [18]. Given the above, Saha et al. [20] argued that individualization in healthcare must take the diversity of patients' perspectives into consideration. Thus, identifying group members shared perceptions may be a first step in improving healthcare that is relevant to the particular group and grounded within their social context [16,17]. The social context also implies that societal structures and norms influence human interaction [21,22], thus it is also important to inquire about healthcare providers' perceptions [23]. In the decentralized Swedish healthcare system, the national government is responsible for overall objectives and regulation. At the two local levels of governmentcounty councils and municipalitiesit is decided how healthcare is to be delivered given the local conditions [24]. Generally, the county councils are responsible for providing high quality healthcare through hospitals, primary careincluding child health servicesand dental care, whereas the municipalities are responsible for care for the elderly, people with physical disabilities or psychological disorders as well as school health [25]. Funding comes mainly from county council and municipal taxes, but also from out-of-pocket fees or national government grants [26]. The purpose of Swedish child health services is to promote children's health, development, and well-being [27]. A local child health center offers voluntary child health promotion programs and free services for all preschool children (newborn to age 6 years) and their parent(s). The responsibility for carrying out the programs rests mainly with the nurse at the center, often a pediatric or district nurse. The nurse's importance as a resource and support for parents has been recognized [28]. The needs of each family determine the frequency of appointments. Typically, there are 10 to 20 health appointments during the child's first year, and then annual appointments until school-based healthcare providers take over these responsibilities [29]. Besides the care and assessment provided by the child health nurse, each center offers additional services, including vaccination programs, language tests, eye examinations, and parental education given in groups. In addition to the nurse, physicians and psychologists are seen regularly or when required. Despite this seeming standardization, there is an increasing recognition of the variation in services provided by the country's child health centers [30]. Inequities have been identified in terms of access to child health services, for example difficulties in attracting fathers [31,32] and unemployed mothers [33] to visit the centers. Other researchers have found that the centers do reach various groups, but do not always adequately meet the diverse support needs of different groups [34]. Inequities are also manifested as less satisfaction with provided services among mothers of low socio-economic status [35], and same-sex parents experiencing heteronormative communication [36]. In an attempt to coordinate services in the decentralized Swedish healthcare system, a government agency published national guidelines in 2014 [27], which emphasized the importance of including and addressing the needs of all parents. Subsequently, efforts have been made to change approaches and attitudes at the local centers [29]. For instance, in addressing the growing number of parents from Somalia, the nurses at one center worked with reflexivity of their own preunderstandings, resulting in better encounters with all families [23]. Research on motherhood in Sweden in the 2000s has often addressed gender equality and both parents' responsibilities of parenting as well as opportunities to do paid work [37,38] independent of the mother's country of birth [39] or specifically focused on Swedish-born mothers [40]. For the latter group, research has found that becoming a mother is more of an individualized life project as compared to mothers with Turkish background living in Sweden to whom becoming a mother was more of a collective project [41]. It has been reported that Sweden's local child health centers often fail to attract immigrant mothers [33]. Sweden's population of nearly 10 million people includes a substantial number (15%) born outside its borders [42]. Of all immigrants to Sweden in 2013, almost 10 % held Somali citizenship; the only groups that were larger were Syrian immigrants and returning Swedish citizens [43]. For two decades, Somalia has been one of the most common countries of origin for immigrants to Sweden [44]. However, according to an official report, Somalis as a group have had particular difficulty integrating into Swedish society [45]. Somalis living in Sweden have among the lowest employment rates [45,46] and educational levels [46] of all groups. Findings of previous studies of Somali immigrants in Swedish healthcare include: dissatisfaction with healthcare encounters [47]; dissatisfaction with childbirth experiences [48,49]; a relatively high proportion of Somali children with autism [50]; and a relatively high risk of vitamin D deficiencies [51]. Overall, Somalis' perceptions of Swedish healthcare have not been well addressed [52]. The current manuscript is directed at examining the perceptions of motherhood as expressed by two groups of mothers -Swedish-born and Somali-born, and exploring the implications for individualized healthcare. This goal includes the recognition that those belonging to a particular group may share perceptions, and that healthcare nurses (the third group) also hold their own perceptions about mothers yet must recognize differing perceptions among various groups. --- Methods --- Setting and participants This study took place in Western Sweden, in the country's second largest region, which has over 1.6 million inhabitants [53]. To enhance understanding of different perceptions among groupsand not only among individualswe asked 20 Somali-born mothers, 50 Swedish-born mothers, and 35 child health nurses to complete two statements about motherhood. The responses of the Swedish-born mothers and nurses were collected at two child health centers in two medium-sized cities in Western Sweden. Responses of the Somali-born mothers were initially collected at one child health center in a mediumsized city. In order to reach saturation [54], additional responses were collected at a local meeting place for parents located in a multicultural area of the region's largest city. The diversity of settings for data collection was thought to contribute to a more robust level of saturation within each group of respondents. The collected, de-identified demographic data revealed that all but two of 50 responding Swedish-born mothers had a partner, and they each had between one and nine children. Among the Somali-born mothers, eight of the 20 respondents had a partner, and they each had between one and seven children. The Swedish-born mothers were 22 to 40 years of age, and the Somali-born mothers were 20 to 34 years old. --- Data collection and analysis Mothers and nurses were informed about the study both verbally and in writing. Verbal and written consent were obtained prior to data collection. All participants were informed about the study's purpose; that participation was voluntary; and that they could withdraw anytime without consequence. Moreover, they were informed that all published information would be anonymous [54]. Using a similar data collection procedure as in previous research [55,56], the participants were asked to complete two statements: 1) "To me, being a perfect mother means..."; and 2) "In everyday life, being a mother means..." The statements were chosen to identify disparities given the mothers' life situations, and to compare their ideal conceptions of motherhood with what they believed was realistic. Each respondent could provide multiple responses to each statement, as shown in Table 1. Analysis was inspired by the qualitative content analysis procedure of Graneheim and Lundman [57]. Analysis can be done on content close to the text (manifest content), or on the underlying meaning (latent content). In this study, the manifest content was analyzed. All responses (n = 543) were read through several times and coded by the research members EE and KE. The codes were clustered into two types of categories: general categories and groupspecific categories (Swedish-born mothers, Somali-born mothers, child health nurses). Of particular interest were the emerging similarities and differences among the three groups. The constructed categories were discussed with the other research team members and further adjusted. --- Results Few differences were revealed between the responses to the statements of everyday and perfect motherhood. Consequently, the perceptions organized in Fig. 1 address responses among the groups that are common to both perfect and everyday motherhood. Perceptions of motherhood commonly expressed and shared among the three groups are presented in the circle in the middle. The three surrounding squares represent three distinct foci of motherhood as derived from the responses: 1) the mother; 2) the community; and 3) the child. The perceptions and quotations are distributed closest to the focus of motherhood mainly addressed. The respondents' group categorization is indicated within brackets: child health nurses, Somali-born mothers, and Swedish-born mothers. For all three groups, a mother was perceived as someone providing the basics, guiding the child, and explaining things, as exemplified in the mid-circle. Here, motherhood was expressed in terms of "Guiding the child in lifeset limits" (nurse); or to make sure "[t]hat the children's most basic needs are met, such as cooking, going to the toilet, and to rest" (Swedish-born mother); or to "love the child" (Somali-born mother). These and similar responses were common among all three groups, irrespective of addressing perfect or everyday motherhood. Despite these similarities, there were also group-specific responses to motherhood. The nurses typically thought of a mother as someone capable of asking others for help. For instance, by "admit[ting] to her surroundings when it's tough and hard work...", or to "understand when she needs to ask for help." Often "the partner"or more rarely, "the father"or parents-in-law were mentioned as providing this help. Moreover, the nurses often focused on motherhood in relation to the characteristics of mothers' themselves, predominately by using positively charged words, such as "being loving", or to be "caring, creative, inventive, friendly." Swedish-born mothers also used "positive" terms to describe the mother's characteristics. Some Swedishborn mothers responded that a perfect mother "felt good." Feeling good was expressed in relation to the child; a perfect mother should take good care of herself so she can take care of the child rather than for her own sake. Unlike the nurses' focus on the mothers, the Swedish-born mothers talked about motherhood almost exclusively from the child's perspective: "mak[ing] sure that one's child is feeling okay", or "... that she develops in every way and that she becomes a safe individual with good self-esteem and self-confidence." Somali-born mothers typically responded that a mother possessed a certain degree of stamina and was Fig. 1 Focus of motherhood perceptions supposed to be active, "never get tired", and always "make an effort." Different from both the nurses' and the Swedish-born mothers' perceptions, the Somali-born mothers typically talked about motherhood as embedded in a community context, as mothers should have "good contact with school, teachers, school nurse, and physicians". To a great extent, the responses for perfect and everyday motherhood overlapped. However, a few differences were identified. The nurses used "positive" terms to describe the characteristics of both perfect and everyday motherhood. Swedish-born mothers were more nuanced when describing everyday motherhood, often using both negative and positive wording in tandem: "Often headache. Often feeling insufficient when the children are screaming. But then there is a glimmer and emotions of joy takes over, the pride and the happiness." More than the other groups, Somali-born mothers emphasized negative aspects of everyday motherhood more than they did for perfect motherhood. Here, motherhood was associated with having a "bad character," "lying," "being angry," "lazy," "tired," "absent," "impatient," or "not taking care of herself." Furthermore, she was not well-educated, did not get by economically and relied on society's help. She was conceived as someone prioritizing her own life over the child's life. --- Discussion --- Motherhood as a social construction Research on motherhood often emphasizes biological aspects, such as pregnancy [58], post-natal depression [59], or breastfeeding [60]. The findings revealed little about these biological aspects. Focusing on biological aspects has been criticized for neglecting interests and power relations that have made women responsible for parenting [61,62]. For example, a labor market in which women are expected to raise children and men are expected to provide [63]. From this perspective, gendered expectations and characteristics are considered to have been socially or culturally constructed [63]. These expectations and characteristics may vary in relation to the social or cultural context. The differences in responses of Somali-born mothers compared with the nurses and Swedish-born mothers may be explained by perceptions being constructed or shaped, in relation to the cultural and social context of the respondent. Family policies and welfare systems differ between societies and shape the responsibilities of women differently [64][65][66]. By the same logic, gendered expectations and responsibilities undertaken may transform for an individual as he or she changes social context; for example, through migration [67][68][69]. --- Focus of motherhood: the child, the mother, the community Various healthcare studies emphasize the community's pivotal role for Somali refugees [70,71]. To talk in terms of Somalis as one group is of course a simplification. However, it is argued that the indigenous philosophies are deeply embedded in Somali societiesboth within Somalia and the Somali diasporaand in governing Somali peoples' way of life [72]. Vital in these philosophies is communalism and the individual's relationship with the community, including social cohesion and collective responsibility [72,73]. Consequently, social networks and the community are crucial for care if one is ill or in an adverse life situation [74][75][76]. Consequently, healthcare information provided by friends and key actors in the community may be just as important as information from healthcare staff [77]. This may be why many Somaliborn mothers' responses concerning both everyday and perfect motherhood focused on the surrounding community rather than themselves or their children. In contrast, the child health nurses generally focused on the mother herself when talking about perfect and everyday motherhood. This is somewhat surprising, given the official position [27,29] that child health centers emphasize the child's central position. Of course, the study's statements explicitly addressing motherhood could have affected the nurses' responses. A related deviation among the groups was their expectations (or not) of a mother's ability to share childcare responsibilities. Despite talking about motherhood within a wider community context, the Somali-born mothers believed that a perfect mother as well as an everyday mother was not supposed to ask others for help when it came to taking care of the child. She was expected to be independent and to manage child-rearing herself. Previous research has suggested that parenting responsibilities are shared to a greater extent between Somali couples living in Finland than is normally the case in Somalia [67]. Studies set in the United States [68] and Sweden [69] argue that the shift toward more equal gender roles between Somali couples in the host country compared to Somalia, may strain relationships; while both partners are expected to work, the husband may be reluctant to share household and parenting responsibilities. The nurses generally responded that mothers should cooperate with others in raising children, not least the parents-in-law. The Swedish-born mothers mentioned cooperation, but limited to the partner. --- The meta-mother Asking the respondents to complete the statement about perfect motherhood may be biased as it could be interpreted as being a stereotypical standard of perfection to which their own mothering should be measured. The stereotype of perfection was frequently addressed by Swedish-born mothers as well as nurses in responses such as, "there's no such thing as a perfect mother." But almost every time the impossibility of perfection was mentioned, it was followed by a statement along the lines of "...but she is doing the best she can," or "...but she always tries her best." In this sense, the ideal of motherhood is argued not to exist in "real life" but rather as a stereotype yet it impacts the expectations of mothering, as exemplified in the previous quotations that she is always, and in every situation, supposed to do her best given her circumstances. From the nurses' perspective, the perfect mother also realized when it was time to seek help, or when she could not manage to take care of the child herself. Previous scholars have focused on motherhood as being filled with stress, ambivalence, frustration, and self-blame [78,79]. Our findings suggest that mothers may be greatly impacted by the concept we termed meta-mothera woman who instinctively knows how and when to act, and who is always giving her maximum in her role as a mother. --- Consequences for the nurse-visitor interaction The differences and similarities described among the groups in this study may affect the individual meeting between the nurse and the visiting mother at a child health center. A good meeting has been identified as a prerequisite for the child health nurse to find out what the mother desires and to fulfill her expectations [80]. However, the nurses at the Swedish child health centers have been found to initiate most of the topics discussed [81], which may contribute to parents' lack of healthcare information related to their own concerns [33]. Moreover, it has been pointed out that normative and gendered perceptions risk being transferred to visitors at the child health centers [28]. Official reports focus attention not only on the centers' visitors, but also on staff and the prevailing norms within the organization [27]. In this study, the nurses both explicitly and implicitly expressed normative perceptions. The findings of this study suggest that increased knowledge about perceptions of motherhood and engagement with the local communities may help to improve equitable access to healthcare through approaches that are embedded in the local community context. Thus, information about child health services should not be limited to the centers, but disseminated in the wider community [16,17]. Given the potentially strong role the community can play, those who already have a position of authority or trust in the local community should be used to disseminate information [16,17,74]. Of the 35 responding nurses, ten indicated the belief that a partner is part of parenting. A few nurses mentioned the sex of this other being as male, but most mentioned a sexless "partner." A reason most nurses did not heteronormatively assume the partner to be male may be due to child health service guidelines putting attention on family constellations other than heterosexual families [27]. However, the nurses did not expect the mothers to be without a partner. Previous research highlights the fact that single-parent families are increasingly common in Sweden, and investigators have reported resulting disadvantages for the health of the child [82], as well as for single-parenting mothers [83]. Some normative perceptions were explicit. For instance, the nurses expected the mother to ask for help concerning her child. However, Somali-born mothers' perceptions of ideal motherhood were of someone capable of taking care of the child herself, not asking for help. Moreover, the nurses' expressions of everyday motherhood were described in rather dashing terms, such as wonderful and beautiful. Swedish-born mothers responded about perfect motherhood in "positive" terms. However, their responses concerning everyday motherhood were much more nuanced. Somali-born mothers rarely mentioned "positive" characteristics of everyday motherhood. There is a risk that the needs of mothers who express fatigue or being irritated will not be met by nurses at the centers. Given the potential impact of the "meta-mother," nurses should be aware of the stereotype of the "perfect mother" and therefore be ready to support mothers where they are. --- Limitations of the study In recent years, Scandinavian scholars have focused attention on fathers' perceptions, and parental experiences [31,32]. Still, most appointments to the child health center are made by the mother [31]. The mother is the norm and has been a major influence on the shaping of centers over the years. Understanding the needs and expectations of visitors other than mothers (fathers, for example) requires that one understand the norm. Focusing on motherhood risks reinforcing the mother's role as the main childcare provider. There is a similar risk of perpetuating stereotypes and generalizations of Somali-born mothers based simply on country of birth. Indeed, these women's experiences may be very different, depending on the reasons for migration, birth at migration, and other factors. These factors were not considered in this study. The intent of the current study was to understand perceptions of motherhood in three groups in Western Sweden with participants completing open-ended statements about ideal and everyday motherhood. As such, the scope was limited and therefore caution should be exercised when generalizing the results and conclusions to other populations. Neither should the results be generalized to all Somali-born or Swedish-born women in Sweden. As with all qualitative studies, there are limitations with smaller sample sizes and nonrandom sampling. In this case, the Somali-born sample was much smaller due to translation resource intensity and logistical constraints of accessing the mothers at a time convenient to them. Data was collected at different sites by necessity in order to access mothers and nurses who were willing to participate in the study. The responses of the Somali-born mothers were translated into Swedish by an interpreter. The Swedish translations were then translated into English by EE. Consequently, there is a greater risk that information has been lost in translation or somewhat distorted for the Somali-born mothers than for the other two groups (for which translation only took place once). No demographic information was collected for the nurses. Previous research [84] has suggested that healthcare providers' backgrounds also affect preconceptions. Previous research [67][68][69] also indicated that gendered responsibilities transform with migration. In the case of the Somali-born mothers, this has not been thoroughly addressed in this paper, which focuses primarily on differences between groups. --- Conclusion With the growing emphasis on individualization in healthcare comes a need to acknowledge the social context, including societal structures and norms. This paper suggests that between groups of peopleand not solely between single individualsthere are differences and similarities in perceptions of motherhood which potentially may have implications for health services access and utilization. In this study, Somali-born and Swedishborn mothers, as well as nurses expressed differences in the focus of motherhood: the community, the child, and the mother herself. Potential convergences and divergences of beliefs by mothers and staff may constitute a source of misunderstanding, and normative or stereotypical encounters. However, recognition of the existence of gendered and cultural constructions may be a first step to avoiding such encounters. Because healthcare encounters do not take place in a social vacuum, healthcare needs to be provided that is relevant to specific groups of the population and that is grounded within their social context. Moreover, group perceptions should be used constructively. When healthcare providers design services that satisfy the needs of a diversity of users, equity in healthcare may be enhanced. --- Authors' contributions EE and KE managed the project. LV proposed data collection methods, KE carried out most of the data collection, and EE and KE jointly analyzed the data. EE drafted the manuscript. All authors helped with revisions of the manuscript. All authors read and approved the final manuscript. --- Competing interests The authors declare that they have no competing interests. --- Consent for publication Consent for publication was obtained from the participants. --- Ethics approval and consent to participate The Research Ethics Committee in Gothenburg, Sweden, decided that this project was not a matter for the Ethical Review Act (registration number 127-15). Consent were obtained prior to data collection.

Background: In light of the growing emphasis on individualization in healthcare, it is vital to take the diversity of inhabitants and users into consideration. Thus, identifying shared perceptions among group members may be important in improving healthcare that is relevant to the particular group, but also perceptions of the staff with whom interactions take place. This study investigates how motherhood is perceived among three groups: Somali-born mothers; Swedish-born mothers; and nurses at Swedish child health centers. Inequities in terms of access and satisfaction have previously been identified at the health centers. Methods: Participants in all three groups were asked to finalize two statements about motherhood; one statement about perfect motherhood, another about everyday motherhood. The responses were analyzed using qualitative coding and categorization to identify differences and similarities among the three groups.The responses to both statements by the three groups included divergences as well as convergences. Overall, biological aspects of motherhood were absent, and respondents focused almost exclusively on social matters. Working life was embedded in motherhood, but only for the Somali-born mothers. The three groups put emphasis on different aspects of motherhood: Somali-born mothers on the community; the Swedish-born mothers on the child; and the nurses on the mother herself. The nursesand to some extent the Swedish-born mothersexpected the mother to ask for help with the children when needed. However, the Somali-born mothers responded that the mother should be independent, not asking for such help. Nurses, more than both groups of mothers, largely described everyday motherhood in positively charged words or phrases.The findings of this paper suggest that convergences and divergences in perceptions of motherhood among three groups may be important in equitable access and utilization of healthcare. Individualized healthcare requires nuance and should avoid normative or stereotypical encounters by recognizing social context and needs that are relevant to specific groups of the population.

Peer support as an organized strategy for giving and receiving help can be understood as an extension of the natural human tendency to respond compassionately to shared difficulty. A widow may offer comforting words, tea, and company to a woman grieving the death of her husband. Someone who has learned to cope with the effects of a serious injury explains how they manage to a newly injured person. Most people who have been through hard times empathize with and have an urge to help when they meet others who struggle with similar problems. It not only benefits the person receiving support, it makes the helper feel valued and needed (Riessman, 1965). Sometimes referred to as self-help or mutual aid, peer support has been used by people dealing with different types of social circumstances, emotional challenges, and health issues, including those with alcohol or drug problems, bereaved individuals, and people living with physical illnesses or impairments (Penney, Mead, & Prescott, 2008). Peer support has a significant history among people with psychiatric diagnoses. This article will review recent literature on peer support among people with psychiatric diagnoses in the United States. It begins by addressing the substantial definitional issues involved and offering a brief consideration of the history of two types of peer support. This will be followed by an examination of recent review articles on peer support in mental health. An ongoing study of a peer-developed approach, Intentional Peer Support, within the context of peerrun programs, is described. Finally, policy and practice implications are discussed. --- Defining Peer Support by and for People with Psychiatric Disabilities In recent decades, there has been increasing attention in the professional literature to the study of peer support among people with psychiatric disabilities. But the ability to conduct a meaningful review of this literature is complicated by the fact that there is no agreed-upon definition of the term "peer support." In the research literature, terms such as "peer support," "peer-delivered services," "self-help," "consumer services," "peer mentors," and "peer workers" are used interchangeably, making it difficult to draw meaningful distinctions among fundamentally different types of interventions (Repper & Carter, 2011;Rogers et al., 2010;Davidson, Chinman, Sells, & Rowe, 2006;Mead & MacNeil, 2005). Despite this confusion, upon examination of the history of peer support, one can differentiate between two major categories that are often conflated in the literature: peer-developed peer support and the practice of employing peer staff in traditional mental health programs. These are defined and discussed below. --- Peer-Developed Peer Support Peer-developed peer support is a non-hierarchical approach with origins in informal self-help and consciousness-raising groups organized in the 1970s by the ex-patients' movement. It arose within an explicitly political context, in reaction to negative experiences with mental health treatment and dissatisfaction with the limits of the mental patient role (Van Tosh & del Vecchio, 2001;Kalinowski & Penney, 1998). While peer support for people with specific medical conditions, like cancer, focuses on coping with illness, peer support by and for people with psychiatric histories has always been closely intertwined with feelings of powerlessness within the mental health system and with activism promoting human and civil rights and alternatives to the medical model that defines extreme mental and emotional states as "mental illnesses" (Chamberlin, 1978). Deegan (2006) sees peer support as a "response to the alienation and adversity associated with being given a psychiatric diagnosis," by which diagnosed people are ostracized from the larger community and work to create their own communities by reaching out to others who share their lot. The development of peer support was influenced by the human and civil rights movements of African Americans, women, and lesbians and gay men in the 1960s and '70s. It was also influenced by the Independent Living (IL) movement of people with physical, sensory, and cognitive disabilities (Deegan, 1992). Peer support was inseparable from human rights activism during the development of the IL movement and is one of four required services of federally funded Centers for Independent Living serving people with disabilities (White, Simpson, Gonda, Ravesloot, & Coble, 2010). The IL movement sees "disability" as the result of physical, attitudinal, and social barriers, rather than the consequences of deficits within individuals with impairments (De Jong, 1979). This formulation resonated with people who had negative experiences in the psychiatric system and used peer support as a means for coping with adverse effects (Penney & Bassman, 2004). Although peer support emerged in a political environment, it is also an interpersonal process with the goal of promoting inner healing and growth in the context of community (Mead, 2003). As a practice, it is characterized by equitable relationships among people with shared experience, voluntariness, the belief that giving help is also self-healing, empowerment, positive risk-taking, self-awareness, and building a sense of community (Budd, Harp, & Zinman, 1987;Harp & Zinman, 1994;Clay, 2005). Peer support, by definition, is "led by people using mental health services" (Stamou, 2014, p. 167;Faulkner & Kalathil, 2012). --- Intentional Peer Support as an Evolution of Informal and Peer-Developed Peer Support In the early days, peer support-more commonly called "self-help" in those years-was often informal and relatively unstructured. People met in apartments, in church basements, and in libraries, but rarely in spaces affiliated with the mental health system (Chamberlin, 1990). But, during the 1980s and '90s, independent, peer-run nonprofit organizations emerged (Chamberlin, 2005). Many of these groups began to offer more structured peer support, generally with some government funding. The development of government-funded peer-run programs meant that these programs needed to more clearly define the vision, principles, and practices of peer support to meet government oversight requirements. Shery Mead has been a pioneer in this work for more than 20 years, developing an approach called Intentional Peer Support (IPS). While IPS grew from the informal practices of grassroots-initiated peer support, it differs from earlier approaches because it is a theoretically based, manualized approach with clear goals and a fidelity tool for practitioners (MacNeil & Mead, 2005). IPS sees its fundamental purpose as helping people unlearn the mental patient role, and defines peer support as "a system of giving and receiving help founded on key principles of respect, shared responsibility, and mutual agreement of what is helpful. Peer support is not based on psychiatric models and diagnostic criteria. It is about understanding another's situation empathically through the shared experience of emotional and psychological pain" (Mead, 2003, p. 1). It is posited as a non-clinical intervention whose benefits are primarily intrapersonal and social in nature (Mead & MacNeil, 2005). In working with individuals with psychiatric diagnoses, the goals of IPS are to move from top-down helping to mutual learning, from a focus on the individual as the locus of dysfunction to a focus on relationships as a tool for growth, and from operating from fear to developing hope (Mead, 2014). IPS is a philosophical descendant of the informal peer support of the ex-patients' movement of the 1970s. What distinguishes it from earlier, less structured peer support is a focus on the nature and purpose of the peer support relationship and its attention to skill-building to purposefully engage in peer support relationships that promote mutual healing and growth. IPS recognizes that trauma plays a central role in the experience, diagnosis, and treatment of people with psychiatric histories, and emphasizes the need for peer support to be trauma-informed (Mead, 2001). Other peer support practitioners have expanded this effort to bring a trauma-informed lens to the practice of peer support through guidebooks and training (Blanch, Filson, Penney, & Cave, 2012). --- Peer Staff Employed in Traditional Mental Health Programs The growth of this approach is illustrated by the recent, rapid expansion in the U.S. of "peer specialists" and similar positions in mental health programs (National Association of State Mental Health Program Directors [NASMHPD], 2012). While there is no standard definition or job description for a "peer specialist," a number of states, provider organizations, and government agencies have such titles, also known as peer mentors, peer support specialists, recovery support specialists, recovery coaches, and a host of other titles that usually involve the words "peer" and/or "recovery." The use of the word "peer" as part of job titles is a topic that deserves fuller discussion than can be offered here. The term is confusing at best; in general usage, a "peer" is an equal, one who shares characteristics or experiences in common with the subject. To employ the word as a euphemism for "service user" or "mental patient" poses both grammatical and philosophical problems. What these job titles have in common is that they apply to employees with psychiatric histories working in paraprofessional roles in traditional mental health programs, often performing the same tasks as non-peer staff (Davidson, Bellamy, Guy, & Miller, 2012). Job descriptions vary: peer staff may provide clinical and/or paraprofessional services that are indistinguishable from those provided by non-peer staff, they may serve as clerical staff or van drivers, or they may have undefined roles that evolve based on the individual's aptitude or the perceived needs of the organization. Peer workers in traditional programs generally do not provide "peer support" as this term is commonly understood by users and practitioners of informal peer support. In fact, peer staff working in traditional programs rarely receive training about or exposure to the principles and practices of peer-developed peer support (Alberta, Ploski, & Carlson, 2012). Peer employees are usually expected to disclose their psychiatric histories and serve as role models for people they serve. Relationships between peer staff and service users are usually hierarchical, similar to staff-service user relationships generally within the mental health system, in contrast to the horizontal relationships that characterize peerdeveloped peer support (Alberta & Ploski, 2014;Davidson et al., 2012;Rogers et al., 2010). An early study of peer specialist services in Bronx, New York, funded by the National Institute of Mental Health (NIMH) in 1990, found that several components of well-being were positively affected by the work of peer specialists (Felton, Stastny, Shern, Blanch, Donahue, Knight & Brown, 1995). Using a quasi-experimental design, the study demonstrated that adding three peer specialists to a team of ten intensive case managers (ICMs) resulted in stronger beneficial effects for service recipients, compared to two control groups (adding three paraprofessionals or no extra staff). The most significant benefits for the group served by the ICM teams with peer specialists were in quality of life, specifically greater satisfaction with living situation, finances, personal safety, and fewer overall life problems (Felton et al., 1995). Based on initial findings of this study, the New York State Office of Mental Health (NYS OMH) established a Peer Specialist civil service title in 1993, the first state to do so. As of 2014, NYS OMH employed about 100 individuals in that title and at least 500 people worked in similar jobs in publicly funded voluntary sector agencies in the state. In both the Bronx ICM study and the Peer Specialist civil service title, the emphasis was initially on bringing the values and principles of peer-developed peer support into paid peer staff roles, but the ability to keep the focus on these values was often compromised by clinicians and administrators who did not understand or support the principles (Stastny & Brown, 2013). The practice of using peer staff in traditional programs has been accompanied by state peer specialist certification programs in 38 states as of 2014 (Kaufman, Brooks, Bellinger, Steinley-Bumgarner, & Stevens-Manser, 2014). These certifications require completion of a state-approved training course, using either a curriculum designed by the state or one of a number of proprietary training programs. There are currently no national standards for peer specialist training, and the length, intensity, and content of the courses vary widely (Kaufman et al., 2014). The expansion of peer staff in traditional programs accelerated when the federal Centers for Medicare and Medicaid Services (CMS) issued a State Medicaid Directors' letter in 2007 clarifying the conditions under which "peer support services" could be reimbursed by Medicaid. As of 2015, 31 states and the District of Columbia offered Medicaid-reimbursable peer support services, and it is likely that, under provisions of the Affordable Care Act, many other states will follow (Ostrow, Steinwachs, Leaf, & Naeger, 2015). The State Medicaid Directors' letter defined "peer support services" as "an evidencebased mental health model of care which consists of a qualified peer support provider who assists individuals with their recovery from mental illness and substance use disorders" (CMS, 2007, p. 1). While this policy clarification spurred an increase in the use of peer specialists, it also added to the definitional confusion, stating that any service provided by a "qualified peer support provider" was, by definition, "peer support." Stastny and Brown (2013, p. 459) 1, below, highlights key features of each review. These are primarily studies involving employment of peer staff in traditional programs, because, while informal and peer-developed peer support has been extensively described in the non-research literature (for example, Blanch, Filson, Penney, & Cave, 2012;Clay, 2005;Mead, Hilton, & Curtis, 2001;Chamberlin, 1990), its effectiveness has not been studied. Chinman et al., 2014 This review looked at the effectiveness of three types of peer support services (peer staff added to traditional services, peer staff in existing clinical roles, and peer staff delivering structured curricula) and found 20 relevant studies between 1995 and 2012 (Chinman et al., 2014). An argument can be made that two of the three types of peer support defined by the reviewers (peers in existing clinical roles and peers delivering structured curricula) are not really "peer support services" in the commonly understood sense of the term. Inclusion criteria included randomized controlled studies, quasi-experimental studies, single-group time-series designs, and cross-sectional correlational studies of peer support services for adults diagnosed with serious mental illness and/or co-occurring mental health and substance use disorders. This review was sponsored by the federal Substance Abuse and Mental Health Services Administration (SAMHSA), which defined "peer support services" as "a direct service that is delivered by a person with a serious mental illness to a person with a serious mental disorder" (Chinman et al., 2014, p. 1-2). This definition is at variance with the definition of the term that grew out of peer-developed peer support; it does not recognize the centrality of equitable, mutual relationships based on shared common experience that is the hallmark of peer-developed peer support. --- SAMHSA's Assessing the Evidence Base Review/ The authors found that peer support services met moderate levels of evidence, and that effectiveness varied across service types, with "peers in existing clinical roles" showing less effectiveness than the other two service types. The review noted that many of the studies had methodological problems. Because the studies under review used disparate outcome measures (e.g., hospitalization days, social support, quality of life), comparisons were difficult. As with most of the review articles discussed in this section, the authors decried the quality of many of the studies, pointing to a need for "studies that better differentiate the contributions of the peer role and are conducted with greater specificity, consistency, and rigor" (Chinman et al., 2014, p. 11). --- Cochrane Review/Pitt et al., 2013 A Cochrane review of 11 randomized controlled studies of what the authors referred to as "consumer services" (Pitt, Lowe, Hill, Prictor, Hetrick, Ryan & Berends, 2013, p. 4) found that the outcomes of such services are neither better nor worse than professionally provided services, although there is some evidence that peer services reduce the use of crisis and emergency services. Among the studies examined, the definitions of "consumer services" varied, making comparisons among and between the studies' findings difficult. For example, some were studies in which peer workers provided services identical to those provided by professionals (usually case management), while others concerned services that were based on peer providers' experiential knowledge. The review looked only at studies that compared outcomes of peer services (e.g., quality of life, hospital days) to outcomes of services delivered by professionals. --- Walker and Bryant, 2013 Walker and Bryant (2013) conducted a metasynthesis of the findings of 27 published qualitative studies and mixed methods studies that examined peer support services provided within traditional mental health programs; studies of peer support provided within peer-run organizations were excluded. Their review reported on the experiences of peer staff and their non-peer colleagues, as well as on the experiences of people using services. Peer staff faced numerous challenges, including low pay, insufficient hours, negative or rejecting attitudes from non-peer staff, and being treated as "patients" instead of colleagues. They also reported positive benefits for peer staff, such as increased self-esteem, larger social networks, and increased community participation. Non-peer staff reported increased empathy for and understanding of people with psychiatric disabilities due to interactions with peer colleagues; however, non-peer staff feared that the presence of peer staff would result in job losses for non-peer staff. People who received services experienced better rapport with peer staff than non-peer staff and reported increased hope and motivation, as well as increased social networks, as a result of working with peer staff. --- Davidson and Colleagues, 2012 Davidson and colleagues (2012) identified three categories of research on peer support that occurred sequentially over the past 25 years. First, there were feasibility studies of peer-provided services in the 1990s, which showed that peer staff could function adequately in ancillary roles and produce outcomes on a par with those of professional services. Second, a number of studies compared peer staff and professional staff providing conventional services in conventional roles. These studies generally reported that peer staff functioned at least as well as professionals, with comparable outcomes. Some studies found that peer staff had better outcomes than professionals on a few measures, including increased engagement among "hard-to-reach" clients, reduced hospitalization rates, and decreased substance abuse rates among people with dual diagnoses. Third, there are nascent investigations into the unique qualities/ contributions of peer services and the outcomes these produce. The authors acknowledge that this research endeavor is in its infancy. They report on two of their own studies in this area. One compared "usual care" with "usual care" plus two different types of peer services, finding that the two peer services conditions resulted in increased participant satisfaction on quality of life measures. The other suggested that peer support may reduce re-hospitalization rates and number of hospital days. --- Rogers and Colleagues, 2010 Rogers and colleagues ( 2010) reviewed 53 studies that met a minimum threshold for research quality as determined by a system developed by Rogers, Farkas, Anthony, and Kash (2008) rating the rigor of disability research and reported outcomes of peer services. They found no evidence that adding peer services to traditional services improved outcomes (neither did it worsen them); some evidence that peer support groups improved a number of outcomes for people who participated regularly (but not for occasional participants); and equivocal findings in other categories, such as one-to-one peer support and residential peer services. The authors noted a number of methodological problems that left them unable to draw firm conclusions related to the effectiveness of peer support and peer-delivered services. --- Davidson and Colleagues, 2006 Davidson and colleagues (2006) examined four randomized controlled studies of peer-delivered conventional services and supports that compared case management teams with and without peer staff members. Two of the studies reviewed found no significant difference in outcomes between the groups. In contrast, one study found that clients receiving services from the team with a peer worker reported increased satisfaction with services, while another found that clients receiving services from the team with a peer worker had fewer hospitalizations and longer community tenure. --- Discussion --- Study Design and Outcome Measures In the aggregate, the reviews and studies described above found minimal to moderate evidence that adding peer-delivered services of various types to traditional mental health services may be effective on a range of outcome measures. However, there are a number of methodological concerns that raise questions about these findings, including the underlying design of many of the studies, the type of peer support studied, and the relevance of the outcome measures selected. Most of the studies reviewed compared the outcomes of peer-delivered services with those of professionally delivered mental health services, and used traditional clinical outcome measures, such as symptom reduction, decreased hospitalization, and reduced substance use (Sledge et al., 2011). Both these factors raise issues about the appropriateness of these studies' designs. Peer support was never conceptualized as a substitute for-or interchangeable with-clinical services (Chamberlin, 1990;Campbell et al., 2006); neither are its goals the same as those of clinical services (Mead, Hilton, & Curtis, 2001;Harp & Zinman, 1994). Peer support staff generally do not have clinical training and are usually paid substantially less than credentialed mental health professionals. Since peer support was a) never envisioned as a substitute for clinical services-and, in fact, arose out of negative experiences with clinical services (Van Tosh & del Vecchio, 2001;Kalinowski & Penney, 1998)-and b) has different goals and thus outcomes than those of clinical services, it is not methodologically sound to compare the outcomes of peer support with those of clinical services. The review articles noted serious methodological problems that interfered with the authors' ability to draw firm conclusions about the strength of the evidence in the research literature for a wide variety of peer-delivered services. Many of the authors had unresolved questions about exactly what types of interventions and services were actually involved in the studies they reviewed. For example, Rogers and colleagues (2010, p. 24) concluded that their review "was hampered by a lack of description of the peer delivered activities, services and supports being provided, a lack of information about the intensity of those services and supports, and little information about the models and contexts of the service delivery.... If the field is to move forward and be adequately reviewed as an evidence-based practice, future research activities should focus on improving the state of our understanding of peer delivered services." It should be noted, however, that one of the review articles discussed above (Walker & Bryant, 2013) looked at qualitative and/or non-clinical outcomes that may have bearing on community participation and social inclusion. This approach shows promise for the development of outcome measures that actually track with the goals of peer-developed peer support as originally envisioned by the pioneers in this field. It should also be noted that none of the review articles-nor the research they reviewed-reflected on cultural considerations in the delivery of peer support services, nor about the development of peer support services in communities of color. --- Definitional Concerns Rogers and colleagues' (2010) statement quoted above is a call for more rigorous studies of peerdelivered services. This call is useful as far as it goes, but, like most of the comments on methodological shortcomings expressed by the review authors, it fails to address the serious definitional issues associated with this body of research. Despite the fact that the studies reviewed used a wide range of confusing and often incompatible definitions, none of these authors addressed the question of whether what they were studying was, in fact, "peer support" at all. The authors do not discuss or take into consideration the history and philosophy of the consumer/survivor/ ex-patient movement or the theories, principles, and practices of peer-developed peer support approaches. Many of the review authors-and the researchers whose work they examined-essentially defined "peer support" as any service or activity provided by a person with a psychiatric history. For example, Davidson and colleagues (2006) defined "peer support as an asymmetric, one-directional relationship" (Fuhr et al., 2014, p. 2), in stark contrast to the mutual, bi-directional relationships conceptualized by Intentional Peer Support (Mead, 2014). The people who have developed and practiced peer-developed peer support over the past 40 years understand it as a specific type of relationship-based approach with a philosophical basis in the potential for mutual growth and healing, and with clear principles and practices reflecting equality and respect. IPS, for example, defines peer support as "connecting with someone in a way that contributes to both people learning and growing... the intention is to purposefully communicate in ways that help both people step outside their current story" (Mead, 2014, p. 8). The development of this type of horizontal relationship is quite different from using peer staff within a traditional program to perform functions such as traditional case management services or driving people to appointments. Simply hiring people with psychiatric histories to do some of the usual tasks of the traditional mental health system is not the same as practicing peer support. By not exploring the true bi-directional relationship of peer support (the peerdeveloped definition), the extent to which the studies above truly identify the effectiveness of peer-delivered services is questionable. --- A New Peer-Led Study of Intentional Peer Support One approach to addressing the methodological and definitional issues discussed above is through peerled research of a peer-developed approach using non-clinical outcome measures that track with the stated goals of peer support. An ongoing three-year quasi-experimental study funded by the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR) is meeting this challenge by examining the comparative effectiveness of Intentional Peer Support in improving community integration, community participation, and quality of life for adults with psychiatric disabilities. This study is led by a principal investigator (PI) with a psychiatric history studying a peer-developed approach (IPS) delivered within the context of peer-run programs. This contrasts with earlier studies, which primarily looked at peer-delivered services (not specifically peer support services) within traditional mental health programs. The study compares outcomes of participants receiving IPS in a peer-run program to those of participants in a peer-run program that does not practice IPS. Outcome data are collected through in-person interviews that assess self-esteem, selfstigma, social connectedness, community integration, community participation, and quality of life at baseline and six months after the initial interview. A new scale developed by the PI, the IPS Core Competencies Scale, assesses the extent to which study participants perceive that peer support staff are practicing the core competencies taught to IPS practitioners. Secondary data sources include staff self-assessments and supervisory assessments, as well as focus groups of staff and service users. Quarterly, the IPS-trained staff at the intervention site complete a selfassessment of their skills and practices using the IPS Core Competencies Scale; supervisors rate staff using this tool biannually. Focus groups with peer support participants and with staff at both sites gather qualitative information on receiving and providing peer support prior to IPS training, 9 months after training, and 12 months later. Randomized regression models and content analyses will be used to examine whether any significant differences in outcome measures occur between the groups. These findings will be supplemented by qualitative findings from the focus groups and staff self-assessments. Study results will provide important information on how an innovative approach to peer support, designed by people with psychiatric histories and delivered within independent peer-run programs, may enhance community integration, community participation, and quality of life for adults with psychiatric disabilities. --- Implications for Practitioners and Future Research As noted above, peer-developed peer support is a non-hierarchical interpersonal process promoting mutual healing in the context of community, characterized by equitable relationships among people with shared experiences and a commitment to growing beyond the limits of the mental patient role. However, in clinical and psychiatric rehabilitation service settings, the term "peer support" has been used to describe activities and jobs that are not necessarily congruent with the peer-developed definition. Peer specialist and similar titles describe staff with psychiatric histories working in paraprofessional roles in traditional mental health programs. These staff may provide clinical and/ or paraprofessional services; work as clerical staff, janitors, or van drivers, or they may have relatively undefined roles that vary based on the perceived needs of the organization. Because peer-developed peer support approaches are not generally available in clinical settings, perhaps it is not surprising that the literature reviewed above conflates a variety of peer-delivered services with "peer support." It is important that policy makers and administrators develop clear job descriptions for a variety of peer-delivered services, and that they understand that these services are not the same thing as "peer support." This will provide administrators, clinicians, and researchers with the opportunity to educate themselves about the distinctions between peer-developed peer support approaches and the varied ways that peer staff are employed in traditional programs, so that they can accurately describe what they are providing and studying. Other peer support-related topics that would be fruitful directions for future research include studying the implementation of Intentional Peer Support with peer staff working in traditional programs, as well as comparing Intentional Peer Support training with some of the state and organizational training curricula for Certified Peer Specialists currently in use. The ongoing study of Intentional Peer Support described earlier is looking at the effects of a peerdeveloped approach to peer support implemented in a peer-run program, using non-clinical outcome measures that correspond to the principles and practices of peer-developed peer support. It is hoped that the results will help the field clarify its understanding of peer support and promote the expansion of services that are congruent with the original, peer-developed meaning of peer support. --- Darby

Broadly defined, "peer support" refers to a process through which people who share common experiences or face similar challenges come together as equals to give and receive help based on the knowledge that comes through shared experience (Riessman, 1989). A "peer" is an equal, someone with whom one shares demographic or social similarities. "Support" expresses the kind of deeply felt empathy, encouragement, and assistance that people with shared experiences can offer one another within a reciprocal relationship.

Introduction There is growing awareness that patient engagement in health research is not only ethically important, but leads to evidence for developing the most effective interventions, policy and practice recommendations, and planning for ongoing research [1][2][3]. Models for patient engagement have been evolving over the past four decades, and research that is grounded within evolved forms of stakeholder participation is typically understood as research as practice. Recently, attention has been shifting towards new forms of interpretive communities known as communities of practice (CoPs) and their potential for developing greater knowledge around participation [4][5][6][7]. CoPs do not privilege research-based evidence over experience-rooted knowledge; therefore, they have the potential to become powerful "venues for bridging traditional rifts in the health sector between research and practice, and among disciplines" ( [8], p.3). CoPs have been found to enhance the performance of interventions through breaking down professional, geographical and organizational barriers, enhancing knowledge sharing, and facilitating the implementation of new processes [9]. There are examples of CoPs that have been successfully assembled in healthcare settings, especially within oncology [8]. Furthermore, training on how to develop and maintain CoPs focused on improving health within different contexts is an important emerging area in the literature around patient support and participation [7]. However, it is not yet well documented how CoPs could influence the development of research for health systems and interventions for youth (i.e., children and adolescents) and their families. In this article we examine patient engagement beyond the traditional hierarchical structures for participation towards developing a greater and more functional understanding of how youth and families are involved through health research CoPs. We do so by exploring the participation of youth and families in health research CoPs created through IN•GAUGE®, an ongoing research program led and coordinated by Dr. Roberta Woodgate. Dr. Woodgate's research engages youth, families and caregivers, service providers, researchers, and policy makers towards building insights into the lived experiences of youth with physical and mental illness. --- Background Patient engagement has roots in several international agreements including the WHO Alma Ata Declaration (1978) in which declaration 10 states, "The people have a right and duty to participate individually and collectively in the planning and implementation of their own health care" [10]. Since the Alma Ata, this ethos has been applied with different patient populations, including children and adolescents. The most formal and largely recognized articulation of children's rights to participate is Article 12.1 of the United Nations Convention on the Rights of the Child (UNCRC), which asserts that "States Parties shall assure to the child who is capable of forming his or her own views the right to express those views freely in all matters affecting the child, the views of the child being given due weight in accordance with the age and maturity of the child" [11]. A few models and typologies for participation run parallel with the UNCRC standards and have been applied to youth engagement in health research. Many of such models have roots in Arnstein's "ladder of citizen participation" [12], which Hart later modified and contextualized as being relevant to young people's participation [13]. Almost a decade later, Shier's "pathways to participation" typology for children's participation in decisionmaking significantly revised the format of the model to five levels operating as a continuum [14]. These frameworks for participation have significantly influenced how children are regarded within research communities. Stewart however, stresses that it is difficult to find a workable definition of participation, and the popularity of incoherent definitions in health research "belies fundamental uncertainties about what [participation] entails and its associated benefits" ( [15], p. 124). Towards overcoming uncertainties around hierarchical frameworks for decision-making, Turner argues that the conceptual frames provided in the growing literature around CoPs can provide a more comprehensive understanding of health research systems aiming to have an influence on practice [16]. Lave and Wenger contributed greatly to the concept of CoPs and focused on informal and situated interactions towards achieving a better understanding of learning that is grounded in practice [17]. Later, Wenger focused on the trajectories of participation, social identities, and the effects of participation within different communities; and developed a set of indicators for CoPs [4]. Wenger defines the three core components of CoPs as the domain, which refers to a "concern, set of problems, or passion about a topic"; the practice, representing the knowledge that the group shares and generates; and, the community, which is the set of interpersonal relationships that are the product of engaging in learning through practice [4]. Wegner, McDermott, and Snyder developed the definition of communities further as "Groups of people who share a concern, a set of problems, or a passion about a topic, and who deepen their knowledge and expertise in this area by interacting on an ongoing basis" ( [18], p.4). When the three elements function well, CoPs become structures that can take on the responsibility of developing and sharing knowledge [18]. le May expanded the CoPs framework to account for the outcomes and benefits of CoPs in health and social care, including knowledge sharing, learning, building social relationships, innovation, and improving organizations [19]. Research communities engaging in research as practice have been studied using the CoPs framework, which has been applied to a variety of contexts (e.g., incarceration, community development, health, and education) with adults and mixed-age populations. Yet, the process of engaging youth and their families in health research communities of practice CoPs is yet to be studied in a systematic way. Furthermore, major funding agencies (e.g., Canadian Institutes of Health Research) equate patient engagement with the adult model of participation and do not give special consideration to how research should be conducted with youth and their families. In the following section we identify how insights about youth and their families' views are taken into account as well as their specific roles in health research CoPs that are created through IN•GAUGE®, a research program focussed on building knowledge through conducting research with youth and their families, rather than on them. --- Approach: Research with youth and their families Turner asserts that framing health research within CoPs should first involve translating research findings into practical implications for organizations and identifying ways of developing and "communicating evidence from social science research that demonstrates its relevance to'real-world' decision-making such that it has maximum impact on healthcare policy and Practice" ( [16], p. 2). Towards achieving the important next step of understanding the roles and outcomes of youth involvement in health research CoPs it is essential to hear from the youth themselves. IN•GAUGE® creates health research CoPs through the implementation of participatory research agendas with co-researchers (we use this term instead of "participants" towards acknowledging the contributions made as well as the power that has been divested according to participatory action research principles). These health research CoPs emerged through both intentional planning as well as the organic development of a web of relationships that results from sustained engagement in an area of research, depending on the particular study. Through IN•GAUGE®, qualitative and arts-based methods (such as photovoice, computerized drawings, body mapping) are applied to help amplify the voices of youth and their families, as well as explore and share research findings from the youth-and family-centred health research program through accessible strategies (such as the development of films, websites and choreography). These strategies all worked to highlight the voices of young people and families with lived experience and to ensure that the best available evidence flowing from the research is in the hands of those who influence children's healthparents, families, healthcare professionals, decisions makers and children and youth. In considering CoPs framework in relation to IN•GAUGE®, the health and well-being of youth with physical and mental illness and their families is the domain. The practice is the development of strategies for improving child and family health and well-being, and the community are the youth and families, service providers, researchers, and policy makers that are brought together into action through the common concern (i.e., the research question). Learning is promoted in IN•GAUGE® health research CoPs through participatory action research protocols involving knowledge brokering and various feedback cycles with youth and families, researchers, service providers, and special knowledge holders [6,20]. For learning to happen within health research CoPs, individuals need to be willing to contribute to the evolution of collective learning through sharing information, developing, and implementing strategies and conducting evaluations [21,22]. Promoting and creating the space for identity, capacity building, leadership, knowledge building, and relationship building --- Identity Youth and families find a sense of their roles and identities within health research CoPs that have been created through IN•GAUGE®. Co-researchers have revealed that they have learned a lot about their own journey with illness through being engaged and finding a space for reflection in the research process. Many co-researchers have reflected on how they felt at ease in the research process and viewed their participation as an opportunity to give back and help others who have similar challenges. Co-researchers often make strong statements about their identity and the ways in which their health condition is intertwined with their identity, and how that identity relates to the particular project. The building of shared identities relating to health research CoPs (i.e., feeling of belonging and being welcomed in the community) is also important, and creates the ability to transcend ways of communicating (i.e., disciplinary, cultural, generational, etc.), acknowledging other's perspectives, and challenging assumptions [23]. Much of this communication is facilitated through the use of highly flexible interview schedules, as well as the creation of a safe space established as a result of the implementation of Youth Advisory Councils (YACs) and Family Advisory Councils (FACs). YACs and FACs contribute knowledge and direction to developing projects through project scoping, giving input on suitable research methods, providing feedback throughout the research, and planning for and participating in knowledge translation (KT) and dissemination (Fig. 1). Within IN•GAUGE®, for those studies that focused directly on youth experiences of health and illness, YACs facilitated the participation of young people in research outside of the direct influence of their families. FACs brought to light the lived experience of health and illness on families and also served to complement the work of the YACs for those projects more focused on youth experiences. Multi-directional communication and critical self-reflection within IN•GAUGE® health research CoPs contributes to connectivity and learning across boundaries and promotes the development of a shared identity and sense of belonging within health research CoPs, one of Wenger's (1998) indicators of CoPs. --- Co-researchers from IN•GAUGE® involved in YACs and FACs frequently state that they enjoyed being engaged through the research process, and state that reflecting through the research process enables them to find different interpretations of their own or a family member's illness and disability. --- Capacity building Capacity building for youth and their families includes skill development and the enhancement of self-esteem and the ability to build social networks. Skill development involves refining communication and advocacy skills within health research systems (i.e., through YACs and FACs), which is important for developing meaningful engagement within health research CoPs more broadly, as well as in health systems more generally [24]. Through promoting and maintaining relational qualities within the research agendas, the researchers themselves provide many opportunities enhancing the self-esteem and abilities of co-researchers to build and extend their social networks [3]. Co-researchers often express that their opinions were being taken seriously and felt empowered to communicate with others through the research. Enhancing this ability to influence social networks is especially important for youth and families who may be disadvantaged through their experiences with illness. In developing IN•GAUGE® projects it is important to be cognizant that belonging to additionally marginalized groups (i.e., Indigenous, female, etc.) can cause enhanced jeopardy for health and social outcomes (Demas,[25]), requiring special attention to how power relationships are addressed and how social relationships are enhanced within health research CoPs. It is especially important when working with such groups that partnerships are fostered, collaboration is promoted, and that shared concerns are explored (i.e., the domain) early on [26]. It is also critical to consider that social relationships may mean something different for each group of co-researchers and can be influenced by factors such as culture and regional norms [27,28]. For example, in an IN•GAUGE® study exploring African newcomer experiences with the Canadian health system, co-researchers developed an ability to extend their influence and knowledge that they then used to advocate for improved access to health services in the context of a change in political system (i.e., government cutbacks for newcomer health services). --- Leadership Co-researchers' learning about identity and the development of new skills eventually lead to greater participation and leadership in projects. A parent of a child with complex care needs talked about their long-term involvement in an IN•GAUGE® study and how being asked to give their thoughts on the research process and dissemination of research findings (i.e., through participating in a video documentary) enabled them to create meaning and engage in deep reflection and learning through the process. Through advisory councils, coresearchers take up a number of leadership roles and have demonstrated commitment and interest in the research reaching its full potential. Co-researchers also report finding spaces within health CoPs where they can directly impact their day-to-day care through finding new pathways for informing service providers about their particular needs. The leadership skills of coresearchers are brought into the initiation and development stages of the research through providing spaces for co-researchers to shape research priorities, project design and methods. --- Knowledge building Patient engagement involves acknowledging that youth and families have certain knowledge and skills, but that they also will gain knowledge and skills through being involved in the health research CoPs [2]. Likewise the experience of knowledge sharing and building holds true to others who may be involved in research (e.g., clinicians), as well as the researchers themselves. Furthermore, direct interactions through health research CoPs are especially important for picking up on social cues and developing critical understanding of the lived experiences of youth and families [29]. Through working with co-researchers it was possible to build knowledge around the topics being researched, as well as knowledge contributing to how to develop approaches within ongoing and future health research CoPs. Co-researchers are keen to be part of the reflexive research design and are asked to give feedback regarding different stages of the research. Innovations in the research process occur through integrating the feedback and striving to find new and better ways to bring youth and families into health research CoPs as fully engaged co-researchers. Through YACs and FACs, coresearchers provide key knowledge for the analysis of data and KT. A few examples include the direct input into content and design for a KT website, involvement in the editing of a video documentary, and feedback on the artistic interpretation of research findings. --- Relationship building Relationship building is an essential component to ethical research engaging youth and families [22], and is central to IN•GAUGE® health research CoPs [3]. Relationship building occurs among youth and families, as well as among youth, families, and different members of the IN•GAUGE® health research CoPs (Fig. 1). Social relationships that are fostered through previous interactions (i.e., through systems of care) between the different co-researchers (i.e., youth and their families, university researchers, service providers, and special knowledge holders) act as foundations for many of the IN•GAUGE® projects. Coresearchers are invited into collaborative spaces where their perspectives are heard and given full consideration and see YACs and FACs as a safe spaces to share experiences with each other and create sense of community. Building relationships in this way lead to the development of respect, knowledge, awareness, and understanding of knowledge within the community and enable youth and their families to contribute in more meaningful ways to health research CoPs. It is also important to acknowledge that being involved in health research CoPs can be burdensome for co-researchers and that the risks associated with tokenistic participation could be managed in part by creating measures to equally value the commitments of co-researchers, such as through adequate remuneration (depending on context, specific research project and contributions, time commitments, etc.). Co-researchers that are part of IN•GAUGE® health research CoPs are given honorariums, as well as other types of compensation (e.g., meals and transportation). Such protocols are put into place to demonstrate respect, value and commitment towards co-researchers. In some situations it is appropriate for co-researchers to be employed as paid staff members on a project as a way of formalizing their roles and compensating them for their knowledge, experience and contributions. Co-researchers involved in IN•GAUGE® YACs and FACs are also often given the option being a coinvestigator or consultant to projects. Such categories came with different benefits and types of payments. --- Limitations This interpretation mainly focused on the engagement of youth and families within health research CoPs through exploring their interactions with university researchers. This paper would have been strengthened by working with co-researchers in its development however time constraints for co-researchers and prior commitments to other knowledge translation activities by members of YACs made such an approach challenging, reflecting broader challenges of working in a participatory manner. Further examination of the internal structures and connections between the other actors (i.e., with service providers and special knowledge holders) within emerging health research CoPs would be advantageous for developing greater understanding and best practices around how health research CoPs function as entire systems [30]. Further investigations into the structures of YACs and FACs would also be beneficial for understanding their impacts of health policy and practice [31]. --- Conclusions The dearth of health literature focusing on patient engagement involves frameworks equating patients with adults, not recognizing that these approaches may not resonate with youth. Through exploring the outcomes of engaging youth and their families through IN•GAUGE®, a sixteen yearlong research program led by Dr. Roberta Woodgate that is focused on working with youth and their families across the spectrum of patient engagement, improving health research and practice, we have gleaned several important insights about the development of health research CoPs for health systems for youth and their families. Promoting and creating the space for identity, capacity building, and leadership is integral to the meaningful engagement of youth and their families in health research. Within such conscious spaces, co-researchers are able to participate, learn, build social capital, and contribute to knowledge and building relationships that are designed to innovate and improve healthcare systems. --- Availability of data and materials Not applicable --- Abbreviations CoPs: Communities of Practice; FAC: Family Advisory Council; UNCRC: United Nations Convention on the Rights of the Child; WHO: World Health Organization; YAC: Youth Advisory Council Authors' contributions All authors contributed to designing the structure and content of the manuscript as well as drafting of the manuscript. All authors have approved the submitted version of the article. --- Ethics approval and consent to participate Not applicable --- Consent for publication Not applicable --- Competing interests The IN•GAUGE® program was trademarked in 2015 by RLW to prevent the use of the same name in other research program. At this point in time there is no plan to sell the trademarked programme. All other authors declare that they have no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The involvement of patients in health research has resulted in the development of more effective interventions and policies in healthcare that respond to the needs of healthcare users. This article examines how working with youth and their families as co-researchers in health research communities of practice (CoPs), rather than just as participants, can benefit all involved. Health research (CoPs) promote an environment in which co-researchers have the opportunity to do more than just participate in the data collection phase of the research process. As co-researchers, youth and their families are able to participate, learn, and contribute to knowledge and building relationships that are designed to innovate and improve healthcare systems. However, in order to ensure engagement of youth and their families in health research that they find meaningful and rewarding, three factors have been identified as important parts of the process: promoting identity, building capacity, and encouraging leadership skills.

Introduction A nearly exponential growth in the scientific output on social epidemiology has been documented 1. However, since Solla-Price's seminal work on the exponential growth of science 2, it has been shown that scientific production doubles every ten to fifteen years 3. Hence, it remains unexplored whether the absolute growth described for general science would also be observed in relative terms for specific areas of knowledge. By examining scientific growth in specific areas, it is possible to highlight emerging themes, so as to indicate possible advances in the near future. The objective of the present study was to assess trends in scientific production of methods and themes in the investigation of social determinants of health, between 1985 and 2007. --- Methods This study consists of a bibliometric analysis of yearly trends in groups of publications indexed in PubMed (http://www.ncbi.nlm.nih.gov/ pubmed) from 1985 to 2007. The main database in PubMed is MEDLINE, but the proportion of non-MEDLINE citations is unknown and a strict search of MEDLINE is not possible. In addition to MEDLINE, PubMed covers in-process citations, some papers before 1950 (OLDMEDLINE), and citations that are out-of-scope from MEDLINE journals. Each journal article added to PubMed's collection is indexed by specialized staff under at least one MeSH (Medical Subject Headings) descriptor of the U.S. National Library of Medicine Thesaurus, which was established in 1960 and has been updated since. The MeSH Thesaurus contains a tree-like cross-referenced structure of controlled vocabulary, used to translate terminology employed in articles in different idioms to a "system language". Each "entry term" encompasses many synonyms, near-synonyms and related concepts, regardless of the idiom, wording and spelling used by authors. When a MeSH term is used, PubMed automatically searches on narrower descriptors further out on that branch of the tree. For instance, the use of the MeSH term "Socioeconomic Factors" automatically searches for "Education Status", "Income", "Occupation", "Inequality" and "Social Class". --- Principles applied to build search strategies First, whenever possible, a MeSH term was employed rather than non-MeSH "keywords" in all search strategies. If an appropriate MeSH term could not be identified, a search strategy was built with text words based on the authors' experience. Second, we favored the earliest MeSH terms -those added before 1985 -and the most general category of the MeSH tree hierarchy. Third, MeSH term definitions were scrutinized in order to choose only those terms with the desired meaning. Finally, based on a preliminary analysis, we excluded redundant terms in the search strategy, so as to keep it as parsimonious as possible. Our searches were conducted on May 17, 2008. --- Search strategies and data analysis We tallied the annual number of publications identified using the following search strategies (details of search strategy can be obtained from the authors): all publications in the PubMed database (search strategy number #1); publications in epidemiology/public health (#2); and selected themes in social epidemiology (#14). We further refined the bibliographic analysis of (#14) by selecting publications focusing on 11 themes: socioeconomic factors (#3), sex (#4), race/ethnicity (#5), prejudice/discrimination (#6), social capital/support (#7), lifecourse (#8), income inequality (#9), stress (#10), behavioral research (#11), contextual factors (#12), and residential segregation (#13). The number of publications identified in the epidemiology/public health search (#2) served as the denominator to calculate the proportion represented by each of the 11 social epidemiology themes (i.e. strategies from #3 to #13 were divided by #2). The total number of publications in PubMed (#1) was used as the denominator to calculate the proportion of publications in epidemiology/public health (#2 divided by #1) and in social epidemiology (#14 divided by #1). We also examined trends for three types of data analysis methods: multilevel modeling (#15), regression-based indices to measure inequalities (#16), and structural equation modeling/causal diagrams/path analysis (#17). The absolute number of publications of these data analysis techniques among the eleven social epidemiology themes were determined for each year between 1985 and 2007 and plotted. Their proportion in relation to the total number of articles in the 11 themes (i.e. the number of publications from strategies #15 to #17 divided by the number of publications from strategy #14) was also determined and plotted for the period 1985-2007. --- Results Between 1985 and 2007, there was a 2.3-fold increase in the annual number of citations added to the PubMed: from 329,263 in 1985 to 759,698 in 2007. In the same period, articles indexed under epidemiology/public health headings (search strategy #2) increased 5.3 times (from 48,719 in 1985 to 256,892 in 2007) and, among the selected themes in social epidemiology (search strategy #14), there was a 5.2-fold increase (from 9,349 to 49,052). In 2007, more than 30% of the scientific output indexed in this bibliographic database had at least one of the descriptors used to identify citations in the area of epidemiology/public health. In contrast, the relative contribution of social epidemiology increased moderately, compared to epidemiology/public health, reaching 6.5% in 2007. Trends in selected themes in social epidemiology are depicted in Figures 1 and2. Absolute frequencies tended to increase over the studied period, and more markedly in recent years. Themes like socioeconomic factors, sex, race/ ethnicity, behavioral research, stress, contextual factors, lifecourse and prejudice/discrimination are good examples in this regard. Due to small numbers, however, we could not be confident about the trend patterns for income inequality (#9) and residential segregation (#13). A different picture emerges when relative frequencies are considered. When the count of all epidemiology/public health publications serves Cad. Sa<unk>de P<unk>blica, Rio de Janeiro, 27(1):183-189, jan, 2011 as the denominator, the exponential growth pattern seen with the absolute count is no longer observed. Increases or stationary trends are seen for behavioral research, race/ethnicity, stress, contextual factors, lifecourse, prejudice/discrimination and social capital/support. Among these, contextual factors showed the steepest relative increase in the period, rising from almost 4% in 1985 to 7% of all epidemiology/public health publications in 2007. In contrast, socioeconomic fac-tors and gender showed declining relative trends. Residential segregation and income inequality each exhibited a fluctuating pattern, which likely reflects the small number of publications. Publications using one of the three analysis methods -multilevel modeling, structural equation modeling/causal diagrams/path analysis, and regression-based indices -all showed steep increases in the absolute number of articles, as well as in their relative frequencies (Figure 2). Among these, multilevel modeling emerged as the most employed method. --- Discussion Our results showed that absolute and relative trends can provide different conclusions, but regardless of how we plotted them, epidemiology and social epidemiology grew over and above the growth of general health science. One could argue that the larger growth of social epidemiologic themes in relation to the total citations in PubMed could be determined in part by an increased indexation of epidemiologic/public health journals in the database over the study period. However, this is not likely to be the case; epidemiologic/public health journals account for only 1% (n = 370) of the 37,665 journals indexed in PubMed and 2.2% of all publications (data from PubMed Journal Database), such that an expressive number of journals would have to be included in the database in order to artificially influence the observed trends. One of the limitations of the present study is that its findings are based only on PubMed; however, this is recognized as the largest and bestknown database in the field of health sciences. Another limitation is that not all retrieved publications necessarily strictly fit the definition of social epidemiology, that is, explicitly incorporating so-cial theory in the article's analytical framework 4. This concern is tempered by our use of MeSH terms, which increased the sensitivity and the specificity of the search strategies. Because the terms used were constant over time, the results are likely to reflect real trends in social epidemiology publications indexed in PubMed. Overall, from the results presented, it can be concluded that the branch of social determinants Y1 number of publications Y2 % in epidemiology/Public health of health has been growing fast, and that this growth was seen in nearly all of the 11 sub-areas. It is important to emphasize that the magnitude of absolute increases in some sub-areas might be misinterpreted as outpacing others, when, in fact, relative figures reveal increases that were actually modest. Although the number of publications in social epidemiology increased more than the average growth of the total publications in PubMed, epidemiology/public health did too, and the only themes in social epidemiology growing over and above trends in epidemiology/public health were those which lent themselves to multilevel or contextual analysis. This is a good example where methodological advances met theoretical needs. However, there is clearly room for wider use of established techniques, and for new methods to emerge and satisfy theoretical needs. --- Resumo Analisam-se tendências bibliométricas de tópicos relevantes na pesquisa epidemiológica de determinantes sociais da sa<unk>de. Realizaram-se buscas no PubMed no per<unk>odo 1985-2007, sobre: fatores socioeconômicos, sexo, raça/etnia, discriminaç<unk>o/preconceito, capital/ suporte social, curso de vida, desigualdade de renda, estresse, pesquisa comportamental, efeitos contextuais, segregaç<unk>o residencial, modelo multin<unk>vel, indicadores de desigualdade baseados em regress<unk>o e equaç<unk>es estruturais/análise de caminhos/diagramas causais. A freqüência absoluta, e n<unk>o a relativa, de publicaç<unk>es cresceu em todos os temas. O total de publicaç<unk>es aumentou 2,3 vezes no per<unk>odo, enquanto que o conjunto epidemiologia/sa<unk>de p<unk>blica e os temas de epidemiologia social aumentaram 5,3 e 5,2 vezes, respectivamente. Efeitos contextuais e modelagem multin<unk>vel apresentaram crescimento relativo acima do observado para epidemiologia/sa<unk>de p<unk>blica. Conclui-se que existe espaço para a ampliaç<unk>o do uso das técnicas de análise existentes e para que novos métodos surjam, atendendo a necessidades teóricas espec<unk>ficas da área. Consumo de Alimentos; Inquéritos sobre Dietas; Métodos Epidemiológicos Contributors R. K. Celeste collaborated in the design, write-up, data collection, analysis and interpretation and final revision. J. L. Bastos collaborated in the design, write-up, analysis and interpretation of data, and revision of the final version. E. Faerstein collaborated in the design and data interpretation, as well as in the write up and revision of the final version.

We analyze bibliometric trends of topics relevant to the epidemiologic research of social determinants of health. A search of the PubMed database, covering the period 1985-2007, was performed for the topics: socioeconomic factors, sex, race/ethnicity, discrimination/prejudice, social capital/support, lifecourse, income inequality, stress, behavioral research, contextual effects, residential segregation, multilevel modeling, regression based indices to measure inequalities, and structural equation modeling/causal diagrams/path analysis. The absolute, but not the relative, frequency of publications increased for all themes. Total publications in PubMed increased 2.3 times, while the subsets of epidemiology/public health and social epidemiologic themes/methods increased by factors of 5.3 and 5.2, respectively. Only multilevel and contextual analyses had a growth over and above that observed for epidemiology/public health. We conclude that there is clearly room for wider use of established techniques, and for new methods to emerge when they satisfy theoretical needs.

Introduction Everyday living in modern Western societies is dominated by mundane activities performed routinely, many motivated by obligations, although needs and desires are also common motives (e.g., Vilhelmson 1999;Michelson 2011;White and Dolan 2009). Since these activities constitute such a large part of everyday life, they are likely to have an impact on people's overall life satisfaction and emotional well-being. This was recently demonstrated by Jakobsson Bergstad et al. (in press) for a subset of routine out-of-home activities. As a further indication of such an impact, research has shown that satisfaction with life domains including work or school, family life, and leisure (associated with performance of particular mundane routine activities) is positively correlated with overall life satisfaction (Pinquart and Silbereisen 2010). Work commutes are in this context a neglected aspect of everyday life. A fact is still that billions of people commute to and from work every workday. An informal literature search of international transportation studies reveals that average commute times vary between 40 and 80 min, with public-transit taking longer than car commutes. An average of 4-10% of waking time on workdays is spent on commutes. Several previous, predominantly US studies have found that work commutes induce stress (see Novaco and Gonzales 2009, for review). Thus, it has been found that long work commutes in congested automobile traffic cause residual stress in the workplace (Novaco et al. 1990). Stress due to work commutes by public transit increases with the complexity of the commute (Wener et al. 2003) and with crowding in vehicles (Singer et al. 1978). In a similar vein, it has been argued that the negative effects of the length of work commutes substantially reduce the benefits of living in attractive places distant from the work place (Stutzer and Frey 2008). In a study using the day reconstruction method to measure emotional well-being, Kahneman et al. (2004) found that the work commute belonged to the episodes that were most frequently associated with negative feelings during the day. The aim of the present study is to add to these research findings by investigating how overall life satisfaction and emotional well-being are affected by the work commute. Our approach is different in that satisfaction with the work commute is measured such that its correlation with overall life satisfaction and emotional well-being can be directly assessed. The analyzed data were collected in Sweden where, as in other European countries, the travel mode split is more even than in the US where driving is predominant. The results therefore also provide a contrast to the previous research in the US demonstrating stress effects. Happiness (also commonly referred to as subjective well-being) has attracted a plethora of cross-disciplinary research in recent years (e.g., see reviews by Dolan et al. 2008;Lyubomirsky et al. 2005). In line with this research, we refer to happiness or subjective well-being as a higher-order construct consisting of a cognitive and two affective components (Busseri and Sadava 2011). The cognitive component consists of a judgment of life satisfaction (evaluations of life circumstances) that is commonly measured by reliable self-report rating scales, for instance the 5-item satisfaction with life scale (SWLS) (Diener et al. 1985;Pavot and Diener 1993;Slocum-Gori et al. 2009) which will be used in the present study. The affective components of happiness include the positive and negative moods and emotional episodes that people experience. Several self-report methods have been devised to measure these affective components. A distinction is whether the methods are on-line such that they assess immediate affects (Stone et al. 1999) or retrospective and memorybased (Schwarz et al. 2009). The positive and negative affect scale (PANAS; Watson et al. 1988) has frequently been used either on-line to measure current mood or retrospectively to assess the frequency and intensity of affects for a specified timeframe. On this measure happiness increases with the frequency and intensity of positive affect (PA), including emotions such as joy and delight, and decreases with the frequency and intensity of negative affect (NA), including emotions such as anger and fear. A measure of emotional well-being (also referred to as the affect balance) is constructed by computing the difference between retrospective assessment of the frequency and/or intensity of positive and negative affect. Other research has shown that affect is related to two dimensions, a pleasantnessunpleasantness dimension labelled valence and an active-passive dimension labelled activation (e.g., see the affect grid, Russell 2003). Diener and Lucas (2000) accordingly proposed that measures of the affective components of happiness should be based on a dimensional description varying in valence and activation. The Swedish core affect scale (SCAS; Va <unk>stfja <unk>ll et al. 2002) that will be used in the present study is such a measure based on the affect grid. Both the cognitive and affective components of happiness may be influenced by the work commute (Ettema et al. 2010). Even though the work commute generally have an intended positive outcome-and would therefore have a positive effect on the cognitive component of happiness-it may still be experienced as stressful. Thus, how commuters react affectively should also be important, whether they are predominantly stressed, relaxed, excited or bored. We have therefore developed the satisfaction with travel scale (STS) to measure a cognitive component and two affective components of the experience of any type of travel (Ettema et al. 2011;Jakobsson Bergstad et al. 2011). The STS that will be used in the present study thus has three components: a cognitive evaluation of the quality of travel, an affective evaluation of feelings during travel ranging from stressed to relaxed, and an affective evaluation of feelings during travel ranging from bored to excited. --- Method The participants were 713 work commuters (41.7% male; age ranging from 20 to 65 with a mean of 41.2 years) living in the three largest urban areas of Sweden (Stockholm population 850,000; Go <unk>teborg population 510,000; Malmo <unk>population 395,000) (for detailed sample characteristics, see Table S1 in the supporting information available online). The participants answered a mail questionnaire that had three consecutive modules consisting of questions about the work commute, overall happiness, and sociodemographics. To minimize memory distortions (Schwarz et al. 2009) the most recent normal commute to and from work was targeted in the questionnaire. In the first module the participants first reported the date, departure and arrival times, intermediate stops, and travel modes. On the basis of the self-reports of mode use, work commutes were classified as made by car if car was used for at least one leg of the commute1 (to work n = 269; from work n = 259), as made by public transit (PT) if PT was used for at least one leg and that no car was used for any other leg (to work n = 251; from work n = 254), and as commutes by slow modes if the commuters walked or biked all legs (to work n = 165; from work n = 164). In the same module the STS was then administered to assess satisfaction with the commute to and from work, respectively. The STS consists of nine seven-point adjective scales; three ). The order between the ratings scales was counterbalanced. In the second module the SWLS (with the time frame last month) was first administered. An average was computed for ratings of agreement on 1 (do not agree) -7 (fully agree) scales to the following five statements: in most ways my life is close to my ideal; the conditions of my life are excellent; I am satisfied with my life; So far I have received the important things I want in life, and; if I could live my life over, I would change almost nothing. A measure of the affective components were thereafter obtained as self-report ratings of the frequency (never = 0; rarely = 1; sometimes = 2; often = 3; very often = 4) during the last month of experiencing three intensities (slightly = 1; moderately = 2; very = 3) of the six positive emotions glad, active, joyful, awake, peppy, and pleased and the six negative emotions sad, passive, depressed, sleepy, dull, and displeased. An affect-balance index was constructed by multiplying the ratings of frequency and intensity for each emotion, then summing with a positive sign for the positive emotions and with a negative sign for the negative emotions. --- Results A composite measure of satisfaction with the work commute was formed by averaging across all nine STS scales. 2,3 As Table 1 shows, on this measure daily commute time (from 10 to 180 min) reduces satisfaction with the work commute. Slow commute modes (walking and biking) also result in more satisfaction than car and public transit (Table 2). A multiple linear regression analysis reported in Table 3 reveals a significant negative effect of daily commute time and a significant positive effect of slow modes (walking/ biking) versus public transit or driving. By dichotomizing the affective components of STS, Table 4 shows that positive or neutral feelings dominate during the work commute. The negative effects of daily commute time are also observed for the affective components. Table 5 summarizes the results of multiple linear regression analyses with SWLS and affect balance as dependent variables, separately for the commutes to work and from work (the full results are provided as Table S2 in supporting information available online). As can be seen, satisfaction with the commutes to and from work directly influences the affect balance as well as SWLS, directly or indirectly through the affect balance. In previous studies (reviewed in Lyubomirsky et al. 2005), socio-demographic factors have accounted for approximately 10% of the variance in SWLS. The present figure is 14% that decreases to 7% when affect balance is partialled out. Satisfaction with the work commute accounts for an additional 2% of the variance in SWLS and an additional 11-12% of the variance in the affect balance. --- Discussion The key finding of the reported survey is that satisfaction with the work commute has a substantial influence on overall happiness, particularly on the balance between positive and negative affect (Table 5). This influence would be negative for participants who are dissatisfied and positive for those who are satisfied with their work commutes. On average satisfaction is high (Table 1), thus a positive contribution is made to overall happiness. In addition the present study fails to show that work commutes are predominantly stressful (Table 4), as previous research has found (Novaco and Gonzales 2009), although negative feelings during the work commute increases with the length of the commute. The present results add to previous findings by suggesting that affect associated with mundane routine activities in different life domains may play important roles for overall happiness. In fact, the role of satisfaction with the work commute is of the magnitude observed for several of the mundane routine activities investigated by Jakobsson Bergstad et al. (in press). We assume here that the causal direction is from satisfaction with the work commute to overall happiness. A reverse direction is however also conceivable. For instance, given the negative effects that unemployment has on overall happiness in the current economic situation, it is possible that the happiness derived from having a job spills over to the satisfaction with the work commute. Why are the present results inconsistent with those of studies of work commuting, predominantly by car, in the US (Novaco and Gonzales 2009)? One factor is that biking and walking, more common in the present study conducted in Sweden (and this would be the same in several other European countries, e.g., The Netherlands), contributes more to Missing values were excluded pairwise satisfaction with the work commute than driving and public transit. That walking and biking provide desirable physical exercise is a reason for their popularity (Lawrence et al. 2006). If commutes are short, as walking and biking commutes usually are, they may also be appreciated as a buffer between the work and private spheres (Jain and Lyons 2008). Particularly over longer distances, satisfaction with the work commute decreases. It is an open question whether public transit leads to more satisfaction than driving when the length of the commute increases. Speaking for public transit is that, more than solo driving does, it allows for engagement in activities such as talking to others, resulting in PAs, and work or entertainment activities that may reduce stress and boredom. The entries in the table are increments (DR 2 ) in hierarchical regression analyses. All of the increments and the full models are statistically significant at p <unk>.01 or less. The full results are given as Table S2 in supporting information available online Other possible factors accounting for satisfaction with the work commute are not directly related to the travel mode per se. Some research has demonstrated adaptation to adverse conditions (Frederick and Loewenstein 1999). The results of the present study do not exclude that some people report positive experiences because they adapt to negative effects of their work commute. These people may be susceptible to adaptation costs (e.g., physiological stress reactions Ng et al. 2009) which the self-report measures in the present study do not fully capture. The findings in the present and other studies (Jakobsson Bergstad et al. in press, 2011) that experiences of work commutes and other mundane routine activities has measurable effects on overall happiness should be a reminder of that engagement in particularly meaningful activities such as practicing generosity, developing social relations or learning to manage stress (Lyubomirsky 2008), even though important, are not the only routes to happiness in life. For many people, being able to make the routines of everyday life work, such that positive feelings dominate over negative feelings resulting from daily hassles, may be equally important for their overall happiness. This insight is particularly significant to convey to policy makers who are responsible for spending tax money to improve municipality facilities.

Research suggests that for many people happiness is being able to make the routines of everyday life work, such that positive feelings dominate over negative feelings resulting from daily hassles. In line with this, a survey of work commuters in the three largest urban areas of Sweden show that satisfaction with the work commute contributes to overall happiness. It is also found that feelings during the commutes are predominantly positive or neutral. Possible explanatory factors include desirable physical exercise from walking and biking, as well as that short commutes provide a buffer between the work and private spheres. For longer work commutes, social and entertainment activities either increase positive affects or counteract stress and boredom. Satisfaction with being employed in a recession may also spill over to positive experiences of work commutes.

INTRODUCTION In 2017, the WHO estimated that approximately 280 000 neonates worldwide died of congenital malformations within 28 days of birth. 1 In China, the incidence of congenital abnormalities is approximately 5.6%, and approximately 1 million infants are born with birth defects every year. 2 With the development of medical ultrasound techniques and improvements in prenatal diagnosis, fetal malformations can be discovered during pregnancy. In China, implementation of the universal two-child policy has resulted in an increase in pregnant women with advanced age and high-risk pregnancies, which in turn has led to an upwards trend in the incidence of fetal abnormalities. 3 Receiving a diagnosis of a fetal anomaly is distressing for expecting parents, who frequently experience intense emotional responses, including shock, grief, anger, uncertainty and fear. [4][5][6] A substantial proportion of expecting parents choose to terminate the pregnancy when their babies are diagnosed with life-limiting anomalies. 7 Termination of pregnancy for fetal anomaly (TOPFA) not only causes great physical pain to women but also brings intense sadness and destructive psychosocial problems, which may last for many years, 8 including anxiety, depression, post-traumatic stress disorder (PTSD), complicated grief and even suicidal thoughts. [9][10][11][12] There is a sizeable body of studies and recommendations associated with women's experiences of grief and support needs. [13][14][15] Chen et al 13 indicated that women with induced labour for fetal abnormality will experience a special coping process identified as 'admitting the child's existence, seeking information and emotional support, avoiding the TOPFA event and looking forward to the future'. Traditionally, fathers often play the role of supporters for their partners, and fathers' experiences of TOPFA have been underexplored in comparison to mothers. 11 16 However, they also experience high levels of --- STRENGTHS AND LIMITATIONS OF THIS STUDY <unk> A qualitative design using face-to-face, semistructured interviews combined with observation is beneficial for improving the accuracy of data. <unk> The inclusion of accounts from mothers and fathers provided a rich insight into parents' experiences and need for social support. <unk> Our population did not distinguish between stillbirth and fetal malformation, and the experiences and needs may be different. Open access grief, anxiety, depression and PTSD, 11 17 18 which requires acknowledgement and validation from healthcare professionals, family and friends, community networks and workplaces. 11 To date, a growing body of studies has focused on men's grief and care experiences following abortion and stillbirth, [19][20][21] but few qualitative studies have exclusively focused on fathers' psychological experiences and need for social support after TOPFA, 22 23 particularly in the context of Chinese culture. In traditional Chinese culture, childbearing is highly valued by many couples, and TOPFA can bring feelings of guilt and inferiority. 24 25 In addition, traditional Chinese cultural concepts such as reporting good news but not bad news and superstition can also have an impact on parents' psychological experiences and coping styles. 25 26 According to the theory of social exchange, social support is a combination of functional support and structural support. Functional support includes emotional, instrumental, informational and appraisal support. Structural support includes formal support (eg, from healthcare professionals) and informal support (eg, from close family members and friends). Social support is often identified as the most critical component for mothers' adaptation to the death of their children, 27 as it can effectively decrease parental grief, anxiety, depression and PTSD. [28][29][30] Therefore, it is necessary to explore the experiences and social support need of parents under the special cultural context of China to help clinical medical staff and social workers better understand and support parents. Given the significant role of social support in reducing the negative psychology of parents, and the lack of recommendations on bereavement care for parents, especially for fathers, this study sought to explore the experiences and social support need of both fathers and mothers following TOPFA against the unique cultural background of China. --- METHODS The study and manuscript were prepared following the Consolidated Criteria for Reporting Qualitative Research guideline. --- Design and setting This study employed a qualitative design using semistructured, in-depth interviews combined with observation. The purpose of interviewing both mothers and fathers was to provide a more comprehensive understanding of parents' experiences and social support need from different perspectives. Because the research topic was sensitive and unfamiliar to the interviewees, the in-depth, semistructured interviews were particularly well-suited for the study. The study was carried out in the Obstetrics and Gynaecology Hospital affiliated to the Medical College of Zhejiang University from March to September 2016. This hospital has set up the Zhejiang Provincial Prenatal Diagnosis Center, which has carried out a variety of prenatal diagnosis techniques, and is responsible for intrauterine diagnosis and intervention of fetal congenital abnormalities and genetic diseases in the province and surrounding areas. --- Participants A purposive sample of women who experienced TOPFA and their spouses was used. Inclusion criteria were that participants were at least 18 years of age, had experienced fetal abnormality (gestational week >14 weeks) confirmed by the Zhejiang Provincial Prenatal Diagnosis Center, decided to terminate a pregnancy for a fetal abnormality and voluntarily participated in the study with signed informed consent. The exclusion criteria were as follows: a history of psychosomatic disease, intellectual disability or illiteracy, and/or inability to understand the interview questions. To ensure a representative sample and capture a wide range of perspectives, the heterogeneity of sample selection was expanded for participants' age, education level, occupation and perinatal loss characteristics. --- Data collection Data were collected using face-to-face, semistructured interviews combined with observation. During the interview, the changes in the interviewees' expressions, speed and intonation were observed and recorded. 31 At the same time, their feelings or opinions were clarified and confirmed in time to ensure the accuracy of the data. The researchers developed the interview outline before the interview, including the following open-ended questions: How do you feel when you received news of fetal anomaly? What are your concerns going forward? What kind of support and assistance would you like? How does the two-child policy affect your current pregnancy and future? The order of the questions in the outline is not fixed and can be adjusted according to the specific situation. All interviews were conducted by the first author, who was specifically trained in conducting qualitative interviews. The interviewer had no prior relationship with parents, briefly introduced herself before the interview and used neutral, objective and non-leading language during the interview to maximise data integrity. Before the formal interview, the interviewer provided participants with a detailed introduction to our research purpose, significance, interview process and privacy protection measures for the interview content, in order to promote the research subjects' familiarity with the research topic and reduce their sensitivity to the research topic. The women and their spouses were interviewed separately in a quiet and undisturbed environment based on the convenience of the interviewees. After a written informed consent was obtained from the participant, interview was recorded with a digital recorder. The sample size of the study was based on the principle of information saturation. 32 The study was discontinued when qualitative Open access data reached saturation. When the number of interviews with mothers reached 15 and the number of interviews with fathers reached 12, our research data were saturated and no new information appeared, so the sampling was terminated. Each interviewee was interviewed 1<unk>2 times, 40<unk>60 min each time. In three cases, the spouse did not appear during the interview, so the interview was missing. To protect participants' privacy, 15 women were numbered 1A<unk>15A, and their spouses were numbered 1B<unk>15B (3B, 13B and 15B were not present). --- Data analysis The interview recordings were transcribed verbatim into textual materials within 24 hours and checked by another researcher. The data were analysed by the same researcher who collected the information following Colaizzi's sevenstep procedure. 33 The specific steps were as follows: the researcher read all transcribed materials carefully, analysed and obtained significant statements, coded recurring and meaningful viewpoints, summarised all encoded viewpoints, developed a detailed and complete narrative, distinguished similar viewpoints and verified the obtained results with the interviewees to ensure the authenticity of the content. Initial themes were developed by the first author and then discussed and refined with qualitative research experts and all authors to avoid subjective influence and ensure the accuracy and objectivity of the results. To ensure the trustworthiness of the data, the following procedures were used: anonymous transcription of each interview; making field notes after each interview and the field notes were examined during the data analysis to help better understand the data; the researchers remained reflexivity to recognise their potential effect on the study findings and maintained faithful to the perspectives of the interviewees. Finally, we conducted the last step of the Colaizzi's seven-step procedure. The participants were invited to respond to the obtained results. The results were presented to them in a general overview table containing quotes, emerged meanings, all themes and subthemes. If the interviewees disagree with the results, the researchers then rechecked the relevant codes to conduct the final analysis. In this step, the research team verified the obtained results with the interviewees to ensure the accuracy and credibility of the results. All interviewees considered that the results represented their perceptions and no significant themes were missed. --- Patient and public involvement None. --- RESULTS Finally, 12 couples and 3 additional women (whose spouses were not present) were interviewed. The demographic information of the participants is shown in table 1, and perinatal loss characteristics are shown in table 2. A total of four overarching themes were identified across the interviews, each with some subthemes (see figure 1). Details of the themes are outlined below. Theme 1: The shock of facing reality Subtheme: Query and verification of fetal abnormalities Without psychological preparation, parents initially showed strong shock and denial in the face of the Subtheme: Abandoning the fetus with reluctance and struggle After identifying fetal abnormalities, bereaved parents often struggled between continuing the pregnancy and inducing labour. In the interview, 11 parents showed obvious entanglement and reluctance when forced to make a choice. Until the last second before induced labour, I hoped there would be a miracle (participant 2A). If I had enough money, I would have still wanted to give birth to him [referring to the abnormal foetus].... As a father, I was still reluctant to give up. I sent him to the operating table with my own hands. It was not that he gave up on me, but I gave up on him (participant 14B). Especially in the face of non-fatal fetal abnormalities, it was more difficult for parents to make the decision to induce labour, and four parents had a strong sense of uncertainty. At present, there are uncertain answers. For us, there was too much uncertainty, so we dared not take this risk to give birth to our children (participant 2A). Due to the uncertain nature of the lump at present, it is difficult for us to make a choice emotionally. We don't know whether the decision to induce labour is right or wrong (participant 5B). --- Subtheme: Compromise with reality and seeking spiritual comfort After objectively weighing the advantages and disadvantages of medical risks, fetal health and potential future economic burden, parents were forced to acknowledge reality, choose induced labour and seek spiritual comfort. In the face of fetal abnormalities, nine men accepted reality more pragmatically than women. I believed in science because the baby was a flawed life and could not survive. That is all we could do. We have tried our best (participant 9A). This was a confirmed fact. If the child would be born with so much pain, we would rather make the choice to terminate in the early stage (participant 12B). Five parents received spiritual comfort and support from their living children, and the living children also made the parents more confident in their decision to induce labour. Some parents sought spiritual comfort by believing in religion. My daughter was very clever and kept asking me about my condition, which is also a comfort to me [expression was comfortable]. At my age, even if the baby was born, his quality of life would not be high, and I do not want to force my daughter to be involved (participant 3A). Buddha has spoken of fate... I comforted myself, thinking that I had no fate with this child, so as to calm my heart, and that I was lucky to find out the foetal abnormality early (participant 13A). --- Subtheme: Intense grief Parents felt intense grief following the death of the fetus, similar to the loss of other relatives. When the doctor said that the baby's foetal heartbeat disappeared, I felt heartrending pain (participant 3A). During the interview, there were five fathers weeping in sadness, indicating that the fathers also had complex grief. However, in traditional thinking, spouses mainly play the role of supporters and need to be strong to avoid Open access aggravating the sorrow of their wives. Therefore, they often hid their sadness in front of their wife and family. I didn't show sadness in front of my wife, I wanted her to feel that I wouldn't care too much about the outcome, but my wife felt that I wasn't sad at all and didn't care about the child, and I was actually very sad [with tears in his eyes, voice trembling] (participant 1B). I really felt a pity in my heart [he expressed regret many times to researcher), my wife could cry, but I couldn't [the corners of his eyes were wet] (participant 4B). Family members, especially grandparents, also had grief, and the grief of family members could aggravate the grief of parents. I didn't want others to know about it [TOPFA event], especially my mother who was also looking forward to the baby. Facing the reality of the abnormal foetus, I was very sad and afraid that my mother's sadness would aggravate my sadness (participant 3A). After the induction of labour, the role of parenthood was completely interrupted, and four pregnant women and their spouses felt at a loss. I felt that this happened suddenly [crying loudly], and it was difficult to accept it for a while. When I saw all the baby-related supplies, I would think about it [TOPFA event]. I was very reluctant to leave this baby. I was pregnant when I was admitted to the hospital, but I came home with nothing. It was even more sad to see others holding their babies (participant 3A). --- Subtheme: The shackles of traditional thinking Traditional Chinese ideas such as 'family succession','son preference' and'superstition' brought pressure to parents. Seven couples in this study were bound by traditional ideas. My dad already had three granddaughters, and was looking forward to having a grandson, which was a little stressful for me (participant 10B). In a country like China, if the neighbours in the countryside suspected that the baby was not developing well, they would certainly speculate, which might cause some gossip (participant 14B). At the same time, due to the conservative idea of'reporting good news but not bad news', some parents were unwilling to share their sad feelings with others to avoid and cover up their inner grief. I didn't want to see anyone when I got home, I wanted to be alone in a small room where no one could disturb me, and I didn't want others to know about it (participant 3A). --- Subtheme: Rumination After accepting the reality of fetal abnormality, 8 parents were still confused about the causes of fetal abnormality and reflected on their own deficiencies during pregnancy. This rumination would lead to parents' strong sense of guilt and self-blame. My wife had a cough in the early stages of pregnancy, and a plaster was applied to her neck... We worked at Taobao and faced the computer for a long time every day. We really did not exercise enough, and our immune systems were not very good... (participant 14B). I wasn't ready for this baby. I was pregnant unexpectedly. Without knowing I was pregnant, I took cold medicine and underwent anaesthesia, so I was an unqualified mother [with tears in her eyes] (participant 12A). Theme 2: Concerns surrounding termination of pregnancy Subtheme: Concerns about induced labour Rivanol amniotic cavity injection is the most commonly used method of labour induction in the middle and late stages of pregnancy. Most parents did not understand the complete process of labour induction and were concerned. In this study, nine pregnant women and their spouses were full of anxiety and fear about the process of induced labour. I was worried about this delivery. I had heard the process was terrible (participant 2A). Subtheme: Concerns about women's physical and mental recovery Induction of labour not only made women suffer great physical pain but also caused great psychological trauma to them. Three spouses expressed strong concerns about maternal physical recovery, and four spouses were more worried about the psychological recovery of the wives. I was worried about my wife's health. After all, I could only talk about the next one [referring to the next pregnancy] after her physical recovery (participant 14B). --- I was worried about my wife's psychological recovery [repeated many times]. If she didn't adjust to this well, having another child would increase her burden (participant 10B). Subtheme: Concern about the subsequent pregnancy TOPFA not only brought intense grief, anxiety, fear and other psychological problems to parents but this traumatic experience also left an indelible psychological shadow on them. Parents who have the need to get pregnant again were especially concerned about the risk of the subsequent pregnancy. There was a psychological shadow, and I always felt that I had experienced a miscarriage. Even if --- Open access everything is normal in the next pregnancy, there will be faint worries (participant 6A). Theme 3: The embarrassment of the two-child policy Subtheme: Contradiction between older age and the two-child policy The comprehensive liberalisation of the two-child policy has aroused countless couples' desire for having children, including older couples. However, in the face of the reproductive risks brought by old age, many families fall into the embarrassing situation of whether to have a second child or not. In this study, nine cases were giving birth to a second child, and six of them were mothers of advanced age. I just wanted to have another child. It was better to have two children. After the second child policy was liberalized, this age [approximately 40 years old] was a concern. The second child policy was embarrassing for us... Our family and economic conditions allowed us to have a second child, but our physical conditions were not suitable (participant 6B). --- Subtheme: Eager to give birth to new life again The reproductive responsibility of women in traditional thinking and the second child policy have effectively aggravated the parents' desire to conceive again. Ten parents expressed their desire to conceive a healthy new life again. Especially when the two-child policy was liberalized and others had two children, maybe only having a healthy baby could truly eliminate the impact of this event [TOPFA] (participant 11B). Theme 4: The urgent need for social support Subtheme: Support from medical staff When parents knew that the fetus was abnormal, they were eager to know the advantages and disadvantages of continuing pregnancy and induced labour, the root cause of malformation, what physical recovery would look like after induced labour and other information from medical staff. In addition, they were eager for understanding and care from medical staff. The views of medical staff play a leading role in our choice (participant 13 A). My wife was still thinking that the child may be good until now, so I wanted the doctor to tell her that the child is definitely bad and make up a white lie to alleviate her guilt, which would also be a balm to my heart (participant 5B). --- Subtheme: Family support In the face of TOPFA, parents need the understanding and support of their families, especially the support of their spouses. My husband's company is the most important thing. If my husband is by my side during childbirth, my heart might be stronger (participant 6A). Spouses who played an important role of supporters would also provide effective support to their wives based on their psychological needs and personality traits. I told my wife from all aspects that the decision to induce labor was the right one... Now my wife was still thinking that the baby might be normal, so I wanted the doctor to tell her that the child was definitely abnormal, fabricate a white lie to alleviate her inner guilt (participant 5B). I took care of my wife, accompanied her, and took her out for relaxation. I would do my best to do well. There are many kinds of support, and I should support her effectively according to her personality characteristics (participant 10B). --- Subtheme: Peer support Peer support refers to making patients with similar diseases, physical conditions or experiences share information, emotions, ideas or behavioural skills through diversified forms. 34 Five women believed that the exchange of experiences and emotional resonance with peers was very beneficial to their psychological recovery. I joined a peer group. They also experienced it [TOPFA]. It was convenient to talk with them. They knew what I wanted to know. Seeing the photos of their babies [healthy babies born later] was very lively and lovely, which gave me a lot of positive energy. They were also more compassionate and gave me suggestions (participant 14A). Peer support was very helpful to my wife. The most important thing was confidence and informational help. They also went through TOPFA step by step and sorted out a set of processes, which was of great significance (participant 14B). Six women were eager to communicate with peers to obtain information and emotional support. I also wanted to communicate with my peers... I wanted to ask them how they came out in the end. We could prepare for the second child together (participant 11A). --- DISCUSSION Fetal abnormalities are serious traumatic events for parents, 35 which can make them face psychological crisis and complex psychological problems. The findings of the present study demonstrate that parents have experienced the following mental processes: denial and verification of fetal abnormalities, abandoning the fetus with reluctance and struggle, an acknowledgement of reality, intense grief, a confrontation with the shackles of traditional Open access thinking, rumination, concerns after deciding to induce labour and a desire for social support. To our knowledge, this is the first study in China to include both fathers and mothers in the exploration of parents' experiences and need for social support following TOPFA. Consistent with previous studies on parents' experiences of TOPFA, 6 23 36 when receiving the cruel fact of fetal abnormality, parents often struggle with doubts, selfblame, reluctance and sadness. We also found that most fathers in our study chose to hide their real emotion to support their spouses and reported more rationalising than mothers, which is similar to prior findings. 22 23 37 This may be because the social role given to fathers as supporters of the mother 38 leads to the suppression of grief, anxiety and stress in fathers, potentially increasing the risk of chronic psychological problems, so that the fathers will experience more anxiety in the subsequent pregnancy. 30 In addition, Chinese parents are bound by traditional ideas such as 'family succession','son preference' and'superstition', which often brought pressure to parents and increased the stigma they felt. However, the surviving children could bring spiritual comfort and support to parents, which is an important predictor of parental grief intensity. 39 Unlike previous studies reported that fathers felt overlooked and marginalised at hospitals while their partners were receiving treatment, 23 36 we found that fathers, like mothers, mainly focused on maternal physical and psychological recovery and the impacts on the subsequent pregnancy, they did not realise that paternal psychological trauma also needs attention. TOPFA can increase the psychological stress of parents, especially mothers, in subsequent pregnancy, which is consistent with previous studies. 40 41 Parents who experience fetal abnormalities have a high degree of anxiety and fear in subsequent pregnancies, and especially worry about the risk of recurrence. 41 Under the influence of the Chinese two-child policy, parents' desire to conceive again was more urgent, but they were still full of doubts about the causes of fetal abnormalities and what to do in the subsequent pregnancy. Therefore, medical staff should establish a longterm follow-up mechanism to continue to pay attention to parents' physical and mental recovery 23 42 and provide them with the necessary knowledge and psychological support, such as information about abnormalities, childbearing and pregnancy examination, to promote them to conceive a healthy new life. Many studies have demonstrated the role of social support in alleviating negative emotions in parents experiencing TOPFA, including anxiety, depression and PTSD. 28 43 44 Parents in this study also showed a strong need for social support. When first informed about the fetal abnormality, parents were full of confusion about the causes and worried about the impacts on any subsequent pregnancy. They urgently need professional guidance and suggestions from medical staff. At this period, the information support from medical staff constitutes the main part of the support system, which can help parents make good decisions; these findings were in line with a prior study. 22 Therefore, medical staff should take the initiative to understand the thoughts and needs of parents with fetal abnormalities and patiently provide them with complete, adequate and appropriate informational support to help them establish a scientific understanding of fetal abnormalities and relieve their confusion and feelings of guilt to facilitate their process of grieving. 45 46 Echoing the experiences of parents in the broader pregnancy loss literature, 47 48 parents in this study also need psychological counselling and empathetic care from medical staff following perinatal loss. Therefore, healthcare professionals should offer parents' bereavement care with empathy and cultural sensitivity. 22 During induced labour in hospitals, parents need care and emotional support from their families, especially the support of their spouses; this finding is consistent with prior studies. 49 Fathers will also actively take on the role of supporters, providing various forms of support based on their wives' psychological needs and personality traits, including appraisal support for recognising their wives' induced abortion decisions and instrumental support such as daily care, company and helping wives shift their attention. However, fathers, like mothers, also need support from their families. 50 Therefore, medical staff should guide family members to treat fetal abnormalities scientifically, provide help and care for parents in life, and give emotional understanding and assistance to make them feel the warmth from their families. Peer support has long been considered an essential component of a supportive network for people facing Open access adversity. 51 Parents in this study also showed a strong need for peer support, and they hoped to obtain informational support and emotional resonance through the experience of sharing and communicating with their peers. Healthcare providers should establish a peer support platform according to the needs of parents, provide a platform for parents to exchange experiences, help them eliminate loneliness and helplessness, release inner pressure and transmit positive energy. Based on the above parents' experiences and social support need, we preliminarily constructed the following social support model (see figure 2) and verified the effects of family support and peer support; see our previous studies for details. 28 52 In the future, we will further verify the effect of the social support model. The study has a few limitations. First, there may be subtle differences in the psychology of parents who induce labour due to stillbirth and fetal malformation, but our population did not distinguish between stillbirth and fetal malformation. Second, we found that most fathers tended to avoid interviews, so we did not collect other demographic information except age to improve fathers' participation. Furthermore, some bereaved parents, especially fathers, refused to participate in the study, and based on the principle of information saturation, we ultimately included 15 mothers and 12 fathers, this small sample size would decrease transferability of our findings. However, the psychology of parents who refused to participate in the study is also well worth attention, and future research should explore essential reasons for their refusal to participate in the study and provide targeted psychological support. Finally, TOPFA will not only bring psychological trauma to parents but also have a negative impact on the psychology of the whole family; therefore, it is necessary to further explore the influence of TOPFA on other family members (such as grandparents, surviving children, etc) and the interaction of emotions among family members. --- CONCLUSION This study contributes to the limited body of international studies on parental experiences and social support need following TOPFA. The findings suggest that TOPFA is an extremely painful experience for parents, characterised by psychological reactions of denial and the need to verify fetal abnormalities, abandoning the fetus with reluctance and struggle, intense sadness, an acknowledgement of reality, a confrontation with the shackles of traditional thinking, rumination, concerns surrounding induced labour and a need for social support. Paternal psychological reactions were often neglected by healthcare providers and the fathers, themselves, because they often played the role of supporters, which requires more attention. Based on the above parents' experiences and social support need, medical staff should provide tailored information and emotional support for bereaved parents and guide family members in support. Finally, medical staff should establish a peer support platform to provide peer support for parents. --- Data availability statement Data are available upon reasonable request. --- Competing interests None declared. Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research. --- Patient consent for publication Not applicable. Ethics approval This study was approved by the ethics committee of the Women's Hospital School of Medicine, Zhejiang University (IRB Number 20150071). All participants were voluntary participation in the study and gave written informed consent prior to participation. All methods were performed in accordance with the relevant guidelines and regulations. Provenance and peer review Not commissioned; externally peer reviewed.

Objective The aim of this study was to explore the experiences and need for social support of Chinese parents after termination of pregnancy for fetal anomalies. Design A qualitative study using semistructured, indepth interviews combined with observations. Data were analysed by Claizzi's phenomenological procedure. Setting A large, tertiary obstetrics and gynaecology hospital in China. Participants Using purposive sampling approach, we interviewed 12 couples and three additional women (whose spouses were not present). Results Four themes were identified from the experiences of parents: the shock of facing reality, concerns surrounding termination of pregnancy, the embarrassment of the two-child policy and the urgent need for social support. Conclusion Parents experienced complicated and intense emotional reactions, had concerns surrounding the termination of pregnancy and an urgent need for social support. Paternal psychological reactions were often neglected by healthcare providers and the fathers, themselves. These findings suggest that both mothers and fathers should receive appropriate support from family, medical staff and peers to promote their physical and psychological rehabilitation.

Introduction Child early marriage is defined as "any marriage carried out below the age of 18 years, before the girl is physically, physiologically, and psychologically ready to shoulder the responsibilities of marriage and childbearing. It therefore has major consequences for public health, national security, social development, human rights, economic development and gender equality 1. Similarly, the age at first marriage is defined as the age at which the respondents began living with her/his first partner 2. The extent of early marriage varies between countries and regions. The highest rates are reported in South Asia and sub-Saharan Africa, where 44% and 39% of girls, respectively, were married before the age of 18 years. Data from 33 countries showed that trends in marriage indicate limited change since the International Conference on Population and Development 3. More over 27 per cent were in East Africa and 20 percent in Northern and Southern Africa 4. As the current estimates of the Convention on the Elimination of all forms of discrimination against women showed, approximately 82 million girls in the world between 10-17 years will be married before they reach 18 years; and of the 331 million girls aged 10-19 in developing countries,163 million will be married before they are 20 years 5. A study in Gojjam and South Wollo zones of Amhara region indicated that early marriage is highly prevalent. The prevalence is higher for women than men. About 49% of women were married before age 15 and about 83% were married before age 18 years 6. By 2015, the prevalence of female early marriage was 76.7% in Amhara region, North Ethiopia. Females who did not know the legal marital age were 12 times more likely to practice early marriage compared to those who knew the legal marital age 7. A study by marital relations and Intimate Partner Violence in Ethiopia showed 70% of respondents had married before age 15 years and 30% had married at ages 15-17 years. Among those who married before 15 years of age, 82.2 % were from rural residences. This study also showed that rural residence added more risk for early marriage. Rural residence was associated with nearly a threefold elevation in the odds of marriage at ages 15-17 years 8. A study by Pathfinder International in Amhara region showed that the prevalence of girls' early marriage was 81.8%. Moreover, about 44% of urban and 53% of rural ever-married women were married between 12-15 years of age. The proportion marrying between age 16 and 17 years was 14.5% in urban and 15.5% in rural areas 9. In Ethiopia, female child early marriage is seen as a way to improve the economic status of the family, to strengthen ties between families, to ensure that girls are virgins when they marry, and to avoid the possibility of a girl reaching an age where she is no longer desirable as a wife. The practice of female child early marriage is now understood to have very harmful effects on the health, psychological, physiological and socio-economic well-being of young girls (as well as for the newborns). None the less, this knowledge is not broadly shared across most of the population 10. In Amhara region of Ethiopia, about 80% of girls are married before they are 18 years old, and the most common age for a girl to marry is 12 years old. Child marriage is rooted in religious and cultural tra<unk>ditions based on protecting a girl's honor since sex before marriage is seen as an extremely shameful act. A girl's worth is therefore based on her virginity and her role of being a wife and mother 11. All relevant laws of Ethiopia established a legal minimum age at marriage of 18 years for boys and girls. In fact, much of the education on early marriage prevention clearly indicates that the legal minimum age for marriage is 18 years for both girls and boys. However, a study in Amhara region shows clearly that the general public's definition of early marriage for girls uses a much lower cut-off than the legal definition. This indicates that the cut-off age for defining early marriage for female adolescents was often ignored with many marriages occurring before age 15.3 years 12. The prevalence of early marriage in the Amhara region is still high. The findings of this study will make an input for the policy makers and planners in the area as well as the regional government to respond to early marriage at all levels of governmental and non-governmental sectors. Furthermore, it will help as an initiative for further investigation and intervention in the area regarding early marriage for those who will be interested in studying its consequences and related issues. Therefore, the goal of this study was to determine the prevalence and factors associated with early marriage in the Amhara region, North Ethiopia. --- Methods --- Study design and sampling The dataset used in this study was obtained from the Ethiopia Demographic and Health Survey (EDHS) conducted by Central Statistics Agency (CSA) in 2016. The survey utilized a multistage cluster sample and was designed to obtain and provide information on the basic indicators of the health and demographic variables. The study design was cross-sectional, in which the data on the independent and outcome variables is collected at the same point in time. --- Target population The target population was all female community members in selected cities of Amhara Region who practiced marriage. --- Dependent variable Female child Marriage (classified as either below 18 or 18 and above). --- Independent variables Age, Education, Religion, income, ethnicity, literacy level and residence area(urban/rural). --- Data analysis The data were extracted, edited, recorded, and analyzed by using SPSS version 23 for Windows. Both bivariate and multivariable binary logistic regressions were done. Bivariate logistic regression was performed and variables with a p-value less than 0.25 were transported into multivariable binary logistic regression analysis to identify the determinants of early marriage among female children. Finally, variables with p-values <unk>0.05 in the multivariable logistic regression modewere taken as statistically significant. The association was reported as adjusted odds ratios with 95 % confidence intervals. --- Binary logistic regression model The binary logistic regression model is used when the dependent variable is a dichotomous and the independent variables are of any kind. This model is mathematically flexible and easily used distribution and it requires fewer assumptions 13. It is presented as follow: llllllllll pp 1 -pp = <unk> 0 + <unk> 1 XX 1 + <unk> 2 XX 2 + <unk> + <unk> kk XX kk (1 ) where p is the probability of being practice early marriage, is a set of independent variables, and <unk> = (<unk>0, <unk>1... <unk>k)T is a vector of unknown coefficients. The quantity to the left of the equal sign is the log of the odds of early marriage in the binary logistic regression. The adequacy of the model was checked by using Hosmer-Lemeshow goodness of fit test. --- Results A total of 1120 married women were selected for this study. The median age at first marriage for women was 15 years and about 48.6% of women got married before their 18th birth day in the selected study area. As shown in Table 1, early marriage was significantly associated with women's educational level (p-value=0.000 <unk> 0.25). The highest proportion of early marriage was observed in the age group 15-19 years (60.6%). The educational status and early marriage cross tabulation also reveals that the smallest proportion of early marriage (21.6%) was observed for those hving higher education while the highest proportion of early marriage was observed for those having no education (51.1%).The proportion of early marriage was 30.3% for women that lived in urban and 52.2% for women who lived in rural. It was also observed that among women that took part in the study, the percentage of early marriage for Orthodox, Protestant, Muslim, and others religion followers were 48.2%, 53.8%, 50%, and 50.0%, respectively (Table 1). The literacy level of women was significant for early marriage in the study area. Women who read only parts of the sentence were 0.49 times less likely to get early marriage than who could not read at all. Women who read the whole sentence were 0.36(AOR=0.36, 95%CI: 0.24, 0.52) times less likely to get early marriage than who cannot read at all. The income level of the family was another significant factor for early marriage in Amhara region. Women who lived in a family with medium monthly income were 0.81(AOR=0.81, 95% CI: 0.60, 1.11) less likely to get early marriage than who lived in the poor family. Women who lived in a family with rich monthly income were 0.57 less likely to get early marriage than who lived in the poor family. The odds of early marriage practice were 2.04(AOR=2.04, 95% CI: 1.88, 2.45) times higher among rural residents compared to urban (Table 2). --- Binary logistic regression --- Assessment of goodness of fit of the model The omnibus tests are used to measure how well the model performs. The chi-square tests are used to measure the difference between the initial model and the regression model in terms of a number of correctly classified subjects or it is the change in the -2log-likelihood from the previous step. Since the omnibus test was significant, the model in the final step was consid-ered to be appropriate (Table 3).If the p-value of the Hosmer-Lemeshow Goodness of fit test statistic is greater than <unk>=0.05, we fail to reject the null hypothesis that there is no difference between observed and predicted values. The value of Hosmer-Lemeshow statistic had a chi-square value of 2.264 and a p-value of 0.645 indicating that the model had a good fit. This shows that there was no significant difference between the observed and predicted model values and hence the model fits the data well (Table 3). --- Discussion This study examined the prevalence of early marriage and the related factors in the Amhara Region of Ethiopia. The prevalence of early marriage practice was 48.57% in the study area. This finding was lower than three previous studies in Amhara region, Ethiopia which were 83%, 81.85 and 76.7% respectively 6,7,9. Education level was a significant factor in early marriage. The percentage of early marriage was highest among women with no education. Females with primary education were less likely to get married before reaching the age of age 18 than those with no education. The higher one's educational attainment, the more knowledge females get and understand the best marriage age, and the effect of having an early marriage. This result was supported by the previous study 12. The odds of early marriage practice were high for those women who did not know the legal marital age compared to women who know the legal marital age. The odds of early marriage practice were higher among rural residents compared to urban. This finding was similar to previous studies in Amhara region which showed that rural residents were more likely to practice early marriage than urban residents 7. The income level was another significant factor for early marriage in Amhara region. Families with medium monthly income were 0.81 times less likely to practice early marriage compared to those having poor monthly income and similarly families with rich income were 0.57 times less likely to get early marriage than poor families. This result was consistent with the previous study which indicates that families with lower monthly income are more likely to practice early marriagthan families with high monthly income 7,14. This study also revealed that, the odds of early marriage practice were higher among rural residents compared to urban. This finding is in line with a previous study in Ethiopia 7. --- Conclusions and recommendations This study established the prevalence and factors as-sociated with early marriage. The authors found that the prevalence of early marriage was still high in the Amhara region, Ethiopia. The study identified education level, family monthly income, literacy, residence area and knowledge of legal marital age as the main factors associated with early marriage practice. Since the education level of women was a significant factor for early marriage, parents and the Ministry of Education should emphasize education of women. In the communities' awareness, regarding the legal marital age, has to be developed. --- Authors' contribution All the authors participated in proposal development, data extracting, analysis and manuscript writing. --- Competing interests The authors declare that they have no competing interests.

Background: Early Marriage is one of the global problems that undermine the personal development and the rights of women seriously. It is delicate among the developing countries such as Ethiopia. It has major consequences for public health, national security, social development, human rights, economic development, and gender equality. Methods: The analyzed data were obtained from the 2016 EDHS and 1120 samples were considered in this analysis. Both bivariate and multivariable binary logistic regression model were used to identify the determinants of early marriage practice. Results: The prevalence of early marriage practice was 48.57% in the study area. The odds of early marriage practice were 2.04(AOR=2.04, 95% CI: 1.88, 2.45) times higher among rural residents compared to urban. The odds of early marriage practice was 0.94(AOR=0.94, 95%CI: 0.57, 1.98) times lower among women who had primary education compared to uneducated women. Those who did not know the legal marital age were 1.61(AOR=1.61, 95%CI: 1.26, 2.07) times more likely to practice early marriage compared to parents who knew the legal marital age. Conclusion: Education level, family monthly income, residence, literacy level and knowledge of legal marital were significant determinants of early marriage practice.

Introduction Sam Ratulangi University has a mission, namely to be at the forefront in carrying out the Tridharma of Higher Education and as a Center for Innovation in Science, Technology and Arts and Culture to Improve the Level and Quality of Community Life, which is described in the words IMANKU (Innovative, Partner, Applicative, Normative, Creative and Excellent). Excellence and Competitiveness in Entrepreneurship is a description of the Mission of Sam Ratulangi University. The Student Creativity Program (PKM) is one of the efforts of the Directorate General for Strengthening Research and Development, Ministry of Research, Technology and Higher Education to lead students to reach a level of enlightenment in creativity and innovation based on mastery of science and technology and high faith. The Social Humanities Research Student Creativity Program (PKM-RSH) is a program that criticizes social humanities phenomena that exist in society. PKM-RSH focuses on elements of creativity and innovation that are useful and provide answers to problems raised by combining the social and humanities fields (Nur Fadhila, 2022). The quality of higher education graduates does not only depend on academic abilities (hard skills) but is also required to have supporting abilities (soft skills) such as thinking skills, management, communication, leadership and working in a team. Lack of soft skills can cause a decline in the quality of graduates. Students who pass the PKM-RSH selection are implementing the Independent Learning Program -Independent Campus (MBKM) and efforts to achieve the Main Performance Indicators (IKU) for universities and study programs (Brahma Wicaksono, 2022). The phenomenological theory of Alfred Schutz puts forward two motifs, namely the "Cause" motif or Because of Motive and the "Purpose" motif or In Order To Motive. Motive "Cause" is what motivates someone to carry out certain actions. Meanwhile, the "goal" motive is the goal that someone who carries out a certain action wants to achieve (Warouw Desie, 2022). This research will reveal the motives for participating in the Social Humanities Research Student Creativity Program (PKM-RSH) at Sam Ratulangi University, which is carried out by students and the student's accompanying lecturers. Data from the Unsrat Dashboard shows that the number of active undergraduate students at Sam Ratulangi University for the 2000/2021 academic year is 535 people, 882 for the 2021/2023 academic year, 1,405 students for the 2022/2023 academic year, spread across 11 faculties. For this reason, it is interesting to research what the students' motives are for taking part in the Social Humanities Research PKM because the numbers are relatively small compared to the total number of students. --- Methods The method used is a qualitative method with a descriptive approach, which is aimed at: (1) collecting detailed actual information that describes existing symptoms, (2) identifying problems or examining applicable conditions and practices, (3) making comparisons or evaluation, (4) determining what other people have done in facing the same problem and learning from their experiences to determine decision plans in the future (Soegiono, 2012). These informants were selected by Purposive and Snowball Sampling. Informants selected purposively are people who are determined because they are deemed capable of providing information and are able to appoint other people as informants who can provide more in-depth information. Meanwhile, informants based on Snowball Sampling are informants who are selected based on clues in the selected field and are rolled out more and more day by day and finish until the saturation point, where the informant has provided the same information as the previous informant (Moleong, 2012). Data collection was carried out using methods commonly used qualitative approaches, namely observation (participant), in-depth interviews (in depth interviews), and document study (18). In qualitative research, data analysis is carried out from the beginning and throughout the research process. In this research, qualitative data analysis will be used with an interactive model developed by Miles and Huberman (2012) --- Results and Discussion Sam Ratulangi University is committed to excellence in the learning process, research and community service as an integral part of the process of forming cross-disciplinary leadership character. Everyone must have the same access to education regardless of their social or cultural background. Our vision and mission reflect the moral responsibility to "humanize" others as aspired by educational figure G.S.J. Samuel Ratulangi in the motto "Si tou timou tumou tou". Sam Ratulangi University has a vision, namely to jointly organize Sam Ratulang University into a superior and cultured university. Its mission is to be at the forefront in carrying out the Tridharma of Higher Education and as a Center for Innovation in Science, Technology and Arts and Culture to Improve the Standard and Quality of Community Life, which is described in the words IMANKU (Innovative, Partner, Applicative, Normative, Creative and Excellent). Students as intellectual actors are expected to be able to develop science and technology that is potential, efficient, and useful in social life. Students are expected to be able to develop research based on observations of social phenomena in the surrounding community, and be able to understand the meaning of research, its objectives and benefits. With this provision, students will be able to take a creative and innovative scientific approach to uncover a phenomenon, discover novelty or prove a hypothesis in the field of social humanities. The Social Humanities Research Student Creativity Program (PKM-RSH) is an activity that provides a forum for student creativity and innovation in the field of research in accordance with scientific principles. In this PKM-RSH, students are expected to be able to criticize social and humanities phenomena that exist in society with a scientific approach, use appropriate methods in searching for information, analyze information using theory, and provide answers to problems that arise from these phenomena. In this way, research results can be published and provide benefits to interested parties. At PKM-RSH, students are expected to be able to explore ideas and develop innovative creative discoveries based on research and development so that they are able to excel in national events. This is the aim of one of the Independent Learning -Independent Campus (MBKM) programs, namely independent studies/projects. Thus, PKM-RSH becomes a form of independent study/project in the MBKM program. PKM-RSH can be a substitute for courses that must be taken or a complement that can be included in the Diploma Companion Certificate (SKPI). Equivalence of Semester Credit Units (credits) from PKM-RSH is calculated based on the implementation time as well as the student's proven contribution and role in activities under the coordination of the accompanying lecturer. The range of credits that students can obtain in all PKM-RSH activities is a minimum of 6 to 10 credits in accordance with the provisions in their respective study programs. In line with Independent Campus learning, PKM activities are expected to provide opportunities and challenges in developing creativity, innovation and capacity, as well as independence in seeking and finding knowledge or solutions through problems and dynamics that exist in society. An explanation of the credit conversion recommendations can be seen in the PKM General Guidelines Book. According to Minister of Education and Culture Regulation No. 3 of 2020 Article 15 paragraph 1, forms of learning activities that are in accordance with PKM-RSH are research and independent studies/projects as seen in the following picture: Copyright <unk> 2023, Journal La Edusci, Under the license CC BY-SA 4.0 The Student Creativity Program -Social Humanities Research (PKM-RSH) is an activity that provides a forum for student creativity and innovation in the field of research in accordance with scientific principles. The general meaning of PKM-RSH is to reveal facts or phenomena through a scientific approach. Meanwhile, the Special Meaning is Innovative in discovering a novelty about a phenomenon or proving a hypothesis in either one scientific discipline or multidisciplinary so as to produce a contribution in the form of information for the progress of science and technology as well as in order to overcome existing problems in society. The aim of PKM-RSH is to foster interest and research skills, understanding of research methods and methods of data analysis. Producing quality research that has the potential to be published in scientific journals and has the opportunity to produce policies that are beneficial to both the academic community and the wider community. PKM-RSH is a combination of social and humanities fields which has research objects on social phenomena and human behavior that can be found in social life. The social sector focuses more on social phenomena of interaction in social life such as economics, psychology, social, education, management and politics. The field of humanities focuses more on basic aspects of behavior in people's lives, such as the development of culture, art, philosophy, customs, history, beliefs or religion, law and values. The combination of social and humanities uses research paradigms in the form of cause-and-effect relationships, conclusive descriptive, phenomenology, hermeneutics, postcolonial, positivistic, historical, structural, development, and so on according to each field of science. Quality research can be seen through the quality of several underlying aspects, namely: intellectual challenge, problem focus, approach methods, theories used, data quality, and outcome impact. Intellectual challenges can be seen from the "state of the art" regarding the topics raised, the use of logic, and the research platforms used. The focus of the problem can be seen from the sharpness in choosing the scope of research, the sharpness in selecting unique problems and the suitability of the virtual or digital approach used. The theory used must be relevant to the problem focus and used in data analysis to answer research problems. The approach method can be measured by the novelty and procedures and completeness of the system used to collect information or data and its analysis techniques. The quality of the data or information collected can be measured from the adequacy and reliability of the data or information collected, including the data sources used. The output impact can be seen from the quality of the output in a logical and systematic manner. Data that can be used in social humanities research is grouped into primary and secondary data. Primary data can be obtained from respondents, participants, sources, artifacts, society (collective memory, myths, folklore, norms, and so on), using questionnaire or survey techniques, interviews, observations, active participation, and experiments. Meanwhile, secondary data can be sourced from archives, literature, reports (data from BPS, companies, etc.), digital data (social media or big data), and written laws or regulations. In 2021 there will be 2 (two) groups that pass the PKM-RSH from the Faculty of Social and Political Sciences, under the guidance of lecturer Dr. Daud Markus Liando, SIP, MSI with the PKM-RSH Team leader being Mineshia Lesawengan. The second team was guided by Dr. Leviane J.H.Lotulung, S.Sos, M.I.Kom, as chairman of the PKM-RSH TEAM is Fabio Y. Lasut. Every year, Sam Ratulangi University opens opportunities for students to take part in the Student Creativity Program (PKM), including Social Humanities Research (PKM-RSH), which is an activity and forum for student creativity and innovation in the field of research in accordance with scientific principles. In this PKM-RSH, students are expected to be able to criticize social and humanities phenomena that exist in society with a scientific approach, use appropriate methods in searching for information, analyze information using theory, and provide answers to problems that arise from these phenomena. In this way, research results can be published and provide benefits to interested parties. At PKM-RSH, students are expected to be able to explore ideas and develop innovative creative discoveries based on research and development so that they are able to excel in national events. This is the aim of one of the Independent Learning -Independent Campus (MBKM) programs, namely independent studies/projects. In 2021 at the Faculty of Social and Political Sciences, Sam Ratulangi University, there will be 2 (two) groups that will pass the PKM-RSH funding. In this socialization, it was explained about the importance of proposal format and writing which must comply with PKM guidelines. Because in making PKM proposals there are still many that do not comply with the guidelines, so many proposals are rejected. Creativity is also something that is valued and important because it needs to be emphasized in the contents of the proposal. Apart from creativity, the local wisdom aspect of PKM products can also be an added value in researching PKM proposals. In this socialization, students received information and explanations about the Student Creativity Program (PKM), especially Social Humanities Research (RSH), which was attended by lecturers and students. In this socialization, students received information on the techniques for writing PKM proposals, as well as important things that must be done so that the proposal can pass, whether it concerns determining the title, creative aspects, should contain elements of local wisdom, etc. This motivates students to make proposals. Students want to take part in PKM-RSH but not many continue to make proposals. Therefore, the leadership encouraged students to continue to make proposals and look for competent supervisors to guide the PKM-RSH team. Student enthusiasm is quite high, but the problem is that many of them do not maintain that enthusiasm until they create and upload quality PKM-RSH proposals. Even though proposal coaching has been provided to perfect proposal writing. To make it easier for students to find out about the --- Encouragement from accompanying lecturers Students are motivated to take part in PKM-RSH because they are continuously encouraged by accompanying lecturers or supervisors. The accompanying lecturer or supervisor continues to accompany the PKM student team and directs and provides inspiration to the students in determining the title and preparing the proposal until its completion. Good collaboration between supervisors or assistants and students produces quality PKM proposals so that they can pass the National Student Science Week (PIMNAS). This PKM-RSH culminates in PIMNAS, so the accompanying lecturer encourages the student team to make good, quality proposals while we continue to discuss to get a proposal that can pass to PIMNAS. The accompanying lecturer encouraged the student team to read the guidebook for preparing the PKM-RSH Proposal, how to write it, understand the completeness of the required files, follow the prescribed writing format, etc., because there are differences in substance between PKM-RSH and other PKM, as well as the forms. After understanding other forms of writing, the student team knows the rules of what can be written and what cannot be written, for example the number of pages and writing sources. Adhere to the writing format and systematics specified in the guidebook, so that the proposal you make can pass the selection. --- As a Challenge to Develop Ideas, Creativity and Innovation Through Research In participating in PKM-RSH, the student and lecturer team was challenged to come up with a title that was interesting, innovative and contained elements of local wisdom. Teams of students and lecturers are challenged to think creatively in expressing ideas in proposals that have scientific quality and make a positive contribution to society. PKM-RSH is also a forum for student creativity and innovation in the field of research in accordance with scientific principles. The student team is expected to be able to criticize social and humanities phenomena that exist in society using a scientific approach, using appropriate methods in searching for information, analyzing information using theory, and providing answers to problems that arise from these phenomena. In this way, research results can be published and provide benefits to interested parties. Through PKM-RSH, students are expected to explore ideas and develop creative and innovative discoveries based on research and development so that they are able to excel in national events. This is the aim of one of the Independent Learning -Independent Campus (MbKM) programs, namely independent studies or projects. --- Opportunity to Work as a Team Students are also motivated to carry out PKM-RSH because they get the opportunity to work as a team with other students. Students not only do research alone but are given the opportunity to work with other teammates. Working as a team creates many thoughts that team members can discuss. The aim of conducting research in the form of a team of students and lecturers is to ensure collaboration. Working as a team ensures that PKM-RSH proposals are completed quickly, and also trains students to exchange ideas and be skilled in communicating or conveying ideas to other team members. When working as a team, togetherness gradually grows with fellow group members, each of whom understands the character of the other members. If there are problems that arise in preparing and completing PKM-RSH, then all teams can provide ideas and solutions so that problems can be resolved Students are motivated to work with student teams in preparing and completing PKM-RSH, because students can get to know each other's personalities and train themselves to respect differences of opinion. Team work also allows students to convey ideas and ideas. And if we pass PIMNAS, it will be an advantage for all of us, not just one person. --- Qualified for the National Student Science Week (PIMNAS) The student's motive for taking part in PKM-RSH is also because they want to get through to the National Student Science Week (PIMNAS), because the PKM program is a prestigious student program which has become an annual program, and getting through to PIMNAS is a proud achievement. Students and lecturers who take part in PKM-RSH are motivated to pass the proposal so that they can continue to PIMNAS. Many other competitors were also good, but the motivation to bring the name of the Department, Faculty and University to the PIMNAS level made the team enthusiastic about participating in PKM-RSH. It would be very encouraging if the PKM-RSH proposal made it to PIMNAS. The student team realized that there were many competitors who were also good, so we had to do our best in making this proposal. In Order To Motif (Purpose) students take part in the Student Creativity Program -Social Humanities Research (PKM-RSH) --- Graduated Without a Thesis Students are motivated to take part in PKM-RSH also because they finish studying to become undergraduates without a thesis if their PKM-RSH proposal passes to PIMNAS. PKM-RSH can be a substitute for courses that must be taken or supplements that can be included in the Companion Diploma Certificate (SKPI). The equivalent of Semester Credit Units (credits) from PKM-RSH is calculated based on the implementation time as well as the student's proven contribution and role in activities under the coordination of the accompanying lecturer. The range of credits that students can obtain in all PKM-RSH activities is a minimum of 6 to 10 credits in accordance with the provisions in their respective study programs. Two students from the Faculty of Social and Political Sciences, Unsrat, who are part of the Student Creativity Program -Social Humanities Research team, finally graduated from college without going through the thesis stage because they succeeded in passing the National Student Science Week (PIMNAS). This is an award by the campus based on the chancellor's decision letter. This is also because students have created scientific work such as in the process of writing a thesis, and this scientific work has been assessed by professors from several universities in Indonesia. Students who only passed the PKM-RSH proposal until the university stage were also not disappointed because they received rewards from accompanying lecturers by giving A grades in seminar courses and certain courses. --- Contribution to Universities and Faculties The --- Difficult to Find Assistant Lecturers Initially the student team had difficulty finding PKM-RSH supervisors, because the supervisors had to be competent in what they were going to research. Also look for lecturers who have the opportunity and time to guide student teams, remembering that lecturers also have the obligation to teach, research and carry out community service with a fairly tight time intensity. There are several lecturers who have been contacted, but are not yet willing to accompany and guide because there are other tasks that must be completed. Initially, students had difficulty finding the right accompanying lecturer for the PKM-RSH team, because the lecturers had busy activities so that mentoring time with the team was quite With guidance from WD3 and the Head of Department, finally the PKM-RSH student team got a companion lecturer who guided the team well so that they produced a good proposal. --- Limited Time to Discuss with the Team and Lecturers There are obstacles that cause the process of preparing PKM-RSH proposals between fellow student teams or with the supervisor. Students are sometimes faced with busy lectures and assignments from lecturers that must be completed on time, also during the Mid-Semester Examination (UTS) and Final Semester Examination (UAS) which require concentration in studying. So the time to meet and discuss with the PKM-RSH Team is limited. In the process of preparing a proposal until the presentation stage, there are obstacles to meeting for discussion, because sometimes it is difficult to match the free time of fellow team members. Especially during UTS and UAS, where we have to concentrate on exams and completing assignments from lecturers, so sometimes meetings with the team to discuss PKM-RSH proposals are delayed. Likewise, meetings with PKM-RSH accompanying lecturers are very busy, because apart from having additional duties or positions, they also have activities outside campus, so that sometimes it becomes an obstacle in holding discussions. It is very clear that the accompanying lecturer has additional duties as a study program coordinator, so sometimes when the student team needs to discuss it turns out he has other activities, so the student team has to wait and adjust to the lecturer's free time. --- Confused about Finding Phenomena and Determining Titles At first the PKM-RSH team had difficulty finding the phenomenon and determining the title. That was the role of the accompanying lecturer to direct the student team to find the phenomenon and determine the title correctly. Students receive very useful input, which stimulates students' thinking so that creative ideas emerge to develop what will be researched. Searching and reading a lot of the latest information and looking at the surrounding environment can help identify social phenomena and determine research titles. From reading you can get ideas and thoughts to determine a topic. Seeing phenomena, problems or social events can also motivate you to determine the right title. --- Difficult to Find Friends to Build a Team Preparing a PKM-RSH Proposal requires a solid team, because there is interaction between one member and another, as well as with accompanying lecturers. Therefore, you must look for teammates who have the same vision and way of working so that there are no debates and you can work cohesively. It is not easy to find friends to build a team, because students come from different backgrounds. (2) In Order To Motive (Goal): -Graduate without a thesis -Contribution to universities and faculties -Find solutions to social phenomena that are beneficial to society. (3) Obstacles encountered: -Difficult to find accompanying lecturers -Limited time to discuss with the team and lecturers -Confused about finding phenomena and determining titles -Difficult to find friends to build a team. As a suggestion in this research, the Student Creativity Program must be disseminated continuously to students so that more students take part in PKM because the results can support the achievement of higher education and faculty IKU. The student team and accompanying lecturers who have passed also provide motivation and assistance to other students who intend to take part in PKM-RSH.

The role of students in supporting creativity programs, especially social and humanities research at Sam Ratulangi University, is very important, because apart from attending lectures, they will become human resources who have academic knowledge, management skills, communication skills, but Students are also expected to have the skills and creativity to carry out research, in order to become productive, superior, competitive, adaptive, flexible, productive and competitive graduates during the industrial revolution 4.0. As well as the achievement of higher education KPI.

individuals within parents' networks influence the way in which they manage their children's asthma medications, and factors driving the development of these networks. Methods: A qualitative research methodology utilizing semi-structured interviews with parents of children with asthma was employed to fulfil the objectives of this study. Results: Twenty-six face-to-face interviews with parents of children with asthma were conducted, recorded, and transcribed. Transcriptions were independently coded for concepts and themes by the research team. Asthma medications was a dominant theme identified, and revealed that parents actively sought advice and support from a series of complex and multidimensional relationships with people and resources in their health network. These not only included health care professionals (HCPs) but also personal connections, lay individuals, and resources. The composition and development of these asthma networks occurred over time and were determined by several key factors: satisfaction with their HCP provider; need for information; convenience; trust and support; self-confidence in management; and parents' perceptions of their children's asthma severity. Conclusions: By exploring parents' asthma network, this study uncovers the complex relationship between HCPs, family and friends of parents of children with asthma, and provides new insight into the intimate and parallel Interventions aimed at optimizing parent's ability to manage their child's asthma could be strengthened by better understanding the networks that influence these parents' choices when managing asthma. There is a need to explore the asthma networks of parents of children with asthma-specifically to gain insights into whom parents select to be within their networks and why; how individuals within parent's networks influence the way in which they manage their child's asthma medications, and factors driving the development of these networks. --- What was learned from the study? It adds to our depth of understanding relating to the sources of information/support within which parents of children with asthma engage, and for the first time articulates the relative importance and influence that parents place on different sources of health information. Emphasizes the major gap in our guidelines, which despite their evolution, continue to be framed in a biomedical model of health care and fail to address the needs of parents. Supports the need for a collaborative approach to the management of pediatric asthma, which involve medical and nonmedical individuals. --- INTRODUCTION Asthma is one of the most common chronic conditions in the Australian pediatric population, affecting up to 20.8% of children at some stage in their childhood, with 11.3% of children having a current asthma diagnosis [1,2]. The burden of pediatric asthma is significant; high rates of unscheduled emergency department visits and hospitalizations, sleep disturbances in one-third of children, 60% reporting absenteeism from school, study or other activities [1][2][3], and considerable extended burden for the whole family through for example missed days of work, sleep deprivation, and anxiety [4,5]. Parents of children with asthma play a key role in their asthma management, however current data suggest that parents are falling short in their abilities to optimally manage the condition [6]. Despite receiving education, in some cases parents still identify a need for more education and information about medication use from professionals [7,8], demonstrating significant gaps in medication knowledge [9][10][11], as well as skills in adherence [12][13][14] and inhaler technique [15][16][17]. This need for information has continued for years, despite parents having access to healthcare professionals (HCPs) [8]. Parents also have a large circle of non-medical individuals/resources to whom they also turn to [18]. Most importantly, poorly managed asthma has important consequences for the family, health care systems, and the child. A lack of parents' knowledge and skills has ramifications on children's understanding of asthma, especially as children grow older and learn to self-manage their condition, and therefore this should be addressed. It is important that children start with a solid knowledge base. Research has established that parents establish their own sources of health information, i.e., their own ''health networks'' that relate to their children's asthma medication management [18]. These networks include professional connections (44% of networks), i.e., their health care professionals (HCP)(such as general practitioners (GPs), respiratory specialists, pediatricians, pharmacists, hospital staff), personal connections (such as spouses, family and friends; 42% of networks), and impersonal connections such as lay individuals (such as school staff, work colleagues; 14% of networks) and resources (such as information resources obtained off the internet or patient support groups including asthma organizations). Exploration of these networks highlights that the parent's health networks are large, complex, and variable with the influence of HCPs being just as significant as the influence of non-HCPs on the management of their children's asthma medications [18]. The asthma networks of participants ranged from two to ten connections, with an average number of five. The most commonly nominated connection was with general practitioners (GPs) followed by family members and the internet. When parents were asked about how influential these connections in their health networks were, professional connections represented 53%, personal connections 36%, and impersonal connections 11% [18]. What this research does not tell us is that although parents have specific preferences for particular health connections, it remains unknown what their role in asthma medication management is, on what basis they selected specific individuals for health advice, and how they may impact the strategies parents employ in the management of their children's asthma medications. We hypothesize that different individuals/resources in parents' networks have different roles within their influence of decisions and choices parents make in management, and this may not be related to the importance parents place on each. Understanding these connections and their importance to parents is critical to our understanding of parent perspectives and decision-making. This will enable HCPs and the community to better support the optimal management of their children's asthma. The overall aim of this study is to gain a deeper understanding of the asthma networks of parents; specifically, to gain insights into: i) The role of individuals/resources within parent's asthma networks, their level of importance to parents, and how they influence the way in which parents manage their children's asthma medications, and ii) Factors driving the development of these networks. --- METHODS --- Study Design The study was conducted between January and May of 2017. It was based on novel empirical data describing the composition of asthma networks parents established within the context of managing their children's asthma medicines [18]. Building on this research, this study adopted a qualitative approach to drive an indepth, qualitative exploration of the connections, relationships, and influences in these asthma networks. This project was approved by the University of Sydney Human Research Ethics Committee (Project No: 2015/762). Methods were performed in accordance with consolidated criteria for reporting qualitative research regulations and guidelines [56]. --- Setting and Sampling Frame The sampling frame included parents of children with asthma who had previously participated in research in this line of inquiry [18]. These parents were contacted as they had previously expressed an interest to be part of future studies and were screened for eligibility. Selection was based on a set of inclusion criteria (Table 1). Once eligibility was established, all parents provided written consent prior to participation. --- Data Collection Data collection occurred in two parts and by one researcher (PSA): i. Participant demographics and baseline data Participant demographic and baseline data were collected and included: age, gender, asthma history, highest level of education, and occupation. The child's level of impairment due to asthma was assessed using the functional severity questionnaire (FSQ) [18]. ii. Semi-structured interview Prior to the commencement of this study, participants had identified individuals and resources with whom/which they engaged around their child's asthma medicines [18]. These individuals and resources identified in parent's asthma networks and were referred to as 'alters' [18]. For this study, participants were required to reflect on the 'alters' that they had previously identified within their asthma network. A semi-structured interview guide (Table 2) was developed based on empirical evidence [18,[20][21][22][23], the theory of self-management and the theory of pediatric medication autonomy [24][25][26][27][28][29][30][31][32][33][34][35]. The interviews were recorded on digital media devices and transcribed verbatim. --- Data Analysis Participant Demographics and Baseline Data Descriptive statistics were used to summarize participant demographic and baseline data. The FSQ [19] consists of six questions with five questions utilizing 5-point Likert scales and one question requiring a 'yes/no' response where 4 points were given for yes and 0 points for no. The raw sum of the six question scores were then calculated and a severity score of low, mild, moderate, or severe was then allocated. --- Semi-Structured Interview Transcribed interviews were reviewed to identify descriptive and contextual information. Deductive and inductive approaches were used to analyze content, identify categories, and arrange them into themes [47]. A deductive approach was used to explore a participant's selection of individuals and resources in their children's asthma medication management, an investigation of their role, relative importance to participants, their impact on their child's asthma medications, and finally on other relations within their networks. An inductive approach was used to identify concepts and themes associated with the development of asthma networks. Data collection and analysis occurred concurrently, enabling further exploration of emerging themes. To ensure reliability and validity, data were independently reviewed (PSA, EA, SBA, LC, and BC) to develop inductive and deductive codes. These included any issues, topics, or ideas discussed and raised by participants. Deductive codes were developed from topics in the interview guide and research literature, while inductive codes were developed --- RESULTS --- Participant Characteristics A total of 26 parents of children with asthma participated in this study. The average time for each interview was 25 min, ranging from 15 to 40 min. Participant demographics and baseline data are summarized in Table 3. Mothers represented 100% of participants (n = 26). The mean age of participants was 42 <unk> 7 (mean <unk> SD) and 10 <unk> 4 (mean <unk> SD) for the participant's children. In this study, 46% (12/26) participants had asthma themselves. When looking at participant's children, 77% were aged between 4 and 12 years and 23% between the ages of 12 and 18. Of these children, 62% were female and 73% had mild asthma as evaluated by the Functional Severity Questionnaire (FSQ) [47]. Reporting of past experience with health care utilization indicated that all children (26/26, 100%) had been hospitalized at least once for their asthma in the past; 69% (18/26) Overall, participants reported wide differences in the nature and level of influence/importance of the different individual/resources identified. Exploration of these networks uncovered a series of complex and multidimensional relationships, and highlighted that some relationships/ individuals truly influenced the decisions made by participants, others filled a gap in knowledge and understanding, others were convenient relationships, and some connections were unrelated to the child's asthma but provided support to parent's continual needs. The specific roles and subsequent influences of individuals/resources are presented under the four categories: healthcare professional (HCP) connections, personal connections, lay individuals, and resources. These relationships and influences are discussed in detail below with examples from participant responses in Table 4. --- HCP Connections HCP connections included general practitioners (GPs), specialists (respiratory and pediatric), pharmacists, and hospital staff. Participants considered the GP to be ''officially in charge'' of their child's asthma and GPs were reported to serve a wide range of roles (18/ 26, 72%). These included the diagnosis of asthma; including physical examinations (inspections of chest and upper airways); being the first point of call in recognizing and confirming any respiratory symptoms and re-confirming a hospital diagnosis of asthma post discharge if the hospital was the first point of call. There was complexity reported around the 'diagnosis' role, as participants reported that GPs were hesitant to confirm a diagnosis of asthma at a young age (under 5 years of age). A diagnosis was often only confirmed by the GP after a hospital visit when symptoms had exacerbated during an acute attack or a flare up of symptoms, resulting in parents often feeling ''frustrated'' (11/18, 61% of parents whose primary provider was their GP). GP's were also described as actively involved in the prescribing of asthma medications, commonly described as taking a ''trial and error'' approach, to determine the most suitable medication for their child. Some participants specifically noted that their GP tended not to provide information about all possible medication side effects or the reasons for prescribing a particular medication. Specifically, GPs were reported to not provide day-to-day management advice, which participants expected would be discussed. A very small number (5/18, 19%) of parents reported that their GPs supplied a written asthma self-management plan and conducted inhaler technique assessments and training. The GP also left participants with many unanswered questions, and concern about treatment options and the medications their children had been prescribed. Participants reported that this impacted their willingness to give their children medication and made them more cautionary in taking on the management suggestions of the GP. Six participants expressed that they only see the GP now for prescription renewal. Specialists who were seen by a proportion (12/26, 46%) of participants (were seen to deliver ''specialized care'' as ''experts in the field''. They were ''respected'' by all participants who were in their care and described their advice as ''valued''. This advice made participants feel ''confident'' that the medication prescribed and management recommended were the most appropriate for their child. They were involved in the diagnosis of asthma, especially when participants did not receive a definitive diagnosis from their GP. Specialist diagnosis of asthma involved monitoring of respiratory symptoms, lung function tests, and trialing of asthma medications for symptom relief. Specialists were reported to consider the role of allergy in asthma and initiated immunotherapy if required. They were also involved in medication management, which entailed medication dose adjustments, providing written asthma self-management plans, trialing different asthma medications, adjustment of medications Table 4 Quotes supporting the perceived role of individuals/resources within parent's asthma networks and how they influence the way in which parents manage their child's asthma medications Theme Quote --- HCP connections General Practitioner''He had his first bad asthma attack and that's when we were told he was an asthmatic. He had to be hospitalized for our GP to finally realize he had asthma! We were petrified and annoyed that that was what it took for him to make up his mind!''(Participant 5) Hospital staff, such as medical practitioners and nurses, provided emergency asthma care during acute exacerbations which participants viewed as ''lifesaving'' (11/26, 46%). Their influence, however, went far beyond the acute management of the child's asthma. In these circumstances, hospital staff were highly influential, and participants reported that they played a role in the way they administered asthma medication to their children post hospital admission, i.e., participants modeled their medication administration behaviors on what they saw in the hospital under emergency circumstances for all future medication administration.'' When mentioned by participants (14/26, 54%), for all, the pharmacist played a fundamental role in the supply of medications. Beyond this, there was a dichotomy of roles described. A majority of participants (10/14, 71%) reported limited potential of the pharmacist to contribute towards their child's medicine management and only turned to the pharmacists for medication supply having infrequent interactions with them. They reported that information about medications, inhaler technique training, and management suggestions were covered by their GP or specialist and did not require any further support. Further participants reported being uninterested in the standard questions and common advice that was provided. For a smaller minority (4/14, 29%), roles reported included medication information and advice, inhaler technique education, and assisting with prescription issues (dealing with incorrect dosages, confirming directions, providing emergency medications when doctors could not be seen). Emotional support and referral to other HCPs were also reported. Some participants described a pharmacist as ''dependable'', making them feel ''confident'' in the way they administered medications to their children and helping to understand the importance of medications and ''taking the orange inhaler [reliever] everyday''. --- Personal Connections Personal connections included family (16/26, 62%) and friends (10/26, 38%), featured throughout participant asthma networks. Participants frequently encountered on-going challenges in the management of their children's asthma medications and while they would interact with HCPs occasionally to rectify these issues when in need of a professional opinion, they would interact with family and friends on a more regular basis, as they lived and socialized with most of these individuals on a daily basis. For family (9/26, 35%) and friends (5/26, 19%) who did not have asthma themselves, they were not reported to be influential in the participants decision-making around their child's asthma medications, however, they were still reported to play an important role in their network. They were involved in physically assisting the participant with the practical aspects of their child's asthma care. During events such an asthma attack or symptom exacerbation, participants turned to these individuals who could assist them in an emergency, watch over their children, and provide continuing assistance with daily life tasks. This included monitoring their child's asthma for acute symptoms when in their care, identifying any increase in asthma symptoms that they may''The doctor talked without any pause to allow me to ask questions. They were not interested to focus on listening or pay close attention to (the) questions asked, (and) when I was finally able to ask, he was always cutting me (off) while I was in the middle of asking or explaining my concerns. I just got so frustrated.''(Participant 23)''Even like doctors sometimes they misjudge, one doctor gave him the Redipred for three days which he didn't need to take and at the end he needed antibiotics for a bacterial infection, I wasn't very happy or pleased. He didn't even explain why he needed it!''(Participant 20)''See he [GP] didn't even tell me to use the spacer at the beginning so there was a little kid trying to take a puffer because when I was a kid it wasn't an aerosol it was the twisty powder so I didn't know about the spacer thing. Trust and Support''In general, ladies I know, their kids have asthma and we talk to each other about it. I tell them that I'm using the Ventolin on my kids and we start borrowing it from each other like it's a toy or something if their kids are sick. Without bothering to go to the doctor because we know it's worked for our own kids and they trust us. They start using it on their kids!''(Participant 12)''With her [friend] I've talked about it a lot; she's seen the worst-case scenario, and I've known her all my life, so I feel comfortable talking with her. So, she's had Davidworst-case scenario-and I know that I can trust what she has to say and she's provided me with great advice in the past that's worked. Also referred me to a great specialist! She was also the first person that I knew who also had a child with asthma.''(Participant 7)''Of course, they [GP] helped me and I will never go to another doctor again because he introduced me to all the medication and this worked for me and my son is very well. He also answers any questions I have whenever he can at any time of the day, he is always supportive' '. (Participant 12) have overlooked themselves, and aiding with the administration of medications. Having family and friends who were always available in all situations was invaluable to participants ongoing management of their children's asthma. They also provided participants with emotional support for feelings of anxiety and fear, which were a result of caring for a child with asthma. The ''support'' of family and friends was described with regard to their influence in triggering the participant to seek professional assistance, especially when their child's asthma symptoms worsened. More importantly, family and friends were found to influence participants' choices of HCPs. Often participants' selection of HCPs was influenced by family and friends' recommendations. When it came to family (7/26, 27%) and friends (5/26, 19%) who either had asthma themselves or a child with asthma, their role was similar to that of other friends (described above), however also related to the sharing of personal lived experiences, stories, and insights. Spouses who had asthma themselves were identified by participants to be role models for their children when it came to administering medications on their own without assistance from parents. Participants reported that it was helpful to turn to these individuals who had similar experiences as they provided realistic advice and valued this sharing as they did not believe they received this from HCPs. They also provided practical advice regarding medication use, such as advice on inhaler device use and technique, as well as medication dosages. Their high level of influence on participant's medication management was conveyed in the way they often relayed management-related information that they had heard from family and friends back to HCPs. They also compared the management advice they had received from their HCPs between each other and took on Lay Individuals Individuals such as school staff and work colleagues were labeled as lay connections (14/26, 54%) in participants asthma networks. The role of school staff was to administer asthma medication following an individualized written asthma self-management plan, especially when a child was experiencing acute asthma symptoms. They were also to inform participants of any symptoms their child has experienced while at school. Outside of that, participants did not perceive them to be influential in any way when it came to managing any aspect of their children's asthma. When it came participant's work colleagues, they played a supportive role for participants. That is, they provided participants with a place to share their experiences, feelings, and emotions. If they had asthma themselves, they shared their stories and insights, however, they reported that this had no impact on their management strategies or decisions when it came to their child's asthma management. --- Resources In addition to professional and personal connections, participants also reported to turn to other resources for additional asthma medication-related information such as the Internet and pamphlets. Participants (15/26, 58%) in this study frequently reported turning to the Internet for health information both prior to and after their interactions with HCPs. Participants used this resource to find practical information regarding their children's medications, their administration, side effects, and to find new and upcoming treatment strategies, particularly when their children experienced increased asthma symptoms. Participants reported feeling empowered as a result of access to quick health information. They reported that this influenced their decision-making independently from HCPs in regard to which medications their child should be on, adherence to medications, and when to initiate or cease medications. --- Factors Driving the Development of Asthma Networks Inductive analysis of the data identified that the development of these asthma networks occurred over time and was driven by six factors: the level of satisfaction with their primary HCP provider; the need for different information; convenience; trust and support; self-confidence in management; and participant perception of their child's asthma severity. These factors are discussed below and supported by quotes from participants in Table 5. --- Level of Satisfaction with Their Primary HCP Providers All participants utilized general practitioners (GPs), often being accessed when their child first started experiencing symptoms of asthma. While GPs were considered highly influential, participants had their own individual expectations of their GP. A large proportion of participants (12/18, 67%) expressed dissatisfaction with their GP, articulating that their needs were not being met. Some participants (10/18, 55%) reported that their GP had poor professional communication. They were not given a chance to ask questions, and if they managed to do so, they were interrupted. Others reported their GP failed to answer their questions, and instead of addressing their concerns, their GP proceeded to ask other questions important to them. Other participants reported that their GP provided inadequate information in relation to at least some aspects of their child's asthma management and treatment or highlighted the poor quality of information provided. While participants did not recognize all the gaps and found it difficult to pinpoint and express their exact needs, they expected to be provided with more detailed explanations on medication side effects (long-term side effects), prognosis of their child's asthma (if the child will ''grow out of it''), potential complications of living with asthma, health management strategies in case of worsening asthma, and any new upcoming treatments (research in the area). Especially in cases where they found themselves in an unfamiliar or critical situation that they did not how to deal with, e.g., being unprepared to recognize or respond to an exacerbation of asthma. Further, participants highlighted that they wanted advice about how to manage their child's asthma on a day-to-day basis. In terms of medicine administration, several participants reported that they were given a demonstration on proper inhaler technique on one occasion, without reinforcement or assessment over time. The majority of participants reported that their child's inhaler technique was never assessed. Few participants reported their GP providing a Written asthma self-management plan, however rarely did participants voice that it was explained clearly or updated regularly. In terms of emotional support, most participants explained that doctors failed to be empathetic and demonstrate an understanding of their ''sense of guilt'', ''anxiety'', and the ''constant worry'' that their child's condition may be causing them. One of the key connections that participants then sourced were other health care professional connections (such as specialists), family, friends, and the Internet. --- Trust and Support Trustworthy and supportive connections were important in shaping participants' asthma networks. Participants reported pursuing management-related advice from those who have previously contributed to their child's asthma or had personal experiences of asthma themselves. Through positive interactions with both professional and social relationships, such as the provision of effective treatment options, quality information and successful recommendations founded a sense of ''trust'' and ''support'' in that connection. Connections that displayed effective communication (through active listening and displaying empathy), honesty, showed respect, and cared for participants and their children helped build trusting relationships. Trusted connections were described to have an important role in expanding participants' networks, which potentially improved their child's asthma medication management. --- The Need for More Information While participants continuously reported wanting more information relating to equipping them to be able to deliver ''the best possible care for their child'', their need for more information was driven by a complex multitude of factors and underlying issues. Some participants wished for more information regarding management strategies, which would enable them to feel involved in the management of their child's illness, more confident, and be able to understand the decisions being made. They reported that feeling that they understood what was happening helped some participants to cope with their child's illness and re-establish a sense of control. Others felt that information provided by their primary HCP was lacking. This was due to physical barriers, such as a lack of time, or that insufficient information had been provided. Immediately after a child's diagnosis of asthma, many participants reported that they experienced difficulties taking in the information that was presented to them and were left with ''many questions'' after consultations with HCPs. This was due to the large amount of information imparted, causing ''information overload''; feeling ''overwhelmed'' as to the realization that their child has a chronic condition; and/or the use of medical jargon. In all of these instances, participants would turn to as many different individuals and resources as they could to answer their questions. --- Confidence in Management Participants who expressed that they did not have an active need to acquire further information from sources other than their HCPs were confident in dealing successfully with the ongoing management of their child's asthma medication. They were ''satisfied'' and ''happy'' with the resources and information they were receiving from their HCPs and felt that they could ''manage all their medications and symptoms'' on their own. These participants reported discussing their child's asthma with fewer people in comparison to others who did not display this same level of confidence. --- Perception of their Child's Asthma Severity The level of interaction and selection of individuals/resources within participant's asthma networks were also influenced by the participant's perceptions of their child's asthma severity. Participants who viewed their child's asthma to be mild in comparison to other children kept their asthma networks small, and rarely interacted with family and friends in regard to medication management of their child's asthma. In fact, they reported a desire to keep all interactions about their child's asthma to a minimum. They reported that they were ''confident'' to manage their child's condition on their own. In contrast, participants who perceived their child's asthma to be ''poorly controlled'' actively sought both physical and emotional'support' from people that were already known to them or formed new connections. In looking for support, they were actually looking for ways to increase their ''confidence'' across all aspects of asthma medication management. These parents sought out additional information from different sources, especially when their child was experiencing an asthma flareup. Consequently, these participants had larger asthma networks. --- Convenience When it came to seeking medication advice, participants made decisions about which individual or resource to utilize based on the level of convenience. The more convenient the source was, the more frequently participants reported to utilize and interact with it. The Internet was an easy and convenient source of medication information both prior to and after their interaction with HCPs, especially when HCPs lacked time during consultations, or participants wanted to re-affirm something that they had heard. By accessing asthma-related websites, participants were able to diagnose and treat symptoms promptly and timely if they were unsure what to do. When it came to HCPs, those who were easily accessible and could provide quick information and advice, such as the pharmacist, were utilized often when in need of reliable information or emergency medication. These HCPs were also held in high esteem in these situations, particularly hospital staff, as they were easily accessible to provide treatment in life-threatening situations. --- DISCUSSION This study explored the asthma networks of parents of children with asthma and has identified the role of individuals within their networks, how they influence the way in which they manage their children's asthma medications, and the factors driving their development. A qualitative exploration of the parents' network utilizing the principles of social network theory [22,48] was employed in this study and has not been conducted previously in this cohort. Important outcomes of this study highlight that when it came to the management of their children's asthma medications, parents actively sought advice and support from a series of individuals with whom they had complex and multidimensional relationships; including HCPs, personal connections, lay individuals, and resources. Some individuals directly influenced asthma medication management decisions made by parents, others provided emotional or informational support for asthma management and some connections were unrelated to the child's asthma but instead provided physical support to the needs of parents. The development of these asthma networks occurred over time and was influenced by their satisfaction with their primary HCP provider; trust and the support provided; the need for different information; convenience; their own confidence in managing the condition; and parents' perceptions of their children's asthma severity. The in-depth exploration of each connection in parents' asthma networks found that parents have a multitude of needs that are fulfilled by different individuals/resources. What parents perceived their children's asthma needed and who/what they felt was capable/available to fulfil these needs strongly influenced their selection of health connections. Support from family and friends was orientated to all aspects of the parents' everyday lives, while friends and family members who either had asthma or a child who did, shared their advice and experiences, filling in gaps in knowledge, helping parents in management decisions and ultimately playing a key influential role. The Internet, a convenient and easily accessed resource for health information, aided in asthma management decisions. Other relationships in parents' networks, such as schools, while not influential, were important to parents, and provided them with the physical support they needed. In contrast, HCPs focused on the diagnosis and treatment of the condition, providing professional advice for their children's asthma. Generally, parents saw GPs and specialists as gatekeepers to their children's asthma medications, truly influencing the asthma medicine management decisions made by parents. Specialists were highly respected and instilled a sense of confidence in parents that the medications prescribed, and management recommended is the most appropriate for their child. While pharmacists had the knowledge and skills to assist parents with asthma management, for most parents, they were relationships of convenience, where they simply supplied what the doctor had prescribed, having little influence on the parent's management decisions. This is not surprising, as other research has shown patients continue to prefer physician-led service [22,36]. For those parents who were more engaged with the pharmacist, it was clear that their level/extent of influence was meaningful. They respected their advice and help and spoke highly of them. In trying to better support the needs of parents, perhaps we need to re-consider where the role of HCPs currently fits in and how we may evolve this role over time to meet all their needs. By exploring these connections, it has provided new information on which HCPs may act on to better engage and utilize patients' existing health connections so that they better meet parents' needs. Parents' health connections were led by experiences of the condition, their experiences with their HCPs, and the child as they grew older and received more autonomy in the management of their asthma. Ultimately, a parent's relationship with their primary provider, such as their GP, played the biggest role in a parent's choices and selection of connections within their health networks. Despite their fundamental and influential role, their satisfaction with their primary provider, the trust they had in the relationship, if they had met parent's information needs, delivered uniformity in education in line with other HCPs and instilled in them a sense of confidence in managing their children's asthma, determined the extent they sought out other connections to meet their needs. Unfortunately, a large proportion of parents were disappointed with their primary providers. This disappointment in primary providers is mirrored in a study by Peterson-Sweeney et al., where parents voiced the lack of education they received from their GPs [37]. However, this study highlights that it has resulted in parents seeking out more health information and support from various individuals or resources to fill in this void and that it is affecting their willingness to communicate with their primary provider and as well as other HCPs. This is concerning, as other connections may not be equipped to provide good-quality professional advice and information. Ensuring these relationships are positive and that HCPs are meeting parents' needs is important. This can be achieved by focusing on the patient experience at every level to increase patient satisfaction, especially at first contact. Positive first interactions with parents as well as patients tend to strongly shape the experiences and emotions that follow, especially when children are first diagnosed at the hospital. Including the child and parent in a triadic discussion as part of a HCPs clinical approach helps foster a positive relationship [38]. When initial consultations with their HCP go well, a positive cycle begins with their HCP; when it goes poorly, as was the case with a large number of parents in this study, it may be difficult to recover. Negative relationships, poor experiences, and a lack of communication have been shown in adults to limit a professional's ability, or wiliness, to identify a patient's health beliefs; education needs; a patient's confidence in managing their own asthma; relevant, non-medical lifestyle factors impacting control; etc. [11,[38][39][40][41][42]. Only by identifying these patient factors through good communication, positive and trusting relations with both parents and children will HCPs be able to present more meaningful, targeted information for patients, which will, in turn, promote better asthma understanding and more effective self-care, and enable professionals to support

Introduction: Interventions aimed at optimizing parents' ability to manage their children's asthma could be strengthened by better understanding the networks that influence these parents' choices when managing asthma. This study aimed to explore the asthma networks of parents of children with asthma-specifically to gain insights into whom parents select to be within their networks and why; how

with parents as well as patients tend to strongly shape the experiences and emotions that follow, especially when children are first diagnosed at the hospital. Including the child and parent in a triadic discussion as part of a HCPs clinical approach helps foster a positive relationship [38]. When initial consultations with their HCP go well, a positive cycle begins with their HCP; when it goes poorly, as was the case with a large number of parents in this study, it may be difficult to recover. Negative relationships, poor experiences, and a lack of communication have been shown in adults to limit a professional's ability, or wiliness, to identify a patient's health beliefs; education needs; a patient's confidence in managing their own asthma; relevant, non-medical lifestyle factors impacting control; etc. [11,[38][39][40][41][42]. Only by identifying these patient factors through good communication, positive and trusting relations with both parents and children will HCPs be able to present more meaningful, targeted information for patients, which will, in turn, promote better asthma understanding and more effective self-care, and enable professionals to support asthma patients appropriately [38]. In short, the poverty of good professional communication, trust, and support leads to poor asthma management for parents of children with asthma and in turn, facilitates network expansion. Further, patients become a sum of all their experiences over time, and it is important that these experiences remain positive, so that they are better equipped to manage their asthma, leading to improved pediatric asthma outcomes. This study highlights the important link between health and social relations in the management of children's asthma medications. When looking at the influence of social support on chronic illnesses such as asthma, they can be both positive and negative [48][49][50][51][52][53]. It is difficult to determine the quality and accuracy of advice and information that family and friends are providing parents. This is especially worrisome if parents are taking this advice on board over HCP advice. Health care professionals need to consider the influential impact family and friends have on parents. Management education should incorporate skills and strategies designed to minimize social influences that hinder the optimal management of their child's asthma and enhance social interactions that facilitate successful management [41,43,44]. Studies have shown that patient behavior is highly influenced by family members and friends, resulting in decreased risk of serious illness or death [48,49]. Adults have reported to turn to family members and friends prior to seeing an HCP or using the Internet to learn from the personal lived experiences of individuals who share similar conditions, especially through participation in online health communities [50,51,54,55]. Through the provision of support and exchange of health information, social connections are proposed to promote healthy eating [52], provide emotional support to help adults with asthma to better cope with their condition [48], and motivate them to participate in preventive care programs [53]. Future research needs to reveal a deeper understanding of the social context in which this occurs in parents of children with asthma, which would allow for the development of tailored interventions that consider the specific roles of family and friends to maximize positive outcomes. Given the complex needs of parents in the ongoing management of their children's asthma, such as symptom monitoring, medication adherence, lifestyle changes, and emotional stress, this study conveys that simply having one health connection alone was insufficient to provide all the support parents needed and required to address their needs. It is clear from our results that HCPs are no longer the sole source of input. HCP solely delivering health care may not be an accurate approach to health-care delivery and providers need to consider the influences of lay advice on parents. It highlights the need to develop pediatricspecific guidelines for asthma management that foster a 'community' approach to management and the need for uniformity in education required between all HCPs. With parents interacting just as frequently with family, friends, and the Internet compared to HCPs, there is much more work to be done to effectively engage parents with HCPs to ensure that they are receiving correct information and are properly supported to provide optimal care for their children's asthma. This is especially highlighted in the fact that all of the children in this study had been hospitalized at least once for their asthma, showing already the potential failure of the primary care system. Given the increasing demands on the time of primary care doctors, particularly in the area of chronic disease management, perhaps referrals to specialists need to be considered more often or a need for greater involvement of community pharmacists or nurses. It has been repeatedly shown throughout the literature that adults with asthma have benefited from pharmacists' interventions in their asthma care [57][58][59][60]. A study by Saini et al. [57] highlighted that pharmacists who deliver specialized models of asthma care to patients are also able to meet their needs. Policies need to enable other health professionals, such as pharmacists, to contribute to optimal chronic disease management if primary providers are not able to meet the demands. However, pediatric asthma is a chronic disease that is often under-represented and under-prioritized by policymakers, government bodies, healthcare professionals, and researchers. While this study has added valuable insights to our understanding of the key influences in parents' management of their children's asthma medications, research limitations need to be considered. That is (i) this study was conducted in only one particular Australian district, (ii) all the parents interviewed were female, although the latest Census data show that fewer than one in 20 families have a father who is the primary carer [45], (iii) the vast majority of the participants had very mild asthma, with the majority symptom-free for the past month, (iv) although interview data were independently reviewed, interpretive bias should be considered. However, interpretation is never completely independent of a researcher's beliefs and preconceptions [46]. To minimize bias, regular meetings were held to critically compare and discuss findings. Further, these findings are based on a cross-sectional case study and parent relationships that may change over time and under different circumstances, leading to different asthma networks. A time series design study or an ethnographic study should be considered to examine parents' asthma networks at different time points during childhood in order to look at the development of these networks over time. This would allow us to gain further understanding of how these asthma networks came to evolve over time. --- CONCLUSIONS In conclusion, this study has uncovered the important and underestimated role of parents' non-medical sources of information/support. It also highlights the complex relationship between HCPs, parents' non-medical sources of information/support, and the intimate and parallel influence they have on a parent's decision-making when it comes to the management of their child's asthma medications. Simply, HCPs are not able to provide all the support parents need and require to address their needs when it comes to the management of their children's asthma medications. This study supports the need for a collaborative approach to the management of pediatric asthma, which involves both medical and non-medical individuals, uniformity in education between all individuals involved, and highlights the need to develop pediatric-specific guidelines for asthma management that foster a 'community' approach to management. for developing educational presentations from Teva and Mundipharma; and has received honoraria from AstraZeneca, Boehringer Ingelheim, and GlaxoSmithKline for her contribution to advisory boards/key international expert forum. Dr. Vicky Kritikos has received honoraria from AstraZeneca, GlaxoSmithKline, and Pfizer. Compliance with Ethics Guidelines. This project was approved by the University of Sydney Human Research Ethics Committee (Project No: 2015/762) and participants provided their informed consent to participate in the interviews. Data Availability. The data are not publicly available as consent was not obtained from participants to share data outside of the requirements of the research process. --- Authorship. All named authors meet the International Committee of Medical Journal Editors (ICMJE) criteria for authorship for this article, take responsibility for the integrity of the work as a whole, and have given their approval for this version to be published.

Introduction: Interventions aimed at optimizing parents' ability to manage their children's asthma could be strengthened by better understanding the networks that influence these parents' choices when managing asthma. This study aimed to explore the asthma networks of parents of children with asthma-specifically to gain insights into whom parents select to be within their networks and why; how

Introduction Racial inequalities in use of dental services and their potential implications are widely documented in the literature [1][2][3][4][5][6]. These inequalities are mostly attributed to socioeconomic and cultural/behavioral attributes [7][8][9], but the question as to how the observed racial gaps in utilization of dental services could be a consequence of racial discrimination is not fully explicated [10,11]. Racial discrimination in healthcare is an individual's appraisal of unfair treatment in a medical setting based on race, colour or national origin [12,13]. Discrimination can occur at individual, institutional, and structural levels [14]. It can operate through different mechanisms, subsequently leading to poor health outcomes and underutilization of services [13,15]. First, the institutional or structural racism arising from racially discriminatory unfair policies and institutional culture can lead to differential educational/employment opportunities, and thus, access to health-promoting resources [15]. Second, racism often leads to the development of implicit racial bias and explicit racial stereotypes, influence clinicians' behavior decision making and communication process, and thus contribute to the differential treatment of members of same institution [16,17]. Third, racism can act as a psychosocial stressor that operates through physiological, psychological, and behavioral pathways which could have consequences for health [16]. Moreover, negative experiences have the potential to influence quality of healthcare, interpersonal trust, medical adherence levels, and treatment delays [18,19]. Finally, the racial discrimination deeply entrenched within the healthcare settings may alter the patient's perception of healthcare interaction and the pattern of their healthcare access [20]. To be more precise, internalization of unfair treatment may engender involuntary responses such as anxiety or increased vigilance and voluntary coping responses like disengagement from situations or environments that negatively stereotype people [21]. This may possibly inhibit certain individuals from using a wide range of needed health services, including dental services. Few studies have evaluated the impact of racial discrimination on oral health, with conflicting evidence. While racial discrimination was associated with experience of toothache among pregnant Aboriginal Australians [22], self-reported dental problems among Canadian immigrants [23] and tooth loss among pregnant Aboriginal Canadians [24], no association was found with periodontitis among Hispanic Americans [25], tooth loss among Brazilian civil servants [26], or oral health-related quality of life among pregnant Aboriginal Canadians [24]. Evidence on the impact of racial discrimination on utilization of dental services is even more limited. Racial discrimination was not associated with a dental visit within the past year among pregnant Aboriginal Canadians, but it was associated with being asked to pay for dental services by dental care providers despite entitlement to free dental care and seeking care off-reserve or out of the community [24]. Among pregnant Aboriginal Australians, racial discrimination was associated with having never visited a dentist before. However, no adjustment for participants' socioeconomic status was attempted [27]. A broader implication of the negative impact of discrimination comes from qualitatively research, which showed that disadvantaged caregivers of Medicaid-enrolled children cited discriminatory behavior attributed to racism as one of the key barriers to accessing dental services [28,29]. Using the well-established Andersen's behavioral model of health services use [30,31] as a theoretical framework to understand factors affecting an individual's decision to use dental services, this study explored the relationship between racial discrimination and use of dental services among American adults. It was hypothesized that individuals experiencing racial discrimination would be less likely to use dental services than those without such experiences. --- Materials and Methods --- Data Source The Centers for Disease Control and Prevention's Behavioral Risk Factor Surveillance System (BRFSS) is an annual, state-based, random-digit-dialled, telephone health survey of the non-institutionalized US civilian population, aged 18 years or older, living in all 50 States, the District of Columbia, Puerto Rico, and Guam. In each geographic region, the BRFSS uses a disproportionate stratified sampling for the landline telephone survey and random sampling for the cellular telephone survey. For the landline telephone survey, interviewers collect data from a randomly selected adult in every participating household. For the cellular telephone survey, interviewers collect data from adults residing in a private residence or college housing who have a working cellular telephone. The data were collected using the BRFSS standardized questionnaire via computer-assisted telephone interviews (CATI). The BRFSS questionnaire consists of core, optional, and state-added questions modules. Interviewers were trained as per the BRFSS protocol, and confidentiality of participants was maintained by ensuring their anonymity to the interviewers [32]. We used data from the 2014 BRFSS because this was the most recent survey, including questions on both oral health and racial discrimination. In 2014, 464,664 interviews were made, representing a weighted median response rate of 47% (48.7% and 40.5% for landline and cellular telephones, respectively). Response rates varied from 25.1% to 60.1% across states. The present analysis is limited to participants in the three states (Minnesota, Mississippi, and New Mexico) that included the optional module on reactions to race in 2014. There were 23,272 participants who completed the optional module on reactions to race and the core module on oral health. After exclusions due to missing values on covariates, the study sample include 11,950 adults. --- Selection of Variables The outcome variable of interest was utilization of dental services, which was assessed with the following item: "How long has it been since you last visited a dentist or a dental clinic for any reason?" with multiple response options (within the past year, 1 year but less than 2 years ago, 2 years but less than 5 years ago, 5 or more years ago and never). Responses were re-coded into those who visited a dentist in the past one year and those who did not visit one (reference group). Two main explanatory variables indicating discrimination were taken from the reactions to race module of the questionnaire. First, perceived racial discrimination while seeking healthcare was assessed with the following question: "Within the past 12 months when seeking healthcare, do you feel your experiences were worse than, the same as, or better than people of other races?" with multiple response options (worse than other races, the same as other races, better than other races, worse than some races, better than others, only encountered people of the same race). For analysis, 'worse than other/some races' responses were coded as having experienced discrimination. The second indicator of discrimination was emotional impact of discrimination, which was assessed using the question: "Within the past 30 days, have you felt emotionally upset, for example angry, sad, or frustrated, as a result of how you were treated based on your race?" with yes/no response options. Based on Andersen's behavioral model of health services use [30,31], several covariates were included as potential confounders of the association between racial discrimination and use of dental services. They were predisposing (age, gender, race/ethnicity, income, education, and smoking), enabling (health insurance), and need factors (number of missing teeth). Self-identified race/ethnicity was categorized into five groups: Non-Hispanic White, non-Hispanic Black, Hispanic, and Asian/Other, and Multiracial. Yearly income was categorized into four groups: Less than $20,000, $20,000 to $34,999, $35,000 to $74,999, and $75,000 or more. A binary variable created for health insurance coverage was based on whether participants were enrolled in a healthcare plan (including Medicare and Medicaid). Educational attainment was categorized as less than high school, high school or equivalent, some college, and college graduate. Current smokers were respondents who reported smoking at least 100 cigarettes during their lifetimes and reported smoking "every day" or "some days". Former smokers were those respondents who reported ever smoking at least 100 cigarettes but reported smoking "not at all". Never smokers were respondents who reported smoking fewer than 100 cigarettes during their lifetimes. The original categories of missing teeth (none; 1 to 5; 6 or more, but not all; and all) were categorized into missing teeth and no missing teeth. --- Statistical Analysis Analyses were weighted to produce nationally representative estimates and accounted for the sampling design (clustering and stratification). Only cases with complete data in all variables were included in the analysis. Analyses were performed in Stata 14 (StataCorp, College Station, TX, USA). We first described the characteristics of the sample according to predisposing (demographic factors, socioeconomic position, smoking status), enabling (health insurance), and need factors (missing teeth). Both indicators of racial discrimination were also included. Secondly, the proportion of individuals who had a dental visit in the past year were compared between categories of the two indicators of racial discrimination and covariates using the Chi-squared test. Thereafter, we constructed two logistic regression models to assess the relationship between each indicator of racial discrimination (racial discrimination when seeking healthcare and emotional impact of racial discrimination) and utilization of dental services in the past year (outcome). Odds ratios (OR) were used as the measure of association. We first tested the association between racial discrimination when seeking healthcare and use of dental services in a crude model, which was subsequently adjusted for predisposing (sex, age, race/ethnicity, education, income, and smoking status in Model 1A), enabling (health insurance in Model 1B), and need (number of missing teeth in Model 1C) factors. No statistical interactions were included in the models. The same set of models was developed when testing the association between the emotional impact of racial discrimination and use of dental services (labelled as Models 2A to 2C). --- Results The characteristics of the study sample are presented in Table 1. The percentage of dental visits within the past year was 68.3% (95% CI 67.1-69.5). In addition, 2.7% (95% CI 2.3-3.1) of adults reported healthcare discrimination and 5.0% (95% CI 4.5-5.6) reported emotional impact of discrimination. Dental visits within the past year were more common among women, older, non-Hispanic White, and non-smoking participants. Those with more education, greater income, health insurance, and no missing teeth were also more likely to have visited the dentist in the last year. Dental visits were less common among those who reported racial discrimination while using the healthcare system or reported emotional impact of discrimination than those who did not (Table 2). Those who experienced racial discrimination when seeking healthcare were less likely to have visited the dentist in the past year than those who did not have such an experience (OR: 0.57, 95% CI 0.41-0.79). Similarly, those who experienced the emotional impact of discrimination were less likely to have visited the dentist in the past year than their counterparts (OR: 0.57, 95% CI 0.44-0.73). The association between healthcare discrimination and use of dental services was fully attenuated after adjustment for predisposing factors (demographics, socioeconomic position, and smoking status) and remained unchanged after subsequent adjustments for enabling (health insurance) and need (missing teeth) factors (Table 3). In the fully adjusted model, the odds ratio was 0.88 (95% CI 0.62-1.25). By contrast, the association between emotional impact of discrimination and use of dental services remained significant after adjustments for predisposing, enabling, and need factors. In the final model, those who experienced the emotional impact of racial discrimination were 25% (OR: 0.75, 95% CI 0.58-0.99) less likely to have visited the dentist in the past year than those without such experience. --- Discussion This study supported our hypothesis that racial discrimination is negatively associated with the utilization of dental services. Those who experienced the emotional impact of racial discrimination were less likely to have used dental services within the past year. This finding was robust to multiple adjustments for known determinants of utilization of dental services, which were carefully chosen according to the Andersen's behavioral model of health services use [30,31]. Most of the current research on racial/ethnic inequalities in use of dental services gives scant attention to the role of racial discrimination by which this association might exist [11]. In this study, both healthcare discrimination and the emotional impact of discrimination were negatively associated with dental visits in crude regression models. Interestingly, the association between healthcare discrimination and dental visits was fully accounted for by controlling for predisposing factors, which emphasizes the role of socioeconomic position in explaining such an association. Having greater financial resources might give individuals more options of healthcare providers and, as such, prevent discriminatory experiences. The above finding also suggests two related points. One is that it might be easier to report emotional impacts of racial discrimination than actual experiences of discrimination. As with other subjective experiences, racial discrimination could be misperceived or overlooked, which can lead to underestimating or overestimating their actual occurrence. The other is that the emotional impact of discrimination can arise from a broader set of life experiences, not only those related to seeking healthcare. This argument explains why the prevalence of the emotional impact of discrimination was higher than that of healthcare discrimination despite the former having a shorter recall frame than the latter (30 days versus 12 months). What is important to remember is that such experiences might prevent people from using other services, thus perpetuating the vicious circle. As a contributory factor to racial inequalities in oral health, discrimination based on race should be viewed as a social determinant of oral health [11]. However, what remains unclear from this line of research is why, how, and to what degree experiences with discrimination in everyday social settings, including a dental clinic or a hospital, influence the lives of racial minorities [2]. One popular explanation that the literature offers is based on the conceptual framework laid by the Major and O'Brien's model of stigma-induced identity threat, which explains how subjective perceptions of discrimination within the healthcare system influence receipt of health services [21]. People may disengage from health services because of discrimination and insensitivity from healthcare staff (on the grounds of gender, race, culture, social class, sexuality or even symptom-related factors such as drug use or homelessness). They may feel alienated by negative discriminatory experiences; being subject to more paternalistic and coercive treatments and cultural or language barriers in assessments [33,34]. Collectively, this would lead to mistrust [19], non-compliance with treatment, and partial or complete disengagement from a dominant cultural institution such as healthcare, further attenuating the unmet need of care. The current findings begin to fill an important gap in evidence regarding racial discrimination and the consequential emotional impact of discrimination on dental visits using quantitative data. In general, the uptake of dental services is conceptualized as issues of acceptability, affordability or availability, and contributory roles of the perceived systemic level and interpersonal discrimination are often overlooked. Dental visits could be a wider experience of racial discrimination prevailing in the society in general and in the healthcare system in particular. It is imperative to account for the accurate impact of discrimination and assess the contemporary policies that pave the way to racial discrimination. Our findings underscore the need to understand the patterns of discrimination and ways in which it influences engagement or disengagement from health services in a longitudinal framework. If understood, this could move policies forward and potentially facilitate developing effective strategies to deliver non-discriminatory and equitable healthcare services. Few caveats should be acknowledged in interpreting the findings. First, the cross-sectional nature of the data limits the extent to which causal inferences can be drawn from our findings. Second, the single-item measure of perceived discrimination in healthcare pertained to experiences during the past 12 months, which intrinsically excluded people who did not seek care in the past year because of access-related issues. Third, the moderate response rate of the BRFSS might raise concerns about selection bias, although the probability weights used to correct for the differential probability of selection and non-response might have corrected it to an extent. Finally, the inconsistent adoption of the optional reactions to race module across the states limits the generalizability of our findings. Characteristics of the states in which the module was used may differ in significant ways from those of states that elected not to use it. Therefore, the present findings represent valid relationships between the variables of interest but cannot be viewed as representing the entire adult nation in the United States. --- Conclusions In a large population survey of adults in the United States, this study shows that the emotional impact of racial discrimination was associated with lower uptake of dental services. This association persisted even after accounting for various predisposing, enabling, and need factors. The preliminary findings from this study underscore the need to understand the patterns of discrimination and ways in which it influences engagement or disengagement from health services.

This study examined the relationship between racial discrimination and use of dental services among American adults. We used data from the 2014 Behavioral Risk Factor Surveillance System, a health-related telephone cross-sectional survey of a nationally representative sample of adults in the United States. Racial discrimination was indicated by two items, namely perception of discrimination while seeking healthcare within the past 12 months and emotional impact of discrimination within the past 30 days. Their association with dental visits in the past year was tested in logistic regression models adjusting for predisposing (age, gender, race/ethnicity, income, education, smoking status), enabling (health insurance), and need (missing teeth) factors. Approximately 3% of participants reported being discriminated when seeking healthcare in the past year, whereas 5% of participants reported the emotional impact of discrimination in the past month. Participants who experienced emotional impact of discrimination were less likely to have visited the dentist during the past year (Odds Ratios (OR): 0.57; 95% CI 0.44-0.73) than those who reported no emotional impact in a crude model. The association was attenuated but remained significant after adjustments for confounders (OR: 0.76, 95% CI 0.58-0.99). There was no association between healthcare discrimination and last year dental visit in the fully adjusted model. Emotional impact of racial discrimination was an important predictor of use of dental services. The provision of dental health services should be carefully assessed after taking account of racial discrimination and its emotional impacts within the larger context of social inequalities.

Introduction This paper gives general information about the experiments performed within the project CroMa (Crowd Management in transport infrastructures) [1]. In addition, experiments from other projects such as CrowdDNA [2] were carried out in the context of this experiment series as well as experiments that cannot be assigned to a third party funded project. The paper includes information about the overall organization, the experimental site, the procedure and timeline, the participants, the data collection technique and gives an overview of all experiments. Further detailed information of single experiments, especially data analysis, will be found in focused papers on these experiments. All data gathered by the sensors used will be made freely accessible in the pedestrian dynamics data archive [3] with the publication of the first scientific results at the latest. General data mentioned in this paper like the overall composition of the test persons based on handed out questionnaires and a measurement course can also be found in the data archive [4]. The CroMa project itself is focused on developing and enhancing different strategies, such as building regulations, crowd management, and innovative action strategies to increase the efficiency of pedestrian facilities in railway and underground stations. These strategies aim to increase the robustness and efficiency of railway stations during peak load and to avoid crushes in the event of critical crowd densities. Research within the framework of CroMa includes the investigation of pedestrian flow in traffic facilities and the study of pedestrian behaviour within dense crowds. These research fields have also been assessed by means of the large-scale experiments described in this paper in which several external (structural) and internal (characteristics regarding the test person sample) parameters have been varied on a controlled basis. --- Preparation of Experiments The CroMa-experiments were conducted from October 8, until October 11, 2021 in the Mitsubishi Electric Hall (MEH) in Düsseldorf, Germany. The MEH is a multipurpose event hall with an interior hall size of 3500 m 2 and an additional main and side foyer. The planning and preparation were divided into two interlocked parts. One being the overall organization and provision of rooms and material and the other being the scientific planning of the individual experiments. Several preparatory meetings were held to discuss issues related to the variety of tested scenarios and statistical significance relative to available time and personnel. A temporal and spatial setup was developed to account for the level of information given to the participants about the aims of the study as well as their learning effect over the course of the day. The experimental plans were simplified, concretized and the times for conducting the experiments, announcements, walking routes, filling in the questionnaires and taking small breaks were calculated to test the feasibility of the setup. Originally, the experiments were planned for March 2020 and had to be postponed due to the growing SARS-CoV-2 (Covid-19) pandemic. When the experiments were conducted in October 2021, the setup was revised regarding compliance with safety measures and expanded to include a hygiene and safety concept. --- General Framework The experiments were performed in a circuit training model. This means that three experimental setups were performed at the same time at three different sites, and participants were guided from one site to the other in designated groups. The three groups were marked with wristband colors: red, green or blue. The experimental sites were labeled alphabetically 'B', 'C' and 'D' (Fig. 1). The experimental sites were separated by black curtains that shielded the view but were not sound proof. To limit the view on to the experiment, the waiting areas within the experimental sites were shielded by curtains as well. Each day (day 1-3) consisted of six experimental time slots lasting 1 hour each, therefore participants attended each experimental site twice a day, but never participated twice in the same experimental setup, as those changed from one time slot to the next. A rough time schedule is shown in Fig. 2. The following experiments were performed in site B: --- Site • Train Platform Experiment (day 1-3; Sec. 3.1) The following experiments were performed in site C: • Crowd Management Experiment (day 1-3; Sec. 3.2) • Single-File Experiment (day 4; Sec. 3.3) • Personal Space Experiment (day 4; Sec. 3.4) The following experiments were performed in site D: • Boarding and Alighting Experiment (day 1-3; Sec. 3.5) • Tiny Box Experiment (day 1-3; Sec. 3.6) • Bottleneck Experiment (day 4; Sec. 3.7) Day 4 was different in that participants were divided into only two groups: Group yellow consisted of 80 people and group red of 120 people. Group red took part in the experiments at site D in all six time slots. Group yellow took part in the experiments at site C for the time slots 1 to 3 and also came to site D for time slots 4 to 6. --- Recruiting Process Participants were recruited by spreading information via various channels including printed and social media as well as e-mail lists of former experiments. The information included a short summary of the project, dates and payment. A QR code and link to a registration website was provided. The website included further information on conditions of participation, days available as well as a registration form. The conditions of participation (originally in German) included: • minimum age of 18 years and recommended age of younger than 75 years • body height of 1.5 m to 2.0 m • not being affected by limited mobility or claustrophobia • wearing dark clothes without lettering and not wearing large bags/backpacks • agreeing to being filmed and the material to be published in a data repository After submitting the registration form, the potential participants received an e-mail that their registration had been received. People were assigned to days based on their statements of availability and evenly divided among the days if they were available for multiple days. People were only able to register for one of the three first days and additionally for the fourth day. After allocation to the days, participants received an e-mail with allocation information. Two weeks prior to the experiments, a reminder was sent including current information on the hygiene and safety concept. The hygiene concept to protect against infection by Covid-19 was a necessary requirement of the authorities and institutions involved. One week prior to the experiments, participants received an e-mail with a reminder of their personal assigned dates and important things to remember and bring along with them (e.g. ID card, comfortable shoes, wear dark clothes). the Mitsubishi Electric Hall and proceed to the registration desk in the main foyer. During registration the identity documents were checked again, participants signed forms that they consented to the conditions of participation, and they were handed a green hat, personal ID code (Aruco Code dict 6X6 1000 [5]) and corresponding number on a wristband (Fig. 3 a) as well as a clipboard with questionnaires and declaration of informed consent to be filled out. --- Registration and Measurement Course The wristbands had three different colors (red, blue, green) and were handed out alternately. That way participants were divided into three experimental groups on arrival. Participants who arrived in social groups were therefore split among our experimental groups, although it cannot be completely ruled out that people who know each other were in the same experimental group. The wristbands were labeled with numbers (Fig. 3 a) referring to the number associated with the personal ID code. The number was used for all questionnaires throughout the course of the day, to allocate sensor information and trajectories to a participant without revealing personal information. After registration participants entered a course that led them to a sequence of stations. At these stations, information was collected and subjects were provided with markers and utensils: • measuring height • applying shoulder markers to top (Fig. 3 c) • putting on green hat and attaching personal code (Fig. 3 c,d) • checking correct fit of hat with code and telling people to leave the hat on for the entire day • time for questions • final check if declaration of informed consent was signed and questionnaires were filled out correctly • targeted addressing of suitable people to ask if they were willing to wear additional sensors (3D motion capturing suit, heart rate sensor) After completing the measurement course participants could check their bags at a cloakroom and proceed to a large waiting area. --- Test Person Sample To perform the experiments we accepted 1500 people for the duration of the four days. Of these, 1038 people attended the experiments. The sample of test persons included people from the age of 18 to 85 years (median=31, <unk> = <unk>17), with 47 % being male, 51 % female and 2 % not specified. Some of the distributions of the demographic data collected via questionnaire are shown in Fig. 4. Data that had to be measured such as body height, body weight and shoulder width were collected by employees during the measurement course (Fig. 3 b). On average, the participants were 1.75 m (<unk> = <unk>0.1) tall, weighed 79 kg (<unk> = <unk>21) and had a shoulder width of 45 cm (<unk> = <unk>4). Female participants were shorter, more lightweight and slimmer at the shoulders on average. Further personal data were collected via questionnaires. Differences in distributions for the different days and experimental groups can be found in the App. A.1. More data can be found in the archive [4]. --- Notes Related to Covid-19 Pandemic At the time recruiting started as well as at the time of the experiments, Germany was at the beginning of a third Covid-19 wave (Fig. 5). Of the enrolled people, 90 % declared that they had been fully vaccinated. Due to the pandemic, a number of precautions were taken: • a hygiene and safety concept was developed by the team and approved by the crisis committee of Forschungszentrum Jülich and the competent regulatory authority of the city of Düsseldorf • participants had to be recovered, vaccinated or tested (referred to as "3G" in Germany) • everyone was tested at the time of arrival and people were only allowed to enter the building with a negative test result • participants had to wear surgical masks at all times (except when eating or drinking at a seat in the waiting area) In the first year of the Covid-19 pandemic, Germany's regulations prohibited people from getting together in large groups and required them to keep a distance of 1.5 m between people not belonging to the same household. During the summer of 2021 these restrictions were dropped for vaccinated people. In public it could be observed that people kept wearing masks and kept their distance to other people to a large extent on a voluntary basis. Because our experiments were designed for situations in which high densities may occur, we attempted to mitigate the behavioural changes described above. In order to get people accustomed to larger groups and higher densities again, we performed an 'icebreaker' experiment. Participants were not informed about the icebreaker experiment, which was performed as part of the walkway to the first experiment. After registration people waited in a large waiting area (Fig. 1 green shading). When the registration was finished up to 100 people were asked to walk to the first experiment at a time (based on the color of their wristband). A person in charge, responsible for an experimental area, walked them into a corridor with two doors (Fig. 1 blue shading). The person in charge asked people to wait until the last participant had arrived in the corridor. The rear door was closed when everyone had arrived (the density in the room was about 1 P/m 2 ). Then the person in charge waited for another few minutes before releasing participants into another open space. The icebreaker was performed once every morning for each group. To assess the extent to which participants were influenced in their actions by the thought of the pandemic, everyone was given a questionnaire about perceptions of various risks (focusing, in particular, on perceptions of risks of Covid-19 infection) and about the potential influence of the Covid pandemic on the experiments, at the end of each experimental day. The questions were answered on a 7-point scale from "strongly disagree" (1) to "strongly agree" (7). The questionnaire started with the general items about whether participants felt uncomfortable in the crowds during the experiments. Participants were then able to rate how much the following seven factors influenced their discomfort: Crowding, concern about contracting Covid, concern about contracting an other illness, unclear instructions, physical exertion, stress caused by the experimenters, being with many people. Two more questions directly addressed the mental engagement with Covid. In addition, participants could indicate in which setting (e.g., in the registration course) and in which type of experiment (e.g. bottleneck) they were most concerned about Covid with a yes/no answering format. In the two final questions, subjects estimated whether they would have behaved differently before the pandemic and indicated whether they had already been in a crowd before the experiments since the onset of the pandemic. N = 1000 participants filled out the questionnaire on the four experimental days. Descriptive statistics are reported for the questionnaire data (Tab. 12). The results of the questionnaire suggest that the Covid pandemic did not have a major impact on the answers of the participants. The concern about infection was reported to be low (mean value M between 1.98 and 2.45). Subjects self-reported that their actions were not significantly different than before the pandemic (M = 2.69). A self-selection effect certainly plays an important role for these results: Presumably only people who were not very concerned about Covid signed up for the experiments. This can also be seen in the last question. The statement that they have already been in a crowd elsewhere was often agreed with (M = 4.02). Furthermore, these results reflect the extensive safety measures that apparently reduced the fear of contagion. In general, a low expression of discomfort was reported in the experiments (M = 2.61). The factors that caused the most discomfort were crowding (M = 3.37) and physical exertion (M = 3.49). Subjects thought most strongly of Covid in the morning registration and measurement course (38 % answered with yes). We explain this by the fact that the registration came immediately after a mandatory Covid test. This meant that the topic of the pandemic was very present at that moment. On days 1-3 participants thought about Covid most often in experiment site C (26.2 %), followed by site D (16.1 %) and B (10.4 %). On day 4 participants thought about Covid quite a lot in the experiment site D (37.9 %) and very little in experiment site C (6.6 %). --- Configuration of Experiments --- Train Platform Experiments In this experimental site two different experiments were performed. The first one investigated the waiting behaviour of people on a simulated train platform under varying physical or social psychological factors, the second one investigated types of social influence in ambiguous situations on train station platforms. The outer dimensions of the platform were 7 m x 20 m x 0.8 m. The ascent and descent was realized by stairs secured with railings and organized in a way that resembled one primary access and three stairs to "board the train" that only a few people could stand on at the same time. The stairs at the narrow side were 3 m wide while the stairs at the long side of the platform were each 1.5 m wide. The smaller stairs were movable and the positions for save attachment to the platform were only visible for helpers moving the stairs. The platform's edge was marked with adhesive tape (width 0.05 m). White adhesive tape marks the safety distance (0.8 m) from the edge of the platform, which is a standard for platforms in Germany. A loudspeaker box with recordings of railroad sounds was placed under the platform during all of the experiments to reduce the influence of sound from the neighbouring experimental sites. Cameras were mounted to record the experiment. They are listed in Tab. 2. Experimental runs in which 3D motion capturing data were recorded are listed in Tab. 11. The mood of the participants (cf. Sec. 4.7) was recorded for all runs. Trajectories were generated as described in Sec. 4.2. The coordinate origin's location was centred in relation to and 0.5 m from the short platform border (Fig. 6a), with the y-axis pointing to the left, parallel to the longer edge and aligned with the midpoint of the large stairs. The data of the Train Platform Experiments are provided online [7]. Waiting Experiments In this experiment instructions were given in the waiting area of the experimental site directly in front of the entrance to the experiment. The waiting area was separated from the experiment by a black curtain and thus the participants could not see the setup during the waiting phase and the instructions. In runs in which questionnaires were completed, this was done after the respective runs in a second waiting area at the opposite side. With the instructions, the participants were informed that the train they intend to board would arrive on the left hand side of the platform. The instructions read as follows, "Imagine you are at a train station. Behind those curtains is the platform, which you will enter through the stairs. You plan to take the train that will arrive in a few minutes at the platform at the left-hand side". The following parameters were varied (a detailed list of performed runs and combinations of parameters can be found in App. A.3): • number of participants: 40, 80-100, (140-180) • obstacle on platform: none (blank), wall, house (Fig. 6 b-d) • inflow: every 2-3 seconds, in groups of ten • waiting time on platform: 2, 4 minutes The degree of familiarity of participants, as well as talking being allowed or prohibited was not manipulated, as in the experiments on decision making described below. The measurements of the wall were 0.6 m x 3.6 m x 2 m and of the house 3 m x 3.6 m x 2 m. Both were aligned symmetrically to the borders of the platform. In runs with 40 people, participants waited either 2 or 4 minutes. The waiting time was counted from the moment the last participant entered the platform. Additionally, the inflow sequence to the platform was varied, for the experimental runs without obstacles. The participants entered individually or in groups of ten. The groups of ten entered the platform with an interval of 35 seconds. For the platform without an obstacle and the setup with the house, additional runs with an larger number of participants were performed. In those runs the group of participants assigned to the corresponding experimental slot entered the platform first. After those participants had positioned themselves on the platform, participants from another group were brought in. Therewith the total number of passengers on the platform was unknown to all participants. Experiments on Decision Making For the experiment, the participants of each run were instructed directly in front of the platform. They entered the platform using the stairs on the long side. On the platform, two areas were marked, one slightly larger than the other. Participants were instructed to wait in the larger area but not instructed on how they should know which area was the larger one (because it was not easily visible). The instructions read as follows, "Imagine you are at a train station and want to leave on the next train. The platform is marked with a white safety strip at the long sides. Two areas are marked in yellow, the so-called 'yellow squares', one on the right and one on the left. You will only be able to board the train from the larger area. Please proceed to the larger of the yellow squares". Depending on the experimental condition the instruction was continued, "Please, do not talk to each other during the whole experiment" or "You are allowed to talk to each other during the experiment". The curtains to the waiting area of the experimental site prevented the participants who did not take part in that specific run from observing the active participants of the run while carrying out the task. Questionnaires were completed for all runs both before entering and after leaving the area. The following parameters were varied (a detailed list of performed runs and combinations of parameters can be found in App. A.3): • number of participants: 10 (small groups), 23-41 (other groups) • special design: 2 marked areas (Fig. 6 e) • inflow: all at once per group • waiting time on platform: up to 5 minutes • degree of familiarity of the participants with each other: no connection at all, short acquaintance before starting the experiment, being in the same group for hours before the experiment • special announcements: talking allowed, talking prohibited The marked areas were placed on the platform asymmetrically (Fig. 6 a) or e)), and had a size of 35 m 2 (left) and 36.7 m 2 (right). --- Crowd Management Experiments This series of experiments investigated the extent to which physical parameters such as the number of line-up gates or the width and the shape of the barrier layouts influence the formation and the density of a queue. For this purpose, an admission situation under the assumption of "admission to the concert of your favorite artist" was simulated using barriers and line-up gates typical to those used at large events. Furthermore, non-physical parameters were considered. The following parameters were varied (a detailed list of performed runs and combinations of parameters can be found in App. A.4): • setup structure grid (narrow): straight, small bend, 90 • -right bend with and without lines (floor markings) (Fig. 7, 8 c-) • setup structure grid (wide): none, lines, signs (above the line-up gates construction) (Fig. 7, 8 a-b) • no. of open line-up gates: 1, 3 • motivation: low (enough time and guaranteed seat ticket), high (the general admission ticket) • norm specification: none, 70%, 85%, 95%, 100% • special announcements: no interruption (nI), with interruption (wI) (along with HRV sensors (Sec. 4.4)) • reference runs to record free walking speed (solo ref) Instructions according to the motivation and the run were given directly in front of the entrance to the experiment before participants could see the experimental setup. At the time of the instructions, participants were separated from the experiment by a black curtain. If a norm specification was given, every participant was given a slip of paper with a note on it saying either "In the following experiment, behave as you always would" or "Imagine that you are a very selfish person. Push yourself to the front during the experiment". The percentage refers to the amount of paper notes prescribing normal behaviour. After each run, participants were directed back to the area where the instructions were given and questionnaires distributed. HRV Sensors were collected (if given) after each group had finished their respective runs. Within the framework of the experiments, a comparison of the estimated and the physically measured density of people was carried out. Some assessors had been previously trained to the Level of Service and had been given further knowledge of density estimation, while others were untrained. Time-dependent densities ranging from low to medium to high were estimated [8] and documented. In addition, positive and negative factors influencing the density estimation of the observers were surveyed using questionnaires. The outer dimensions of the experimental area were 7 m x 20 m. The line-up gates construction was 2.5 m x 3.3 m x 1.18 m (length x width x height) with a passage width of 0.5 m each. The line-up gate construction at the 90 • bend setup was 30 cm wider. Police barriers with dimensions of 2.0 m x 0.94 m x 1.1 m were used to set up the structure grid. Cameras were mounted to record the experiment and are listed in Tab. 3. Experimental runs in which 3D motion capturing data were recorded are listed in Tab. 11 and runs in which HRV data were recorded in Tab. 10. The mood of the participants (cf. Sec. 4.7) was recorded for all runs. Trajectories were generated as described in Sec. 4.2. The coordinate origin was located where participants enter the experimental area, with the y-axis pointing in walking direction and aligned with the midpoint of the middle entry gate (Fig. 7). The data of the Crowd Management Experiments are provided online [9]. --- Single-File Experiments This series of experiments investigated how walking speed and density affect physiological arousal. For this purpose, subjects were equipped with electrodermal activity (EDA) and heart rate variability (HRV) sensors (cf. Sec. 4.3, 4.4). Additionally the effect of gender on walking speed and density was investigated. The experiments were performed in the setup of classical single-file experiments, where people walked in ovals behind each other. Overtaking was prohibited. The instructions read, "Please walk one behind the other in the oval until a signal to stop is given. Do not push or overtake". The following parameters were varied (a detailed list of performed runs and combinations of parameters can be found in App. A.5): • number of participants • gender: male, female, mixed • running order by gender: random, alternating Instructions to start or to stop were given by a person standing between the two experimental setups without using technical amplification. Participants walked in the oval for at least 2 minutes or until they had walked one round at very high densities. The width of the ovals walking paths was 0.8 m with a circumference of 14.97 m as measured from the middle of the indicated walking width (Fig. 9a). The course was indicated by colored markers on the floor. Two oval experiments were performed at the same time. They were separated by a wooden wall (Fig. 9b). Cameras were mounted to record the experiment and are listed in Tab. 4. Experimental runs in which EDA and HRV sensors were recorded are listed in Tab. 9 and 10. Trajectories were generated as described in Sec. 4.2. The coordinate origin was located at the lower side of the two ovals' in the axis of the screen wall (Fig. 9a). The data of the Single-File Experiments are provided online [10]. --- Personal-Space Experiments This experimental series investigated physiological arousal when personal space is violated at low densities. Seven participants were positioned within an area marked on the floor and then passed by ten other participants (individually or several simultaneously) from all directions without being touched. Passing Participants were instructed to "Enter the area and walk around until a signal is given to leave the area" and standing participants to "Stand on one of the floor markings. Then remain in place until a signal is given to leave the area". All participants assigned to be standing in the designated spots were equipped with electrodermal activity (EDA) and heart rate variability (HRV) sensors (cf. Sec. 4.3, 4.4). Instructions were given without using technical amplification as the groups were small. Eight runs of four minutes each were performed in total (a list of performed runs can be found in App. A.6). The dimensions of the experimental area were 12.1 m x 5.3 m. The experiment was performed next to the Oval Experiments outlined in the previous subsection. There was no visual shielding between the two experiments and it was possible for participants to pass unhindered between the two experimental sites. Participants with EDA and HRV sensors were placed at positions indicated as red dots in Fig. 10a. Passing participants were participants who were not currently on runs in the neighbouring Oval Experiment. Questionnaires were completed at the end of the whole experiment set including the runs of the Oval Experiment. Cameras were mounted to record the experiment and are listed in Tab. 5. Experimental runs in which EDA and HRV sensors were recorded are listed in Tab. 9 and 10. No trajectories were exported for this experiment. The data of the Personal Space Experiments is provided online [11]. --- Boarding and Alighting Experiments This series of experiments investigated how different parameters influence the boarding and alighting process of a train car. For this purpose, the boarding area of a local train was mimicked. Sliding doors could be opened from outside the experimental area via ropes without interfering with participants. Different parameters were varied (a detailed list of performed runs and combinations of parameters can be found in App. A.7): • number of persons boarding/alighting/staying in train car • groups: none, pairs, groups of 3, groups of 5, mix • norm: none, 70%, 80%, 100% Instructions indicating 'the arrival' and 'the departure' of the train as well as the opening and the closing of the doors were given without using technical amplification from a person standing behind the waiting/boarding pedestrians. Special announcements were made via slips of paper or by directly addressing individuals by the investigators to make individual, targeted announcements when necessary to achieve the study objective. These kinds of announcements as well as handing out luggage were done in the waiting area to the left of the experimental area. Persons assigned to a group got sticky dots of the same color and had the instruction to stay together during the boarding process. For the variation of norm, the percentage in the list above refers to the amount of paper notes prescribing 'normal', considerate behaviour whereas the remaining persons got the information that pushing is allowed. Where questionnaires were completed, this was done after the respective runs in the waiting area. The outer dimensions of the experimental area were 20 m x 20 m. The inner dimensions of the train car were approximately 9.2 m x 3 m (the exact dimensions can be extracted from Fig. 11a) and aim to mimic a typical local train in Germany with the measurements w door = 1.2 m, w const1 = 0.5 m, w const2 = 0.8 m, w const3 = 4.0 m, w aisle1 = 0.9 m, w aisle2 = 2.2 m as indicated in Fig. 11a. Cameras were mounted to record the experiment and are listed in Tab. 6. Experimental runs in which 3D motion capturing data were recorded are listed in Tab. 11. The mood of the participants (c.f. Sec. 4.7) was recorded for all runs. Trajectories were generated as described in Sec. 4.2. The coordinate origin was located on the axis of the front of the bottleneck in the middle between the two bottleneck sides. The data of the Boarding and Alighting Experiments are provided online [12]. 6 5 4 3 2 1 0 1 2 3 4 5 6 X [m] 0 1 2 3 Y [m] --- Tiny Box Experiments This experimental series investigated the relationship between density and physiological arousal while waiting. For this purpose up to eight participants waited in 'tiny boxes' under different conditions. All participants were equipped with electrodermal activity and heart rate variability Sensors (cf. Sec. 4.3, 4.4). The following parameters were varied (a detailed list of performed runs and combinations of parameters can be found in App. A.8): • number of people in box • communication: speaking allowed, speaking prohibited The tiny boxes were four wooden boxes of 1 m <unk> 1 m and a height of 1.5 m (Fig. 12). Participants could enter and exit the boxes through a one-sided swinging door. The experiments were performed in a delivery channel close to experimental site D (cf. Fig. 1) to shield participants as well as possible from acoustic and visual influence of the experiments performed at the same time. Participants were chosen based on age and gender from the respective groups taking part in the experiments at site D. Instructions were given without technical amplification as the groups were small. Where questionnaires were completed, this was done after the respective runs in an area in front of the delivery channel. Cameras were mounted to record the experiment and are listed in Tab. 7. Experimental runs in which EDA and HRV sensors were recorded are listed in Tab. 9 and 10. No trajectories were exported for this experiment. The data of the Tiny Box Experiments are provided online [13]. --- Bottleneck Experiments This series of experiments investigated different physical and social-psychological aspects in a bottleneck scenario. The following experimental parameters were varied (a detailed list of performed runs and combinations of parameters can be found in App. A.9): • bottleneck width: 0.6 m, 0.7 m, 0. The experiments were performed on day 4 at experimental site D. The announcer and further observers were standing on a scissor lift that was parked behind the bottleneck construction and raised to a height of several meters to have a good overview. All announcements were made with a microphone connected to a portable loudspeaker. To increase the initial density, the participants in the first row were asked to stay in place while everyone else was asked to take one step forward, before each run. The intended initial density was 1 P/m 2. Whenever special targeted announcements were necessary to achieve the study objective, they were made via slips of paper or by the investigators directly addressing individuals. In 'normal' condition runs participants were instructed as follows: "You are in a crowd where people walk through a door at a normal pace. You yourself move purposefully, but without haste." In 'hurry' conditions the instructions were "You are in a crowd where people are in a hurry to pass a door. You yourself are also moving briskly", and in 'full commitment' conditions "You are in a crowd where everyone wants to pass through a door as quickly as possible and pushes their way through. You yourself do everything you can to get to the front and through quickly as well." Fig. 13 shows snapshots of two example experiments and sketches of the bottleneck construction. The outer dimensions of the experimental area were 20 m by 20 m. The bottleneck construction consisted of an 4 m x 2 m x 1 m aluminium frame with gray plastic panels, weighing 250 kg per side. They were each visually extended to be 6 m long by adding trade fair walls. Each side was secured against slipping with anti-slip mats and 750 kg concrete blocks which were bolted to the bottleneck construction. Participants were to maintain their motivation until they crossed a finish line 8 m behind the bottleneck. The way to the finish line was marked with barrier tape. Beyond the finish line, participants could return to the line-up area by turning to either sides. Where questionnaires were to be completed, this was done after the respective runs in an area to the left of the experimental area. To take safety precautions, a practised crowd manager was present at the experiments equipped with an air pressure horn. The horn was activated whenever a participant indicated discomfort during an experiment by calling out'stop' aloud or if the crowd manager himself identified a situation as critical or potentially harmful. At the beginning of the day, all participants were trained in what was to be done in the case of the horn being activated. The procedure included immediately stopping in the current position without further movement. Designated helpers that were close to the crowd at all times started tapping people at the shoulder once the crowd had come to a full stop. On the signal of'shoulder tapping' participants were allowed to turn around and move to the far back of the experimental site. The procedure was continued until every person was tapped at the shoulder. Apart from the test run where the horn was activated on purpose, two runs (4D250, 4D280) were aborted and resolved in the way described above. Cameras were mounted to record the experiment and are listed in Tab. 8. Experimental runs in which 3D motion capturing and pressure sensor data were recorded are listed in Tab. 11. Pressure sensor data (cf. Sec. 4.5) and the mood of the participants (cf. Sec. 4.7) were recorded for all runs. Trajectories were generated as described in Sec. 4.2. The coordinate origin was located on the axis of the front of the bottleneck walls in the middle between the two bottleneck sides. The data of the Bottleneck Experiments are provided online [14]. The scientific content of some of the bottleneck experiments in this series is part of the CrowdDNA project. --- Sensors Different combinations of sensory systems were used in the different experiments. This included camera recordings, electrodermal activity sensors, heart rate variability sensors, pressure sensors (at the wall as well as on participants), 3D-motion capturing systems and mood buttons. The design and use of the individual sensors as well as their synchronization in time will be described in the following sections. --- Time Synchronization Between Sensors Accurate time synchronization is required for reliable connection of data from multiple sensor sources. Depending on the sensors' technical settings, different technical solutions need to be adopted to enable synchronization. In an ideal setup, all sensors operate with the same frequency, are connected to the same metronome to capture the exact same instance in time and have the same time code. However, in reality different sensors

Understanding pedestrian dynamics and the interaction of pedestrians with their environment is crucial to the safe and comfortable design of pedestrian facilities. Experiments offer the opportunity to explore the influence of individual factors. In the context of the project CroMa (Crowd Management in transport infrastructures), experiments were conducted with about 1000 participants to test various physical and social psychological hypotheses focusing on people's behaviour at railway stations and crowd management measures. The following experiments were performed: i) Train Platform Experiment, ii) Crowd Management Experiment, iii) Single-File Experiment, iv) Personal Space Experiment, v) Boarding and Alighting Experiment, vi) Bottleneck Experiment and vii) Tiny Box Experiment. This paper describes the basic planning and implementation steps, outlines all experiments with parameters, geometries, applied sensor technologies and preand post-processing steps. All data can be found in the pedestrian dynamics data archive. Keywords CroMa project • controlled experiments • train platform • crowd management • single-file • personal space • boarding and alighting • bottleneck • tiny box • 3D motion capturing • electrodermal activity • heart rate • luggage Collective Dynamics 8, A141:1-57 (2023) Licensed under

for all runs. Trajectories were generated as described in Sec. 4.2. The coordinate origin was located on the axis of the front of the bottleneck walls in the middle between the two bottleneck sides. The data of the Bottleneck Experiments are provided online [14]. The scientific content of some of the bottleneck experiments in this series is part of the CrowdDNA project. --- Sensors Different combinations of sensory systems were used in the different experiments. This included camera recordings, electrodermal activity sensors, heart rate variability sensors, pressure sensors (at the wall as well as on participants), 3D-motion capturing systems and mood buttons. The design and use of the individual sensors as well as their synchronization in time will be described in the following sections. --- Time Synchronization Between Sensors Accurate time synchronization is required for reliable connection of data from multiple sensor sources. Depending on the sensors' technical settings, different technical solutions need to be adopted to enable synchronization. In an ideal setup, all sensors operate with the same frequency, are connected to the same metronome to capture the exact same instance in time and have the same time code. However, in reality different sensors operate on different frame rates, are not pairable with a metronome or might show a drift in time. In order to keep the deviations as small as possible, a global time was introduced and distributed by Tentacle timecode generators [15] and as many sensors as possible were attached to submodule metronomes. --- Camera and Trajectories Camera recordings can be used for experiments in many ways. On the one hand, they make it possible to get a qualitative view of behaviour and to reconstruct the actual execution of the day and any deviations from the plan that may have occurred, as well as to reconstruct announcements and their intonation via the audio track. On the other hand, cameras can be used specifically to obtain measurement results such as extracting walking paths (trajectories) or documenting facial expressions of participants in response to the experiments. In total, 21 cameras were mounted in order to perform the above tasks. Cameras intended for extracting trajectories and serving documentation purposes were mounted under the ceiling (<unk> 8.65 m) facing straight downwards. Camera views were overlapping and mounted in such a way that the occlusion of people was minimized. All cameras used for trajectory extraction were backed up since image loss would have been fatal for the experiments. The approximate fields of view at head height for cameras mounted in the main experimental sites are shown in Fig. 14 and listed in Tab. 2 to 8. The camera types and settings are listed in Tab. 1. Trajectories The trajectory extraction was performed with the pedestrian tracking software PeTrack [17,18] for cameras indicated in Sec. 3.1 to 3.7. Cameras operated with the Simplylive system produced frame drops, double and black frames. The reason could not be reproduced unequivocally. These artifacts were detected and treated before continuing with trajectory extraction. Black frames were detected by applying a binary filter on the grey-scale frames of the video and checking if all pixels were black. Duplicated frames were detected by computing the difference between each frame with the previous one in greyscale. On these differences, DBSCAN [19,20] was used to detect clusters with camera-based parameters. Each frame which did not belong to a cluster was considered to be a duplication of the previous one. Afterward, the videos were reencoded with ffmpeg skipping these erroneous frames. Before exporting the trajectories from PeTrack, the results of the tracking were used to further improve the output data to interpolate the movement between dropped frames. For this, the displacement of each pedestrian in a frame was computed using the Lucas-Kanade method [21,22]. Computing the ratio between these displacements and the average displacement of the previous frame gave the number of missing frames. To implement the mapping from pixel to real world coordinates two types of calibration [23] had to be performed. Intrinsic calibration was performed to take into account the distortion of the lenses and internal hardware combinations. Extrinsic calibration was performed to create a transformation map between the camera and real world coordinate system and was performed every morning with a ranging pole and attached levelling unit. The resulting mean re-projection error for all calibration points for all days and cameras was 1.1 cm with a standard deviation of 0.6 cm and a maximum error of 2.2 cm. However, values differ greatly depending on the camera. The values for the individual cameras are shown in the appendix in Tab. 20. For cameras used for code reading, recognition in the software PeTrack was performed with the code marker method using Aruco Code dictionary dict 6X6 1000 [5]. All other cameras' recognition was performed using the multicolor marker method within PeTrack. After the automatic extraction of trajectories, all runs were manually corrected. To handle the perspective distortion of the cameras for a correct head position in space, the individual heights of each person were accounted for and if a code could not be read a default height of 1.75 m was applied. The different camera views of each experimental area were combined into one single dataset by linear interpolation from the trajectory of one camera view to the trajectory of the other camera view in the overlap region. --- Electrodermal Activity Ambulatory sensors (EDA Move4) from the Movisens company [24] were used for measuring electrodermal activity. A total of 28 sensors were used, which were activated every morning and their data saved every evening. The EDA Move recorded electrodermal activity using the exosomatic method at a constant voltage of 0.5 V. The measurement range is 2 -100 micro Siemens and the sampling frequency is 32 Hz. The sensor was attached to the non-dominant hand of the subjects using a wristband. There were two cables attached to the wristband. These cables connected the two measurement electrodes to the sensor. The electrodes were structural non-woven electrodes with special gel/solid gel and a diameter of 55 mm, which was cut to size if necessary. The electrodes were glued to the palm of the hand below the little finger so that the gel surfaces of the electrodes did not overlap. If the electrodes did not hold well, they were fixed with leukotape. The EDA sensors were always attached by the experimenters and worn for a maximum of one hour. Between different subjects, the sensors were not read. The separation of the data was done in the aftermath by cutting up the individual experiment blocks. The sensor number and the subject number for the day were noted and thus the data of the sensor and the remaining data of the subjects could be linked. EDA data was recorded in runs listed in Tab. 9. --- Heart Rate Variability Movisens ambulatory sensors (ECG Move4) [25] were used for heart rate measurements. A total of 28 sensors were used, which were activated every morning and their data saved every evening. The ECG Move records the heart rate with a resolution of 12 bit. The input range is 560 mV (CM), <unk> mV (DM) and 3 db bandwidth from 1.6 -33 Hz. The sampling rate is 1024 Hz. In addition to the ECG sensor, the ECG Move contains a number of other sensors. These include a 3D-acceleration sensor, which records with 64 Hz and has a measuring range of <unk> 16 g, and a rotation rate sensor with a measuring range of <unk> 2000 dps and a resolution of 70 mdps, with an output rate of 64 Hz. It also has a pressure sensor with a range of 300 -1100 hPa at a resolution of 0.03 hPa, a sampling rate of 8 Hz and a temperature sensor that measures ambient temperature at a frequency of 1 Hz. The sensor was placed below the chest using disposable electrodes. The electrodes contained a highly conductive wet gel and a high quality Ag/AgCl sensor. They had a decentred connection to reduce motion artifacts. The ECG sensors were frequently attached by the subjects themselves. Between different subjects, the sensors were not read. The separation of the data was performed by cutting up the individual experimental blocks. The sensor number and subject number for the day were noted and thus the sensor data and the rest of the subject data could be linked. Heart rate data was recorded in runs listed in Tab. 10. --- Pressure During the bottleneck experiments on day 4 (Sec. 3.7), two pressure sensors from Tekscan (Pressure Mapping Sensor 5400N [26]) were employed to estimate normal forces within a crowd. Each sensor consists of 1768 measurement cells covering an area of 57.8 cm <unk> 88.4 cm. Before the actual data recording, the sensors must be calibrated. For this purpose, the sensor was placed horizontally on a table and successively loaded with 10 kg, 20 kg, 30 kg, 40 kg, 50 kg, 95 kg, 110 kg, or 120 kg in total. Corresponding pressure values were measured with a sensitivity of S-40 and used for a multi-point calibration. On either side of the bottleneck, a pressure sensor was attached vertically at a height of 0.97 m for the lower edge (Fig. 15 a). The short side of the sensor was bent around the corner to place 10 cm of the measurement area inside the bottleneck and 47.8 cm in front of it (Fig. 15 b). Teflon foil was spread over the pressure sensors to reduce shear forces and ensure a secure attachment. Each sensor was connected to a laptop recording pressure with the I-Scan software using a sampling rate of 60 fps. Furthermore, two participants were equipped with flexible pressure sensors on their body, each with two upper arm sensors and one sensor for the back. Xsensor LX210:50.50.05 [27] has 2500 measuring cells providing pressure measurement on an area of 25.4 cm <unk> 25.4 cm on the participant's back. The arm sensor (Xsensor LX210:25.50.05 [28]) covers an area of 12.7 cm <unk> 25.4 cm with 1250 measuring cells. All sensors were calibrated in advance by the manufacturer resulting in a pressure range of 0.14 Ncm -2 -10.3 Ncm -2. For easy wearing, the three pressure sensors were tucked into designated pockets of a specific T-shirt (Fig. 16) and connected to a tablet. The tablet, which was carried in the chest pocket throughout the experiments, used the Software Xsensor Pro V8 to capture pressure at a sampling rate of 25 fps. In order to receive as much pressure as possible at the central part of the back, the volunteers who wore the shirts were 1.91 m and 2.04 m tall. Unfortunately, no pressure data from the Xsensor sensors in the T-shirts were recorded during the experiments. --- 3D-Motion Capturing We used the 3D motion capturing (MoCap) system MVN Link by Xsens to track the full body motion of a person in the crowd [29]. While optical MoCap Systems need a free line of sight between the tracking points on the body and a set of cameras, the Xsens Mo-Cap system uses inertial measurement units (IMU) as sensors. These IMUs measure the acceleration, the angular rate and the magnetic field strength and a line of sight between the body and a camera is not necessary. Therefore, it is possible to capture the full body motion even in dense crowds. Each MVN Link suit (Fig. 17) is equipped with 17 IMU sensors on predefined independently moving body segments. The measurement can be triggered manually and the recorded data is stored locally on a body-pack in the suit. Thus, the measurement is self-contained and the data can be downloaded afterwards. After doing a calibration procedure and taking detailed body dimension measurements, the MVN Analyze software then calculates the full body motion based on a biomechanical model from the measured data set. The processed data includes the orientation, position, velocity, acceleration, angular velocity and angular acceleration of each body segment as well as the angles of joints and the location of the centre of mass. The data can be exported either as xml file or as biomechanical c3d file. Because the IMU based motion capturing is self-sufficient and based on relative measurements only, the absolute positioning in space suffers from a drift which can accumulate over time. The head trajectories extracted from camera recordings, however, have a small positioning error. Therefore, we used a hybrid tracking algorithm [31] to combine both data sets. In particular, that means that the position of the biomechanical model was shifted and rotated to match the head position and orientation of the camera trajectories. On all days, we equipped 20 people with an Xsens MVN Link Motion Capturing system. On experiment days 1-3, these persons were part of the red group (Fig. 2), and on day 4 they took part on experiment site D, namely the bottleneck experiment. 3D motion capturing data were therefore recorded in runs listed in Tab. 11. --- Mood-Buttons In order to be able to classify the mood of the test subjects over the course of the day in the individual experiments, we installed simple mood button terminals (Happy-or-not [32]). The terminals consist of four smiley-faced buttons with a sign saying "how did you feel in the last run?" (Fig. 18) that participants were invited to press after every run. The system saved a time stamp for each pressed button. The terminals were attached to man-rails and placed so that participants passed them after each run. Care was taken to ensure that the grids were positioned in such a way that the walking path was affected as little as possible and no backlog was created. Participants were actively asked to press a button after each run. In the Train Platform Experiments, Figure 18 Terminal showing Mood Buttons placed in some experiments to compare the well-being of the participants between the experimental runs. The question on top translates as "How did you feel in the last run?" from German. the terminal was placed in the corridor leading participants from the area where they filled out questionnaires back to the waiting area in front of the experiment. In the Crowd Management Experiments, the terminal was positioned 15 m after passing the entry gate on the way back to the line-up area. In the Boarding and Alighting Experiments, the terminal was placed next to the waiting area (behind the train car for runs marked as'reversedirection'). In the Bottleneck Experiments, one terminal was placed at each side of the bottleneck. Participants passed the terminal on their way back to the line-up area, regardless of whether they turned right or left after passing the finish line. No mood buttons were placed in the Tiny House, Oval or Personal Space Experiments. --- Summary and Discussion This paper presents pedestrian experiments conducted as part of the CroMa project aimed at increasing the robustness and efficiency of transport infrastructure. Even though the planning and execution of large-scale experiments requires far-reaching planning and organizational steps that go far beyond the scientific content, experiments under laboratory conditions offer the opportunity to control factors and can therefore be worth the effort involved. This publication provides an overview of the individual experiments carried out as well as descriptions of sensor techniques applied, as the contents and goals of the experiments were planned and evaluated by different disciplines and had to be coordinated and combined with each other. Furthermore, it presents the context in which the individual experimental runs and experimental sites were intertwined. The results of the scientific analyses will be published in subsequent content papers. Even though the experiments took place during a global pandemic, the questionnaire results as well as the evaluation of well-being during the experiments (mean value of mood buttons over all days) show that the overall concept of communication, hygiene and safety measures as well as the slow acclimatization to density (queuing, measurement course, waiting area, icebreaker) led to the participants feeling confident. As a result, they felt good and the thought of a potential infection seemed to have no meaningful influence on their actions. This is consistent with the impressions of the organizers regarding the mood of the subjects during the experiments. For each of the conducted experiments, the goal of the study is described along with which parameters were varied, how participants were approached and which dimensions the experimental areas and geometries had. The description is supplemented by impressions of the experiments given through sketches and snapshots. In the chapters about the sensors, the technical specifications are listed. Furthermore, it is documented how the sensors were synchronized with each other, how many of the sensors were used, with which settings they were operated and which basic processing steps were carried out if necessary. For each sensor there is an overview of the runs in which the sensors were used. For each experiment, a link to the data archive is given under which the respective complete data will be made freely available after publication of the respective content paper. days, without whom it would not have been possible to fulfil all tasks and ensure a smooth process on all days. Ethical Review The application of ethical approval for the experiments "Crowd Management", "Single-File", "Personal-Space", "Train Platform" and "Boarding and Alighting" were submitted by A. Sieben to the German Psychological Society (DGPs, the Society) and approved in December 2019 (file reference SiebenAnna2019-10-22VA). The "Bottleneck" experiment was submitted to the ethical review committee of the University of Wuppertal (German: Bergische Universität Wuppertal) by A. Seyfried and was approved in January 2020 (file reference MS/BBL 191213 Seyfried). Funding --- A. Appendices --- A.1. Test Person Sample: Statistics per Day Figure 19 Panel of different histograms showing demographic data of the participants for each day. Row one refers to data of participants on day 1, row 2 to day 2, row 3 to day 3 and row 4 to day 4. Column 1 shows age, column 2 gender, column 3 body height, column 4 body weight and column 5 shoulder width. Data shown in grey includes data from all participants of the respective day, data in green data of participants belonging to the green experimental group and data in red, blue and yellow to data of participants belonging to experimental group of the respective color accordingly. Respective Medians are shown in the same color as the data. --- A.5. Experimental Configurations Oval Experiments --- A.8. Experimental Configurations Tiny Box Experiments

Understanding pedestrian dynamics and the interaction of pedestrians with their environment is crucial to the safe and comfortable design of pedestrian facilities. Experiments offer the opportunity to explore the influence of individual factors. In the context of the project CroMa (Crowd Management in transport infrastructures), experiments were conducted with about 1000 participants to test various physical and social psychological hypotheses focusing on people's behaviour at railway stations and crowd management measures. The following experiments were performed: i) Train Platform Experiment, ii) Crowd Management Experiment, iii) Single-File Experiment, iv) Personal Space Experiment, v) Boarding and Alighting Experiment, vi) Bottleneck Experiment and vii) Tiny Box Experiment. This paper describes the basic planning and implementation steps, outlines all experiments with parameters, geometries, applied sensor technologies and preand post-processing steps. All data can be found in the pedestrian dynamics data archive. Keywords CroMa project • controlled experiments • train platform • crowd management • single-file • personal space • boarding and alighting • bottleneck • tiny box • 3D motion capturing • electrodermal activity • heart rate • luggage Collective Dynamics 8, A141:1-57 (2023) Licensed under

Introduction Historically, non-farmers were not allowed to settle in rural agricultural villages in Israel. According to the decision of the Israel Land Administration, residence in an agricultural settlement was allowed only for those owning and working the land (Alterman, 2017). Resolution 737 passed in 1995 was novel and groundbreaking in that it allowed the agricultural village to construct a number of residential homes for sale to those who do not intend to engage in agriculture. This decision was made as a result of political and other pressures to alleviate the effects of the economic crisis that had devastated agriculture by heavily indebting the farmers to banks and suppliers. Subsequent migration of young people from the rural to the urban space left an increasingly aging population in these villages (Sofer and Applebaum, 2006). This decision as part of the economic recovery plan allowed agricultural communities to benefit from changing the designation of the land from agriculture to construction, enabling farmers to cover, at least in part, the debts they had accumulated (Hananel, 2012). In the kibbutz villages (pl. kibbutzim) whose unique way of life was characterized by social and financial equality, equal partnership of members in the production and the service industries, and a Z. Greenberg (*) Tel-Hai College Faculty Social Sciences and Humanities, Yiftah, Israel e-mail: [email protected] high level mutual responsibility, the economic crisis resulted in debts to banks and private suppliers shouldered by the members of the kibbutz who were also left without savings for the future (Russell et al, 2011). Thus ensued the negative migration of young people from the kibbutz to the urban areas, as they realized that the future of the kibbutz was no longer secure. The social crisis in the kibbutzim manifested in a lack of trust in the kibbutz leaders and managers, a phenomenon of withdrawing into the family unit and abandoning the mutual relations that had so typified kibbutz society, as well as a full-blown ideological-value crisis where the foundations and philosophy of kibbutz society were questioned. In reality, the economic crisis resulted in negative migration and an aging kibbutz population, fewer services offered to the residents, especially those that were not profitable, and an atmosphere of despondency and loss of ideology and future direction (Ben-Refael, 2011). The solutions for easing the crisis in the agricultural villages were handled at the State level and led by the National Kibbutz Organization, and included a recovery plan involving economic, organizational and managerial changes that would allow the continued existence of the kibbutzim and their future growth. The aforementioned Resolution 737, which allowed the kibbutzim to bring in new residents who had no part in the for-profit industries cooperatively owned by kibbutz members, was made with the understanding that the arrival of a new population would relieve the kibbutz and its members of their debts and allow them to continue life in the rural space (Arnon & Shamai, 2010;Greenberg et al, 2016). The construction of new neighborhoods allowed the arrival not only of young urban families searching for a rural quality of life in the revitalized community but also allowed the kibbutz memberfs' own children to return and live in the kibbutz environment without the oppressive burden of the settlement's obligations on their shoulders (Charney & Palgi, 2014). Three groups of residents arrived in these expansion neighborhoods built adjacent to the original kibbutz borders: families from urban centers who wanted to live in a rural environment, families from towns and small cities located near the kibbutzim, typically of lower socioeconomic status, who wanted to improve their quality of life, enjoy services at a high level and improve their defined status, and grown-up children of kibbutz members who wished to return to the kibbutz environment but not as members (Greenberg et al, 2016). The arrival of the new residents led to major changes in the kibbutz: the average age in most of the kibbutzim decreased, the number of children increased dramatically, re-vitalizing the kibbutz day care centers, schools and children's activity centers, the social fabric of the village received a muchneeded injection of energy, and the rural space gained professionals in the areas of management, sales and the independent professions. Some of them were employed by the kibbutz's businesses and industries but many work outside the kibbutz boundaries. In this way, gentrification came to the rural settlements in Israel (Charney & Palgi, 2013), but the arrival of new residents was not without conflicts and disagreements. The kibbutz wanted to independently manage the service systems. The question also arose of the statutory status of the newcomers regarding the management of the community that now included two population groups that sometimes had opposing interests. The kibbutz members viewed the newcomers as mere consumers, those who are required to pay for the services they receive from the kibbutz, and less so as partners involved in the management of the newlyenlarged community. Any attempt by the new residents to share control was seen as questioning their confidence in the kibbutz administration to manage the system. The current study contends that kibbutz members treated the space that would become the expansion neighborhoods in a similar way: the structures and open spaces within and adjacent to the original kibbutz were defined as part of the historical social and cultural capital that they sought to preserve and perhaps even promote to those who were not members of the kibbutz, to come and live in the settlement. The status of the kibbutz members and historical regulations allowed them to influence the planning of the new neighborhoods in addition to planning topography and traffic routes. In this study, we examined the characteristics of these open spaces and the impact of planning on the development of rapport and relationships between the kibbutz members and the new residents, as well as their common social capital. The results were based on an analysis of 67 planning maps of the kibbutzim and their expansion neighborhoods, as well as tours and interviews with kibbutz members, non-member residents and officials in the field of planning. We maintain that villages with spaces designed as meeting areas promote closeness and familiarity and shared social capital. Villages without such common meeting spaces will see more feelings of disappointment, anger and strained relations between those who live in the original kibbutz and those residents of the expansion neighborhood. First we shall present the structural development of the kibbutzim throughout the years and the reasons that led to the planning of expansion neighborhoods. These will be examined with reference to the theoretical terms of social and economic capital and cultural and social space as defined by Bourdieu (1984Bourdieu (, 1989)). The Results section will present findings from the analysis of the maps, tours and interviews; we will propose a typology for defining the demarcation between the two parts of the village. Finally, we shall offer an interpretation based on Bourdieu's theory of the elements of space as a proposal for understanding the characteristics of specific types of planning. --- The theoretical framework for the study --- Open spaces Open spaces are unbuilt areas with natural features of surface and ground rock, as well as flora and fauna, and in which human activity is at a low level. Open spaces contribute to maintaining the level of 'naturalness' in a region that is under construction and development pressures; they often enable the continuing functioning of ecological systems and the survival of local nature and landscape features. There are two approaches to defining open spaces: 1) ecological, according to which the natural spaces importantly preserve abiotic conditions such as water sources, soil, plant types and endemic animal species (Handy & Maulana, 2021) utilitarian, according to which open spaces have social, economic and political significance, since human beings need these areas for the benefit of leisure, sports, and quiet, peaceful spaces where the pace of daily life can be slowed-parks, green rings and green areas that separate different neighborhoods (Firth et al, 2011;Moran, et al, 2017;Tesler et al., 2018). Another distinction concerning open spaces regards those located within or adjacent to large cities vs those outside the cities. In planning the former, the need to preserve natural treasures combines with meeting human requirements for gardens, parks and green corridors in built-up areas as outlets for recreation, active leisure and sports activities, pleasure and environmental quality (Moran et al, 2017;Turner et al, 1993). The latter, those outside the cities, include national parks, nature reserves, and natural landscapes and natural treasures and species where the emphasis in on a future vision of preservation of an ecological balance (Frenkel & Orenstein, 2012). --- Social capital Social capital is an aggregate expression of the joint activity of a group whose members share common values and goals. This type of capital contains familiarity, interpersonal relationships and partnerships, and is a product of social interaction. In the daily life of the community, it is expressed in interpersonal relationships, the degree of the individual's commitment to the group and its goals, sharing, trust and reciprocity (Coleman, 1988). Putnam defined social capital as "features of social organization, such as trust, norms, and networks that can improve the efficiency of society by facilitating coordinated actions" (Putnam et al., 1994, p.167). Social capital affects the individual's sense of well-being. Briggs (2004) distinguished between private social capital expressed in relationships and types of social and personal connections, and public social capital which is common to all individuals and reflects all the relationships, the strength of the shared social cohesion and the group's ability to reach agreements and decisions and promote joint processes. Portes (1998) contended that a group characterized by a high level of social capital can utilize it to promote common interests and accumulate additional capital over other weaker groups located close by. Hence, social capital is significant in building power relations between groups occupying a certain area (Alon-Mozes, 2020). Bourdieu (1989) maintained that social capital symbolizes the group's significance not only for its members but for others who are outside it, observe it and are even influenced at times by its activities. According to Bourdieu (1984), the joint activity of the group members is required to maintain the existing capital that lends the group its status and to accumulate new capital that will strengthen the group's status in the region. Group capital is the outcome of joining economic, social and cultural capital. Bourdieu contends that the connections between these three components reflect the group's history and identity and contribute to the cohesion and mobilization of its members to promote common interests. Pavin (2007) viewed the significance of social capital as a means of strengthening the community and enabling it to survive, asserting that deep and meaningful personal relationships among community members contribute to solidarity and cooperation and enable the whole group to successfully deal with multi-system crises. Gallent (2013) and Medha, and Ariastita, (2017) cited the significance of collaborations in creating external connections and mobilizing resources that promote the future development of the community. However, no validated indicators have yet been found to define social capital as a justified variable; this lack of definition makes it difficult to understand its empirical significance when examining the community and the processes taking place in it (Amoyal & Carmon, 2011). There is another criticism, this one of political significance, according to which the very use of the term'social capital' originates from a socio-economic logic that accepts privatization and reduces the State's commitment to its citizens. This approach dictates that social capital is capital that grows from below, from residents' initiative and activity, and migrates upwards, as a response to the downsizing and privatization that have been occurring since the 1980s (Ferragina & Arrigoni, 2017). An examination of social capital reveals three definitions of such capital. 'Bonding' social capital is built from social networks of those with a strong common identity which may be ethnic, cultural, socio-economic or another. 'Bridging' social capital refers to shared norms and connections of people from different ethnic, racial or social groups whose capital is created during meetings and joint activities with the purpose of promoting common interests of the group members. 'Linking' social capital is capital that represents more all-inclusive connections, less stringent but broader in scope. 'Linking' social capital is weak but important pattern of social capital, which content together different groups living side by side and making up a wider entity such as a nation. Connecting capital is an example of connecting the state and its institutions with civil bodies, associations and communities in order to promote common goals. This capital symbolizes relationships that cross class and social boundaries and is significant in networks that connect actors from different networks, such as joint activity of representatives of civil society and the state to promote common issues (Abbas & Mesch, 2018). Social capital is therefore a consequence of familiarity, connections and activity, all of which have an impact on the development of the space in which they exist (Arisoy and Paker, 2019). Amoyel and Karmon (2011) perceive the significance of social capital as a source of social stability in weak neighborhoods undergoing accelerated gentrification. To date, most writing has focused on the development of social capital in urban neighborhoods undergoing regeneration (Medha, & Ariastita, 2017) Matsuoka & Urquiza, 2022); the current article examines the development of the new social capital with the arrival of new residents in the rural settlements. --- Kibbutz planning as a means of expressing social and cultural capital Since the establishment of the first kibbutz, this specific type of village has been a planned settlement where the conditions of the area, the topography, and the location have all been taken into account, together with the unique, cooperative way of life of its members that expressed values of equality, sharing, providing centralized services of all kinds and being satisfied with few material goods (Khana, 2011). CHyutin and chyutin (2010) noted that the planning of many kibbutzim was entrusted to a few planners who specialized in this type of community and who were employed by the National Kibbutz Movement that oversaw the development of kibbutzim in Israel. They outlined the planning features that can be seen repeatedly in many kibbutzim and that expressed significance values and articulations in the daily life of the settlement (Karniel, & Churchman, 2012). Thus, in the center of each village are several central buildings around which the kibbutz continued to develop: the all-important dining room that served as a meeting place for all kibbutz members three times each day. In addition, single-story buildings whose rooms had central functions -offices of managers and directors, meeting rooms for committee decision-making-and the club-the focal point for social gathering. Outside of these buildings was the large park-like square that served as an informal meeting place and turned this group of buildings into the kibbutz center (Kahana, 2011). The traffic routes inside the kibbutz were pedestrian paths, movement was on foot or by bicycle. All of these gave the kibbutz the sense of a contained area where all the activities of management and daily life could take place in just a small area located in the center of the kibbutz (Chyutin et al., 2010). Lawns and open spaces were planned in a ring around the center of the village where ceremonies, major events and holidays were celebrated. Another space beyond that ring was dedicated to the children's houses where the children lived and were educated (Epstein-Poloush and Levin, 2016). Yet another space was dedicated to the kibbutz members' homes, neighborhoods with modest houses in one-or two-story buildings, depending on the availability and the type of land. Another part of the kibbutz was reserved for the farm structures-where animals were raised, tools and machines were stored, warehouses stood, and in later stages, where kibbutz industries were established (Kahana, 2011). The planning of the new expansion neighborhoods was different from all that had been familiar until now in the kibbutzim. The new neighborhoods were planned outside the borders of the existing village, sometimes adjacent to it and sometimes at a distance from the original kibbutz. They were planned in the style of residential suburbs: private houses, walking paths and vehicle roads throughout the neighborhood. The kibbutz planning committees were often involved in the decisions concerning the location of the new neighborhood, the type of buildings that would be allowed (whether 'free' construction or according to specific models of homes decided upon by the kibbutz), and the planning of the common, open spaces between the kibbutz and the expansion neighborhood. There are three models for the general planning of expansion neighborhoods. While all three place the new neighborhood outside the perimeter of the original kibbutz, each has a different interface between the new neighborhood and the kibbutz. Type A is the peripheral model, where the new neighborhood is adjacent to the original settlement along one perimeter, therefore sitting alongside the kibbutz. In this scenario many of the houses in the new neighborhood are close to those in the original settlement. Type B is the expansion neighborhood that is nearby the kibbutz but has no apparent connection to it because of the large open spaces that separate them. Type C is a lengthwise neighborhood were only a few houses are adjacent to the original kibbutz settlement and the remainder stretch further and further away from it (Fig. 1). In all three models open spaces can be found where there are monuments belonging to and reminiscent of the old kibbutz. The planning of these open spaces and their significance in creating the connection between the kibbutz members and the new residents are the subject of this article. Vol:. ( 1234567890) --- Rural gentrification Gentrification describes the renewal of a neighborhood through the migration of people of a high socioeconomic level into depressed neighborhoods or urban areas where an aging population lives at a lower socio-economic level (Chung, 2021). The new arrivals to this neighborhood find it attractive, take over the space by purchasing apartments and houses, open new businesses that characterize their lifestyle, and develop services that suit them. All of these lead to an increase in the value of the neighborhood's real estate but also increases the price of services in the neighborhood and the general cost of living, making it extremely difficult for the original population to remain there. They are forced to leave and move to another, weaker and more affordable neighborhood (Brummet & Reed, 2019). Levine and Aharon-Gutman (2022) point out that gentrification changes the conceptual framework by which local residents once viewed their neighborhood as an extension of their home, a space of stability, security and belonging. The neighborhood now takes on additional economic dimensions, becoming a source of income and a magnet for investment, and leading the local population to act in an entrepreneurial manner that can profit from the gentrification process. These patterns of migration of a strong population to weak and aging villages is also well known in the rural area (Gosnell et al., 2011) as one result of an economic crisis that affected so many rural populations worldwide and brought the migration of an urban population seeking quality of life to small communities (Phillips & Smith, 2018) far from the big-city lifestyle (Guimond & Simard, 2010;Nelson, 2018) and closer to nature (Phillip, 2005;Schwake, 2021). Rural settlements in Israel also underwent similar processes, as detailed in the Introduction section of this article. The national program for land designation had historically targeted certain rural settlements for agriculture, allowing only those willing to work the land to take up residence there. This declaration prevented gentrification from occurring in rural communities for many years, until Resolution 737 in 1995 that opened the gates of the rural village to a new population, most of it urban, who wanted to enjoy the quiet, close-to-nature life offered in these settlements (Arnon & Shamai, 2010). The establishment of the expansion neighborhoods in long-existing villages such as the kibbutzim brought a relatively large number of new people to the village in a short period of time, people of medium-high socioeconomic status and with managerial and economic knowledge, skills and experience (Greenberg, 2012). The processes that occurred in the Israeli rural space were similar to those happening in cities and other populations. --- Methodology The current study used a mixed-method type of research in which we combined analysis of planning maps with interviews. We analyzed 67 settlement planning maps that had been approved by the official committee for expansion neighborhood plans. Analysis of the maps was carried out using a GIS analysis (Mahmoody and Jelokhani-Niaraki,2021). We examined the distances between the outer edge of the original settlement and the houses of the expansion neighborhood, the number of kibbutz homes and new neighborhood homes adjacent to each other, the location of the open spaces and the uses assigned to these parcels of land. We examined the visibilitythe ability to see the old kibbutz from the new neighborhood and vice versa. All of these were gathered in a table that displays the data collected about each one of the villages examined. This table was the basis for the empirical analysis of the distance between the two parts of each village and of planning and functional characteristics of the open spaces, forming the foundation for the typology that we present in Table 1. In the second phase of the research, we conducted tours and observations in order to better understand people's own perceptions of the significance of the spaces, their daily function and their effect on people's lives. We conducted semi-structured interviews with both kibbutz and expansion neighborhood residents (Flick, 2018). The choice to include both was based on the findings of Arnon and Shamai (2010) on the motivation to connect, as seen in both the members and the newcomers. The interviews were conducted in the participants' homes and in some cases by Zoom due to the Covid-19 quarantine periods. The questionnaire included demographic questions and questions related to the way of life in the village: how often the respondent accesses the open spaces, Vol.: (0123456789) what activities are carried out in them, to what extent they feels that the spaces between the new neighborhood and the original settlement are significant in establishing relationships between local residents, the respondent's degree of participation in the activities of the entire community, what compels them to take part in an activity and what deters him/her, and a request to relate an example of a meaningful experience from a meeting with other residents in his/her early days in the village, and what kind of social connections they have with kibbutz members and with residents of the expansion neighborhood. --- Results --- Findings of the map analysis Three patterns of distance were found between the original kibbutz boundaries and those of the new neighborhood: 1) Up to 37 m between the original kibbutz and the closest point of the new neighborhood. This planning included a road inside both parts of the village settlements and a sidewalk on both sides of the road; 2) A distance of up to 60 m between the original kibbutz and the new neighborhood with an internal road, a sidewalk for pedestrians on both sides and plant beds next to the sidewalk. We considered both of these groups to be in the category of'soft barriers' because they allow motorized or foot traffic; an individual engaging in physical activity could easily cross from one part of the community to the other. In the third group of kibbutzim, the expansion neighborhood and the kibbutz have larger areas separating them, a distance of over 60 m. Table 1 shows the average distances between the kibbutzim and their expansion neighborhoods. In over 60% of the planning maps, we found a relatively short distance between the original kibbutz and the expansion neighborhood, and that the connecting road conformed to the regulations for transportation on urban roads-a two-lane road with sidewalks on both sides which must be at least 35 m wide. For over one-third of the kibbutzim the distance was slightly greater. In 23% of the planning maps, longer distances of more than 90 m were found between the new neighborhood and the original kibbutz. --- Characteristics of open space planning in the maps In this analysis we examined the planned functional characteristics of these spaces. Our questions included: Did the planning propose to change their function? What were their functions prior to the building of a new neighborhood and what did the map expect them to be in the future? Then we examined the types of traffic possible in these spaces and the visibility between the two parts of the settlement. Based on the analysis of the maps, we defined three types of barriers: • Soft barriers-found in about 40% of the planning maps examined. The plan allows movement and spontaneous interaction between residents of Vol:. ( 1234567890) both parts of the village. Local roads, lawns, playgrounds and public buildings that are accessed by the entire population are shared in these open spaces. • Semi-hard barriers-found in about 15% of the planning maps. These appear in the planning maps as public open spaces or as agricultural areas. This type of barrier allows direct visual contact between the expansion neighborhood and the original kibbutz, but does not allow pedestrian traffic. • Hard barriers-found in approximately 45% of the planning maps. Active industrial and agricultural buildings and spaces owned by the kibbutz's agricultural cooperative separate the original kibbutz from the new neighborhoods. These may be buildings dedicated to industry, orchards or groves, or wooded areas. It can be concluded that these are permanent and will not be moved from their place. These barriers do not allow direct visual contact or pedestrian movement between the two residential areas. --- Results from the interviews The interviews revealed that the planning of the open spaces had relevant effects on the feelings, involvement and sense of familiarity of the residents of the new neighborhood. Interviewee #9 from Kibbutz #4 referred to his expectations before arriving in the village: "Between us and the kibbutz there is a wooded area. On the planning map it appeared as a nature reserve and it was like such a dream to live close to nature." [But] "...in my day-to-day life it creates disconnection and alienation, it doesn't create connection. I sit on the balcony enjoying the view from the woods. It's amazing and very different from where I came from. At the same time I know that on the other side is the kibbutz... I'm not really a part of it. Between me and them there is a forest, we are not connected to the kibbutz itself. It hurts me that it is like this because that is not what I intended when I came here." Interviewee #1 lives in the expansion neighborhood of Kibbutz #15. He described his relationship with the kibbutz: "From the moment we arrived there was the feeling that we are not them. We are the new people, the others who don't really belong to the kibbutz. I have to go through a gate and cross the highway to enter the kibbutz that I am a part of. It doesn't make sense. All the public buildings are there, I have to drive in order to go to every meeting, it affects my motivation to go there, to participate." Interviewee #3 lives in the new neighborhood of Kibbutz #7, and described the lack of new social capital: "It's them and us -each in a different part of a settlement. They live in a kibbutz and I live in a neighborhood, the neighborhood is not part of the kibbutz, ours is a village in itself..., it's really a different place. For me, my connection with them is only that I pay them taxes..." Interviewee #5 from Kibbutz #9 described the effect of the neighborhood plan on her motivation: "When we got here I was motivated to be a partner, to take part, that's why we came here. Today it no longer exists for me, I don't feel part of this settlement"... Interviewee #7, a kibbutz member from Kibbutz #4 also expressed a situation of lack of shared social capital with the new residents: "Now in the expansion neighborhood there are some thirty or forty families that we don't even know. Don't know what they look like at all. So we are trying now these days once again to see how we can connect them to here. They see all the announcements [for activities]. They choose not to come. They don't feel a part or a connection." Interviewee #13, a kibbutz member from kibbutz #2: "We hardly go there, there is no reason to. Why go all the way there? To see houses and streets? I don't know anyone there anyway. You hardly see them, there is no real reason to go there." These interviews express the meaning of the way in which open spaces are incorporated into the planning of expansion neighborhoods in rural settlements and their effect on familiarity, the motivation to be involved in activities and the formation of new common social capital in the settlement that has received Vol.: (0123456789) new residents. Long distances and movement barriers suppress the motivation to make connections with the veteran residents who reside in the original kibbutz, to participate in activities, and take part in kibbutz life. This situation makes it difficult to create common social capital. In contrast, in kibbutzim where the distances between kibbutz and expansion neighborhood are small and where there is a soft buffer, processes of getting to know each other and building a sense of belonging between the old and new residents are more likely to occur. Interviewee #13 lives in an expansion neighborhood of Kibbutz #6. She described the importance of the new spaces and their meaning in building a shared sense of belonging: "The large grassy area and the playground contribute to the feeling that we are part of the kibbutz, it began with going there to let our children play together immediately after taking them out of the kindergarten. We parents slowly got to know each other, we spend a lot of time together." Another example of the planning that enabled informal meeting between the two populations came from interviewee #15 who lives in the new neighborhood of kibbutz #5: "At noon I see everyone playing together, and the parents sitting together and talking. For the parents of the younger children, it doesn't matter who someone is or where they live... everyone does everything together." Interviewee #11 from Kibbutz #6 recounted: "The elderly couple who live next door came in and brought a cake. I didn't understand at all what they wanted, I'm not used to it. We were shocked. But this is my memory of arriving here, that they are interested in us, that we are welcome." Interviewee #9, a member of Kibbutz #4 said: "In the beginning there were concerns and we didn't know how it would be with the expansion neighborhood. We were also frightened by the intensive construction. But we got to know each other little by little, you meet them on the walk-ing path, in the grocery store, there is a motivation to get to know each other and make contact and then you realize these are people who came here and want to be part of this place. Today it is a large kibbutz and there are many children once again, and a kind of joy has returned to this place. Interviewee #2, a member of kibbutz #1, said: "I like to walk in the afternoon in the grassy area and the playground. It's nice to see a lot of young people here, little kids. There was a time when it was sad and old and abandoned. The new people brought life and events and culture here, as far as I'm concerned, I feel that I've benefited from them and the kibbutz is renewed." --- Discussion Planning open spaces in a settlement is significant in creating quality of life for its residents (Moran et al, 2017). Areas that promote meeting opportunities between people and populations contribute to increasing friendliness, the motivation to meet, and to advancing joint community activities that contribute to shared social capital (Firth et al, 2011;Tesler et al, 2018). In many of the kibbutzim we found planning of areas that promote members' meeting between themselves and with different population groups in the settlement. Aharon-Gutman (2014) teaches about the significance of open spaces in a city as enabling meeting, rituals and group events that contribute to the construction of a distinct group identity. The findings of the current study show that a certain planning of open spaces also allows meeting between groups; specifically in these cases, between the veteran population that considers itself 'owners' of the settlement and the newcomers who arrived with an aspiration and ideal of being absorbed into the community (Arnon & Shamai, 2010). Shared activities promote a sense of belonging and motivation to participate in both the new residents and the veteran kibbutz members. We would like to emphasize that the new relationships created between the new and old populations have a great impact on the lives of both groups. For the new residents, living in a small settlement while Vol:. ( 1234567890) having left family and friends behind in the urban center lends great importance to personal relationships and local friendships which impart a sense of belonging, security and personal well-being. As for the veteran population, the flight of their younger generation to the cities and a different way of life during the economic crisis, created a 'brain drain' in terms of future kibbutz leaders and few young children in the community's classrooms and playgrounds. The expansion neighborhoods brought a new supply of youthful energy. Designing spaces that encourage some level of relationships contributes to the promotion of the main goal for which these neighborhoods were established (Moran et al, 2017). The characteristics of new social capital-the development of acquaintance, interpersonal relations and partnerships, are all the product of social interaction. Social networks that promote motivation to participate in the community are beginning to take shape; these will also contribute to the developing construction of the common capital of the renewed community's members (Putnam, 1994). Private social capital was the prominent issue in our interviews and to a lesser extent group social capital. These findings correspond to Briggs (2004) definition of group social capital as being composed of relationships, social connections and personal ties. The new social capital that can potentially be created in the shared community of a revitalized kibbutz meets the definition of social capital according to Pavin (2007)deep and meaningful personal ties that contribute to solidarity and enable the entire group to successfully deal with multi-system crises (Greenberg et al, 2016). Phillips and Smith (2018) discussed the symbolic significance of buildings and monuments for the preservation of local identity in the wake of gentrification in small communities. Our findings showed that not only monuments are significant but also the empty spaces that can symbolize the social and cultural capital of the existing population vis-a-vis the newcomers joining them. In our study, we found settlements that built their expansion neighborhoods at a distance from the original village and with nonresidential spaces acting as buffers between the two sections, areas that make it difficult to establish physical contact between the two parts of the same village. These 'hard' barriers affect the degree of familiarity between kibbutz members and new residents, as well as the motivation of both populations to participate in shared community activities. Our interpretation of this situation is that this type of 'out-of-sight, out-of-mind' planning is evidence of the complexity of the decision to create the expansion neighborhoods in the first years of their establishment. The construction of those neighborhoods was portrayed as the best and perhaps the only way to dig the kibbutz out from under its heavy debt after the economic crisis of the 1980s shocked the agricultural villages into a new reality. They realized that kibbutz

This article examines the impact of open space planning on relations and cooperation between locals and new immigrants in rural settlements. In recent years kibbutz settlements have transformed agricultural land into residential neighborhoods for migration of previously urban populations. We examined the relationship between residents and newcomers to the village, and the effect that planning a new neighborhood adjacent to the kibbutz has on creating motivation for veteran members and new residents to meet and build common social capital. We offer a method of analyzing planning maps of the open spaces between the original kibbutz settlement and the adjacent new expansion neighborhood. Analysis of 67 planning maps led us to define three types of demarcation between the existing settlement and the new neighborhood; we present each type and its components and offer their significance in the development of the relationship between veteran and new residents. The active involvement and partnership of the kibbutz members in deciding the location and the appearance of the neighborhood about to be built allowed them to determine the nature of the relations that would be forged between the veteran residents and the newcomers.

and monuments for the preservation of local identity in the wake of gentrification in small communities. Our findings showed that not only monuments are significant but also the empty spaces that can symbolize the social and cultural capital of the existing population vis-a-vis the newcomers joining them. In our study, we found settlements that built their expansion neighborhoods at a distance from the original village and with nonresidential spaces acting as buffers between the two sections, areas that make it difficult to establish physical contact between the two parts of the same village. These 'hard' barriers affect the degree of familiarity between kibbutz members and new residents, as well as the motivation of both populations to participate in shared community activities. Our interpretation of this situation is that this type of 'out-of-sight, out-of-mind' planning is evidence of the complexity of the decision to create the expansion neighborhoods in the first years of their establishment. The construction of those neighborhoods was portrayed as the best and perhaps the only way to dig the kibbutz out from under its heavy debt after the economic crisis of the 1980s shocked the agricultural villages into a new reality. They realized that kibbutz real estate was their greatest asset and that bringing in new families would assure the settlement's future growth and the opportunity to keep the kibbutz afloat financially, socially and ideologically. Faced with this solution to the woes of the kibbutz, they were afraid of changing their unique way of life, afraid of newcomers with different ideologies taking over the administrative systems in the kibbutz, afraid that savvy real estate companies would take over and change their lifestyle, and they were afraid of losing their unique social capital they had accumulated over the years. After all, with the kibbutz's communistic ideology of no private material possessions, they owned nothing else, except the kibbutz itself. Therefore, their statutory 'right' to influence the planning of the expansion neighborhood affected its location, the distance between it and the original kibbutz, the functional characteristics of the open spaces between the two parts of the settlement, and basically, the nature of the relationship between the kibbutz members and newcomers. Theoretically, the open spaces perceived as barriers separating the two parts of today's settlement can be interpreted as how the kibbutz members marked their identity anew as a group with unique ideology, values and lifestyle. The planning of these spaces allowed them to redefine the cultural and social capital they had accumulated and then lost during the crisis years, when the social and political changes in Israel altered the status of the kibbutz members from an elite and storied group to a group facing a longterm multidimensional crisis. The values and ideas had that characterized the lifestyle of the kibbutzim and its glorious history of 'turning a wasteland into farmland' were no longer revered (Amit-Cohen & Sofer, 2016). This may be the manifestation of the real estate view of the locals in the gentrification process. In contrast to the individual thinking in the Vol.: (0123456789) urban neighborhood offered by Levin and Aharon-Gutman (2022), we offer in the case of a kibbutz a common social thinking of the kibbutz members who wish to improve their real estate and they do this through the planning of the open spaces. Faced with their inability to change reality and restore their status as an elite group in Israeli society in general, kibbutz members had the opportunity to mark their uniqueness at the local level, with the planning of the new kibbutz neighborhood. Thus, the re-marking of their capital leaves a stamp that is perhaps the final chapter in the legacy of the once-glorious kibbutz. --- Limitations of the study This study focused on an examination of open spaces and their significance in building the new social capital. There are other dimensions of capital-building such as participation in organized activities, in committees and common forums. Future studies can examine the degree of participation of the new residents in community forums and committees as another means of building the new social capital of the shared community. Future research can also utilize the method of analyzing the planning maps proposed here to examine the same in other types of villages in rural area. --- Conclusions The expansion neighborhoods that began to appear in rural kibbutz settlements beginning in the late 1990s attracted a different population than that of the traditional kibbutz -young urban families seeking quality of life and a back-to-nature existence, but with none of the original kibbutz ideology of working the land and living a frugal socialistic lifestyle. This gentrification of kibbutz land was met with differing attitudes from the kibbutz members, ranging from fear of admitting a new and untested population into their protected life to gratitude for the infusion of a new young generation that would keep the kibbutz economically and socially relevant. The kibbutz members' opportunity to participate in the planning of the new neighborhoods gave expression to their attitudes toward the arrival of a new population. By analyzing the original planning maps of the new neighborhoods we detected three types of demarcation between the borders of the original kibbutz and its expansion neighborhood(s) -soft, semi-hard and hard barriers, each one signaling the degree to which the two populations should function as one unit or remain separate, and whether a new combined social capital could be expected. --- activity substance abuse psychosomatic symptoms and life satisfaction among adolescents. International Journal of Environmental Research and Public Health,15(10), 2134. https:// doi. org/ 10. 3390/ ijerp h1510 2134 Turner, R. K., Turner, R. K., Pearce, D. W., and Bateman, I. (1993). Environmental economics: an elementary introduction. Johns Hopkins University Press. ISBN 0-8018-4862-8 Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

This article examines the impact of open space planning on relations and cooperation between locals and new immigrants in rural settlements. In recent years kibbutz settlements have transformed agricultural land into residential neighborhoods for migration of previously urban populations. We examined the relationship between residents and newcomers to the village, and the effect that planning a new neighborhood adjacent to the kibbutz has on creating motivation for veteran members and new residents to meet and build common social capital. We offer a method of analyzing planning maps of the open spaces between the original kibbutz settlement and the adjacent new expansion neighborhood. Analysis of 67 planning maps led us to define three types of demarcation between the existing settlement and the new neighborhood; we present each type and its components and offer their significance in the development of the relationship between veteran and new residents. The active involvement and partnership of the kibbutz members in deciding the location and the appearance of the neighborhood about to be built allowed them to determine the nature of the relations that would be forged between the veteran residents and the newcomers.

QOL. They found that sex, age, education level, marital status, occupational status, monthly income, drug side effects, comorbidities, body mass index (BMI), type of TB, phase of treatment, stigma, depressive symptoms, and social support were associated with QOL 2,6,16,17. Social support refers to the amount of perceived and practical care received from family, friends and/or the community 18. Previous studies have shown that social support affects the QOL of TB patients 9,19. Patients with adequate social support from family, friends and community are likely to have better QOL 20. Furthermore, social support was also an important predictor of stigma 21. Patients with poor social support are more likely to be isolated and alienated, with manifestations such as being denied shared utensils and food by family members and losing their jobs, which may lead to stigma 22,23. Additionally, good social support will increase life satisfaction and social confidence, enabling patients to adapt to a crisis and reducing the pressure of the patient's role change, thus also reducing the risk of psychological distress 24. Previous studies have also demonstrated that perceived social support is associated with psychological distress in TB patients during treatment 25. Because TB is transmitted by droplets and is highly contagious, patients with TB often experience great stigma, whether at home, in the workplace or in the community 26. Studies of patients from a variety of backgrounds have indicated that between 42 and 82% of TB patients report stigma 27,28. Research has suggested that social stigma may affect life satisfaction in TB patients during and even after treatment 15, and TB-associated stigma is one of the most important aspects affecting QOL 29. Stigma disrupts patients' social interactions with others and reduces social functioning and ability to fulfil daily roles, ultimately endangering patients' QOL 2. In addition, studies conducted in rural China and Ethiopia have shown that experienced stigma is significantly associated with psychological distress 30,31. TB patients who feel stigmatized may less frequently use health services and conceal their illness because of low self-esteem and social isolation. Moreover, studies have reported that TB-associated stigma is associated with psychological stress disorders. These factors can increase the risk of mental health problems, such as psychological distress 21,32,33. The main factor that affects the QOL of patients with TB is psychological distress 34. Studies have indicated that once TB is diagnosed, a wide variety of psychological responses are observed, for example, 51.9% to 81% of TB patients suffer from psychological distress 30,35,36. Studies have also reported that the presence of mental health problems is the strongest predictor of decreased QOL 8, and depression is also believed to be an important cause of poor QOL in patients with chronic diseases 37. Notably, psychosocial burdens may have a greater impact than clinical symptoms in TB patients 34. Psychological distress may interfere with an individual's immune response system and affect adherence to anti-TB treatment, which may lead to poor QOL and exacerbate mortality from the disease 38,39. QOL of TB patients is generally neglected in existing national TB control programs 17, and the lack of research on influencing factors of QOL may be one of the key reasons. As mentioned above, previous studies mainly relied on regression analysis and mostly explored only the direct relationships among variables. The pathways reflecting social support, experienced stigma, and psychological distress effects on QOL remain unclear. Without this understanding, it is difficult to determine precisely which variables should be the primary target of QOL priority interventions. Structural equation modelling (SEM), however, aims to decompose the direct and indirect effects of variables, discover the potential and important associations, and produce a more complete picture of causal effect mechanisms to understand the mechanisms and pathways that might explain these relationships 40,41. In addition, the SEM incorporates measurement errors into the research model, which is more robust than the regression model 42,43. The use of SEM enables us to untangle the complex relationships among social support, experienced stigma, psychological distress and QOL. Understanding the mechanisms and pathways of the relationship among social support, experienced stigma, psychological distress and QOL can help accurately determine the intervention objectives to improve QOL for TB patients, improve the effectiveness of intervention measures, achieve better clinical management, and ultimately increase the possibility of obtaining the best treatment outcomes and achieving the WHO's strategy to end TB. Based on the above theory and empirical research results, we proposed a hypothetical model (Fig. 1). As illustrated in Fig. 1, the current study aimed to test the following hypotheses: (1) social support, experienced stigma, and psychological distress are associated with QOL (H1); (2) experienced stigma mediates the relationship between social support and psychological distress (H2); (3) psychological distress mediates the relationship between experienced stigma and QOL (H3); and (4) experienced stigma and psychological distress are sequential mediators from social support to QOL (H4). According to these research hypotheses, we suggest ways to improve the QOL of TB patients. --- Methods Study design and setting. A cross-sectional, questionnaire-based survey was carried out between November 2020 and March 2021 at three TB medical institutions in Dalian, Liaoning Province, Northeast China. The three medical institutions were selected based on the number of patients attending, type of patient and location. The first is the only tertiary specialized hospital for TB prevention and control in Dalian and is divided into northern and southern parts located in Ganjingzi District and Pulandian District, respectively. It currently has nearly 500 beds, serving the whole city's TB patients, especially critically ill TB patients, as the main medical institution for TB patients in Dalian. The other two institutions are TB dispensaries located in Lushunkou District and Zhuanghe City (a county-level city), which serve only local TB patients with a milder instance of the disease. Participants. TB patients who attended the selected TB medical institutions between November 2020 and March 2021 were recruited as participants. The inclusion criteria were patients with a definite TB diagnosis according to national TB program guidelines, aged 18 years or older and with a new or relapsed case of TB undergoing treatment. The exclusion criteria were patients with psychosis, communication problems, or difficulty understanding the questionnaire and completion of treatment. A total of 481 patients were recruited and completed a structured questionnaire. Of the 481 questionnaires obtained, eight were excluded due to logical errors or large amounts of missing data. Ultimately, this study included 473 TB patients, with a participation rate of 98.34%. Ethics procedure. Ethical approval was provided by the Ethics Committee of Dalian Medical University, Liaoning Province, China. Before participating in the study, each participant was informed of the purpose of the study and how the results would be presented and received guarantees that their personal information would not be disclosed. Each participant voluntarily signed an informed consent form to participate in our study. All methods in our study were conducted in accordance with relevant guidelines and regulations (Declaration of Helsinki). --- Measurement. A structured questionnaire consisting of questions concerning demographic characteristics, treatment status, social support, experienced stigma, psychological distress, and QOL was developed by reading a large amount of relevant literature and consulting experts in related fields. Demographic characteristics included sex, age, marital status and education level. Treatment status included the category of TB treatment, phase of treatment and self-assessed disease severity. Social support. Social support was measured using the Oslo 3-item social support scale, a 3-item questionnaire that is commonly used to assess social support-related issues in clinical and community settings 44. This questionnaire contains questions that ask patients about the number of people they feel close to and on whom they could count on for serious problems, how much people cared about them and the ease with which they could receive practical help from neighbours. Its overall score ranges from 3 to 14, with a high score indicating a high level of social support. In the current study, the scale's Cronbach's <unk> reliability coefficient was 0.718. Experienced stigma. Experienced stigma was assessed using a 9-item stigma questionnaire developed in accordance with Chinese social and cultural contexts 45. The questionnaire assesses the stigma experienced by patients on the three dimensions of prejudice, discrimination, and rejection. Responses to each item were rated on a 4-point Likert scale, ranging from strongly disagree (= 1) to strongly agree (= 4). The scores of each item were summed to obtain the total score (range 9-36). Higher scores indicate higher levels of stigma experienced by TB patients. The scale showed good reliability and validity, and its Cronbach's <unk> in this study was 0.946. Psychological distress. The Kessler Psychological Distress Scale (K-10) questionnaire was used to assess psychological distress in TB patients 46. Numerous studies have demonstrated the reliability and validity of this scale 30. The scale is composed of 10 items divided among four subscales: nervousness, agitation, fatigue and negative affect 47. Negative affect includes hopelessness, low mood, sadness and a sense of worthlessness. An example of such an item is "How often did you feel hopeless in the last 30 days?". The frequency of each item was recorded on a 5-point Likert scale, ranging from none of the time (= 1) to all the time (= 5). The overall score ranged from a low of 10 to a high of 50, indicating an increase in psychological distress. In this study, the scale had high internal consistency (Cronbach's <unk> = 0.929). QOL. The QOL is an index of the satisfaction levels of the body, spirit, family and social life and the overall evaluation of life 48. In the current study, a 6-item quality of life scale (QOL-6) developed by Phillips in 2002 was used to measure QOL in TB patients 49. This scale consists of six items covering physical health, psychological health, economic circumstances, work, family relationships and relationships with nonfamily members. Patients were asked to rate the extent to which the six traits reflected their actual life situation over the past month. Each item was recorded on a 5-point Likert scale, ranging from very poor (= 1) to excellent (= 5). The overall score ranged from 6 to 30, with a high score reflecting good QOL. This scale has been used to assess the QOL of differ-ent populations 17,49,50. In the current study, the scale had acceptable internal consistency (Cronbach's <unk> = 0.792). QOL was parcelled to produce three categories using the item parcelling method for the final model analysis 51. --- Statistical analysis. The complete and correct questionnaires were inputted into the database established using EpiData 3.1 software (EpiData Association, Odense, Denmark) by double entry to ensure the accuracy of the data. The data were exported to SPSS 21.0 (IBM Corporation, Armonk, State of New York) for preliminary statistical analysis. Descriptive statistical analysis included the frequency and percentage of classified data and the mean and standard deviation (SD) of continuous data. T tests and analysis of variance were used to compare QOL scores among different groups. Pearson correlation analysis was used to evaluate bivariate correlations. All comparisons were two-tailed, and P <unk> 0.05 was considered statistically significant. When multiple potential mediating variables and complex relationships were considered in the research model, we used AMOS 23.0 software (IBM Corporation, Armonk, New York, USA) to conduct structural equation modelling (SEM) to test the hypotheses. Confirmatory factor analysis (CFA) was carried out to test the reliability and validity of the constructs and combined with SEM to improve the research model 42. The maximum likelihood method was used to estimate the parameters. Additionally, the 95% confidence interval (CI) was calculated using bootstrapping with 5000 resamples for all effects 52. The bootstrapping performed was a non-parametric test that does not rely on assumptions of normal distribution, and the effect was considered statistically significant if the 95% CI did not include zero. The goodness-of-fit index (GFI), comparative fit index (CFI), Tucker-Lewis index (TLI), standardized root mean square residual (SRMR), and root mean squared error of approximation (RMSEA) were calculated to examine the fit of the model. GFI, CFI and TLI values were greater than 0.900, and SRMR and RMSEA values were less than 0.080, indicating adequate goodness of fit 53. --- Results Participants' demographic characteristics and treatment status. Among the 473 participants, the mean age was 48.36 (SD = 17.58) years, and most participants (60.04%) were aged 45 years or older. There were more than twice as many male participants (69.13%) as female participants (30.87%). There were slightly more participants with a high school education or above (34.88%) than those with a middle school education (33.19%) or a primary education or below (31.92%). Nearly two-thirds of the participants (65.33%) were married, and only 72 (15.22%) had relapsed. More than half of the patients (59.20%) were in a continuous phase of treatment, and nearly one-third (29.60%) felt that their current condition was severe. Among the respondents, the average QOL score was 20.41 (SD = 3.65). Age, marital status, education level, treatment category, treatment phase and self-assessed severity were significantly associated with QOL (P <unk> 0.05) (Table 1). Correlations of the variables. The mean scores of social support, experienced stigma, and psychological distress were 9.71 (SD = 2.27), 18.86 (SD = 7.14), and 19.62 (SD = 7.49), respectively. Social support was negatively correlated with experienced stigma (r = -0.263, P <unk> 0.01) and psychological distress (r = -0.151, P <unk> 0.01) and positively correlated with QOL (r = 0.579, P <unk> 0.01). In addition, experienced stigma was positively correlated with psychological distress (r = 0.453, P <unk> 0.01) and negatively correlated with QOL (r = -0.429, P <unk> 0.01). Psychological distress was negatively correlated with QOL (r = -0.480, P <unk> 0.01) (Table 2). Reliability and validity of the constructs. Through factor analysis, the unstandardized estimates of each item were significant (P <unk> 0.001), and the standardized factor loadings of each item were > 0.5, which met the physical requirements of factor analysis, indicating that each item has a substantial effect on the measurement of latent variables. The CR value represents the internal consistency of the construct. The higher the CR is, the greater the internal consistency of the tested factors. In this study, all CR values were > 0.7, indicating that the constructs exhibited acceptable internal consistency. Moreover, AVE is the average of the explanatory power of the calculated latent variable to the observed variable. The higher the AVE is, the higher the convergent validity. The value of AVE is recommended to be greater than 0.5. In this study, the AVE ranged from 0.500 to 0.819, which implied that the interpretation degree of latent variables with respect to the observed variables was good and the convergent validity of the constructs was high. The values of <unk> AVE s in the diagonal were greater than or slightly lower than the Pearson correlation coefficient of other related constructs. This fact indicates that the discriminant validity among factors is significant and that each factor can be well separated. Overall, these constructs exhibited good reliability and validity 54,55 (Tables 3,4). 2 shows the research model with unstandardized path coefficients. Age, marital status, education level, treatment category, treatment phase and self-assessed severity acted as covariates. Education level was positively associated with QOL (<unk> = 0.131, P <unk> 0.001), while selfassessed severity was negatively associated with QOL among TB patients (<unk> = -0.095, P <unk> 0.05). As shown in Table 6, the total effect of social support on psychological distress was -0.154 (95% CI (-0.245, -0.068) and (-0.243, -0.067)). Social support significantly predicted psychological distress via experienced stigma (95% CI (-0.187, -0.075) and (-0.184, -0.074)). However, the direct effect of social support on psychological distress was nonsignificant (95% CI (-0.117, 0.052) and (-0.114, 0.054)). Therefore, experienced stigma fully mediates the effect of social support on psychological distress. The total effect of experienced stigma on QOL was -0.163 (95% CI (-0.246, -0.088) and (-0.249, -0.090)). Experienced stigma significantly predicted QOL via psychological distress (95% CI (-0.156, -0.095) and (-0.156, -0.059)). However, the direct effect of experienced stigma on QOL was also nonsignificant (95% CI (-0.137, 0.012) and (-0.137, -0.011)). Thus, psychological distress fully mediates the effect of experienced stigma on QOL. The total effect of social support on QOL was 0.524 (95% CI (0.435, 0.635) and (0.435, 0.635)), and the direct effect was 0.463 (95% CI (0.386, 0.562) and (0.384, 0.561)), accounting for 88.36% of the total effect. In addition, social support significantly predicted QOL via the sequential mediation variables of experienced stigma and psychological distress (95% CI (0.017, 0.058) and (0.016, 0.057)), whose estimated multiple indirect effect was only 0.033, accounting for 6.30% of the total effect. In sum summary, all the hypotheses were supported (Fig. 2, Table 6). --- Fit indices of the overall research model. --- Discussion Patients with TB often have symptoms such as cough, chest pain, low fever, fatigue, and loss of appetite. In addition, the treatment of TB is a complex and lengthy process, requiring many medications and a long period of treatment. These factors significantly affect the QOL of patients 17. However, to date, the complex relationships among social support, experienced stigma, psychological distress, and QOL in patients with TB have not been fully explored. To our knowledge, this study was the first to use SEM to explore the interrelationships among social support, experienced stigma, psychological distress, and QOL and to examine whether experienced stigma and psychological distress play mediating roles. In the current study, factor analysis indicated that each construct displayed good reliability and validity, which further verified the stable structure of the scale in TB patients and provides a basis for future studies to measure the social support, experienced stigma, psychological distress, and QOL of TB patients. More importantly, the fitness indices exhibited good model fit, indicating that our proposed research model is reasonable and provides key information for improving the QOL of TB patients. Moreover, this study found that education level was associated with QOL in terms of demographic characteristics. Previous studies have also demonstrated that education level is an important predictor of QOL, such that a higher education level has a positive effect on the QOL of TB patients 2,9. Patients with higher levels of education contribute to greater knowledge about TB from the outside world. Knowledge of TB can improve health-related behaviors such as taking anti-TB drugs on time and seeking care in a timely manner 57. This will contribute to the effective control of the disease and reduce the patients' stigma, thus reducing psychological distress and improving the QOL. However, patients with low levels of education may lack a correct understanding of TB. This often leads to doubt about the ability to cure TB and reduced self-efficacy 58. Patients with low self-efficacy also have stronger stigma experience 59, which increases the risk of psychological distress and affects the QOL of patients. This study also found that patients with perceived severe illness had worse QOL than those with mild illness. Previous studies have also demonstrated that worse physical symptoms are associated with lower physical health-related QOL and higher mental health-related QOL among TB patients 60. Understandably, patients with more severe disease have more complex clinical conditions and longer treatment times, as well as increased patient concerns, which may be particularly damaging to QOL. In addition, it is understandable that the more severe the illness are, the more obvious the symptoms. Obvious symptoms, especially a prolonged cough, may lead to a greater degree of accidental disclosure of the illness. This will have a negative impact on access to social support, increase stigma and psychological distress and threaten QOL. The results also showed that social support demonstrated a significant, direct effect on the QOL of patients with TB. Social support helps improve patients' QOL 20, which has also been found in studies on patients with traumatic brain injury 61. A possible explanation is that patients who receive adequate social support might have improved health outcomes. Moreover, consistent with previous studies, stigma was a predictor of QOL 62. Stigma can damage patients' self-esteem and self-efficacy, lead to patients' isolation from society and self-concealment, and ultimately endanger patients' QOL 63. In addition, psychological distress exerted a direct effect on QOL in our study. Studies have reported that untreated depression is independently associated with poorer QOL 39. Another study also demonstrated that mental distress had a significant effect on QOL 6. Patients with psychological distress were less likely to adhere to treatment regimens, thus eliminating the chance of successful treatment, impairing their function, and reducing QOL 39,64. Previous research has demonstrated that patients who receive an adequate amount of social support are likely to have the best mental health outcomes 20. Our results suggest that social support can also have an indirect negative effect on psychological distress through experienced stigma. Sufficient social support can increase patients' self-esteem and make patients more likely to be diagnosed in a timely manner and to comply with treatment, thus reducing the occurrence of psychological distress 65,66. In addition, our results confirmed that psychological distress moderated the relationship between experienced stigma and QOL. It is not difficult to understand that the experience of stigma will lead to patients' feelings of inferiority, lack of confidence and low emotional well-being, which threaten patients' emotions and cause psychological distress, thus affecting their QOL 33. Our results also indicated that experienced stigma and psychological distress are sequential mediators from social support to QOL, a relationship that has not been demonstrated in previous studies. However, the findings seem logical because patients with better social support have more emotional and financial resources, which makes them face less discrimination and stress, and are likely to use drugs with fewer side effects; thus, they may have improved QOL 65,67. In the current study, SEM was used to test the mediating variables. In epidemiological studies, the assessment of mediation has been widely used to open up the "black box", allowing us to discern complex relationships between variables 68. In practice, understanding the interrelationships among social support, experienced stigma, psychological distress and QOL provides an opportunity to intervene effectively in QOL among patients with TB, and it allows for interventions to be tailored to these specific pathways. Specifically, interventions aimed at improving the QOL of TB patients should focus on increasing social support for patients. At the same time, the role of experienced stigma and psychological distress should also be understood and addressed. Given that experienced stigma and psychological distress mediated the effect of social support on QOL in TB patients, interventions should be combined with measures to eliminate stigma and reduce psychological distress. This was essential for improving the QOL of TB patients. Previous studies have demonstrated that family functions and doctor-patient communication are the most important sources of social support for patients 69. The attitude of family members has an important influence on TB patients. There are widespread psychological burdens among TB patients, such as lack of confidence in a cure and fear of treatment failure 70. Constant encouragement from and care by family members can increase patients' confidence and their feelings of being taken care of. Therefore, family members can be educated and trained to provide better support for patients. Doctors also play an important role in the treatment of TB, and a good doctor-patient relationship is the fundamental factor to ensure the normal operation of the treatment process. It is necessary to require medical staff to establish the concept of patient-centred service and to show respect and humanistic care in the process of medical service delivery 71. It is also important to provide more financial support for patients. Although the country has established some free TB treatment policies, some items, such as the cost of expensive adjuvant drugs, are not included in the free package. Previous studies have also found that TB clubs, composed of health workers and TB patients, have been successful in reducing stigma among TB patients 72. In addition, community awareness and patient education may contribute significantly to a reduction in stigma 73. This study has several limitations that need to be addressed in future studies. First, although SEM was applied, the causal relationship between the variables could not be inferred due to the cross-sectional nature of the data. Therefore, longitudinal studies are needed to validate the current findings. Second, the study sample only included TB patients from Dalian, Liaoning Province, Northeast China, which limited the ability to generalize the results to individuals from other regions with different social and cultural backgrounds. Future research should expand the study area to determine the suitability of our study model. Additionally, the study was limited to TB patients who already had access to health care, while those who did not seek any care were not recruited. The latter are probably the most marginalized and affected by TB, and their participation could enrich our findings. The study also did not include healthy people as controls. Therefore, the results may not capture the impact of TB on patients alone. Finally, only quantitative analysis was conducted in this study, and data were collected through patient self-reports. Patients may hide certain facts, which may cause our results to be underestimated. Extensive interviews and qualitative analysis are needed for a more comprehensive assessment. --- Conclusion This study empirically explored the interrelationships among social support, experienced stigma, psychological distress, and QOL and tested whether experienced stigma and psychological distress played a mediating role. Using the SEM method, we found that (1) social support, experienced stigma and psychological distress affect the QOL of patients with TB; (2) experienced stigma mediates the relationship between social support and psychological distress; (3) psychological distress mediates the relationship between experienced stigma and QOL; and (4) experienced stigma and psychological distress are sequential mediators from social support to QOL. Understanding and managing the QOL of TB patients may lead to better outcomes, and the results of this study provide useful information to help TB patients achieve better QOL. --- Data availability The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request. --- Author contributions L.Z. and X.C. conceived and designed the research and advanced the whole research. X.C. analysed the data and drafted the manuscript. X.C., J.X., Y.C., R.W., H.J., Y.P., Y.D., M.S., L.D., M.G. and J.W. were involved in data collection, entry, and verification. All authors read and approved the final manuscript and agreed to take responsibility for all aspects of the work. --- Competing interests The authors declare no competing interests.

The complex relationships among social support, experienced stigma, psychological distress, and quality of life (QOL) among tuberculosis (TB) patients are insufficiently understood. The purpose of this study was to explore the interrelationships among social support, experienced stigma, psychological distress, and QOL and to examine whether experienced stigma and psychological distress play a mediating role. A cross-sectional survey was conducted between November 2020 and March 2021 in Dalian, Liaoning Province, Northeast China. Data were obtained from 473 TB patients using a structured questionnaire. Structural equation modelling was used to examine the hypothetical model. The research model provided a good fit to the measured data. All research hypotheses were supported: (1) social support, experienced stigma and psychological distress were associated with QOL; (2) experienced stigma fully mediated the effect of social support on psychological distress; (3) psychological distress fully mediated the effect of experienced stigma on QOL; and (4) experienced stigma and psychological distress were sequential mediators between social support and QOL. This study elucidated the pathways linking social support, experienced stigma, and psychological distress to QOL and provides an empirical basis for improving the QOL of TB patients. Tuberculosis (TB) is a major infectious disease that poses a serious threat to human health and has significant negative social and economic consequences 1,2 . It leads to poor health for millions of people every year and is a major public health problem 1 . In 2019, there were an estimated 10 million new cases of TB worldwide, of which approximately 833,000 were in China, accounting for 8.4% of the global total, ranking third 1 . TB is also the leading cause of death from infectious diseases globally, and approximately 1.41 million people died of TB in 2019, of whom approximately 33,000 died in China, accounting for 2.4% of the global total 1 . The burden of TB remains high in China. Although the suffering caused by TB has been acknowledged for thousands of years, most current TB programs and research have primarily focused on detection, microbiological treatment, prevention, and control, while the quality of life (QOL) of TB patients has been neglected 3,4 . Although effective anti-TB drugs are available and TB patients have access to effective treatment, TB infectivity, chronic progression, long-term drug treatment over a period of at least 6 months and drug side effects have significantly affected patients' daily lives, thus affecting their QOL [5][6][7] . Research has confirmed that TB patients tend to have poor QOL, demonstrating QOL significantly worse than that of the general population 8,9 . The World Health Organization (WHO) defined QOL as an individual's perception of their position in life within the cultural context and value system in which they live and in relation to their goals, expectations, standards, and concerns 10 . In addition, QOL refers to a person's subjective assessment of their life's satisfaction and meaning 11 . QOL can affect treatment adherence in TB patients, while non-adherence to TB treatment is thought to be an important reason for the gap between high financial inputs and poor performance in TB control [12][13][14] . More importantly, impairments in QOL are associated with poor treatment outcomes, which can increase TB mortality and morbidity and negatively impact TB control 15 . Therefore, it is necessary to explore the factors that influence the QOL of TB patients and to improve the QOL of TB patients. Previous studies have analysed factors associated with

INTRODUCTION The formulation of national education goals in Law Number 20 of 2003 mostly discusses forming attitudes and values. Multicultural diversity (customs, culture, language, arts, local skills, etc.) in Indonesia serves as a basis for shaping and enriching these values for every individual in the country. On this ground, such diversity must be preserved and developed without taking out the noble values of the Indonesian nation through education. Introducing the students to the environmental, social, and cultural conditions enables them to understand the values and identity of Indonesian citizens in-depth. In the end, this process is expected to bolster the improvement of human resources and the student's cognitive, psychomotor, and affective skills. The government plays the integration of local culture into the national education curriculum through Ministerial Regulation (Permen) No. 79/2014 concerning the Local Content of the 2013 Curriculum. This regulation underscores the role of local content as teaching materials at various levels of education to educate students about local potential and characteristics. This relevance is closely related to research variables that discuss the development of local culture in the basic education curriculum. The enactment of Law of the Republic of Indonesia Number 32 of 2014 concerning Regional Government provides a new nuance in education management. Regional autonomy in the implementation of education is the embodiment of the law, which gives authority to the regional government to create education management that fulfills the needs and conditions of the region. With the authority, the regional government is expected to develop a local content curriculum as one of the compulsory subjects in every educational unit, especially primary school. As mentioned in Regulation of the Minister of Education and Culture Number 057 of 2014, "learning content and guideline of group B general subjects (art, culture, and craft and physical, sports, and health education) is national-scale subjects developed by the government and can be enriched with local content by the regional government and educational units." Banggai Regency is among the regions with distinctive and diverse local cultural characteristics. The people living here come from various tribes with different languages and customs. Generally, the indigenous people of the Banggai Regency come from the Banggai, Balantak, and Saluan tribes. Such diversity is shown in their daily activities, whose languages are Saluan and Balantak. The language the indigenous people use becomes one of the slogans sounded by the local government regarding cleanliness, namely "PINASA (Pia Na Sampah Ala, Pile Na Sampah Ala, Po Kitayo Sampah Alayo). In addition, such cultural diversity is linked with formal events, e.g., weddings and welcoming guests. Traditional clothes and music of the Banggai regency also color the events. Besides language and customs, the Tourism Office of Banggai Regency develops the local culture's potential through the handcrafts of the local community. The results of the preliminary observation performed in several primary schools in the city of Luwuk showed that most educational units, specifically primary schools in Banggai Regency, had incorporated local languages (Saluan and Balantak) into the local content curriculum. This aligns with the Regulation of Banggai Regent Number 56 of 2017 concerning the Curriculum of Local Content of Local Languages (Saluan and Balantak). According to a teacher in one of the primary schools in Luwuk, the primary school teachers taught the aforementioned local languages in songs, fairy tales, speeches, and the like. It is interpreted from the information that a small part of the local cultural context of Banggai regency has been implemented in basic education. However, it has not been carried out optimally, and the results have not been achieved by the learning objectives contained in the curriculum. For such reasons, a study was conducted to analyze deeply the local cultural content mentioned previously and the factors influencing the achievement of the curriculum and the learning objectives. The curriculum structure of primary/Islamic primary schools is under Ministerial Regulation Number 57 of 2014, consisting of Group A general subjects and Group B general subjects. Group A general subjects are curricular programs intending to develop students' attitudes, knowledge, and skill competencies to strengthen abilities in the life of the people, the nation, and the state. Group B general subjects are curricular programs aiming to develop students' attitudes, knowledge, and skill competencies regarding the environment in the social, cultural, and artistic fields. Utilizing local culture in the curriculum through the concept of local content aligns with BNSP guidelines. Local content develops competencies according to regional characteristics, includes elements of excellence, and is an independent subject. Each educational unit formulates Competency Standards and Basic Competencies, enabling the integration of local culture into the basic curriculum according to research variables and strengthening local content by the Education Unit Level Curriculum Operational Guidelines, which emphasize curriculum design responsive to local social and cultural realities. This is in line with the Explanation of Law Number 20 of 2003 concerning the National Education System, which underlines the role of local content as a student's understanding of the potential of the area where they live. Mulyasa's perspective (2009) describes the Local Content Curriculum as competency development that adapts to regional characteristics, including regional advantages determined by the education unit. Nonetheless, the goal is to link the needs of families and communities with national education goals. This perspective is by the research focus on developing local culture in the basic education curriculum. It is also important to note that this subject gives opportunities to the students to develop the skills considered necessary by the region. As a consequence, the local content subject must contain the characteristics of local culture, skills, and local values and discuss social and environmental issues, which can ultimately equip students with basic life skills. The local content curriculum is a set of plans according to the conditions and needs of the region and methods employed as a guideline in implementing the teaching and learning process. Specifically, the local content curriculum aims to a) introduce and familiarize students with their natural, social, and cultural environment; b) equip students with abilities, skills, and knowledge of their regions that are useful for themselves and the community in general; c) make students behave in harmony with the values or rules that apply in their regions, and preserve and develop the cultural values to support national development; d) make students aware of the environment and the problems and can help find solutions (Walisman, 2007: 94). A study conducted by Achmad Basari in 2014 entitled Penguatan Kurikulum Muatan Lokal dalam Pembelajaran di Sekolah Dasar concludes that the school's limitation in developing local content curriculum by regional potential in preparing students to have abilities is due to the following factors: lack of human resources to develop local content curriculum following regional potential; lack of understanding of teachers/school members in developing local content curriculum; the previous local content curriculum is considered adequately effective for students. Additionally, Rofiq, in his 2010 research, elaborated that the wealth of local culture can be used as a repertoire of local content in the educational unit-level curriculum. With the inclusion of local culture as local content, it has a dual function at once. First, a manifestation of demands for guidelines for developing school curriculum to eliminate the inferiority of religious education in educational arrangements. Second, an effort to eliminate culture shock is mostly experienced by the students, making them find compensation in other cultures, which often resist religious values. Drawing upon those previous studies, local culture can be integrated into the basic education curriculum, particularly in the local content subject. The purpose of this integration is to shape the characters/attitudes of the students in harmony with religious values and filter Western cultural values and the negative effects of globalization that increasingly affect the young generations' personalities. --- METHOD The following stages were carried out in analyzing local culture in the basic education curriculum. --- a. Curriculum context identification Identifying the curriculum context included the analysis of characteristics, potential, excellence, local wisdom, and needs/demands of the region. The method of identification and analysis was adjusted to the team's capability. b. Analysis of local culture to be developed in the basic education curriculum This kind of local content encompassed four clusters of local content serving as the interface between local culture (social, cultural, and political dimensions), entrepreneurship, pre-vocational (economic dimension), environmental education, and other local specialties (physical dimension). During this stage, there was also an analysis based on the following criteria: a) conformity with the development level of students; b) teachers' abilities and availability of educators; c) availability of facilities and infrastructure; d) no conflict with religions and nation's noble values; e) no social and security vulnerability; f) appropriateness related to the implementation in educational units; g) characteristics by the conditions and situation of the regions; h) local content needs analysis component (characteristics, potential, excellence, and needs/demands); i) developing basic competencies referring to the core competencies; j) compiling local content syllabus. The data came from community leaders/traditional leaders, school members (principals, teachers, and students), and cultural observers. For school members, this study took a sample of several schools that would become research subjects. Sample taking was based on the number of population (overall area of primary schools) or, in other words, all primary schools in Banggai Regency (340 schools) from the data on the kemdikbud.go.id. On the other hand, the accessible population was primary schools in the districts that lived by most indigenous people of Banggai Regency, i.e., Balantak and Saluan, and those in Luwuk as a comparison. Accordingly, primary schools becoming accessible to the population were those in Balantak District (with a majority of the Balantak Tribe), Kintom District (with a majority of the Saluan Tribe), and some schools in Luwuk, including in Luwuk District, Luwuk Utara District, and Luwuk Selatan District. Moreover, data collection techniques relied on observation, questionnaires, interviews, and documentation. The collected data from interviews, observation, and documentation were analyzed using a qualitative method. Meanwhile, the survey or questionnaire data were analyzed using descriptive statistical analysis. --- FINDINGS AND DISCUSSIONS --- Curriculum Context Identification Identifying the curriculum context according to the Regulation of the Minister of Education and Culture of the Republic of Indonesia Number 81 A of 2013 includes the analysis of characteristics, potential, excellence, local wisdom, and needs/demands of regions. Below is the description of the characteristics, potential, excellence, and needs of Banggai Regency viewed from geographical, natural resources, human resources, and cultural and historical potential a. Analysis of Local Content Curriculum Context Viewed from Geographical Potential Banggai Regency is an autonomous region in Central Sulawesi Province, with Luwuk as its capital city. The regency is located at coordinates between 122023' and 124020' East Longitude and 0030' and 2020' South Latitude. It has a land area of <unk> 9,672.70 km2 or around 14.22% of the total area of Central Sulawesi Province and a sea area of <unk> 20,309.68 km2 with a coastline of 613.25 km. Administratively, Banggai Regency is divided into 23 districts, 291 villages, and 46 sub-districts. The boundaries of Banggai Regency are as follows: 1) To the north, it borders the Tomini Bay. 2) To the east, it borders the Maluku Sea. 3) To the south, it borders the Banggai Kepulauan Regency and Banggai Laut Regency. 4) To the west, it borders the Tojo Una-una Regency, Morowali Regency, and Morowali Utara Regency. Regarding morphological conditions, Banggai Regency has natural diversity with mountains, rivers, and small islands. Villages in Banggai Regency are generally located at an altitude of less than 500 m above sea level, with the land surface dominated by the mainland and hills. Mountains surround the villages, yet less in number than the mainland and hills. Hills or mountains generally dominate villages 500-700 m above sea level. In contrast, villages located at an altitude of higher than 700 m above sea level are commonly dominated by mountains. This geographical potential has led us to the fact that the Banggai Regency is surrounded by neighboring regencies, i.e., Kepulauan, Banggai Laut, Morowali, Tojo Una-una, and Maluku Utara. This offers good opportunities for economic turnover, entrepreneurship, and tourism. The strategic location of the capital city of Banggai Regency also gives opportunities for information, communication, and transportation access between districts in the regency. With such natural diversity, Banggai has excellent potential to develop many natural resources, including agriculture/plantation, fishery, industry, and tourism. Nevertheless, some obstacles arise along the way, i.e., lack of cooperation between school members, especially the local content curriculum development team, with local government agencies regarding implementing the local potential in developing local content curriculum. The schools and the agencies are not far in the distance, not to mention the reliable communication and information access supported by local content potential across Banggai Regency. Hence, it is essential to have intensive communication and cooperation between school members through the local content curriculum development team and local government agencies concerning the utilization of Banggai Regency's potential in developing the curriculum mentioned earlier. --- b. Analysis of Local Content Curriculum Context Viewed from Natural Resources Potential Banggai Regency is rich in natural resources, such as agriculture, plantation, forestry, fishery and marine, mining, energy, and tourism. This natural wealth is a potential for developing the local content curriculum. Elaborated below is the detail of each natural resource. --- 1) Agriculture The agricultural sector still plays a significant role in determining the economy of Banggai Regency since farming is the primary source of livelihood for most residents. This is consistent with the data that approximately 93.98% of all villages/sub-districts in this regency are areas with potential for food crop agriculture, as displayed in the following table. --- 2) Plantation The commodity of plantation crops in Banggai Regency is quite strategic. It contributes to the region and creates jobs in its production activities (business and marketing), reducing the unemployment rate. The leading plantation production in this regency is oil palm plants, with a production of 70,719.55 tons. Crops that need improvement in planting and production are flat-leaved vanilla, producing 1.50 tons. --- 3) Plantation Banggai Regency is a large meat supply area in Central Sulawesi Province, as presented below. --- 4) Fishery and Marine Banggai Regency has a sea area of <unk> 20,309.68 km2 and is located in Zone II of Tomini Bay and Zone III of Tolo Bay. The length of the coastline of Banggai Regency is <unk> 613.25 km. The sustainable potential of marine fisheries is <unk> 48,621.1 tons per year, consisting of pelagic fish with <unk> 39,387.9 tons/year and demersal fish with <unk> 9,239.2 tons/year. The regency also possesses an aquaculture area of <unk> 8,825 Ha. The cultivated types are tiger prawns (exports to the USA, the European Union, and Asia), Vannamei, and milkfish. A freshwater pond with an area of <unk> 260 Ha is for cultivating carp and Nile tilapia. For mariculture, the existing potential is <unk> 6.396 Ha, of which <unk> 78,800 m2 has only been utilized for seaweed cultivation (<unk>44,750 m2), cultivation of groupers/floating cages (<unk> 430 m2), Scylla crabs (<unk> 30,300 m2), and cultivation of pearls. Most areas in Banggai Regency are in the coastal area; therefore, they are dominated by coastal villages with a total of 131 villages (data before restructuring) with 14,067 fishing households. --- 5) Mining and Energy Banggai Regency has abundant mining potential, consisting of oil and gas and metal mineral mining materials, such as nickel and gold, non-metallic mineral mining materials, or non-metallic mineral materials (mine c). --- 6) Industry The industrial sector in Banggai Regency is engaged in large, small, and home industries. Given in the table below are the types of industries. --- 7) Tourism Banggai Regency owns a variety of natural tourist attractions and tourism potentials spread across several districts. However, all still needs improvement to make it a tourism icon of Banggai Regency that attracts local, domestic, national, and international tourists. Here is an overview of tourist attractions and tourism potentials of Banggai Regency. The above potential of natural resources indicates that Banggai Regency is rich in its various resources. Extensive agricultural/plantation land, great productivity, marine fisheries containing different types of fish and marine biota, a variety of types of community craft industries, and natural tourist attractions across districts in Banggai Regency provide comprehensive knowledge and information in the development of local content curriculum that bolsters the regional development of this regency. Nevertheless, the shortcoming of this curriculum is that the learning process in schools has not incorporated the information of such natural resource potential. It is essential to rearrange local content in the basic education curriculum by incorporating information related to natural resource potential in agriculture/plantation, fisheries, industry/crafts, and tourism. This process is expected to provide knowledge to the students in utilizing their natural resources, e.g., improving the use of agricultural/plantation land, marketing fishery products, entrepreneurship industry, and tourism promotion. --- c. Analysis of Local Content Curriculum Context Viewed from Human Resources Potential Human resources refer to people with potential that can be utilized and developed to become adaptive social beings (being able to adapt to natural challenges, development of science and technology, and socio-cultural changes) and transformative social beings (being able to understand, translate, and grow all of their experiences and social contacts for the benefit of themselves and the environment in the future). Therefore, they can use the potential of the natural surroundings in a balanced and sustainable manner. Human resources are key to the success of all aspects/potentials of local content. This aspect can positively and negatively impact the quality of local content to be developed, depending on the people's paradigm, culture, and work ethic. Local content is not realized and implemented without involving and placing humans as a central aspect. To support the implementation of local content curriculum, there are four private colleges, 25 vocational schools, 51 senior/Islamic senior high schools, 142 junior/Islamic junior high schools, 385 primary/Islamic primary schools, and several groups of art and culture in Banggai Regency. --- d. Analysis of Local Content Curriculum Context Viewed from Cultural Potential The dominant cultural identity in Banggai Regency is the culture of Banggai, Balantak, Saluan, and Andio. Tribes in this regency include Balantak, Saluan, and Andio tribes, all of which have their characteristics in terms of language and art. The Balantak tribe is located in Banggai Regency, which has been present with its customs for a long time. In its history, this tribe used to be a fusion of seven community groups in the family of pitu bense tompotika staple bondolong, who spoke the same language, namely the Gombe language. The language has been preserved and used as an everyday language better known today as the Balantak language. The Saluan tribe is a tribal community with a distinctive name, the Loinang people. Most Loinang people work as farmers who grow crops on the land or create their fields near the villages. The language used by the tribe with a population of 172,670 people (based on the 1985 population census data, in cultural ideas) is the Saluan language. Meanwhile, the Audio tribe occupies the Tompotika mountain valley and inhabits the Masama district, Banggai Regency. 80% of the population work as farmers and gardeners and 20% as fishermen and office workers. This tribe communicates with the Andio language, different from the one used by the other three tribes in the Banggai Regency (Banggai, Balantak, and Saluan languages) Moreover, the Banggai tribe mostly inhabits the islands, namely the Banggai Laut and Banggai Kepulauan regencies. Previously, these regencies were part of the Banggai Regency. Only then did they experience restructuring into two regencies. The language used by most of the Banggai tribe is the Banggai language. However, there have been many changes due to the cultural fusion with various tribes in the regency, including the Bugis, Gorontalo, Jawa, and Buton tribes, making the native language of the Banggai tribe only spoken in the Bulagi area, the Sea-sea tribe. Apart from the language, tribes inhabiting the Banggai Regency have cultural diversity that still exists today, such as Tumpe, Sumawi, Ande-ande, Osulen, Tontila, Umapos, Balatindak, and others. One frequently performed culture is the Molabot Tumpe traditional ceremony held annually in September, coinciding with the laying of the maleo bird (Sulawesi endemic bird living in the Bakiriang area, Batui District). This ceremony is a series of past customs of the Banggai Kingdom that had historical ties to the establishment of Banggai Kepulauan and Banggai Laut regencies. All in all, Banggai Regency has a variety of cultures with characteristics from its linguistic and artistic aspects. However, more or less the same as the abovementioned natural resource potential, the cultural potential in Banggai Regency has not been implemented as part of school lessons except for the language --- e. Analysis of Local Content Curriculum Context Viewed from Historical Potential The historical potential is in the form of relics and heritage traditions that are still preserved presently. If their management is optimized, historical concepts will be a tourist attraction that can also become an asset or a local advantage of a certain region. Consequently, preserving traditional values by giving a new touch to combine traditional and modern interests is important, meaning that the historical asset or potential can be part of local content. Given below are the data on the historical potential of Banggai Regency. Similar to the cultural potential, Banggai Regency also has historical potential in the form of historical heritage. Unfortunately, the people of Banggai Regency, especially the younger generation (students), have not widely recognized the historical heritage. This historical heritage can be part of local cultural content in the basic education curriculum to provide knowledge to the students regarding the history of their region, along with becoming part of the tourism promotion of Banggai Regency in supporting the region's progress. --- Analysis of Local Culture to be Developed in the Basic Education Curriculum Based on the identification of local culture described previously, an analysis is carried out to determine local cultural content to be developed in the basic education curriculum. --- a. The Conformity of Students' Development Level The stages of student development have distinctive characteristics and separate developmental tasks useful as directions for normal development. On the contrary, success in performing developmental tasks creates a feeling of pride and happiness. Universally, maturity ends in optimizing the development of the human soul, which can only be reached when it goes through a process toward the ultimate goal of human personality development (Ilahi, 2012;25-26). Education in the contemporary context is an effort to develop and encourage humans to appear more progressively based on high values and noble life to create a perfect being related to reason, emotion, and action. The culture of Banggai Regency has significant relevance to Law Number 05 of 2017 concerning the Advancement of Culture. Thus, the government of Banggai Regency prioritizes education as a vital facility to empower the local culture to create civilized younger generations of Banggai Regency and become heirs to local culture who can maintain and preserve their genuineness in social life. --- b. Availability of Educators Implementing the curriculum of local cultural content must be supported, which are considered essential elements for applying local content in an educational unit. As stated in the Regulation of the Minister of Education and Culture of the Republic of Indonesia Number 81A of 2013 concerning the Implementation of Curriculum of Local Content Development Guidelines Point D, teachers are the supporting capacity for implementing local content (item 2). The teachers assigned to the local content subject must have a) ability or expertise and graduated from the relevant field; b) experience in the covered field; c) a strong interest in the covered field. Local content teachers can come from outside the educational units, such as the nearest educational units, community leaders, socio-cultural actors, etc. --- c. Oriented to Local Culture Introduction As affirmed in Law Number 05 of 2017 concerning Advancement of Culture Article 28 section 1). Central and regional governments must publish information on the inventory, security, maintenance, and rescue objects of cultural advancement; 2) Everyone can play an active role in publishing information relating to the inventory, security, maintenance, and rescue of advancement of cultural objects; 3) Publication is done to disseminate information to the public within the country and abroad with various forms of media. The matter mentioned above is part of the elaboration of local cultural potential that emphasizes protection, development, utilization, and supervision so that culture grows and develops, making it recognized by younger generations through school subjects. Hence, the use of the local cultural content curriculum as a subject is a concrete step for the government of Banggai Regency through Regent Regulation Number 56 of 2017 concerning the Curriculum of Local Content of Local Languages (Saluan and Balantak). It functions as part of 1). introduction to the local culture of Banggai Regency on the national and international level, 2). Regional development on the national and international level, 3). As a form of devotion to the local potential that is explored from the identity of the ancient people in terms of language, art, and traditional games. --- d. Covering Dimensions of Knowledge, Attitude, and Skills From a human's perspective, education is fundamental to students' learning process. In a universal setting, education positively affects developing and cultivating the student's creativity, skills, and personality. Consequently, the function of education is believed to mainly cover the needs of students to get motivation and stimulation in performing their learning activities. This motivation and stimulation have a significant impact on achieving the growth of three basic components as basic educational functions, namely cognitive aspect (knowledge), affective aspect (personality), and psychomotor aspect (skills). --- e. Availability of Facilities and Infrastructure To integrate the development of local cultural content into a subject in primary schools, it is necessary to have facilities and infrastructure that meet its minimum criteria. The minimum criteria of facilities comprise furniture, educational equipment, media, books, other learning resources, information and communication technology, and other equipment that every school must have. Meanwhile, the minimum criteria of infrastructure consist of land, buildings, rooms, and power installations and services that every school must have. --- f. Having No Conflicts with Religious Values Education positively affects the development and cultivating the students' creativity, skills, and personality. Law of National Education System Number 20 of 2003 Considering National Education System Article 1 section (1) describes the substance of education as a conscious and planned effort to create a learning atmosphere and learning process so that students actively develop their potential to have religious, spiritual strength, self-control, personality intelligence, morals, and the skills needed by themselves, society, nation, and state. Through the development of this local cultural content, education in Banggai Regency is expected to be a pathway to carry on culture or instill the ability to behave, alongside teaching skills and knowledge and playing a role in enhancing the personality of students to be religious. This requires the creativity of teaching the development of local cultural content as an applicable and targeted part of guiding great and noble students in everyday life at school, at home, and in the community. Through local cultural content subjects, primary school teachers are expected to create students with noble characters and be helpful to the nation, state, and religion. The curriculum of local cultural content development integrated into a subject, namely arts, culture, and crafts, in Group B, is expected to lead the process of moral actualization and developing spiritual, noble, and ethical characters with life values as a guide for us to a better and more meaningful life. Thus, this can bring forth moral nobility through piety in religious rituals, building human interaction, and respecting the environment. --- g. Being Able to Counteract Social and Security Vulnerability Banggai Regency has a lot of local wisdom and cultural content, as stated in the "Document of Cultural Ideas of Banggai Regency in 2018", including the tourism potential of nature, beach, history, music and dance, and traditional games and sports. It can be part of the resolution to prevent the region's social vulnerabilities and security threats if managed properly. We have a cultural philosophy, "Mompo Saangu Tanga Mombulakon Tano," meaning Put Our Hands and Thoughts Together to Build the Nation/Region. Moreover, the morality movement "Pia Na Sampa Ala," which unites elements of the life of Banggai Regency people from Luwuk to the border of the Obo Balingara (Nuhon) and Keramat Valley (Toili Barat) areas, means "see trash, take it." This movement does not only teach a concept of "taking" but also moral awareness in the sense that any trash we see is picked up and disposed of in the trash bin. This philosophy comes from the cultural characteristics of Banggai Regency with the capital city of Luwuk called "BERAIR" (Bersih or clean, Aman or safe, Indah or beautiful, and Ramah or friendly). If managed properly, the very multicultural areas, including Saluan, Balantak, Banggai, Java, Bugis, Bali, Buton, Muna, Padang, Lombok, and others, will be able to compete with other regions, such as Raja Ampat, Bali, Bunaken, Wakatobi, Lombok, and even Yogyakarta. --- h. Developing Basic Competencies by Referring to the Core Competencies In every syllabus, a basic concept that needs special attention is integrating each core competence 1, 2, 3, and 4 within one unit or topic discussed. From the integration of all core competencies, the basic competencies are created (Kurniasih and Sani, 2014;46). Learning process planning includes a syllabus and lesson plan that contains the subject identity, competence standards, basic competencies, competence achievement indicators, learning objectives, teaching materials, time allocation, learning methods, learning activities, learning assessment, and learning resources (Regulation of the Minister of National Education Number 41 of 2007 concerning Process Standards for Primary and Secondary Educational Units). --- i. Compiling Local Content Syllabus The syllabus contains the subject identity or theme, competence standards, basic competencies, learning materials, learning activities, competence achievement indicators, assessment, time allocation, and learning resources as a reference for developing lesson plans. The syllabus is developed by educational units based on content standards, graduate competence standards, and guidelines for preparing the school-based curriculum. In its implementation, the teachers can develop the syllabus independently or in groups in a school or several schools, a group of subject teachers or a teacher activity center, and the Education Board. This process is under the supervision of the Education Board of the regency or city responsible for primary and junior high schools, senior and vocational high schools, and the department in religious affairs for Islamic primary, junior, senior, and vocational high schools (Regulation of the Minister of National Education Number 41 of 2007 concerning Process Standards for Primary and Secondary Educational Units). Chamsiatin (in Akbar, 2017;28-29) points out that syllabus development follows steps: 1) fill in the identification column and 2) review competence standards. Reviewing competence standards needs to pay attention to (1) the hierarchy of scientific disciplines or material difficulty levels, (2) the linkage of competence standards and basic competencies between subjects, (3) the review of basic competencies, and (4) identifying the main materials. Reviewing main materials needs to pay attention to (1) the level of students' physical, intellectual, emotional, social, and spiritual development, (2) benefits for students, (3) scientific structure, (4) the depth and flexibility of materials, (5) the relevance to students' needs and environmental demands, and (6) time allocation. Next is (5) develop learning experiences; learning experiences contain learning scenarios that highlight students' learning experiences, giving them opportunities to construct their knowledge, develop their life skills, and be meaningful to their lives. Accurately opting for learning approaches, models, methods, techniques, and tactics greatly determines students' learning experiences. 6) Formulate indicators; indicators are the elaboration of basic competencies showing signs of students' actions or responses. Indicator development should consider regional characteristics, educational units, and students, using structured and observable operational verbs. Choices on operational verbs can be formulated by the teachers and used as a basis for developing assessment instruments; 7) determine assessment types; the assessment is done by using a test or a nontest in a writing, oral, performance, product, attitude, project, portfolio, self-report, and other relevant forms; 8) determine time allocation; determining time allocation in each basic competence is based on the effective amount and time allocation of subjects per week by considering the number of basic competencies, flexibility, depth, difficulty levels, and basic competence importance levels, according to the needs of students to master basic competencies; 9) determine learning resources; learning resources cover reference books, objects, materials, information sources, events, physical-social-psychologicalcultural environment, and other relevant things. Learning resources should align with competence standards, basic competencies, indicators, and learning objectives. --- CONCLUSION The present work draws the following conclusions regarding the analysis result of local culture in the basic education curriculum in Banggai Regency. First, identifying and analyzing the context of the local content curriculum involves an in-depth investigation of relevant factors, including geographical potential, natural resources, human resources, and cultural and historical dimensions that form the foundation of the area concerned. Survey and interview methodologies are implemented as the main instruments of primary data collection, complemented by secondary data sources that complement and enrich the understanding of the contextual framework. Second, an in-depth analysis of the context and potential of local culture produces cultural choices that can be developed, such as the Development of Tourist Destinations to increase environmental awareness, the Development of Regional Languages for appreciation of local linguistics, the Development of Regional Skills & Crafts Industries for economic growth, and Preservation of Regional Arts & Sports to pass on and appreciate cultural expressions. This finding is relevant to research variables on the development of local culture in the basic education curriculum, linking academic and practical aspects in maintaining the sustainability of local culture. Considering the findings related to implementing local culture in the basic education learning process and curriculum, this study suggests considering the following matters for the effectiveness of local cultural content learning. First, collaborate with external teaching staff such as community leaders and cultural actors, enrich students' perspectives, and increase the authenticity of local content. Second, ensure the availability of adequate learning facilities, including educational media, visual aids, and teaching materials, to support a variety of approaches and enrich students' understanding of local culture. Third, ensure that the budget is sufficient for implementing local content learning, including obtaining teaching materials, training teachers, and providing facilities by developing local culture in the basic education curriculum.

This study explores the diversity of local cultures in implementing the curriculum in elementary schools in Banggai District, an area rich in distinctive and diverse local cultures. To maintain the noble values of the Indonesian nation, this cultural diversity needs to be preserved and developed through education. One way to make this happen is through utilizing local culture in the curriculum structure, known as local content. This research uses a local cultural analysis approach in the development of the basic education curriculum, with the following stages: first, identification of the curriculum context, which includes an analysis of characteristics, potential, advantages, local wisdom, and regional needs/demands. Second, an analysis of local culture, which will be developed in the basic education curriculum, includes four clusters of local content covering socio-cultural-political, economic, environmental, and other local-specific dimensions. The results of the study indicate that several elementary schools in Banggai Regency have implemented local content subjects in accordance with Banggai Regent Regulation No. 56 of 2017 concerning Local Language Local Content Curriculum (Saluan & Balantak). Based on an analysis of the context and potential of local culture and aspects of basic education curriculum development, several local cultures can be developed, including developing tourist destinations, developing regional languages, developing regional skills and craft industries, and preserving regional arts and sports.

as community leaders and cultural actors, enrich students' perspectives, and increase the authenticity of local content. Second, ensure the availability of adequate learning facilities, including educational media, visual aids, and teaching materials, to support a variety of approaches and enrich students' understanding of local culture. Third, ensure that the budget is sufficient for implementing local content learning, including obtaining teaching materials, training teachers, and providing facilities by developing local culture in the basic education curriculum.

This study explores the diversity of local cultures in implementing the curriculum in elementary schools in Banggai District, an area rich in distinctive and diverse local cultures. To maintain the noble values of the Indonesian nation, this cultural diversity needs to be preserved and developed through education. One way to make this happen is through utilizing local culture in the curriculum structure, known as local content. This research uses a local cultural analysis approach in the development of the basic education curriculum, with the following stages: first, identification of the curriculum context, which includes an analysis of characteristics, potential, advantages, local wisdom, and regional needs/demands. Second, an analysis of local culture, which will be developed in the basic education curriculum, includes four clusters of local content covering socio-cultural-political, economic, environmental, and other local-specific dimensions. The results of the study indicate that several elementary schools in Banggai Regency have implemented local content subjects in accordance with Banggai Regent Regulation No. 56 of 2017 concerning Local Language Local Content Curriculum (Saluan & Balantak). Based on an analysis of the context and potential of local culture and aspects of basic education curriculum development, several local cultures can be developed, including developing tourist destinations, developing regional languages, developing regional skills and craft industries, and preserving regional arts and sports.

INTRODUCTION The number of deaths worldwide due to the COVID-19 pandemic has exceeded 3 million as of April 20, 2021 (1). In this pandemic, the grief response may become more complex and prolonged, leading to psychological problems among people. During the same period, the pandemic caused 9,629 deaths in Japan, a 2% mortality rate (1). Thus, the number of COVID-19-related deaths has been relatively lower in Japan than in the Western countries, where the infection has exploded (2). As the number of deaths due to the pandemic increases, there is growing focus on the suffering of the bereaved families and their support systems in the wake of the COVID-19 pandemic. Pre and post-face-to-face family contact with loved ones is limited in in order to prevent infection (3). The suddenness and unpredictability of bereavement due to COVID-19 makes it challenging to establish advance care planning, a key component of effective terminal care (4). Bereavement due to COVID-19 infection may also interfere with the adaptive mourning process in terms of disruption of social norms, rituals, and mourning practices for death, as seen during past epidemics of infectious diseases (5). Because of these peculiarities, there is concern that bereavement due to coronavirus infections may increase psychological risks, such as complicated grief and depression (5,6). For this reason, safer funeral practices have been explored in infectious disease pandemic disasters through the modification of funeral rites and the introduction of traditional ritual techniques specific to each culture (6,7). In the early stages of the pandemic, public awareness regarding COVID-19 deaths was low because there was no rapid increase in the number of deaths in Japan. However, public awareness increased considerably following the mass media broadcasts of the deaths of two celebrities from COVID-19-a popular comedian and a famous actress-in March and April 2020, respectively. Around the same time, there were infections among funeral workers, which raised public concern about conducting funerals in these difficult times. It became apparent that bereavement during the COVID-19 pandemic was different from that caused by other types of death in Japan. We addressed two issues related to bereavement. First, family members of persons succumbing to COVID-19 were not allowed to bid goodbye to the deceased before death or see them face-to-face even after death. Second, the impact on end-of-life care and bereavement due to restrictions on visits to hospitalized patients, which had significant ramifications given the fact that a majority of people die in hospitals in Japan (8). We shall discuss efforts to address these issues in Japan. --- POTENTIAL PSYCHOLOGICAL CONCERN REGARDING BEREAVEMENT DUE TO COVID-19 In early April 2020, during the initial stages of the pandemic, a statement regarding its psychological impact was released by the Chairman of the Disaster Preparedness and Disaster Response Committee of the Japanese Society for Traumatic Stress Studies. The statement pointed out that bereavement due to COVID-19 could lead to psychological problems such as prolonged grief and ambiguous loss, one which is without closure or clear understanding of why it occurred (9). In order to prevent the spread of COVID-19 in Japan, it was recommended, in principle, that visiting a person at the end of life with COVID-19 infection should be prohibited (10). Not being present at the death of patients with COVID-19 infection impedes the usual farewell rituals conducted for the deceased by the bereaved. For example, touching the deceased's body helps those bereaved to realize that their loved one has actually died. Not being present at the end of a person's life and not being allowed to touch the body potentially interferes with the mourning process and increases the risk of prolonged grief (9). In the Netherlands it was noted that grief levels were higher among people bereaved due to COVID-19 than those who had lost loved ones due to natural causes (11). Traumatic bereavement is more likely to occur when the death of a loved one is sudden or unnatural, such as when a loved one's body is damaged (12). It is accompanied by regret, anger, and guilt over death (For example, could the end have been prevented? Is the loved one suffering? Was death or dying unjust?) (12). Traumatic bereavement is a risk factor for PTSD and depression, as well as prolonged and complicated grief (13). These symptoms after bereavement often co-occur and share common features, but many epidemiological studies support their distinctiveness (14,15). Since bereavement due to COVID-19 shares the common feature of traumatic bereavement (6), systematic psychosocial support is needed. Psychological therapies, such as cognitive-behavioral therapy, optimized for each symptom, have been shown to improve PTSD, grief, and depression after bereavement (16,17). Although Japan has yet to report any empirical studies on grief after bereavement due to COVID-19, considering its cultural practices regarding funerals, it is reasonable to expect increased rates of prolonged grief. In Japan, it is customary to hold a funeral that is widely attended by family members, locals, or business partners. Such bereavement ceremonies are essential to facilitate the grieving process of the bereaved by allowing them to share their feelings and memories of the deceased. Traditional Japanese funeral ceremonies include sharing meals and alcohol with all the people who attend. However, most funerals have now been restricted in Japan, following instances of people becoming infected with COVID-19 after attending funerals. Being in close physical proximity with friends or others may produce feel-good hormones such as oxytocin, dopamine, and serotonin (18). When they are not physically present to say goodbye and grieve with a loved one, they may be more likely to experience a sense of ambiguous loss (18). An ambiguous loss is an indefinite loss that persists without resolution or closure, such as when a loved one is missing (e.g., kidnapped or swept away by a tsunami and never found) or physically present but psychologically absent (e.g., the former personality is still intact due to dementia) (19). Ambiguous loss differs from ordinary bereavement in that there is no definitive information or finality (19). In Japan, guidelines regarding COVID-19 patients have been in place right from the beginning of the pandemic. For example, the number of visits to critically ill patients should be limited. Existing restrictions relate to both end-of-life care and contact with the body after death. However, there is the need to take care of family feelings at each stage, which highlights the usefulness of communication through social networking services and online tools in the COVID-19 Nursing Practice Guide for Critically Ill Patients, Version 1 (published in April 2020) (20). In response to such recommendations, efforts at the grassroots level are underway to encourage communication between patients and their families through online communication tools such as videophones. For example, a physician has launched a crowdfunding campaign to purchase tablets in hospitals to facilitate online communication between critically infected patients and their families. He reached his goal in just half a day (21), and by the end of the campaign, he had raised more than five times his goal of more than 16 million yen, enabling distribution of the tablets to approximately 80 facilities (22). Similar efforts-making videoconferencing possible in end-oflife care settings, when patients cannot see their families faceto-face because of the pandemic (23)-are expected to alleviate patients' suffering in a way that medical personnel by themselves cannot. There is also a growing focus on comforting patients in their dying days, with families asking medical personnel to show the patients their favorite pictures and play their favorite music (24). From the perspective of preventing infection during transportation and cremating the bodies of those who have succumbed to COVID-19, the Japan Medical Association's Implementation Manual (6th edition) (25) requires that crematorium workers and mourners do not touch the body. Under these circumstances, to arrange the farewell ceremony close to the conventional one, flowers and photographs are placed on top of the coffin (which usually mourners set inside the coffin). In case there are restrictions on the number of people who can be present at the funeral, the cremation service provider can take pictures of the deceased before the funeral and show them to the family later. The family could also ask the service provider to place photographs, flowers, and other items related to the deceased on the coffin. Such acts at the funeral ceremony perhaps reflect the bereaved family members' sentiments that they are not leaving their loved ones alone at the time of their death. In addition, there is a custom of wiping and cleansing the body and applying makeup on the face after death (so-called angel care), which nurses generally perform when patients die in Japan (26). However, Version 2 of the COVID-19 Nursing Practice Guide for Critically Ill Patients published in July 2020 points out the importance of family members' participation in angel care in terms of grief care (26). The guideline recommends explaining the risk of infection to family members. If they still wish to participate in angel care, one recommendation is that they take the same preventive measures as do the medical personnel like wearing protective clothing, and touching a safe body area where they are not exposed to bodily fluids. Continuous efforts are needed to make mourning rituals safer with technology, and changes in funeral practices acceptable in each local culture (19). --- THE IMPACT OF RESTRICTED END-OF-LIFE CARE VISITS FOR NON-COVID-19 INPATIENTS Japan has been a super-aged society since 2007. In 2019, 28.4% of the population were 65 years and above and 4.7% were 85 years or older (27). In the 1950s, more than 80% of people died at home in Japan (28). In 2017, more than 80% of patients died in hospitals or institutions (8), despite nearly 70% of them wanting to die at home (29). Fewer people die at home because of the increasing trend toward nuclear families, which has led to a decline in family relationships (28). Many older relatives are moved into homes for the elderly because it may be a burden on family members and others who care for them (29). It is necessary to improve home medical care in present-day Japan to achieve end-of-life care at home, but only 5% of all medical institutions could support it in 2014 (8). For these reasons, many patients choose to receive endof-life care for diseases other than COVID-19 in a medical facility or palliative care. However, many hospitals now restrict visits to non-COVID-19 inpatients-a necessary and natural measureto prevent nosocomial infections (infections caused by pathogens in the hospital). There is a concern that these measures will result in a situation where terminally ill, non-COVID-19 patients, will not be provided with adequate end-of-life care. One of the risk factors for prolonged grief and PTSD after the death of a COVID-19 patient is the bereaved family's inability to say goodbye to the deceased. Since family members of non-COVID-19 inpatients have restricted visitation, the same psychopathological risk can be assumed for them (3). In light of the philosophy of palliative care, it is desirable for the psychological health of patients and their families to spend time with each other so that the patients are taken care of at the end of their lives, by their own. Therefore, there has been a move to provide a flexible response so that patients can be involved in end-of-life care while taking measures to prevent infection. In response to this situation, the Japanese Society for Palliative Medicine has suggested in a pamphlet that families should consider caring for patients at home (30). In one such case, based on a nurse's suggestion, a leukemia patient who was prepared to die in the hospital could go home and spend the next 10 days with his family (31). In some areas, the number of individuals switching to end-of-life care at home has nearly doubled compared to previous years (32). Cases of end-of-life care at home have also been reported in the UK and Portugal (33,34). In Portugal, most families encourage terminally ill patients to stay at home for an extended period (34). Although, in some cases, family members were unable to visit their hospitalized relatives freely during the COVID-19 pandemic, they asked individuals, called end-of-life caregivers, from some organizations such as hospices, to provide endof-life care for terminally ill patients living alone in Japan (35). Conversely, some physicians involved in palliative care have expressed concern about suggesting end-of-life care at home. It is necessary to consider the situation of families who have difficulties or anxieties about administering end-of-life care at home and seek gentle and heartwarming end-of-life care at hospitals for their loved ones, knowing that they will have to observe all the required measures against infectious diseases during their restricted visits (36). The UK has increased opportunities for bereaved institutional support, including the issuance of guidelines by the NHS to allow only one family member to visit patients who are unlikely to recover or who are days or weeks away from their death (37). Although end-of-life care has been discussed as an issue for medical care in super-aging Japanese society, COVID-19 could be considered an opportunity to think about dying at home. As a country with one of the highest life expectancy levels, Japan has focused on care for the elderly. It is precisely for this reason that it is expected to lead the world in establishing a system of advanced end-of-life care. --- RESPONSE TO BEREAVEMENT DUE TO COVID-19 IN JAPAN: LESSONS LEARNED FROM PAST MAJOR DISASTERS Japan experienced an extremely high level of loss in the Great East Japan Earthquake of 2011 (38), which led to the development of academic and public health efforts to deal with grief; these were also deployed during the COVID-19 pandemic. For example, the Japan Disaster Grief Support project established after this earthquake and implemented in May 2020, provided grief-related psychoeducational materials for the bereaved (39). Musashino University, the National Center of Neurology and Psychiatry, and others developed several treatment/prevention programs for prolonged grief, which have been shown to be effective in other countries too, based on empirical evidence. These include complicated grief treatment (40) and its Japanese version, along with group cognitive behavioral therapy for the bereaved with the distress of grief less severe than complicated grief (41). Since face-to-face therapy is limited during the COVID-19 pandemic, to provide such services, it would be necessary to devise programs including web-based grief treatment and videoconferencing psychotherapy (42,43). We provide a modified treatment program for patients in whom face-to-face treatment at an institution had to be interrupted as a preventive measure against COVID-19. Even before the COVID-19 pandemic, the U.K. and U.S. had established guidelines and training methods for telepsychological interventions and developed laws; this area has not yet been developed in Japan. In the early stages of the pandemic, some academic volunteers translated these guidelines into Japanese. The infrastructure for telepsychological interventions has been developed; however, its growth is not sufficient to meet the demand. Professionals must work together to ensure that grief support continues without interruption. To ensure that those who need help do not suffer, we must provide more flexible support, including online programs that can be implemented for bereaved families in remote areas, in preparation of the post-Corona era. --- DATA AVAILABILITY STATEMENT The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s. --- AUTHOR CONTRIBUTIONS YM and YT conceived the ideas. YM developed the draft paper. MI and SN verified the draft paper as experts in grief-related research and treatment in Japan and supervised this work. YT encouraged YM to investigate the research or guidelines related to bereavement due to COVID-19 in Japan. All authors have discussed the contents and contributed to the final manuscript. --- Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

This commentary discussed the psychological issues related to bereavement in the wake of the COVID-19 pandemic. Specifically, we addressed two aspects in the context of Japanese culture. The first relates to the psychological distress of members of the bereaved family who could not visit their loved ones who had COVID-19 before or after their death. The second relates to the bereavement experience of those who were unable to be with their loved ones when the end came, even though they did not have COVID-19, because of restrictions on visiting hospitals. We seek to focus on the need for a support system for bereaved families to help them through the grieving process, and discuss end-of-life care in such circumstances, and in the post-COVID-19 era, as in current day Japan.

The past four decades have witnessed a plethora of studies on gender-role portrayals in advertising on television (Cheng 1997;Furnham and Voli 1989;McArthur and Resko 1975;Milner and Collins 2000;Pacilli et al. 2016;Paek et al. 2011). From the perspective of marketers and advertisers, gender is a primary segmentation variable in developing marketing strategies and defining target groups (An and Kim 2007;Milner and Collins 2000). Also, gender portrayals have possible effects on corporate images and on purchase intentions of consumers (Ford and LaTour 1996). From the perspective of activists, advertising councils, or policy regulators, however, the stereotypical depiction of men and women in today's ads is problematic for many reasons (Coltrane and Messineo 2000). It is argued that advertisers create and perpetuate gender stereotypes, which may erode gender equality and harm society at large (MacKay and Covell 1997;Oppliger 2007). The accumulated empirical evidence clearly suggests that gender roles are highly stereotypical in television ads (Eisend 2010;Furnham and Mak 1999;Furnham and Paltzer 2010). However, despite this strong research tradition around the world, it is less clear how a country's culture shapes gender roles in TV ads. This research deficit can be traced back to the research designs used in previous research. In fact, most research on gender stereotypes in television advertising is based on single-country studies (Das 2011;Kim and Lowry 2005;Uray and Burnaz 2003). Because such studies work with a specific sample at a specific time of the year, we cannot use them to compare findings across countries. For instance, if we see a difference between two studies from two different countries, we may not be sure whether the observed dissimilarities reveal cultural differences or whether they can be traced back to differences in time frame, sampling, codebook, or other characteristics of an individual study (Matthes et al. 2012). Comparative research is therefore urgently needed. Only a very few studies have analyzed gender roles in a comparative fashion allowing for a test of country and cultural differences (An and Kim 2007;Gilly 1988;Milner and Collins 2000;Paek et al. 2011). Most of these studies have relied on one or two comparative gender indices (such as Hofstede's), and more recent indices have not been employed. This limitation makes it difficult to come to a definite conclusion about the role of cultural variables in explaining the portrayal of men and women in TV advertising. Even more importantly, almost all the relevant studies have looked at television ads in only two or three countries (Gilly 1988;Milner and Collins 2000), which is a clear limitation in terms of detecting cultural patterns. In fact, only Paek et al. (2011) looked at television advertisements across seven countries and used Hofstede's masculinity dimension and the gender development index to explain gender-role portrayals. Although Paek et al.'s (2011, p. 204) study is pioneering in many aspects, they only tried to predict the gender of the primary character and the voiceover and sampled the material from different months (i.e., July vs. November), Bwhich may reduce the generalizability and the comparability of [their] findings within and across the countries.<unk>Also, their data are from 2002, and it remains unclear-as with most content analytical research-how the findings can be generalized to more recent years. Even more importantly, all the known prior comparative studies have not empirically modeled the influence of culture on gender stereotypes. It is not the same to interpret differences between countries by drawing on cultural differences as it is to empirically measure and confirm the role of culture in an adequate statistical model. If a specific culture shapes gender stereotypes, then a country's score on a cultural dimension should help to explain the degree of gender stereotyping. Unfortunately, such a multilevel analysis (see Raudenbush and Bryk 2002) is not known to have been conducted in extant research. Therefore, the aims of our study were to observe gender-role portrayals around the world, using comparable measures, sampling strategies, and classic as well as recent gender indices to explain country differences. We analyzed a total of 1755 television advertisements from 13 countries, which were selected based on their scores on various gender indices: Austria, Brazil, China, France, Germany, Japan, Netherlands, Romania, Slovakia, South Korea, Spain, United Kingdom, and United States. To date, our study is the largest known study on gender portrayals ever conducted, and we are the first to employ the concept of gender egalitarianism from Project GLOBE (House et al. 2004) and various other gender indices at the same time. --- Gender-Role Portrayals in Advertising Research on gender role portrayals in U.S. television advertisements started in the 1970s (Dominick and Rauch 1972;McArthur and Resko 1975), followed by research in Great Britain and Europe in the 1980s (Furnham and Voli 1989;Manstead and McCulloch 1981) and in Asia in the 1990s (Cheng 1997;Sengupta 1995). Generally, this research has led to the consensus that gender roles are highly stereotypical in television advertisements (Eisend 2010;Furnham and Paltzer 2010) across a range of commonly explored variables. The most frequently analyzed variables are the gender of the primary character, the gender of the voiceover, the age of the primary character, and the setting associated with the primary character. Analyses of the gender of the primary character led to rather mixed results in previous research, with some studies showing male predominance, others showing a female predominance, and still others finding almost no difference. Nevertheless, the majority of studies showed a predominance of male primary characters (see Eisend 2010;Furnham and Paltzer 2010). Some research has indicated a relation between Hofstede's masculinity index and the gender predominance of the primary character (Milner and Collins 2000), whereas others found the predictive role of Hofstede's masculinity index to be minimal (Paek et al. 2011). We thus hypothesize that there will be more men than women depicted as primary characters in our analyses (Hypothesis 1). The age of the primary character is another widely studied variable. Most studies report a predominance of women in the younger age segment (under 35), whereas more men were found in the middle and older age segments (Furnham and Mak 1999;Furnham and Paltzer 2010). In a meta-analysis, Eisend (2010) found that the odds of women being younger are three times higher than for men. Hence, there are strong grounds to expect that female primary characters will be depicted as younger compared to male characters (Hypothesis 2). Furthermore, the predominance of male voiceovers (which was often interpreted as the Bvoice of authority<unk>) is one of the most consistent findings in the literature-a predominance that is even more pronounced in Asia compared to other regions (Furnham and Mak 1999). Previous research has indicated that a higher score in Hofstede's masculinity index increases the odds of a male voiceover (Paek et al. 2011). Thus, we predict that there will be more male voiceovers compared to female voiceovers in television ads (Hypothesis 3). In terms of product categories associated with a specific gender, there are relatively few consistent findings; this may be because different studies often employ different product categories. One finding that was confirmed in most studies, however, was the association between women and body products, or as other studies called them, Btoiletries,<unk>Bbeauty products,<unk>and Bpersonal care products,<unk>as well as Bhousehold and cleaning products<unk> (Furnham and Paltzer 2010). For men, product associations were less clear, but some studies found associations between men and television advertisements for cars, telecommunications, electronics, technology, and computers (Ganahl et al. 2003;Royo-Vela et al. 2008). Based on these findings, we expect that female primary characters will be more likely to be seen in ads for toiletries, beauty products, and cleaning products (Hypothesis 4a), whereas male characters will be associated more with telecommunications, electronics, technology, computers, or cars (Hypothesis 4b). The setting is another variable that has often produced clear gender divisions. Most often cited is the association of women with a home setting (Das 2011;Uray and Burnaz 2003;Valls-Fernández and Mart<unk>nez-Vicente 2007). A meta-analysis shows that the odds of women being depicted at home (vs. at work) are approximately 3.5 times higher than for men (Eisend 2010). Another finding that was similar across most of the literature is that more men than women are shown in a workplace setting (Prieler and Centeno 2013;Valls-Fernández and Mart<unk>nez-Vicente 2007). Therefore, we hypothesize that female primary characters will be more likely to be depicted in a home setting (Hypothesis 5a), whereas male characters will be more likely to be shown at work (Hypothesis 5b). In addition to these commonly analyzed variables, we focused on the working role of the primary character (no working role, high status worker, and lower status worker). Although categorizations differed, several studies investigated whether the primary character was working (Coltrane and Messineo 2000;Das 2011;Uray and Burnaz 2003) or was a homemaker (Uray and Burnaz 2003;Valls-Fernández and Mart<unk>nez-Vicente 2007)both leading to highly stereotypical results that suggested more men than women are working and more women than men are depicted as homemakers. Thus we expect female primary characters will be more likely to be seen in lower status working roles compared to male primary characters, who will be more likely to be seen in higher status roles (Hypothesis 6). --- Cultural Models, Gender Indices, and Advertising To fully understand gender-role portrayals, scholars have repeatedly pointed to the important role of cultural differences. The most widely applied cultural model in advertising research is Hofstede's cultural dimensions (Okazaki and Mueller 2007), including four dimensions (he later added two additional ones): power distance, individualism/collectivism, uncertainty avoidance, and the masculinity dimension (De Mooji and Hofstede 2010;Hofstede 2001). The most relevant dimension for the purposes of the present paper is Hofstede's masculinity dimension, which has been used in cross-cultural content analyses on gender in advertising, including studies comparing two countries (An and Kim 2007;Huang 1995;Moon and Chan 2002;Odekerken-Schröder et al. 2002) and some comparing three or more countries (Milner 2005;Milner andCollins 1998, 2000;Paek et al. 2011;Wiles et al. 1995). However, only a few studies have confirmed an association between the masculinity index and gender portrayals (Huang 1995;Wiles et al. 1995), whereas more studies have led to mixed results (An and Kim 2007;Milner and Collins 2000) or to results that were mostly opposite from those predicted by Hofstede's masculinity index (Milner 2005;Moon and Chan 2002;Odekerken-Schröder et al. 2002;Paek et al. 2011). Considering the mixed results of previous research in general and criticisms of Hofstede's (2001) study in particular-as being rather outdated (An and Kim 2007;Okazaki and Mueller 2007) and for its masculinity dimension that mixes two sub-dimensions (i.e., the characteristics of a society and the gender role distinctions) (Emrich et al. 2004;Hofstede 2001)-it is crucial to use additional indices, such as the more recent framework from the GLOBE project (House et al. 2004). Its theoretical importance and promise have been mentioned in several articles on theory in advertising research (Okazaki and Mueller 2007;Taylor 2010). Still, the GLOBE study has been used rarely in gender stereotyping research to date. In contrast to Hofstede's study, the GLOBE project differentiates between societal practices and values. Practices are measured through questions regarding Bwhat is,<unk>while values are measured through questions regarding Bwhat should be<unk> (House et al. 2010). We have decided to use societal practices in our study because gender portrayals are about the way a society actually performs, whereas values are about how a society should perform (House et al. 2010;Okazaki et al. 2010). Although Hofstede's (2001) masculinity index influenced the GLOBE project (the GLOBE dimensions include: Performance Orientation, Uncertainty Avoidance, Human Orientation, Institutional Collectivism, In-Group Collectivism, Assertiveness, Gender Egalitarianism, Future Orientation, and Power Distance), the latter separated this dimension into gender egalitarianism and assertiveness. Gender egalitarianism was defined as Bthe degree to which a society minimizes gender role differences while promoting gender equality<unk> (House et al. 2010, p. 118). Gender egalitarianism reflects society's beliefs about whether biological sex should determine roles in society; gender egalitarian societies rely less on biological sex to make those determinations (Emrich et al. 2004). In such societies, there is less occupational sex segregation, more women in the labor force and in positions of authority, and generally a higher status for women. Thus, gender egalitarianism is clearly related to our study. Despite the call to use Project GLOBE's dimensions in advertising research (House et al. 2010;Okazaki and Mueller 2007), only a few studies to date have employed them. These studies have included the dimensions of assertiveness (Okazaki et al. 2010;Terlutter et al. 2010), performance orientation (Okazaki et al. 2010), and humane orientation (Diehl et al. 2012), and they have indicated an association between GLOBE dimensions and advertising evaluations. Thus, no known study to date has used gender egalitarianism as a theoretical framework or used Project GLOBE dimensions in content analysis. In addition to testing the predictive power of cultural models for gender stereotyping, we take into account other dimensions related to gender development because previous research has indicated that gender role portrayals may be influenced by a country's gender development (Eisend 2010;Paek et al. 2011). Paek et al. (2011) found that the predictive role of the Gender-related Development Index (GDI) by the United Nations Development Program (UNDP) seems to be minimal for the gender of the prominent character in an advertisement. However, they also reported that the odds of using a male voiceover significantly increased as GDI scores dropped. In his meta-analysis, Eisend (2010) used another gender index created by the UNDP-namely, the Gender Empowerment Measure (GEM)-and found a correlation between the GEM and gender stereotyping in advertising. Our research extends these studies by testing whether the GDI (UNDP 2014a), the UNDP's Gender Inequality Index (GII) (UNDP 2014b), which replaced the GEM due to criticism, and the World Economic Forum's Gender Gap Index (GGGI) can predict gender stereotypes (Hausmann et al. 2014). All of these indices are based on demographic data. The GDI is based on gender gaps in life expectancy, education, and incomes; the GII on reproductive health, empowerment (such as share of parliamentary seats and higher education levels) and labor market participation; and the GGGI is the most inclusive, being based on gender gaps on economic, political, educational, and health criteria (Hausmann et al. 2014;UNDP 2014a, b). Furthermore, besides Hofstede (2001), our study uses GLOBE for the first time in the extant research on gender portrayals as a theoretical framework (House et al. 2004). Because all indices differ and previous evidence of their impact on gender portrayals is mixed at best, we ask as an exploratory research question how the five indices (Hofstede's masculinity index, the Project GLOBE's gender egalitarianism index, the GDI, the GII, and the GGGI) predict gender-role portrayals in television advertising. --- Method We analyzed gender role stereotypes across Asia, America, and Europe (Austria, Brazil, China, France, Germany, Japan, Netherlands, Romania, Slovakia, South Korea, Spain, United Kingdom, and United States). We selected the countries based on the following criteria. First, we wanted to have a broad range of countries that have different scores on various gender indices. For that reason, we have included countries with a high score on Hofstede's masculinity index (Japan, Slovakia, and Austria) and with a low score (Netherlands), as well as countries with a high score on GLOBE's gender egalitarianism index (United Kingdom, Netherlands, and France) and a low score (South Korea). Second, we wanted to include the countries that were frequently sampled in previous studies, for instance, the United Kingdom and the United States. Finally, the selection was driven by practical reasons, such as access to TV channels and the language qualifications of our students/coders. The most recent data from all gender indices were used. In May 2014, 15 h of prime-time TV programming were recorded from those broadcasters with the largest shares of viewers for each respective country. We focused on private channels because we were not interested in the effects of country-specific regulations of public-service broadcasters. For Austria, we took the largest private channel as well as the largest public service channel because the audience share of the private channel is rather low. For China, we took CCTV-1 as the clearly dominating TV channel. The recording time was split into three typical weekdays of the same week, that is, 3 <unk> 5 h per channel. The sample included weekdays and weekends to ensure a higher diversity of advertisements (May 14/Wednesday, May 16/Friday, and May 18/Sunday). We started the recordings after Mother's Day in most of the recorded countries to ensure that no large country-and/or region-specific events occurred during the days of recording. We focused on prime time because it was commonly used in previous studies (Paek et al. 2011). Because the definition of prime time varied by country, we used the most inclusive definition (i.e., from 6 PM to 11 PM). Like in most prior research (see Kim and Lowry 2005), duplicate ads, political ads, ads for films and CD's, public service announcements (PSAs), and ads with children, animals, or comic figures as dominant actors were not considered in our study. After 150 ads were collected in one country, we stopped additional data collection; however, the 3 days sampled produced less than 150 unique advertisements for several countries. Our sample consisted of 1755 ads. These included ads from Austria (ORF2, n = 98; Puls4, n = 124), Germany (RTL, n = 144), the United Kingdom (ITV1, n = 149), the United States (ABC, n = 149), France (TF1, n = 150), Spain (Telecinco, n = 146), Brazil (Rede Globo, n = 123), Netherlands (RTL4, n = 149), Romania (ProTV, n = 115), Slovakia (TV Markiza, n = 118), China (CCTV-1, n = 137), Japan (NTV, n = 150), and South Korea (SBS, n = 127). --- Coding Procedure and Reliability The codebook was adopted from prior research (Prieler and Centeno 2013;Furnham and Paltzer 2010;Nassif and Gunter 2008). The following categories were included in the analysis: primary character, product category, voiceover/narrator, age, dominant setting, and working role. These are standardized measures in this research (see Appendix Table 6 for codes). The coding team consisted of 30 coders from two major universities in South Korea and Austria. Coders were one of the authors and 29 undergraduate students receiving compensation/course credit for their work. Students were unaware of the hypotheses and were native or bilingual in the language of the country they coded. Training sessions on using the codebook were held with all coders prior to testing reliability. Ads were randomly selected for the reliability test but had to involve a primary character so that all codebook categories could be included. We performed three reliability tests using Krippendorff's Alpha (Hayes and Krippendorff 2007; n = 10 each, from 2 to 30 coders). One test was between coders for different countries on English-language advertisements, one test was between coders coding the same country prior to coding, and one test used the coded material without coders knowing they were tested. Although 65 % of the reliability measures in these three tests were above <unk> =.80, the lowest intercoder results for the numerous intercoder reliability tests were for primary character (<unk> =.80), voiceover/narrator (<unk> =.89), age (<unk> =.72), dominant setting (<unk> =.72), and working role (<unk> =.63). This was still above the recommended chancecorrected agreement of.60 by Neuendorf (2011). --- Results --- Primary Character and Voiceover In our first hypothesis, we stated that there will be more male than female primary characters. Among all spots with a primary character, 50.7 % of the characters were women. As can be seen in Table 1, in some countries (Brazil and Korea), the share of male primary characters was slightly higher than those of female primary characters. In other countries, however, the share was almost identical. Thus, we cannot find a substantial male predominance of primary characters. We also hypothesized (Hypothesis 2) that female primary characters would be depicted as younger compared to male primary characters. As Table 1 reveals, this pattern was the case in seven countries (Austria, Germany, France, Spain, Slovakia, Japan, USA). For the other countries, however, there was no significant effect. Based on previous research, we assumed in Hypothesis 3 that we would find more male than female voiceovers. When looking across countries (see Table 1), the share of male voiceovers (61.8 %) was, in fact, significantly higher (p <unk>.01) than the share of female voiceovers (32 %). Confirming Hypothesis 3, this result was mirrored in most countries. The opposite effect, however, was found for France, with 58 % female voiceovers. --- Product, Setting, and Work Role In line with prior research, we expected that female primary characters would be more likely to be seen in ads for toiletries, beauty products, personal care, and cleaning products, whereas male characters would more likely be associated with telecommunications, electronics, technology, computers, or cars. The findings are reported in Table 2. In all countries but Japan, the association of female primary characters with toiletries, beauty products, personal care, and cleaning products can be confirmed. The association of male primary characters with products related to technology and cars (see Table 2) was observed for Brazil, Germany, the Netherlands, Spain, and the United Kingdom. The effect was not significant by the conventional p <unk>.05 level for Romania, Slovakia, Austria, USA, China, Japan, and South Korea. Another prominent finding in previous research is that female primary characters are more likely to be shown in a home setting, whereas male characters are more likely to be associated with a work setting. Table 3 shows the findings for each country. The stronger depiction of female primary characters in home settings compared to male primary characters can be confirmed in Brazil, China, Germany, the Netherlands, Romania, South Korea, and Spain. There was no significant effect for six countries: Austria, France, Japan, Slovakia, USA, and United Kingdom. When it comes to the dominant depiction of male primary characters in work settings (see Table 3), we observed significant associations for Austria, Germany, France, Japan, the Netherlands, and the United Kingdom, but no significant effect for South Korea, Brazil, the United States, Spain, Romania, Slovakia, and China. As can be seen in Table 4, it was not the case that male characters are depicted in higher status jobs compared to female characters. This association was significant only for Japan. However, when interpreting our findings, the very small numbers of depicted working roles need to be taken into account. Therefore, we also checked whether or not a female or male character was depicted in any working role at all. One could anticipate that male characters will be more likely to be shown in any working role compared to female characters. As Table 4 reveals, the association between male primary characters and the depiction of a working role was statistically significant for Austria, France, Japan, the Netherlands, Slovakia, and the United Kingdom. The association was not significant by conventional levels for Germany, Brazil, China, South Korea, Spain, Romania, and the United States. --- Multilevel Analyses So far, we have observed how gender is related to the depiction of primary characters, and we have looked at single countries. Although such an analysis is useful, we are unable to explain why an association is found in one country and not in another. Thus, the question we want to ask is whether variations in the association between gender and character depiction can be explained by cultural differences between countries. In order to answer this question, hierarchical linear models (i.e., multilevel analyses) are needed (Raudenbush and Bryk 2002). Multilevel models are warranted when cases are clustered within countries. The advantages of multilevel analysis are that we can explain the individual-level variation in the dependent variable while statistically controlling the variation across levels of analysis and that we try to predict the variation of regression slopes by including constructs at the country level. In particular, we will examine whether the country differences in the relation between gender and character depiction can be explained by the five indices introduced previously. Because our outcome variables are binary, we ran a logistic hierarchical non-linear model with the Logit-link function using PQL estimation (distribution at level-1: Bernoulli) with the statistical package HLM 7. The level-1 model includes the gender of the primary character. The level-2 model includes the respective index (i.e., Hofstede's index, GLOBE, GDI, GII, or GGGI). Because the level-2 variables are correlated, we ran a separate model for each index. In the first step, we computed the variance components in order to examine whether there is a significant amount of variance between the classes. The upper-level variance is significantly different from zero for the age of the primary character (<unk> 2 = 55.29, p <unk>.001), the product category body and cleaning products (<unk> 2 = 79.03, p <unk>.001), the product category technical products and cars (<unk> 2 = 27.97, p <unk>.001), working role shown (<unk> 2 = 30.79, p <unk>.001), and the status of the average level of a gender index (e.g., the effect of gender on depicted age at the average level of Hofstede's masculinity index). For most outcomes, the level-1 effect of gender was statistically significant, confirming hypotheses Hypothesis 2 (character age), Hypothesis 4a (products for women), Hypothesis 4b (products for men), Hypothesis 5a (home setting), and Hypothesis 5b (work setting). However, we could not confirm the assumption that female primary characters are more likely to be seen in lower status working roles compared to male primary characters, who were theorized to be seen in higher status working roles (Hypothesis 6). Yet, female primary characters were less likely to be depicted in any working role compared to their male counterparts. Beyond the effects of the level-1 predictor, the main focus of the multilevel model lies on the cross-level interactions between the gender of the primary character and the gender indices. This interaction tests whether the variations between countries in the effects of gender on the outcome variables can be traced back to variations in the five gender indices. In order to answer this question, we first looked at the random-coefficients model. For the dependent variable age of the primary character, no systematic differences in the regression slope between the countries were observed (<unk> 2 = 9.50, p =.39). The same was true for the dependent variables of product category body and cleaning products (<unk> 2 = 7.08, p >.50), setting: home (<unk> 2 = 10.11, p =.34), setting: work (<unk> 2 = 5.05, p >.50), presence of a working role (<unk> 2 = 7.19, p >.50), and status of the working role (<unk> 2 = 12.45, p =.19). However, for the category of technical products and cars, there was a statistically significant variation that could be explained by level-2 variables (<unk> 2 = 25.23, p <unk>.01; not shown in Table ). Thus, for most outcomes, there were no differences in the regression slope that can be explained by a specific culture. As can be seen in Table 5, we found no substantial cross-level interactions for all five gender indices. That is, none of the five gender indices was able to explain why there was a stronger (or weaker) association between the gender of the primary character and the outcome variables in a given country. Put formally, an increase in a gender index did not lead to an increase in the association between gender and the outcome variables. --- Discussion Our study is the largest known to date on gender-role portrayals in advertising, and it is the first known with sample equivalence, using comparable TV programs to illuminate the effect of culture on gender-role portrayals. In addition, it is the first study known to test the role of gender indices for genderrole portrayals in advertising using multiple gender indices as independent variables and multiple gender-role variables as dependent variables. We have found significant differences among the countries investigated. Although every country showed traditional gender-role portrayals for some variables, some countries showed non-traditional gender-role portrayals for several variables and thus seem to use a more gender-equal approach toward gender roles. For example, in the United Kingdom, no significant age differences were found between the share of male and female primary characters as well as male and female voiceovers, and similar shares of men and women were shown at home. Similarly, in the United States, men were not stereotypically associated with car/ electronic products, approximately the same number of men and women were shown both at home and at work, and men and women were portrayed in work roles about evenly. These findings align with those of several previous research studies showing that in the United Kingdom and several European countries, gender portrayals have improved in recent years (Furnham and Mak 1999). By contrast, advertisements in several countries, such as Germany, were identified as highly traditional in the use of gender roles for almost all investigated variables. Across all countries, some variables tend to indicate more traditional gender-role portrayals than others. For example, for more than half the countries, the variables of voiceover, age, toiletries/household products, and setting produced significant gender differences. Although we were able to observe differences between countries, we found that these differences cannot be explained by cultural or gender indices. The effect of a specific culture in shaping advertising messages is, therefore, smaller than commonly thought. This finding stands in contrast to Eisend's (2010, p. 436) meta-analytic results showing that Bgender stereotyping in advertising depends on developments related to gender equality in society rather than the other way around.<unk>There are many potential explanations for why our study comes to a different conclusion. First and foremost, it is important to stress that no study of which we are aware has modeled the influence of culture in a multilevel model. As should be apparent, when looking at Tables 1, 2, 3 and 4, one could easily pick two or three countries and explain the observed differences by different scores on gender indices. However, such an analysis strategy does by no means confirm that an increase in a gender inequality index leads to an increase in gender stereotyping across countries. For this, a multilevel model is necessary. Second, it is possible that we have found no relationship between the cultural variables and stereotyping because gender portrayals in advertising are lagging several years behind actual developments in society (Eisend 2010;Kim and Lowry 2005). Unfortunately, this cannot be sufficiently tested with the present data because we would need to draw a sample of TV ads over time. Third, because gender stereotypes in advertising are decreasing over time (Eisend 2010), our findings may be different simply because our sample is the most recent one, reflecting a declining influence of culture. That is, although advertising and its gender-role portrayals may still vary across cultures, in some cases, they may become more universal due to global markets and networked publics (Paek et al. 2011). Fourth, our study used equivalent samples in all countries which is not possible in a meta-analysis such as Eisend's (2010). Fifth and finally, gender stereotypes might not be measured sufficiently by gender indices, and it may be that another index should be considered as more appropriate for measuring stereotypes (Williams and Best 1990). --- Limitations and Future Research Directions As with every research project, our project has several limitations. These include a rather small sample size in some countries, which was also the case in several previous studies (Furnham and Paltzer 2010). More importantly, our sample includes countries from Asia, Europe, and the Americas, but none from Africa or Oceania. Related to this point, we were not able to sample all countries that we would have liked because of a lack of access to the TV channels. Generally, the number of countries was very small for a multilevel model. Larger and more diverse samples are therefore needed in future research. In addition, our sample was drawn in May 2014. Thus, it is able to represent only this specific period; additionally, seasonal variations might be a concern in research on advertising. Furthermore, although we strived to choose comparable TV networks, this is a daunting task because various countries have different broadcasting systems. In addition, analyzing only one TV channel for each country, even the most dominant one, might not be fully representative of the pool of TV ads from each country. Finally, we suggest that future research studies should attempt to analyze gender-role portrayals in television advertising over time using longitudinal approaches. --- Practice Implications Practitioners in the countries we analyzed are called to raise their awareness for gender stereotypes in television ads. Even when practitioners reside in countries with high gender equality, gender stereotypes still prevail in television advertisements. Obviously, current (self-) regulatory efforts do not seem to be successful in implementing an unbiased representation of men and women in television ads. We suggest that advertising councils as well as advertising professionals should work toward a clearly defined set of recommendations about how men and women should be presented. We also suggest that advertising educators should sensitize students about gender role depictions in commercials, how they are observed using scientific methods, and what effects they may have on women and men. --- Conclusion Our study was able to show that there appears to be a global pattern of gender stereotyping still at work. This finding is significant for two reasons. First, it is well known that gender stereotype

Although there are numerous studies on genderrole portrayals in television advertising, comparative designs are clearly lacking. With content analytical data from a total of 13 Asian, American, and European countries, we study the stereotypical depiction of men and women in television advertisements. Our sample consists of 1755 ads collected in May 2014. Analyzing the gender of the primary character and voiceover, as well as the age, associated product categories, home-or work setting, and the working role of the primary character, we concluded that gender stereotypes in TV advertising can be found around the world. A multilevel model further showed that gender stereotypes were independent of a country's gender indices, including Hofstede's Masculinity Index, GLOBE's Gender Egalitarianism Index, the Genderrelated Development Index, the Gender Inequality Index, and the Global Gender Gap Index. These findings suggest that gender stereotyping in television advertising does not depend on the gender equality prevalent in a country. The role of a specific culture in shaping gender stereotypes in television advertising is thus smaller than commonly thought.

systems. In addition, analyzing only one TV channel for each country, even the most dominant one, might not be fully representative of the pool of TV ads from each country. Finally, we suggest that future research studies should attempt to analyze gender-role portrayals in television advertising over time using longitudinal approaches. --- Practice Implications Practitioners in the countries we analyzed are called to raise their awareness for gender stereotypes in television ads. Even when practitioners reside in countries with high gender equality, gender stereotypes still prevail in television advertisements. Obviously, current (self-) regulatory efforts do not seem to be successful in implementing an unbiased representation of men and women in television ads. We suggest that advertising councils as well as advertising professionals should work toward a clearly defined set of recommendations about how men and women should be presented. We also suggest that advertising educators should sensitize students about gender role depictions in commercials, how they are observed using scientific methods, and what effects they may have on women and men. --- Conclusion Our study was able to show that there appears to be a global pattern of gender stereotyping still at work. This finding is significant for two reasons. First, it is well known that gender stereotypes in advertising can influence gender-role stereotypes in society, further perpetuating gender roles and gender inequality (MacKay and Covell 1997;Oppliger 2007). Second, our findings clearly suggest that gender stereotypes in TVadvertising can be found around the world, independent of a given gender equality status in a particular country. It follows that more progressive countries do not necessarily depict women-in terms of gender equality-in more progressive ways in television advertising. We hope our research helps to spur a discussion among scholars, advertisers, and regulators on the global dominance of gender stereotyping in advertising. --- working role (<unk> 2 = 27.28, p <unk>.001). There was no significant amount of variance for the depicted setting: home (<unk> 2 = 16.15, p =.06) and for the depicted setting: work (<unk> 2 = 14.43, p =.11). The results of the multilevel model are presented in Table 5. Because we included the grand-mean-centered terms for the gender indices, the effects of gender on the outcome variables must be interpreted as the effect of gender at the

Although there are numerous studies on genderrole portrayals in television advertising, comparative designs are clearly lacking. With content analytical data from a total of 13 Asian, American, and European countries, we study the stereotypical depiction of men and women in television advertisements. Our sample consists of 1755 ads collected in May 2014. Analyzing the gender of the primary character and voiceover, as well as the age, associated product categories, home-or work setting, and the working role of the primary character, we concluded that gender stereotypes in TV advertising can be found around the world. A multilevel model further showed that gender stereotypes were independent of a country's gender indices, including Hofstede's Masculinity Index, GLOBE's Gender Egalitarianism Index, the Genderrelated Development Index, the Gender Inequality Index, and the Global Gender Gap Index. These findings suggest that gender stereotyping in television advertising does not depend on the gender equality prevalent in a country. The role of a specific culture in shaping gender stereotypes in television advertising is thus smaller than commonly thought.

Background Workplace violence is defined as any event that results in harm caused by work-related assaults, threats and abuse, whose impact on the victim may lead to a deterioration in health, safety and well-being [1]. Workplace violence (WPV) is often associated with the type of occupation, with higher incidence among professions involving interactions with many individuals; therefore WPV is a matter of considerable concern for the health sector. A US report notes that on average, 20 workers are murdered and 18,000 are assaulted each week while at work [2]; similar figures are provided by European reports [3]. Forty-eight per cent of non-fatal workplace violence incidents take place in the healthcare sector [4]. About 50% of healthcare workers experience workplace violence in the course of their career [5,6]. Healthcare workers have a 16 times greater risk of suffering workplace violence than workers in other sectors [7]. Nurses are more at risk [8,9] and female workers, both nurses and doctors, are at even higher risk [10]. In a sample of 1826 health professionals, about 11% had suffered physical assault, 5% on more than one occasion, while 64% had received threats or verbal abuse [11]. Saeki et al. [12] report a prevalence of 15%. Data from the US National Crime Victimization Survey for the period 2005-2009 show a rate of workplace violence of 5.1/1000 for all occupations, 10.1/1000 for physicians and 8.1/1000 for nurses. For mental health workers the violent victimisation rate was 20.5/1000, second only to the rate of law enforcement officers (47.7/1000) [13]. In Belgium, a study on patient-physician aggressions [14] conducted by means of an online questionnaire which was completed by 4930 participants, found that, in the preceding 12 months, 37% had been the victim of aggression: 33% verbal aggression, 30% psychological, 14% physical and 10% sexual. Psychiatric and emergency departments were the settings where violence most commonly occurred. In Israel, Carmi-Illuz [15] compared the risk of violence between a sample of hospital-based physicians and a sample of community-based physicians, finding a substantially comparable risk. A particular form of violence, homicide, is infrequent but extremely disturbing. In 2006, a leading schizophrenia specialist, W. S. Fenton, was killed at his office by one of his patients [16]. In the healthcare professions, homicide is a malicious and intentional event and is a very rare violent circumstance. We have carried out a retrospective analysis, to explore the main features of all the reported cases of work-related homicides of physicians in Italy over the past 32 years. --- Methods We relied on the main national statistical databases (ISTAT, EURISPES, EU-RES) [17][18][19] and on PubMed to trace all cases of work-related homicides of doctors in Italy from 1988 to 2019. The group was completed using the documentation from the forensic psychiatric work of the Psychiatric Clinic of the University of Sassari, Sardinia, Italy. The Italian Institute of Statistics (ISTAT) provides annual bulletins on homicides and other type of violences (e.g. intimate partner violence) disaggregated by determinants such as Region, gender and age, and type of work. We accessed the annual bulletin-specific databases to obtain te aggregate number of health-worker related homicides [17]. The EURIPSES on the other hand, is a national private agency that operates since 1982 on research in three main fields: social, political and finance. A specific national bulletin is released every year and deals with many aspects of the Italian society, including violence disaggregated by place and type of work [18]. Finally, the EURES represents another national Institute that since 1990 is intended to make research on socio-economic areas with studies at national and subnational level. The EURES also contains a specific database on homicides and related aspects like relationship between victims and assaulters, motivations, and risks analysis [19]. Based on all the available information from the aforementioned database, we first triangulated the information on figures related to homicides of medical doctors under our study period. Additionally, we searched PubMed for relevant articles. The search terms included: "homicide"; "doctor"; "medical"; "assault"; "aggression"; "nurse"; "health workers"; "hospital"; "health-care"; "kill"; "death"; "Italy"; "murder"; "physician". Both articles published in English and Italian were considered for this review. The criteria for selecting reports included the presence of murder or other closely related synonims and the exposures of interest (e.g. being a medical doctor). Epidemiological studies of any health outcome and of any study population as well as of any design, including cross-sectional, case-control and cohort studies, were considered. Two reviewers (LL and SB) evaluated the eligibility of studies. In case of discrepancy a third reviewer (AN) provided arbitration. The initial search provided 13 non-duplicate records, of which 12 full texts were assessed for eligibility. After exclusion of 5 records that did not meet the preestablished inclusion criteria, 7 studies were retained for qualitative synthesis. Of these 7 studies, 5 combined exposure and outcomes. While the literature mostly focuses on workplace violence, particularly in hospital settings, and is often primarily centred on psychiatric patients, our study has broadened the scope beyond hospital settings and psychiatric patients to include all cases of work-related doctor homicides. Specific variables were extracted from each dataset and their value as a percentage of the total was calculated. The victims' and the perpetrators' ages were grouped in 10-year intervals while the geographical location was divided into "Northern Italy" "Central Italy" and "Southern Italy and the Islands". The method of killing has been classified as "cold weapons" (referring to any weapon that does not involve fire or explosion), "firearms" and "other", while the locations of the homicide have been classified as "community-based clinics", "hospitals", "street", "perpetrator's home" and "victim's home". In addition to their gender (male/female), the offenders have also been classified according to whether they were psychiatric patients or not. A further distinction concerned the type of patient ("regular patient", "occasional patient", or "family member") and motive ("revenge", "crime of passion", "other"). --- Results From 1988 to 2019, 21 physicians were killed in workrelated circumstances in Italy, which means around 0.7 physician killed per year; slightly more than 20,000 overall homicides have been carried out in the same timeperiod in Italy. The victims were more often male (n = 15.7%), with a higher concentration (Fig. 1) in age group 50-60 (n = 7; 33.3%). Geographically, the homicides were more prevalent in Southern Italy (South and Islands) with 57.1% of cases (n = 12), while only 9.5% were recorded in Central Italy (n = 2) and the remaining 33.3% in Northern Italy (n = 7). As to method, 47.6% of the homicides were committed with firearms (n = 10) and 42.9% with sharp instruments/knives (n = 9). In one case (4.8%) a blunt weapon was used; in another case, the method was poisoning (4.8%). With regard to location (Fig. 2) about half of the homicides took place outside hospitals, mainly at GPs' surgeries or out-of-hours primary care services. The most common locations, accounting for 28.6% (n = 6) of cases, were GPs' and out-of-hours surgeries. Next come mental health outpatient clinics, making up 19.0% (n = 4) of locations. A further 19.0% of murders occurred in the street (n = 4) while 14.3% of cases occurred at the victim's home (n = 3). Three cases took place in hospitals (including adjoining locations such as the car park). One homicide took place at the offender's home (4.8%). With regard to the victims' medical specialties, the most common was psychiatry with an incidence of 28.6% (n = 6). This is followed by 4 murders of GPs (19%), 3 of primary care doctors in the out-of-hours service (14.3%) and 2 murders of urologists (9.5%). The remaining cases involved a forensic doctor, an anatomopathologist, a general surgeon, a gynaecologist, an oncologist and a neurosurgeon. In 52.4% (n = 11) of cases, the perpetrators were patients under the victim's care, while 19.0% (n = 4) were occasional patients. The remaining 28.6% (n = 6) were family members of the victim's patients. These latter six cases were all driven by revenge, specifically 4 of the perpetrators who were family members (19.0% of the total) sought revenge for the patient's death. In 2 out these latter 4 cases, the perpetrators were two fathers seeking revenge for the death of their little girls. In 66.7% of the cases recorded (n = 14), the motive for the murder was revenge without previous threats. In 2 cases, the motive was a worsening of the murderer's health. A further 2 cases were crimes of passion.. In only one case did the claimed motive originate after the victim had stopped treating the perpetrator. In another case, the motive for the murder is not entirely clear even though the offender, a psychiatric patient on probation, blamed both his victim and all the other doctors at the mental health clinic for the regime he was subjected to (i.e. for being required to report to the Mental Health Centre, MHC, every 3 days to receive treatment). In the last case, the murderer was a medical psychologist who was both a patient and a colleague of the victim, and who took revenge on his colleague for having been subjected to two compulsory mental health hospitalisation orders at the healthcare facility where he had previously been employed. In 6 cases (28.6%), the motive was revenge preceded by the offence of stalking/threats. Specifically, 2 cases were passion-related: the victims were two women doctors who had long been subjected to threats and stalking by their patients. In one case the murder was preceded by a caution issued by the police. However, 23.8% of cases (n = 5) are not explained by either passion or revenge. Indeed, 3 murders were committed in the out-of-hours general medical facilities by occasional patients (2 drug addicts and 1 intoxicated person) while 2 other cases took place at MHC outpatient clinics. Analysis of the data collected also shows that in 14.3% (n = 3) of cases the homicide was followed by the offender's suicide and in 9.5% (n = 2) the perpetrator killed more than one person. Lastly, 42.8% (n = 9) of the perpetrators had a psychiatric disorder while the remaining 57.1% (n = 12) had no diagnosed disorders at the time of the homicides (Fig. 3). --- Discussion Based on analysis of the data, we would like to highlight that murders on doctors are an extremely rare event; however, some relevant remarks can be made on the characteristics of the phenomenon and possible preventive measures. Firstly, the literature and statistical data confirm that the medical profession is a dangerous one. In addition to the constant risk of professional liability due to an increasing number of malpractice claims, doctors are exposed to the risk of physical assaults at work and, in extreme and fortunately rare cases, of being killed in connection with their profession. The US BLS (Bureau of Labour Statistics) (OSHA 4) reported that 69 healthcare workers (HCWs) were killed between 1996 and 2000 [20]. In Italy, in the period between 1988 and 2010, 17 physicians were killed at work [20]. The literature highlights the aggressiveness and violence of patients in hospitals and particularly the risk associated with psychiatric patients [21][22][23][24][25]. We have analysed the cases of physician homicides linked to the doctor-patient professional relationship, both within and outside clinical settings. The sample we examined supports more complex reflections. The first reflection concerns the location of the homicide: in most cases it was not the hospital. Out of 21 cases, only 3 homicides were committed in the hospital or in nearby places (e.g. car park), while the other murders took place outside the hospital. In particular, a significant number of murders occurred in outpatient settings (community clinics), while still others occurred at the patient's home, at the doctor's home or in casual places (on the street). In a particularly striking instance, the patient took a revenge on the doctor: he sent a poison-laced bottle of wine as a Christmas gift to the victim's home for the Christmas holidays, causing the doctor's death. This unusual and horrifying case confirms that the homicidal intent may play out beyond healthcare settings and reach the victims elsewhere, including in their own home. The significant number of homicides in places other than hospitals calls into question the limited scope of preventive measures, which so far have focused on hospital settings. Some authors [26,27] have reported a sharp decline in violent behaviour against healthcare workers after the introduction of specific security measures in hospitals. Similar systems would certainly also be effective in community clinics, where almost 50 homicides have taken place [17]; these outpatient settings as a rule have insufficient security arrangements and would require the implementation of security protocols. Another observation concerns the type of offender. In our sample, the perpetrators included not only the doctor's regular patients, but also patients' family members and occasional (drop-in) patients. With regard to the authors of violent acts in the workplace, Rippon [28] identifies several categories: patients, family members, visitors and co-workers. In Turkey, a study found that 64.5% of the attacks were carried out by the patients' family members [29]. This data also exposes the limitations of violent behaviour risk assessment and management systems in reducing the risk of violent behaviour by patients [24,30]. While these tools are certainly useful when dealing with patients registered with healthcare facilities, they may fail to prevent homicide in cases where the perpetrator is a patient's family member or an occasional/drop-in patient, as these types of individuals cannot be subjected to risk assessment. In the sample analysed, almost half of the perpetrators of homicide were patients' family members or occasional patients. The sample reveals a "wide scope of danger" both in terms of location, which extends beyond the hospital setting and may also include the doctor's or the patient's home, and in terms of offenders, who may also be patients' family members. Some remarks can also be made about the motives for the murder in the sample. In a large number of cases the motive was revenge against the doctor for a claimed error in diagnosis, surgery or treatment causing harm to the patient. In a number of cases the murder did not occur as an escalation of an outburst of violence but was the outcome of planned and premeditated revenge against the doctor. Often the revenge was preceded by threats. In some cases the murder was preceded by stalking. While the instances of stalking seem to be relatively few, their number may be an underestimated also due to the fact that this behaviour has been recognised as an offence only in recent years. Therefore, it would be appropriate to highlight the role of stalking as a risk factor for homicide, and to raise public awareness of this behaviour to improve its management and prevent its escalation into violent acts. While of course stalking behaviour is not necessarily a precursor of murder, a useful recommendation is to report all stalking behaviours and take specific precautionary measures. Finally, psychiatric patients warrant specific remarks. Large and Nielssen [30] have analysed homicides by psychiatric patients in psychiatric hospitals and proposed a classification into three patient categories: acute psychiatric patients soon after hospital admission, patients not receiving medication with a history of serious violence, and patients with dementia or intellectual disability, held in low-security inpatient settings in contact with vulnerable patients (victims of the homicide). In this study, the victims were both healthcare workers and fellow patients. In the sample we analysed, 43% of the perpetrators had a diagnosed psychiatric disorder, while 57% did not. We do not have data on how many offenders with psychiatric disorders were found to be not prosecutable due to mental impairment. A study by Knable [22] describes a sample of mental health workers who were killed by patients in the US. The study highlights the low frequency of these events (about one case per year) and describes the characteristics of victims and perpetrators. The victims were most likely young female caseworkers, with little work experience, killed during unaccompanied visits to residential treatment facilities. The perpetrators were mostly male, diagnosed with schizophrenia, with poor adherence to medication. The debate about the link between mental illness and violent behaviour is still open and very controversial. In accordance with the literature and in light of the data examined here, we recommend that psychiatric patients be assessed for their risk of violent behaviour, particularly if they have a history of violent behaviour, current substance use and non-compliance with their medication [24,[31][32][33][34][35]..However, it should also be noted that individuals (patients or family members) who are not mentally ill can also be offenders. --- Conclusions This paper is significantly limited by the small size of the sample group, which makes it difficult to draw general conclusions. However, with this caveat, we can highlight six things to know about the homicides of physicians. 1. Physicians may be attacked not only in hospital settings but in other locations too, even at home. Isolated outpatient clinics are at high risk and would benefit from improved surveillance and security systems. 2. The perpetrator is not necessarily a patient, as doctors may be targeted by a family member of the patient seeking revenge. 3. Many cases stem from allegations of medical malpractice against the doctor. Conflict mediation systems should be set up for the parties' use. In Italy, the Gelli-Bianco Law, on the safety of healthcare and the professional liability of healthcare providers, has introduced compulsory mediation before legal action can be taken. --- 4. Stalking is a risk factor that should not be underestimated; doctors should be aware that it might precede violent behaviours. Physicians who are victims of stalking should always contact the police. 5. Psychiatric patients, especially those with a history of violent behaviour, substance use and poor adherence to their medication regimen should undergo a specific risk assessment for violent behaviour. However, the sample analysed suggests that the threat is not limited to psychiatric patients. 6. Physicians should receive specific training in effective communication skills and conflict management with patients and their family members, including non-violent communication techniques, listening skills and conflict identification and recognition. While each case in our sample has unique characteristics, each can be considered a piece of a broader puzzle, which requires more studies in order to move beyond prejudice and silence. --- Availability of data and materials The datasets generated and/or analysed during the current study are available in the ISTAT (https://www.istat.it/), EURISPES (https://eurispes.eu/) and EURES (https://www.eures.it/) websites. Public access to the databases is open. --- Declarations Ethics approval and consent to participate Ethics approval and consent to participate were not required, as this is an observational study based on anonymized data routinely collected by the ISTAT, EURISPES and EURES, which are public Italy bodies reporting on health and socio-economic determinants (according to Italian law, the study was exempted from approval by an ethics committee because all the data used were deidentified and publicly available). This study is also ethically compliant with the National Law (D.Lgs. 101/2018) and the "General Authorisation to Process Personal Data for Scientific Research Purposes" (nos. 8 and 9 of 2016, referred to in the Data Protection Authority action of December 13, 2018). --- Consent for publication Not applicable. --- Competing interests The authors declare that they have no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Healthcare workers have a 16 times greater risk of suffering workplace violence than workers in other sectors and around 50% experience workplace violence in the course of their career. The objective of this study is to explore the characteristics and circumstances of work-related killings of doctors. Methods: Work-related homicides of doctors over the period 1988-2019 were identified retrospectively through the Italian national statistical agencies. Variables such as perpetrator, motive and location of the crime were obtained through forensic psychiatric work. After classification, the absolute and percent values of the main characteristics of the homicides were calculated. Results: Over the period considered, 21 doctors were killed in Italy in connection with their professional activity. In 52% (n = 11) of cases, the killer was one of the doctor's patients, in 29% (n = 6) of cases it was a patient's relative, in 19% (n = 4) an occasional patient (first consultation). The location of the homicide was a community clinic in 48% (n = 10) of cases, the street in 19% (n = 4) of cases, the doctor's home in 14% (n = 3), the hospital in 14% (n = 3) and the patient's home in 5% (n = 1). In 57% (n = 12) of cases the perpetrator was not affected by any mental disorders. The motive for the homicide was revenge in 66.7% (n = 14) of cases; in 28.6% (n = 6) the revenge was preceded by stalking. Conclusions: Doctors should be aware that the risk of being killed is not limited to hospital settings and that their patients' family members might also pose a threat to them.

Introduction "Why should we cherish "objectivity", as if ideas were innocent, as if they don't serve one interest or another? Surely, we want to be objective if that means telling the truth as we see it, not concealing information that may be embarrassing to our point of view. But we don't want to be objective if it means pretending that ideas don't play a part in the so-alt.chi: Confronting Power in HCI #chi4good, CHI 2016, San Jose, CA, USA cial struggles of our time, that we don't take sides in those struggles." "Indeed, it is impossible to be neutral. In a world already moving in certain directions, where wealth and power are already distributed in certain ways, neutrality means accepting the way things are now. It is a world of clashing interests -war against peace, nationalism against internationalism, equality against greed, and democracy against elitism -and it seems to me both impossible and undesirable to be neutral in those conflicts." Howard Zinn, Declarations of Independence: Cross-Examining American Ideology. [30] "A phone, an Internet mobile communicator, a strangle hold on modern technology... A phone, an Internet mobile communicator, a strangle hold on modern technology... A phone, Do you get it? These are not three separated devices. This is one device, and we are calling it iPhone." Steve Job's Satirical Misquotes, 2007 This paper is a polemic. As such it makes no pretense at remaining neutral. As engineers and designers it's often not really in our interests to take sides in social and political debate. Our livelihoods depend on people giving us money to build things. Within HCI this has led to a passive acceptance of a middle class aspirational world where technology assists and entertains us. For example: "Let's age Tommy to 3 years old. Tom and Sara take him skiing for the first time. Tommy's SmartPhone, now version 23.0, downloads the 'Virtual Skiing Coach,' which uses accelerometers sewn into Tommy's clothing to sense his posture and then offer suggestions for maintaining balance; when it foresees an impending collision, it quickly blurts out instructions on how to stop." Generation Smartphone [25] Is this it? Is this the best we can do with all our technology and resources? How about making the world a better place? Has Candy Crush made the world a better place? Has it? Can you stop playing Candy Crush for a minute and answer me? Oh forget it. The smartphone, just like the personal automobile, has a big impact on society for good and ill. However, professional organisations who represent the technologists who are building and developing these systems are reluctant to take a position on how this technology is used or developed. SIGCHI has been described as "scrupulously apolitical" [11] p.57, whereas IFIP, the organisation that organises the yearly Interact conference in HCI, proudly describes itself as "The leading multinational, apolitical organization in Information & Communications Technologies and Sciences." http://ifip-tc13.org/about-us/ However, as Zinn notes, neutrality itself is a position, in general one supporting the status quo. Therefore, whereas one may laud the difficult balancing act required to satisfy the many interested parties in HCI, there may also be some discomfort with the extent the HCI community may seem to distance itself from the political and social consequences of the technology it researches, builds and promotes. It would be extreme to describe the HCI community as politically supine, playing lip service to social good while queuing up to take cash from commercial and government interests, but an outsider could get this impression. So wake up! The world is changed by technology. It's time to get off the fence. If you think smartphones will lead to a utopian information sharing society then make that case, if you think they are an evil scab representing a commercial strangle hold on Western society (and shortly all society) then let's hear you say it. In this paper we explain why the smartphone is one of the most important pieces of technology to affect our world since gun powder, the mechanical timepiece, and the petrol engine. We then consider its affect on complimentary and competing technology, and conclude by arguing that the smartphone is the ultimate technology to support an American flavour of radical individualism, and a means of concentrating power into the hands of the few. Almost all, new, high impact technology is accompanied by a shrill warning of doom: The novel, a product of cheap and ubiquitous printing, will undermine the morality of young ladies; TV, then video, then computer games, then social media, will corrupt the innocent, undermine society and lead to a dystopia, and so forth. Such warnings are often based on the claim that this or that technology is fundamentally different from technology that has gone before. Such differences are always a matter of degree, a mechanical time piece is similar to a sundial but allowed time keeping to become universal and pervasive, video was similar to TV but removed the control of what was seen and when from the broadcasters, a smartphone is similar to a desktop but you can use it while mobile and use it as a telephone. Key to the impact of technology is its role within society, its interaction with commercial and political forces, and the relationship individuals have with the technology. The smartphone is special because everyone has one (or will have one), it's the dominant form of interpersonal communication, and it is the dominant means of accessing data services. These services are not built by a benevolent set of magicians living in a fairy kingdom. They are built for the primary means of commercial exploitation. You are a data point and you are valuable. This is not the desktop you used to log onto once a week to write a letter or play a computer game, this device is with you all your waking hours recording everything about you. Commentators such as Morozov [10,18] and Naughton [19] elegantly articulate the importance of data ownership as well as critiquing the influence of Silicon Valley, and its offerings, on everyday lives and culture. Morozov in particular is well known for reprimanding the drive to solve problems -using arbitrary technologies -that do not actually exist, even calling out work published at CHI for this (i.e. "Bincam" in Thieme et al 2012 [27]). A more thoughtful, and constructive, treatment of this problem space is given by Raghavan [23] in his discussion of how Sevareid's Law ("the chief source of problems is solutions") can provide useful motivation to understand current technology design. However the importance of the device that is used to provide data to, and extract data from the individual, is underestimated. The smartphone is powerful, not just because of this role as a gateway, nor because of the commercial power behind it, but also because it has a critical psychological hold on its users. But we have seen nothing yet. Media streaming services have already killed off the high street music store, and the smartphone has pretty much killed the iPod [28]. In the future credit and debit cards can be replaced with the smartphone, car information systems can be replaced by the smartphone, remote controls for household appliances can be replaced by the smartphone. The trick is to lock users in to a specific system, and skim revenue from all these new activities. Currently tech companies like Apple are dwarfed in revenue by companies like Walmart, oil and gas are still the big revenue generators, but the smartphone offers unparalleled opportunities to control and make money out of a users everyday behavior. That is why Apple is valued at over $700 billion. Just as credit and debit cards began as a convenience to avoid carrying cash, to manage monthly expenditure, now it's very hard to live without a credit or debit card in modern western society. It is only a matter of time until having a smartphone will be a requirement for our day to day lives (and for many users they already are). --- The Psychological Importance of the Smartphone "Cluley and Dunne (2012) argued that a psychoanalytic account of commodity fetishism is needed to contend with the contradictory ways people consume. Retreading Marx with Freud, they conclude that narcissism rather than fetishism is truly what is at stake because the commodity form is not merely a masking of social relations of production; it is also, if not more so, a means of identification and selfaggrandizement." p.114, Reyes et al [24] We can argue that important commercial factors drive the ongoing development and take up of the smartphone, but it is our psychological relationship with the technology that makes it so powerful. Reyes et al [24] French TV, the approach offers us some insights into why the smartphone is so powerful. From Lacan they use the term look and the term gaze. Critical is their contention that the smartphone is more than just another fetishised commodity 1. "In brief, our position is that the look involves phenomena of self-presentation and copresence, seeing and being seen as a social subject in real, brick-and-mortar space. The gaze, on the other hand, entails disruptive moments of disengagement with that first domain of the look. The former level of analysis is most congruent with Marxist conceptions of commodity fetishism, whereas the latter level of analysis is necessary for contending with cell phones as more than just another fetishized commodity and coming to terms with their unique position as communications technologies embedding consumers within new mobile media ecologies." -p.115, Reyes et al [24] AN Our interpretation of this is that the way the smartphone controls our consumption of media; mass, personal and in terms of social interactions, makes the device special. The smartphone connects us to another realm, as such it transcends technology and becomes a gateway, like a mirror, to the other, as well as the device being a traditional commodity. This affects the way we use and relate to the device. As an ethnographic study Reyes et al [24] relate this psychological framework to concrete observations, for example: The tension between the smartphone as a normal commodity (the look), in contrast with it's role as a gateway to the other (the gaze), can help explain our relationship with the device. Reyes et al [24] list conflicting behaviors caused by this tension: Individuals customise there device but must conform to the constraints of a mass produced device; They stay physically connected to their smartphones, even when disconnected from the information the smartphone is designed to access; Individuals are concerned with the impact the device has on their privacy but use and display the smartphone very publicly; Individuals exhibit discontinuous behavior, alternating between engaging with technology and the people around them; Individuals find themselves disregarding established social norms in order to use their devices. "I The smartphone, as a gateway to the other stains the environment (not necessarily in a negative way), but in a way that cannot be ignored. This is key to the psychological importance of the smartphone and leads to enormous social and personal impact much of which is extensively examined in modern literature. For example from the effect of the smartphone on feelings of isolation [22], the creation of digital identities [20], to the etiquette of sharing a smartphone (or not) [13]. As with the wrist watch and the automobile, the smartphone extends from the commercial deeply into the psychological. Here we need to distinguish between concepts like the cult of the individual -its bad to have the state oppress individuals because everyone matters -and radical individualism -Its all about me, and everything important and useful comes from individuals with the vision making it all happen (unless their strength is sapped by red tape and no-hopers supported by state handouts). Barbrook and Cameron [3] Within this context, the smartphone is a political tool; something that fights authoritarianism such as in the Arab Spring (Well, okay Egypt is looking pretty unsprung now but at the time the smartphone was heralded as a tool that would change political landscapes forever); something that can regenerate local communities with ideas such as time banking; or in contrast fight unionised labour and organised capitalism for example with Uber and AirBnB. But the true contribution of political change that we derive from the smartphone is the ability to monetise everything (monetise by the second, by the kilobytes, by the individual) and know everything about individuals (their interests, their movements, their routines, their friends and contacts). De Unamuno [6] offers a detailed exploration of the political impact of the smartphone from a Marxist perspective in order to "understand how the use of contemporary technology and smartphones in particular, enable an advanced form of exploitation, where smartphones are not only used to extract surplus value from workers' personal time-space, but also marketed as essential to the workers' cultural identities, something the workers must have but must not question." p.9-10 [6] However, the smartphone has also been promoted as something that can subvert formal power structures, encourage local community engagement and support the disenfranchised. It has been actively argued that the smartphone is a force for social good. For example Han et al [12] argues that "mobile technology suggests new opportunities for community informatics". In this work they use two community web services, one focused on digital cultural heritage, the other on local volunteer efforts using time banking. Han et al's work show how "mobile technology transcends the limitations of time and place, it expands the ways of accessing and interacting with local community information and lowers the barrier to participation". We applaud the work of researchers attempting to explore the use of technology for social good. However, believing that smartphones offer a route to achieving this goal is mistaken. Take Han et al's [12] study on time banking using hOurworld. If we find the Penn time bank we discover 189 current members with a total of 58 exchanged hours. The total staff and student population of Penn is approximately forty thousand so pretty much no one at UPenn uses time banking. We see this same story again and again within HCI Academic research: the potential for social good is outlined, a study is conducted, the results are promising, no take-up occurs, no interest is shown by commercial players. There's an irony that a device ostensibly for communicating with others is all about the person using it. The smartphone rivals the personal automobile as an expression of the self. You can't share them, they store all your personal data (cleverly in the cloud so you are locked into a specific manufacturer), and they are like little mirrors that you can stare dead eyed into and hope to see a better more aspirational you instead of a sad wage slave stuck in a soul destroying commute. Uber and AirBnB are the ultimate expressions of a laissez-fair system. By using the smartphone to monetise time by the minute and services by the kilobyte we can dispose of regulation, unionisation, collectivism in general. The phone is yours and it empowers you as an individual. However, the politics of the individual is, by it's nature, a politics that attempts to deny the existence of politics. It's not about the group, it's not about society, it's not about justice, its about you. In a dazzling slight of hand powerful organisations and interests convince users that only their individualism matters, thus disempowering the powerless, and generating vast quantities of personal data that empower the powerful. Dominant technologies have a significant impact on related fields; steam power was extinguished by the combustion engine, the typewriter by the desktop computer, the telegram by the telephone. Sometimes this is progress, but sometimes, like the effect of the personal automobile on pedestrian access, or cycling, the dominant technology can also impede progress. The smartphone is dominant for three reasons; 1. The commercial, psychological and political drivers behind the technology are very powerful 2. The ability of the smartphone to absorb multifunctionality, and 3. The ability for devices to become peripherals of the smartphone. The iPod was one of the first casualties [28], The desktop is pretty much the second attacked on two sides by tablets as well as smartphones. Aylett and Quigley [2] argue that the field of pervasive computing has been subsumed in many ways by the smartphone, and strongly argue that the smartphone, although ubiquitous, is not ubiquitous computing because "Devices have to be sold as personal, beautiful toys that sit squarely in the centre of your attention." p.433, [2]. The potential for wearables is also heavily compromised by the smartphone. Google Glass was presented as a new way to engage with computing, but apart from being a massive mistake in terms of understanding how people outside the Bay Area saw technology 2, why use something like Google Glass when most of the functionality is already on your beautiful smartphone? The Apple watch of course requires an iPhone to work. Do you really think Apple want to sell a product to replace the iPhone? Buy another product to go with your smartphone sure, make sure both have to be up to date and you have even more reasons to throw out a 2 year old phone, but replace? Are you crazy? Meanwhile the smartphone is likely to subsume the credit card, audio storage 2 Sorry Google is not cool anymore. devices, in car entertainment, sat navs, and with the right peripherals fitness devices. A lot of designers implicitly understand the danger of the smartphone gobbling up anything new they design. With a significant number of project intentionally not using an app to realise the design objectives. For example the datacatcher project [7] where bespoke technology was manufactured rather than sticking some software on a smartphone. The smartphone is the new QWERTY keyboard, its going to be around for a long, long time. --- Conclusion So the world will move onwards and no one will care how many Facebook posts you made, or what you retweeted, or how groovy your choice at Spotify. You will get old and you will die. The data you generate over your life staring at your beautiful little technological mirror will be gobbled up by machine learning algorithms dedicated to maximising profit. Currently the dominant response to this within the HCI community is to make a prettier mirror. This is in direct contrast to the many researchers in the humanities producing incisive comment (if in a rather wordy and over florid style for us engineers) on modern technology. We acknowledge that the CHI community increasingly draws upon literature, the humanities as well as the social, political and communication sciences in its efforts to understand the current-and future -societal impact of interactive technology. However, these efforts are all too often relegated to the alternative venue of alt.CHI (for example see Kirman et al [15], Baumer et al [4]) or squirreled away as sexy sounding -but all too niche -design fiction workshops (for example see Linehan et al [16]). In perhaps the most relevant main track CHI paper to our rant Thus, in the context of the CHI 2016 theme, CHI4Good 3 with the focus on the "under-served, under-resourced, and under-represented", we hope this paper may encourage a few CHI attendees to read these commentaries and perhaps incorporate or cite them in their subsequent research output. We hope that our underlying political position may provoke engineers who disagree with us to do so openly. Politics, is after all, very much about debate. However, this so called neutral, apolitical position of HCI in the face of massive potential social upheaval and change is just not tenable. We may not be able to do much about the social impact of our technology, but if we can't then who can? The paper draws solely on a Western understanding of technology; the smartphone is portrayed as taking shape through two single bodies: the (Apple) industrial designer and the user. The laboring hands that assemble devices for consumption on a factory floor somewhere in Asia or Sub-Saharan Africa are rendered as invisible as these regions' designers. "The smartphone rivals the personal automobile as an expression of the self," the paper concludes. The comparison with the automobile could have been productive, if the authors had followed through; the making of the car and the mobile phone unfold not only through the making of consumers, but also drastic restructurings of work from Fordism to outsourcing, automation and digital labor. Just as the car haunts the American dream of middleclass, so the mobile phone haunts the knowledge economy and the idea that information technology would lead to the elimination of the factory. The authors feed into what they critique: a dominant view of tech innovation that equates it with Silicon Valley, masculinity, and Western capitalism. The paper itself is not a critique of HCI but highlights why so much of HCI fails to engage its critical turn; because it's easy to launch one critique after another. By ignoring prior work one can more easily stake out new territory: the person who brought Lacon to HCI, the paper we now need to cite when we write about smartphones. The sensationalism of the rant will generate conversations at the conference much like a Fox News headline, a mode of "scholarship " that is unlikely to have the kind of impact in education, technology and society that we should have as the technical and social make each other in ever more complex ways. --- Commentary For alt.chi paper

YAFR (Yet another futile rant) presents the smartphone: an unstoppable piece of technology generated from a perfect storm of commercial, technological, social and psychological factors. We begin by misquoting Steve Jobs and by being unfairly rude about the HCI community. We then consider the smartphone's ability to kill off competing technology and to undermine collectivism. We argue that its role as a Lacanian stain, an exploitative tool, and as a means of concentrating power into the hands of the few, make it a technology that will rival the personal automobile in its effect on modern society.

Introduction The overrepresentation of students from a migrant background in special education constitutes a phenomenon of special interest that refers to the demonstrated tendency among students from a migrant background to be identified and placed in special education programmes at disproportionately higher rates compared to their native peers in the education system [1,2]. In the field of education, this phenomenon has been international [3,4]. The case of the USA is significant, as this is one of the most recurrent research topics in the last fifty years [5,6], since Dunn [7] highlighted the overrepresentation in special education of not only migrant children, but also children from social minorities and non-middle-class family environments. However, in Spain, the phenomenon of overrepresentation has received rather insufficient attention, despite the important migrant movements recorded, especially in the south, and in spite of the warnings given already in 2005 by the European Monitoring Centre on Racism and Xenophobia [8] about the large number of students from a migrant background in special education centres in Europe. The phenomenon of overrepresentation of students from a migrant background and students from minorities in special education poses significant concerns that cannot be ignored, as they affect Human Rights and Children's Rights, which are related to educational equity, social justice and inclusion, and these concepts are strongly related to each other within a desirable culture of peace in educational centres, as has been recently stated by A<unk>a<unk>os et al. [9]: "The movement for a culture of peace and non-violence [...] conceives peace as a construction, as a project that concerns all citizens. The ingredients to achieve this are: democracy, dialogue, solidarity, justice and a spirit open to difference and diversity. The latter are the ingredients of inclusive education in schools, where educational work is an instrument of transformation" (pp. [24][25] In this sense, and within this context, the aim of this study was to analyse the phenomenon of overrepresentation of students from a migrant background in special education centres, specifically in the stage of primary education, in the region of El Campo de Gibraltar (Southern Spain), which is an area characterised by the arrival of an important migration flow. --- Inclusive Education to Transform Education and Society The United Nations Convention on the Rights of Persons with Disabilities [10] underlines the importance of the full and effective inclusion of people with disabilities in all aspects of life. This convention establishes fundamental principles that advocate for non-discrimination, equal opportunities and respect for human diversity. Similarly, it recognises people with disabilities as full rights-holding individuals and promotes their active participation in society. Within this framework, inclusive education emerges as an essential objective, aiming to guarantee an equitable access to education for students with disabilities in conventional schools [11,12]. However, this goal cannot be met in the short term, as it implies a process of evolution and development that requires time. Studies reveal that inclusive education not only contributes to the social development of all students, but it also improves educational opportunities as a whole [13]. Additionally, it promotes understanding and tolerance, playing a key role in the eradication of prejudice and discrimination against students with disabilities. Studies such as that of Achamrah [14] demonstrate that students with disabilities at conventional schools reach high performance in both the academic and social scopes compared to their peers at special education centres. This perspective finds consensus in the belief that the education of students with disabilities can be easily adapted to the general curriculum. Therefore, it is obvious that most students with disabilities do not attain an optimal progress in their education when they are isolated, thus emphasising the importance of inclusion as a means to ensure the full exercise of their right to education [15]. Therefore, it is considered that inclusive education not only represents a fundamental right supported by international agreements, but it also stands as a means toward a fairer and more equitable society. To achieve the full inclusion of students with disabilities in conventional schools, it is essential to implement inclusive policies and provide the necessary training and resources for teachers to create a welcoming and enriching educational environment for everyone, regardless of their differences. This approach not only benefits students with disabilities, but it also enriches the educational experience of all students and promotes greater understanding and tolerance in society as a whole [12]. --- Overrepresentation of Students from Ethnic Minorities in Special Education Centres Diversity in the classroom, in terms of cultural origin, language skills, or disabilities, is an increasing reality that educators must face [16]. Inequality in the educational success of students from a migrant background of ethnic minorities persists as a serious problem [17,18]. These students have a greater representation in special education compared to their Caucasian peers [19], which is a concern that has prevailed for decades [7,[18][19][20]. This overrepresentation in special education centres demands immediate attention and action [21]. With the aim of improving the success rates of these students, it is essential to implement programmes that promote early intervention, respect the different cultures and address the specific challenges of the latter [22]. In Spain, special education is governed by Organic Law 3/2020, also known as the LOMLOE, which establishes the structure and services for students with special educational needs (SEN). Since each group of students has specific characteristics and needs, which require adequate curricular adaptations [23], the LOMLOE emerges as a legal framework that sets measures of attention to diversity, curricular adaptations and specific resources to guarantee an inclusive education. For a student to be referred to this system, she/he must have gone through a process of identification and valuation of her/his educational needs by competent professionals. However, the decision of referring the student to a special education centre is jointly made by parents, teachers and professionals. Although the ultimate goal is to benefit the student, there is some criticism of the referral of these students to special education systems, as it can lead to the segregation of the students and the stigmatisation of their educational needs, thereby limiting, for instance, their opportunities to interact with peers without disabilities, and thus affecting their social inclusion [24,25]. Therefore, educators must be ready to attend to different needs and capabilities, since classroom diversity demands a constant adaptation of the pedagogical strategies. In this sense, it is necessary to provide specific and adequate training to educators, in order to guarantee that they can attend to diversity, especially in centres with students and families of particular characteristics [26][27][28]. In the case of students from a migrant background with disabilities, they often face unique challenges, such as language and cultural barriers. The current lack of diversified training makes it urgent to develop specific training in each educational centre, adapting it to their realities [29]. Consequently, specific training may help educators to effectively respond to these needs. This fact is not only necessary to guarantee equal opportunities, but it also contributes significantly to improving the academic performance and general wellbeing of all students, in general, and migrant students with disabilities, in particular [30]. In turn, language barriers represent a significant challenge for students from ethnical minorities in special education [31,32]. These barriers are often mistaken for learning difficulties, while the actual problem lies in a lack of understanding of the language [33][34][35]. The monolingual perspective of schools may influence the evaluation and support of these students, contributing to the problem of unequal representation [1]. To address this, it is crucial to adopt a holistic and intersectional approach, promoting cultural sensitivity and early intervention strategies [17]. Moreover, families, educators and health professionals must collaborate with each other to provide adequate support to the students [18]. Consequently, to fight the disproportionate representation in special education, it is fundamental to adopt culturally responsive educational systems that value and use the culture, language and experiences of all students [36]. This requires a profound transformation in the assumptions and practices of the educational structure, from classrooms to decision making in the school system [37]. Collaboration and transforming change are essential for the attainment of a more equitable education system that receives cultural diversity. --- Situation of Students with Special Educational Needs from Morocco Although the intersection between cultural diversity and special education has not been thoroughly explored in the Spanish context (where this study was conducted), the topic of migration received considerable attention in the 1990s [38]. From the mid-1980s, the economic growth in Spain led to an increase in the number of migrants who settled in this country, resulting in a population of migrants that represents 14% of the total population of Spain. Thus, during the 2000s, Spain emerged as a relevant destination for international migration. According to the data provided by the Spanish Institute of Statistics, in 2021 [39], 5,375,917 migrants lived in Spain, of whom 775,936 were from Morocco, representing 14.4% of the migrant population, i.e., the highest percentage of migrant citizens. Specifically, the Report on the Integration of Foreign Students in the Spanish Education System [8] states that, in the education scope, Spain has 30% of students from a migrant background from Africa, followed by the European Union (24%) and Latin America (21%). In the academic year 2021/22, a total of 882,814 migrant students were registered, representing 11% of the total body of students and almost 15% of the total migrant population in Spain. Regarding special education, in the academic year 2020/2021, 5135 migrant students were registered in public and private educational centres in Spain, of whom 1154 students were in Catalonia, 934 in the Community of Madrid and 670 in Andalusia [8]. These data indicate a significant increase compared to previous years, underlying the importance of understanding the migratory phenomenon between Spain and Morocco [40]. Two of the factors that can be identified in the migration of Moroccan families in Spain are [41]: (1) academic training and (2) the educational attention given to their children. During the last two decades, the education system of Morocco has faced a series of significant challenges that have led it to the limit of its capacity [14,42]. Moreover, the academic level of Moroccan migrant students is notably lower than that of Spanish students, and even that of the rest of the migrant community [38]. In this sense, and according to the National Immigrant Survey Report [43], 57% of individuals aged 16-64 years had not completed primary education (in contrast with 29% of all non-community migrants and 20% of migrants born in Spain), and only 3% had completed higher education (compared to 13% and 19%, respectively). Measures such as the Spanish Letter of Education and the Emergency Plan were developed to correct the deficiencies of the Moroccan education system [42]. However, despite these efforts, the results obtained to date appear to have been insufficient. This situation is demonstrated in the disproportionate representation of Moroccan students in other countries, as is the case of Spain. Moreover, although different factors have been reported to contribute to this disproportion, little effort has been made [44,45] to place these factors in the context of broader social and sociological phenomena, such as the cultural construction of disability, categories of disability and conceptualisations of individual difference. In addition, the challenges identified in the communication processes due to the language and cultural differences may, in certain cases, hinder mutual understanding [31,38,46,47]. Due to the scant and outdated literature on this topic, we consider that this study is very relevant, not only for the knowledge generated, but also for the new research lines that can be developed from its results, which are necessary, real and up-to-date. Thus, with the aim of contributing to the findings of other authors, the aim of this study was to analyse the causes of the overrepresentation of students from ethnic minorities (including both students born outside of the study country and students born in the study country with a migrant family background), mostly Moroccan, in special education centres in the region of El Campo de Gibraltar (Southern Spain). The following research questions guided the analysis: --- • What is the experience and training of teachers and specialists who work in special education centres in the region of El Campo de Gibraltar? --- • What are the main reasons that lead families to migrate and school their children in a different country? --- • What barriers do professionals identify with the families during the schooling period of students from a migrant background? --- Materials and Methods The present study is framed within the project entitled "Study on the disproportionate representation of students from a migrant background in special education centres in the province of Cádiz" (MigraNEE Cádiz), funded by the University of Cádiz-Convocatoria Proyectos Puente (REF. PR2022-056). The project analyses the different education stages (primary education, secondary education and basic vocational training); however, the data presented in this study are specifically focused on the stage of primary education. The aim of the present work was to analyse the overrepresentation of students from a migrant background in special education, specifically in primary education, in centres of the area of El Campo de Gibraltar (Southern Spain). To this end, a qualitative methodology was followed, based on the interview technique. This study was conducted from an ecological-cultural and inclusive approach [48,49], justified by the need to know and delve into the different discourses provided by the participants included in this investigation. Thus, with the use of thorough interviews, the aim was to generate knowledge through the voices of those involved in this phenomenon, with the participation of the community. --- Participants A total of 16 interviews were conducted in two special education centres, an early childhood and primary education centre and an Educational Guidance Team (EGT) of the area. The anonymity of all participants and educational centres of reference was guaranteed throughout the process. Table 1 shows the final sample. When selecting the educational centres, it was necessary to establish a profile of their students and families who met the characteristics of this study. Consequently, the selection of participants was based on the following criteria: --- • Education centres with a larger number of students from a migrant background with disabilities compared to other centres (this information was provided by the counsellors of the EGT); --- Instrument For the gathering of information, two ad hoc semi-structured interviews were designed, based on relevant questions about the profile of the educators, the situation of the students and their families and aspects that may explain the phenomenon of overrepresentation of students from a migrant background in special education. The interviews were revised and piloted by the research team of the project. After making the proposed changes, the script of the interviews was revised and the final changes were made. The interviews were designed inductively, with open-ended questions. Individualised interviews were developed for the counsellors of the EGT, the principals, the teachers and specialists in therapeutic pedagogy (TP) and hearing and speech (HES), and the social integration technician (SIT). --- Procedure To access the potential participants, the EGT of the area was contacted, and the nature and goals of the project that encompasses this study were presented to the members of the EGT. Their collaboration was requested, as well as establishing contact with the centres that met the characteristics of this study. Once the information of the centres was obtained, these were contacted, agreeing to a date and time for the interviews, taking into account their availability. Each participating group had an interview script specifically adapted to it, with approximately 40 min each. The interviews were conducted individually and in groups, face-to-face, and they were guided by the members of the research team. At the beginning of each interview, the participants were asked to provide their informed consent to have their conversations recorded, in order to gather every detail, notifying them that these would only be used for academic purposes, guaranteeing their confidentiality and anonymity. Lastly, all interviews were transcribed. --- Data Analysis Based on the data analysis systematisation model of Miles and Huberman [50], after gathering the data and transcribing the interviews, the information was reduced. This process consists in "selecting, focusing and abstracting the raw data into units of meaning, known as content categories" ([51], p. 358). Regarding the categorisation of the content, Rodr<unk>guez et al. [52] pointed out that this process implies "judging, valuing whether certain units can or cannot be included in a certain code, and making decisions in that respect" (p. 210). Taking into account the indications for this process, from the present study, a code was assigned to each unit of meaning; each of these codes refers to the categories in which they are included. This coding process was carried out inductively, thus the categories emerged from the testimonies of the participants. Once the coding process was performed, three general categories were defined, which refer to the professionals who had been interviewed, as well as to the information they provided in relation to the students and their families. Likewise, subcategories were established for each of the categories identified. Table 2 shows the categories and subcategories extracted from the analysis of the interviews and their corresponding codes. Language LIN Barriers that appear due to the existence of different languages between the school and the family. --- Communication COM Difficulties in the communication process between the school and the family. --- Cultural CUL Differences in the reference culture between the school and the family. --- Conflict CON Situations of conflict generated by language, communication and cultural barriers between the school and the family. After performing the coding process, the text fragments that corresponded to the same category and subcategory were grouped, and the entire content was revised. Then, the main ideas were synthesised, with the aim of presenting them orderly and clearly in the results. Lastly, textual quotations from the different testimonies of the participants were extracted to illustrate the results obtained in the present study. To perform the analysis of the data extracted from the interviews, the NVivo 12 software was used. --- Results The results are presented in three sections that analyse, mainly, the key elements of the overrepresentation of students from a migrant background in special education in the stage of primary education. Firstly, the previous experience of the education professionals is presented, as well as the initial and continuous training received in terms of attention to diversity and inclusion. Secondly, we show the motives that led the migrant families to migrate to Spain with their children. Lastly, this section addresses the main language, communication and educational barriers that the participants identified between the educational centre and the migrant families. --- Experience and Training of the Professionals That Attend to Migrant Students with Disabilities in the Educational Centres The professionals who work in the participating centres had certain experience in the diagnosis and detection of educational needs. They are teachers who had received specific initial training to attend to diversity and educational inclusion: 1. Special Education; 2. Physical Education; and 3. Therapeutic Pedagogy. All participants had received continuous training on attention to diversity and inclusive education. Likewise, a SIT was also interviewed, who also had a Bachelor's Degree in Primary Education and experience in centres with migrant students. However, this teacher did not mention any specific initial or continuous training on attention to diversity. "I began with social education. Then, I continued with psychopedagogy, teaching with pedagogical therapy, and a master's degree in early intervention. Finally, I entered an early intervention centre in Malaga as an interim, and that's the training I have regarding education". (C3) "Afterwards, I studied conflict resolution... but just a little bit. Master's degrees are poorly guided... at least in my experience... to work with this type of student". (TP1) Secondly, the participants stated that they had received no specific training with regard to migrant students. They highlighted the need to acquire knowledge in this scope (language, conflict resolution, etc.) due to the great presence of migrant people. "Generally, few courses related to migrant students are offered, and these are focused on teaching Spanish". (TP3) "Not much, really. Last year I had some trouble, because it was my first year with... most of my students were migrant students, and I received help from my colleagues in the centre. But regarding training, I would say I barely have any; there was some at university during my degree, but I don't really remember any of that". (TP2) Thirdly, the participants pointed out the lack of specialised staff to attend to the students and families with different languages, stating that translators are needed in centres with a large presence of Arabic students (Spanish area of the Strait of Gibraltar). "The language specialist comes once a week. Thus, with so little time, this specialist can't work adequately with these students, and their tutors can't be counselled about working with them. In fact, the work with the language specialist begins in Year 4 of primary education". (TP3) "We need some sort of resource, such as a translator, in these centres where there is a high percentage of migrants". (PT3) --- Reasons Why Families Migrate from Morocco to Spain: The State of Students from a Migrant Background When They Arrive in Spain Most of the students registered in the participating centres migrated from Morocco to Spain due to the scant or null educational attention that their children receive in their home country. Most of the students are Maghrebian and, in their home country, many children with disabilities are not treated in schools; in some cases, they are attended to in associations. However, their educational needs are not covered in any case. "The thing is that, in their country, mainly Morocco, they are extremely marginalised, and they try to bring them here so that they receive the attention they need". (TP2) "A person living in Morocco, for example, with no resources for their child, after visiting associations and seeing that their learning does not improve, eventually will search for whatever there is, wherever that is... if you gotta go to France, you go to France... if you hear there are resources in Cádiz, then you go to Cádiz". (SIT) In the case of Spain, the participants indicated that these families not only find that their children are schooled in special education centres in the mixed or pure modality, but they also receive the appropriate curricular adjustments and implementations free of charge. This is another reason why families decide to leave their home country, in some cases with neither a job nor residence, simply to ensure that their child is attended to. In other cases, they move to the homes of relatives who live in Spain, and then look for housing. "[...] They arrive here and the centre gives them everything, because specific centres are free of charge for all children. They have specialists, physiotherapists, lunch room, transportation... they have everything; and their children are attended to". (C8) The participants also highlighted that most of the students arrive without a previous diagnosis or with the wrong diagnosis. Therefore, when they arrive at the Spanish educational centre, the evaluations and diagnoses are initiated by the different teams of specialists, resulting, in some cases, in incongruent assessments. Furthermore, with regard to the proportion of Spanish and migrant students in special education, although the participants did not provide exact data, they mentioned that an increasing number of migrants are referred to special education. "In the case of students from a migrant background, they sometimes come with the wrong diagnosis, and the counsellor wouldn't establish a totally different diagnosis; even the coordination with social services is poor, especially in Year 2 and Year 3 (primary education), where I have the most students". (TP2) "I didn't count them... I don't know how many they are, but I know that there is one in every classroom. In fact, there is an increasing number of them in the classrooms". (HS1) --- Main Barriers Identified between the Education Professionals and the Migrant Families The participants underlined that there are numerous barriers in the schooling and educational process of students from a migrant background. Firstly, barriers such as language and the migrant parents' lack of knowledge hinder communication and their counselling. Similarly, there is a lack of willingness in some migrant families to adapt to the culture and learn the language. In some cases, this barrier makes it difficult to establish an accurate diagnosis of the needs of the students. "But then, if you can't ask the children questions, if you can't interview them, if you can't more or less see their traits... for example, the girl we have in the specific classroom was diagnosed with ASD, but she has characteristic traits of Down syndrome. It's very difficult to make a correct diagnosis". (SIT) However, the participants mentioned that they use diagnostic tests focused rather on visualising the student than on the language. Nevertheless, the interviews with the tutor and the parents and the observation of the child in different contexts are affected by language; therefore, the language barrier was identified as an important limitation in the evaluation and communication of the results with the parents. "When they just arrive at the centre or are very young, it is very difficult due to their lack of knowledge of the language and their lack of understanding". (TP3) The participants usually faced this difficulty using visual tools, requesting help from translators, associations or some relatives living in Spain who knew both languages. "They always come accompanied by a relative living here who knows both languages and translates for them, or they may come with associations with people who know Arabic". (C8) Along with the language barrier, the participants also pointed out the cultural barrier. In this sense, migrant parents are usually more reluctant to refer their children to early intervention. Some of the participants identified the stereotype of gender as a possible barrier that may limit the diagnosis and impartiality. Some students may show distrust or pay no attention due to the gender of the professional, although the participation and attitude of the parents with centres is positive, showing gratitude for the treatment received. "It is very visual, exactly. Many pictograms, and that, in the end, is a universal language, right? But, with the families, we sometimes have problems. In some cases, none of the two parents speaks Spanish, they don't understand well... In many cases, the mothers attend the tutorials less frequently, and the fathers are usually the heads of the families...". (HS1) "For example, last year, in secondary education, I observed this stereotype toward women; students aged 13, 14 and 15 years who literally told me 'you're a woman, what can you say about this?' When I evaluated them, they didn't really pay attention to me, because I'm a woman. They have that rooted in them. Even when some mothers attend, if the father is not also present, they don't want to listen to me". (C1) Although it was reported that some families are reluctant to follow the guidelines provided from the centre to attend to the needs of their children, in general, the participants stated that the families usually engage with the school and are very grateful for the treatment, help and education received from the education professionals and the centres. "Yes, they usually collaborate in most cases [...] moreover, they are often very grateful for the centre and all. They show respect toward the centre, us and our culture. They are collaborative people". (Pri1) "Yes, they are very grateful for having their children in the centre. They are very happy". (C8) --- Discussion Through interviews with different educators, the results of this study help to improve the understanding of the overrepresentation of students from a migrant background in special education, showing the formative situation of educators who are in charge of attending to the needs of these students. Similarly, the results also help to understand the motivations of the families to migrate to Spain and the impact of these migratory phenomena on the educational centres. Moreover, the results allowed identifying both language and cultural barriers between the different agents present in this context, students and families, which may lead to biased evaluations and subjective decisions made by education professionals, resulting in the referral of students from a migrant background to special education, thereby favouring the overrepresentation of students from a migrant background in special education [6,21,33]. Regarding the first research question, about the experience and training of teachers and specialists in special education centres, specific training on inclusive education is important for professionals who work with migrant students with special educational needs. In this sense, the experience and qualification of counsellors and SITs play a crucial role in the quality of the education that is provided to these students. However, there is a lack of specific training on attention to diversity, especially in the case of SITs. This underlines the need to invest in the professional development of educators to guarantee that they are adequately prepared to cover the needs of this group of students. Previous studies have reported this situation, highlighting the lack of training among professionals and of a clear profile that attends to the needs that are present in their context [1]. The study of Arroyo and Berzosa [28], among the main problems in educational attention to migrant students, identified that the training received by the teachers is homogeneous, without a specialised training that considers the needs and characteristics of their students, as well as the specific situation of each centre. This shows differences in the training of professionals, especially in terms of attention to diversity. In some cases, this lack of specialised training contributes to the exclusion of students from a migrant background with disabilities. In this sense, it can be asserted that the homogeneous training of educators reflects a dynamic of institutional power that does not adequately recognise the specific needs of these students. The results obtained in the present study reveal an important challenge in the education system, showing the urgent need for the competent administration and institutions to develop and implement specific actions that facilitate the training of educators in addressing the complications associated with the diversity of students from a migrant background with disabilities. Moreover, it is necessary to develop specific training in each educational centre, adapted to the real contexts and far from the homogenising propositions that come from educational administrations [27]. In regard with the second research question, which tackles the reasons for migrating and the impact of the latter on education, this study reveals that most of the migrant families from Morocco moved to Spain to find an adequate educational attention for their children with disabilities [14,43]. This shows the importance of guaranteeing that the education systems in both countries provide adequate support and resources to these students [39]. Moreover, these results highlight the need for effective collaboration between social and educational services to facilitate the successful transition of these students and their families. Thus, schools must lead this transition process toward a new context and a new education system, ensuring that both students and their families can thrive in this new environment [28]. Furthermore, this requires the participation of school associations that favour aspects such as intercultural mediation between the families and the schools. Lastly, to address the third research question, which refers to the barriers identified between professionals and families during the schooling period of the migrant student (i.e., language and cultural barriers), the participants identified these as significant challenges in the interaction with the migrant families. Previous studies have already identified some of these barriers, such as language barriers and family-school relations [34][35][36]50], in line with the results of the present study. The lack of mastery of the Spanish language among some families hinders communication and mutual understanding. This underlines the need for implementing effective communication strategies, such as the use of visual tools and the participation of translators, to overcome these barriers [45,46]. In addition, the participants mentioned that the cultural barrier may influence the way in which parents interact with the education professionals and how they perceive the diagnoses and recommendations [6]. Despite the barriers identified, the interviewees pointed out that most of the migrant families show a collaborative and grateful attitude toward the school and the education professionals. This shows the value that these families attribute to the education of their children and the importance of establishing relationships of trust and respect between the schools and the families [19]. The results of previous studies are in line with those of the present work, as is the case of Santos and Lorenzo [51], who reported that, in general, the participation of Moroccan families is lower than that of other families whose mother tongue is Spanish, although they show high levels of satisfaction with the tutor. In view of this situation, it is necessary to improve the training of professionals, in order to help them understand the language and cultural diversity of students from a migrant background. The lack of specialised professionals reflects a structural gap that must be addressed in order to guarantee a specific and effective support for migrant students with disabilities. In this sense, the demand for greater presence of trained and qualified professionals reflects the need to recognise and approach language and cultural diversity in contexts with significant migration. Considering that it is a challenge for the professionals of these centres to master all the languages of the students [6,53,54], it is fundamental to find alternatives that respond to this problem. There are experiences in which the educational community has engaged in a common project to learn the language of the students through extracurricular workshops [55]. Regarding the specialisation of teachers, it would be interesting to promote specific actions in teacher training centres (TTC) to attend to the realities of this type of centre [1,28]. The participants also demanded a greater presence of qualified professionals, especially in contexts characterised by migratory movements, as is the case presented in this study. The role of the teacher in temporary language adaptation classrooms (ATAL in Spanish) for migrant students is important for ensuring a specific attention to migrant students and their families. Therefore, a larger number of professionals of this area are required, especially in regions with a significant percentage of students from different cultures and languages [6,27,37]. The dates obtained here on the overrepresentation of students from a migrant background in special education in this study has also been reported in different countries other than Spain [8,33], motivated by the existence of biased evaluations and subjective decisions made by educators [1,19,20]. However, the factors that contribute to this phenomenon are complex and multifaceted, and no resolution has been developed to fully address it [6,7,19,21,45]. Therefore, it is necessary to apply an integrated approach that includes the culturally sensitive training of educators, the implementation of inclusive teaching practices, the active participations of the families and the collaboration between schools, communities and governmental agencies [6,38]. Lastly, we propose recommendations and future research lines on this topic. In this sense, the study provides a solid foundation for future work and recommendations for educational policies. It is crucial to address the divides in the training of professionals, especially for SITs, and to promote the continuous training on inclusive education and attention to diversity. Furthermore, it is fundamental to work on the development of effective strategies to overcome the language and cultural barriers, including the access to translation and cultural support services. Finally, this study underlines the importance of an inclusive approach focused on the student in education, recognising and respecting the diversity of the students, regardless of their origin or condition, as well as their educational needs [12]. This approach not only benefits migrant students with disabilities, but it also enriches the educational experience of all students and contributes to building a fairer and more inclusive society [11,13]. --- Implications and Proposals for Improvement With regard to the main implications and propositions for improvement derived from this work, it is important to promote the participation of the educational centres with social entities that can provide services to improve the transition and adaptation of students from a migrant background, such as academic support, language reinforcement and intercultural and leisure activities. This allows both students from a migrant background and their parents to socialise with other families [56]. Another practical recommendation to improve family integration would be to establish a mentorship programme within the educational institutions [57]. The aim is to develop a continuous feedback process of support and counselling between the minor and the mentor; the latter can be a family that had already gone through a similar situation, or a native family. This would allow attending to the needs of the migrant family in terms of integration into the education system and into society. Finally, it would be beneficial to organise training and sensitisation sessions within the school [58,59], addressing interculturality with the entire educational community, in order to bring the different cultural realities closer and facilitate the transition and adaptation of migrant families. --- Highlights --- • The overrepresentation of students from a migrant background in special education centres is due to different factors; --- • The professionals of these centres are not trained to specifically attend to migrant students; --- • The lack of specialised staff to attend to non-Spanish-speaking students and families contributes to their overrepresentation in special education centres; --- • The lack of specific training among professionals may result in wrong diagnoses and biased evaluations; • Families migrate from Morocco to Spain in search of adequate and

This study analyses the overrepresentation of students from a migrant background in special education in the stage of primary education, in the region of El Campo de Gibraltar (Southern Spain). This phenomenon poses significant concerns in terms of educational equity and discrimination, as well as questions about the possible underlying causes. It was decided to address this issue from a qualitative methodology through semi-structured interviews with 16 educators of educational centres that have a significant number of students from a migrant background. The results show the need of educators for specific training in attention to diversity. Likewise, many migrant families come to Spain due to the lack of adequate attention that their children receive in their home countries. Lastly, language and cultural barriers were identified, which allow for biased assessments and subjective decisions that result in the referral of students from a migrant background to special education from a multidimensional perspective. Therefore, it is fundamental to provide training to professionals, overcome the language and cultural barriers and ensure that the necessary resources are available, thus guaranteeing an inclusive and equitable education for all students, regardless of their origin.

a mentorship programme within the educational institutions [57]. The aim is to develop a continuous feedback process of support and counselling between the minor and the mentor; the latter can be a family that had already gone through a similar situation, or a native family. This would allow attending to the needs of the migrant family in terms of integration into the education system and into society. Finally, it would be beneficial to organise training and sensitisation sessions within the school [58,59], addressing interculturality with the entire educational community, in order to bring the different cultural realities closer and facilitate the transition and adaptation of migrant families. --- Highlights --- • The overrepresentation of students from a migrant background in special education centres is due to different factors; --- • The professionals of these centres are not trained to specifically attend to migrant students; --- • The lack of specialised staff to attend to non-Spanish-speaking students and families contributes to their overrepresentation in special education centres; --- • The lack of specific training among professionals may result in wrong diagnoses and biased evaluations; • Families migrate from Morocco to Spain in search of adequate and specialised educational attention for their children; --- • To address overrepresentation in these centres, it is essential to guarantee the availability of support resources and promote intercultural collaboration in the education system. --- Data Availability Statement: The authors state that the data explored in this study are completely original and have been analysed by a team of experts in the area, including the authors of this article. This study is framed within a competitive research project that received funding. With the aim of ensuring the confidentiality of the participants, the data have not yet been disseminated or shared in any repository. In conclusion, all the information that supports this study will be made available to the public when this article is published. --- Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. --- Conflicts of Interest: The authors declare no conflict of interest.

This study analyses the overrepresentation of students from a migrant background in special education in the stage of primary education, in the region of El Campo de Gibraltar (Southern Spain). This phenomenon poses significant concerns in terms of educational equity and discrimination, as well as questions about the possible underlying causes. It was decided to address this issue from a qualitative methodology through semi-structured interviews with 16 educators of educational centres that have a significant number of students from a migrant background. The results show the need of educators for specific training in attention to diversity. Likewise, many migrant families come to Spain due to the lack of adequate attention that their children receive in their home countries. Lastly, language and cultural barriers were identified, which allow for biased assessments and subjective decisions that result in the referral of students from a migrant background to special education from a multidimensional perspective. Therefore, it is fundamental to provide training to professionals, overcome the language and cultural barriers and ensure that the necessary resources are available, thus guaranteeing an inclusive and equitable education for all students, regardless of their origin.

Need for Transgender Gynecological Care Death rates from cervical cancer have dropped dramatically in the United States due to the effectiveness of the Pap test in detecting cervical cancer, which has allowed clinicians to treat abnormal and precancerous cells. 1,2,3 Yet transgender men (individuals assigned female at birth but with a male gender identity) obtain cervical cancer screening less frequently and are less likely to be up-to-date on Pap tests than cisgender women (individuals assigned female at birth and with a female gender identity). 4 According to the 2015 US Transgender Survey, although transgender men are vulnerable to chronic undetected human papillomavirus (HPV) infections, only 27% report having had a Pap smear in the past year compared to 43% of women in the general population. 5 Barriers like the ones presented below may prevent transgender men from scheduling appointments and accessing life-saving screenings. First, however, I offer a personal perspective on the need for cancer screening. Testosterone Therapy and Pelvic Pain During the summer of 2006, I celebrated the completion of my first year on testosterone. Over the course of that year, I documented many physical changes as my body morphed from a female to male shape. I had undergone chest surgery (a bilateral mastectomy with nipple grafts), but I still had other body parts-uterus, ovaries, fallopian tubes, vagina, and cervix-that required screening typically marketed only to female patients. The last time I had had a pelvic exam and Pap test, I was 19 years old, and I was now turning 26. Although cervical cancer is most frequently diagnosed in patients ages 35 to 44, 1 it was important to me to schedule another exam, not only to make sure that my tissues were healthy but also to address my ongoing extreme pelvic pain. After initiating hormone therapy-in my case, testosterone-transgender male patients can experience cramping and pain that can last more than 6 months. 6 My pelvic pain had been ongoing day and night for almost 4 weeks. At the time, I didn't know that this was a common symptom of testosterone use, 5 so I was concerned. Although motivations and reasons for seeking gynecological care vary, in my case, I had dysphoria about my reproductive organs; I was tired of dealing with pelvic pain; and I was concerned about uterine, cervical, and ovarian cancer. (Please note that there is currently no evidence that testosterone therapy increases risk for ovarian, uterine, or cervical cancers among transgender men. 7 ) Not having a desire to retain my uterus, I hoped my exam would render me eligible for a laparoscopic total hysterectomy and an oophorectomy. If so, I would be joined by a low percentage of transgender men who have had a hysterectomy. In a 2015 survey that included more than 8000 trans male respondents, 14% reported having had a hysterectomy and 57% reported wanting one someday. 5 Barriers 1: Gynecological care for men. Scheduling an appointment for gynecological care, including a Pap test and pelvic exam, and then following through with the appointment can be emotionally difficult for transgender men. I wanted to see a clinician about my concerns, but I was afraid to call and schedule an appointment because I didn't know if anyone would take me as a patient. I'd had negative experiences with clinicians and staff in the past, so even if someone would see me, I was worried about how I'd be treated. I gained courage by having a female friend call her obstetrician's gynecology office and ask if anyone would see a trans man as a patient. The office responded, "We've never worked with a trans man before, but send him our way!" It was a relief to hear this kind of response in 2006, given the lack of clinical education and training in how to respond to transgender patients' obstetrics and gynecological needs. 8 This lack of training persists. A 2015 survey of obstetrics and gynecology clinicians found that 80% of respondents reported not having received training in transgender care during residency, but almost 89% reported that they would be willing to provide routine Pap tests for transgender men. 8 In my case, knowing the clinic was open to me, even if its staff didn't have experience, diminished my anxiety-a little. 2: Is the target population gender and gender identity inclusive? When I called the clinic to make an appointment, the receptionist sounded surprised to hear a man's voice ask for a pelvic exam. And as I walked into the clinic, I had to ignore that women was the only word on the sign. The door closed behind me, and I kept my head down as I approached the front desk. I wasn't sure how people would react to seeing a man in a waiting room full of women. I was relieved that the front desk staff greeted me with a smile and treated me like any other patient. 3: Gynecological clinical encounters with trans men. Transgender patients' positive experiences in health care settings increase the likelihood of their remaining compliant with recommendations, including for screenings. If you're a clinician or a staff member, the upshot here is that patients' first encounters could be their last if they feel terrible about what happened to them there. So, the first time a trans man patient has a Pap test, it is important to talk with him about the procedure, including speculum use, swab insertion, and total time it typically takes to complete a physical examination. Approaching patients using a trauma-informed care model can help alleviate transgender men's anxieties about having their body parts examined and their experience of dysphoria or discomfort. 9,10 Although I went to my appointment alone, I accompanied a transgender male friend to his and joined him in the exam room. While my friend was still fully clothed, the clinician reviewed his health history. She informed him that if he had discomfort with any of the language she used, he should let her know his preferred terms. Trans men might be uncomfortable with clinically accurate terminology associated with body parts, so when patients prefer different language, clinicians should mirror their language. Testosterone causes atrophy and dryness of vaginal tissue 6 ; for trans men or cisgender women who do not engage in penetrative vaginal sex, a speculum can be especially uncomfortable. My friend was extremely anxious about speculum insertion, due to never having had penetrative sex, so this particular clinician did well to assure him that she would use an appropriately sized speculum for his anatomy. Pediatric-sized speculums, however, are not always helpful, and over lubrication should be avoided. 10 The clinician continued to communicate each step she would take and described what my friend might feel. She then asked him to disrobe from his waist down. Upon completing the exam, she left the room to allow him to put his clothes back on and later returned to answer his questions. This was a positive encounter for my friend and also for me as an observer. This clinician became a trusted caregiver in our community of trans men, and we all began supporting each other in making and attending appointments. 4: Gendering and body parts. Another barrier to care that transgender men tend to experience, if insured, is receiving notice of denial of claims coverage by an insurer. Clinicians can help prevent this occurrence with clear billing communication. After an appointment, for example, a clinician should note the patient's gender in his health record and notify stakeholders that this might be different than what is on his insurance card. If gender markers, like pronouns or names, are mismatched for a gender-linked procedure, like a Pap test, or gender-linked body parts, like a cervix, it's helpful to trans men patients when clinicians explain to the billing department the organ-specific services rendered. Taking these steps might decrease the number of insurance claims that a trans man patient is denied. After my clinician determined that a hysterectomy would be the best treatment for my symptoms and concerns, I informed her that all of my documentation-including driver's license, birth certificate, and insurance card-affirmed my identity as a man. She stated that she would note specific organs present in my health record and submit forms for precertification to my insurance company. Although my procedure was precertified, 4 months later a postpayment audit flagged the claim due to my being a man, prompting my insurer to request a refund. My clinician helped me appeal by writing a letter noting the medical necessity of the service she provided to me, and the appeal was eventually approved. In my case, the clinician advocating for me increased my trust and desire to return for future care. 5: Inadequate lab results. Inadequate tests or samples are more common among transgender male than cisgender female patients, requiring return office visits and repeat screening. 11 Often, the lab notes list "inadequate sample," "atrophy," or "dysplasia" as a reason for abnormal or inconclusive results. Not wanting to undergo another exam, a trans man patient receiving such results might not return to a clinic but might remain concerned about the results. In order to decrease the odds of an inadequate test, clinicians should inform the labs they use that a patient on testosterone (which causes thinning of vaginal tissue) had a cervical swab and also note whether the patient is amenorrheic. 10 Taking these steps can decrease the chance of abnormal results, confusion, error, or sample disposal by the lab. For example, the tissue sample taken after my hysterectomy came back with a note about cervical dysplasia, but I didn't find this alarming because my clinician explained that my long-term testosterone use caused cervical epithelial atrophy, which can mimic dysplasia. 11 6: Sex practice diversity and risk awareness. Both transgender men patients and clinicians can be misinformed about screening guidelines and risks. Some trans men lack not only understanding of risk factors for HPV but also general gynecological knowledge. Some clinicians might assume that trans men are less likely to be at risk for HPV because they might also assume that trans men don't have penile/vaginal penetrative sex. This assumption is wrong. There are a wide range of sexual practices in which trans men might be interested, including penetrative vaginal, anal, or oral sex with partners who have penises that produce sperm. Minority stress 12 due to gender-related discrimination and victimization has a negative impact on health; alcohol use, a history of psychosocial distress, and a history of sex with men only are risk factors for sexually transmitted infections in trans men. 13 Anyone With a Cervix Transgender men need to see themselves reflected in data, research, and cervical cancer screening guidelines published in authoritative, reliable sources. When conducting research for this article, I noticed that many sources continued to use anatomical words and pronouns intended to apply narrowly to cisgender women only. 14,15,16,17 It is important to include trans men in cervical cancer screening recommendation language, for example, by stating clearly that screening is for women, transgender men, or anyone with a cervix. Patients who identify as trans men find it acceptable and preferable to test for HPV with self-collected vaginal swabs. 18,19 Clinicians should offer and provide self-collection swabs as an option to trans men patients, along with education about risks and benefits of HPV vaccination. Clinicians should also offer other forms of screening, such as urine tests for sexually transmitted infections, to increase the likelihood that trans men patients follow screening guidelines while being spared the discomfort of a pelvic exam. The more frequently trans men have positive experiences in gynecological health care settings when seeking routine screening, the more likely they will be to practice regular screening and illness prevention. While I no longer need pelvic exams, I still visit my clinician's office for hormone therapy monitoring and other health care because I have established trust and had positive clinical encounters. Everyone deserves equality in enjoying this level of quality and trust in their health care. --- Conflict of Interest Disclosure The author(s) had no conflicts of interest to disclose. The viewpoints expressed in this article are those of the author(s) and do not necessarily reflect the views and policies of the AMA. Copyright 2020 American Medical Association. All rights reserved. ISSN 2376-6980

Cervical cancer is most frequently diagnosed in patients ages 35 to 44, but risk persists as individuals age. Among patients who are regularly screened via the Pap test, cancer is rare and death rates have dropped dramatically in the United States. Nevertheless, access to regular screening can be difficult for transgender men (individuals assigned female at birth but with a male gender identity) due to misinformation, discomfort scheduling appointments, fear of being mistreated or of refused services, lack of insurance, and clinicians' lack of knowledge. This narrative explores 6 barriers to cervical cancer screening for transgender men and offers recommendations for eliminating cervical cancer inequality.

Introduction Since 2000 China has experienced rapid aging of its population and is a trend that is predicted to continue well into the future (1). According to the seventh national population census data, by the end of 2020, 18.70% of the population in China will be aged 60 years and older. The "Fourteenth Five-Year Plan" of the National Plan for the Development of the Aged Cause and the Care Service System clearly emphasizes the need to continue to expand the "balanced, reasonable, high-quality, efficient, urban and rural coverage of care services, benefiting the entire population. " At present, family pensions, community pensions and institutional pensions are the three main types of urban pension models in China, and a "9,064" pension service model has been formed (90% of the older adults live at home, 6% live in community, and 4% live in institutions). This shows that community care services play a crucial role in China's pension system (2). As an old industrial base province in Northeast China, Liaoning Province ranks first in terms of population aging due to the influences of industrial structures, economic levels, social culture and other factors. Since 2020, Liaoning Province has taken the lead in carrying out reform pilot work of community care services for older people in China. A series of policies and regulations have been promulgated, including the Regulations of Shenyang Municipality on Home-based Care Services. It has carried out reform pilot work of community care services for older people in 200 communities in nine cities, including Shenyang and Dalian, and formed a series of working models with demonstration effects. With the trend of aging, the health of the older population is of concern. Older people tend to suffer from more physical and mental diseases and are prone to psychological problems, inducing the risk of suicide (3). According to 2018 Report on Older Adults' Psychological Health in China, 95% of older people have different degrees of psychological disorders, and they are prone to emotional problems. In June 2019, the National Health Commission also pointed out that changes in living conditions, social relations and physical conditions associated with older age, such as the weakening of the body and the death of a partner, bring negative feelings of inferiority and worthlessness. If these feelings are not appropriately managed, psychological problems such as depression and/or anxiety may result. According to the World Health Organization, "health is not only the absence of disease or weakness in the body but also the physical, psychological and social integrity of a person. " The psychological health of older people is not only a medical and health problem but also a public problem affecting social stability and development (4). At present, child support, institutional care and community care services focus on meeting the survival and physiological needs of older people, while overlooking or neglecting their spiritual needs and psychological problems (5). Recently, the Chinese government has placed greater importance on the issue of the care of older people in policies, focusing on their living conditions, and demanding the improvement of the quality of care services to meet their diverse needs. However, while most policies tend to guarantee basic living needs, such as food, housing, and basic medical care, and emphasis on resource support for service institutions, there is a lack of support for psychological services. Similarly, current research on community care services mainly focuses on which service items should be established (6), how processes and quality can be improved (7), and how older people's satisfaction with services can be enhanced (8). However, there is limited research on the impact of community care services on older people's health, especially their psychological health, and there is a lack of investigation into the demand for services and service effectiveness, which directly affects the improvement and development of community care services. From this point of view, the psychological health of older people in China receives little attention but requires consideration. Given this, this study is committed to studying, within the context of actively promoting community care services for older people in China, whether the community can become the real spiritual home for older people, provide them with high-quality and convenient services, and thus positively impact their psychological health. An additional question was whether this impact differs for sub-groups of the population. This is an important issue affecting the construction and improvement of the community care service system for older people in China. This study explored the impact of community care services on the psychological health of older people using the research results of 200 community care pilots for older people in Liaoning Province. --- Literature review 2.1. Community care services for older people Population aging and the care of older people are social issues faced by all countries globally. Welfare countries have implemented models of community care for older people since the 1960s. In the UK, the Department of Health and Social Security (DHSS) proposed that "use community facilities and resources to provide long-term care and assistance for people in need of care in the community, especially the very vulnerable older adults ", and an important objective of community care services is to reduce loneliness and abandonment experienced by older people. Ayers and Lyman (9) emphasized the critical role of psychological counseling and emotional assistance in caring for older people. Connelly (10) divides the content of community care into four types, the first is life care (including home services, apartments, etc.), the second is material support (including providing food, tax relief, etc.), the third is medical support (including disease treatment, health care, etc.), and the fourth is overall care (including improving the living environment, promoting social participation, etc.). --- Definition of psychological health The World Federation for Mental Health (WFMH) defines psychological health as "developing one's mood into the best state within the range of physical intelligence and emotion that is not inconsistent with other people's mental health. " The World Health Organization (WHO) (2014) summarizes three characteristics of psychological health: "good personality, good ability to handle affairs, and good interpersonal relationship (11). " Wu (12) reported that as older people experience physical aging and social role changes, their psychological health also presents different characteristics, and attention should be paid to their cognitive abilities and adaptability. --- Measurement of psychological health In terms of psychological health measurement, early researchers mainly analyzed older people's psychological health from a single perspective of subjective psychological feelings. Siebert et al. (13) reported that loneliness, depression, subjective well-being and life satisfaction are sensitive indicators that can be used to quantify older people's mental health. Kozma Based on the research results of scholars from China and other countries, the measurement of the psychological health of older people using subjective emotional indicators only is limited. A comprehensive evaluation index system including loneliness, cognitive ability, life satisfaction and interpersonal activity participation as the main dimensions should be established. --- Interfering factors of psychological health of older people In order to fully understand the root causes of psychological problems and diseases of older people and improve their psychological health, scholars from all countries are committed to analyzing the factors that influence their psychological health, which can be roughly summarized into three aspects: physiological condition (sleep quality, illness, etc.), social relationships (family relationships, interpersonal relationships, etc.), and economic conditions (income levels, social security). Regarding physiological factors, Jorunn (19) studied the relationship between the ability to perform Activities of Daily Living (ADL) and loneliness in older people through structured interviews in a rehabilitation center, and concluded that ADL functions could reduce loneliness. Jacobs et al. (20) analyzed the relationship between physical health and loneliness in older people and found that those with physical dysfunction were more likely to feel lonely. In terms of social relationships, Wu (21) found that the social support of family, neighbors and friends affects the subjective well-being and mental health of older people by affecting their self-esteem, and reducing loneliness. Regarding economic conditions, Saber (22) reported that the level of economic income significantly impacts older people's happiness index. --- Methods --- Sample data sources The data for this study were derived from the results of a survey conducted by the Institute of Urban and Rural Community Construction of Northeast University in 200 community care service pilot communities in Liaoning Province from 2021 to 2023. The survey adopted a multi-stage stratified cluster sampling method and was aimed at people aged 60 years and older. Step 1: The sample size was determined to be 800 people by calculation; Step 2: The sample size for each city was determined on the basis of the size of the population in each of the 14 cities in Liaoning Province; Step 3: Sampling communities were identified using a systematic random sampling method; Step 4: Using cluster sampling method, six samples of older people in each community were randomly selected and asked to complete a survey. A total of 852 questionnaires were distributed, and 741 valid questionnaires were returned, yielding a response rate of 87%. The data have strong universality and representativeness. --- Variable selection and measurement Independent variables: community care services for older people. According to the relevant regulations of the Ministry of Civil Affairs and other departments, community care services for older people include eight elements: living care services, meal and cleaning services, medical and health services, visiting and chat services, emotional counseling services, emergency rescue services, cultural and sports activities services, and legal aid services. Values were assigned to each of service item for the purpose of analysis, with a value of '1' assigned if the community provided this social service, while a value of '0' was assigned if the service was not provided. Dependent variables: older people's psychological health. According to the measurements of psychological health in the literature review, this paper aimed to build a comprehensive and objective measurement index of older people's psychological health. Therefore, the four indicators of loneliness, cognitive ability, life satisfaction and participation in activities were selected to measure older people's psychological health. "Loneliness" is a sensitive indicator of psychological health with which it is significantly negatively correlated, and the higher an individual's psychological health, the lower the perceived loneliness (23). The question "Do you often feel lonely?" was included in the questionnaire which was answered using a 1-5 scale where "1 = always, 2 = often, 3 = sometimes, 4 = rarely, and 5 = never. " "Cognitive ability" refers to the ability of the human brain to process, store and extract information. Based on the Mental Health Scale for the Older Adults (Urban Version), this study designed seven questions that reflected seven aspects: memory, spatial recognition, learning ability, understanding ability, expression ability, calculation ability and reaction speed. Each answer was assigned the value of "1 = correct, 0 = wrong or unknown, " and the scores of the seven questions were added to compute a total score. "Life satisfaction" is a key indicator of older people's subjective well-being and Song et al. (24,25) showed that the life satisfaction index of older people could be used to reflect their psychological health status. Life satisfaction was measured by the question "Are you satisfied with your current life?" which was answered using a 1-5 scale where: "1 = very dissatisfied, 2 = dissatisfied, 3 = average, 4 = satisfied, and 5 = very satisfied. " "Participation in activities" belongs to the dimension of interpersonal communication, which is measured by the frequency of engagement in social activities. Scholars from all countries regard interpersonal activities as a measure of older people's psychological health. --- "Participating in activities" was assessed by the question "Do Frontiers in Public Health 04 frontiersin.org you participate in the following activities now?" including "music or dance, sports, playing cards or mahjong, communicating with friends, voluntary activities, and other outdoor activities. " The question was answered using a 1-5 scale where: "1 = not participating, 2 = not every month but at least once a year, 3 = not every week but at least once a month, 4 = not every day but at least once a week, and 5 = participating almost every day. " The answers to the six items were summated to compute a total score. Control variables: referring to the existing research about the factors that influence older peoples' mental health, data were collected regarding the participant's personal characteristics, behavior habits, economic status and family status, including age (1 = 60-65 years old, 2 = 65-70 years old, 3 = 70-75 years old, 4 = 75-80 years old, 5 = 80 and over), gender (0 = female, 1 = male), living conditions (1 = family, 2 = living alone, 3 = nursing home), marital status (1 = married and living with a spouse, 2 = married but not living with spouse, 3 = divorced, 4 = widowed, 5 = unmarried), physical exercise (0 = no physical exercise in the past, 1 = physical exercise in the past), disability (1 = no disability, 2 = mild disability, 3 = moderate disability, 4 = severe disability) and economic situation (1 = very poor, 2 = relatively poor, 3 = average, 4 = relatively wealthy, 5 = very wealthy). --- Empirical models 3.3.1. Confirmatory factor analysis In this study, four variables: loneliness, cognitive ability, life satisfaction and participation activities were used as indicators of "older people's psychological health. " In contrast "older people's psychological health" was a latent variable, which is the common part of the four measurement indicators -a common factor. In order to test whether the whole factor model was significant, confirmatory factor analysis (CFA) was conducted on the four indicators of loneliness, cognitive ability, life satisfaction and participation in activities to verify whether the factor load coefficient of "older people's mental health" was significant, and to verify whether the measurement of "older people's mental health" was valid. --- Structural equation model (SEM) As a confirmatory method, the structural equation model can explore the structural relationship between variables, and is widely used in psychology, sociology and other fields (26,27). In this study, the explanatory variable "community care service" was an explicit variable, and the explanatory variable "older people's psychological health" was a latent variable measured by four indicator variables: loneliness, cognitive ability, life satisfaction, and social activities. The structural model equation of community care services and older people's psychological health is shown by the formula (1): <unk> <unk> <unk> = + + <unk>X(1) The measurement model is shown in formula (2): y <unk> = + <unk>(2) In formula (1), <unk> Is a latent variable, indicating older people's mental health; <unk> Indicates the change of the latent variable after each unit change of the indicator variable; X refers to community care services that affect older people's psychological health, which are composed of living care services, meal and cleaning services, medical and health services, visiting and chat services, emotional counseling services, emergency rescue services, cultural and sports activities services and legal aid services; <unk> Represents a random perturbation term. In the formula (2), y = (y1, y2, y3, y4) is a group of indicator variables for older people's psychological health. This study used four indicators: loneliness, cognitive ability, life satisfaction, and social activities; <unk> indicates the measurement error item (Table 1). --- Analysis of empirical results --- Descriptive statistical results Descriptive data and participants' characteristics are summarized in Table 2 which shows that the average score for loneliness was 4.03, indicating that respondents had a low sense of loneliness as a whole and do not feel particularly lonely. Cognitive ability is composed of memory, spatial recognition, learning ability, understanding ability, expression ability, calculation ability and reaction speed, and the average score was 6.52, indicating that respondents had high overall cognitive ability and normal intelligence. The average life satisfaction score was 3.88, indicating that participants were relatively satisfied with their current life as a whole and had a positive and optimistic attitude toward life. However, the average number of people participating in activities was 9.63, which indicates that participants had less contact with people and were not fully involved in various social activities. According to the social withdrawal theory, people's abilities will inevitably decline with increasing age, and they gradually lose their social roles, thus reducing social interaction and participation in activities. This view is consistent with the results of this research. In terms of community care services, the average value of each service is ranked from high to low as follows: meal and cleaning services, medical and health services, cultural and sports activities services, emergency rescue services, legal aid services, emotional counseling services, visit and chat services, and living care services, with the average value of 0.83, 0.57, 0.37, 0.28, 0.21, 0.17, 0.15, and 0.12, respectively. Overall, the service items and supply coverage of the existing community care services are seriously insufficient, with the focus on physical health and basic living needs while social and spiritual needs are overlooked. According to Maslow's hierarchy of needs theory, community care services should not only focus on the first level of physical needs and the second level of safety needs, but also focus on the third and fourth levels of social and respect needs. --- Model estimation results Before running the model, the data in this study were first tested for reliability using SPSS software. The results showed that the Cronbach's <unk> of each study variable was above 0.6, indicating that the data had high reliability. Subsequently, AMOS software was used to draw a model path map and perform data operations, and after debugging, the model was optimized. The final output is shown in Figure 1. The fitting index RMSEA = 0.067 and SRMR = 0.039 indicates that the model fitness of the constructed factor model was good. According to the model estimation results in Figure 1, community care services have a significant impact on older people's psychological health with an impact coefficient of 0.72. The three services with the most significant impact were cultural and sports activity services, visiting and chat services, and emotional counseling services, with coefficients of 0.38, 0.26, and 0.22, respectively. This result indicates that within the community, the involvement of professional social workers and psychological counselors, carrying out diverse cultural and sports activities to enhance the social participation and communication circle of older people can enrich their leisure life, provide a sense of collective belonging and existence, and maintain a positive and healthy mentality. Visiting and chat services are a direct way to comfort people. Service personnel can play the role of relatives or friends of empty nesters to a certain extent and alleviate older people's loneliness through communicating with them. At the same time, due to personal, family, or neighborhood factors, older people are prone to suffer from depression and anxiety. Counseling services provided by professional psychological consultants can help older people reduce internal misunderstandings and distress and lower their risk of mental illness. In addition to these three items, the service items that ranked fourth to fifth in terms of impact were: meal and cleaning services and living care services. These service categories can improve the quality of life of older people, ensure that they live with dignity in their familiar community environment, thereby making them feel the meaning and value of life, and enhancing their life satisfaction. While the other three service elements -medical and health services, emergency rescue services and legal aid services were found to have no significant effects, and may even have a negative impact on older people's mental health. This is mainly due to the relatively short development time of the three services, which are still in the initial stage of development, insufficient experience of practitioners, high service costs, and insufficient coverage, resulting in a lower overall service level (Figure 2). --- Robustness check Considering that older people's psychological health is also affected by other factors, this study also added control variables to the following models (1) and ( 2), including age, gender, living conditions, and whether participants undertook physical exercise. The regression coefficients and significant changes in the models were evaluated to assess the robustness of the empirical models. In Table 3, Part A shows the estimated results of the structural model for various factors affecting older people's mental health, and Part B shows the estimated results of the measurement model, which is essentially consistent with the estimated results of the structural equation model. Part C shows the fitting indicators for the entire model, and of them, RMSEA<unk>0.08, SRMR<unk>0.05, and R 2 <unk> 0.7312, indicating that the overall fit of the model is good, and the empirical results are robust and acceptable. --- Heterogeneity analysis The data were further analyzed to assess heterogeneity and whether the impact of community care services on older people's psychological health varied according to whether they live alone, whether they are disabled, and their economic level. Table 4 shows the results of the heterogeneity analysis. Results showed that living care services have a differential impact on disabled versus non-disabled older people, and have a significant positive impact on the psychological health of disabled older people, but no significant impact on non-disabled older people. The reason is that disabled people are unable to complete daily life independently due to physical impairments, and need the help of professional service personnel (28). Meal and cleaning services were also found to have a differential impact on the psychological health of solitary older people, disabled older people, and poor older people. This is likely due to the greater vulnerability of these groups who have a greater need for these services. In addition, the coverage and quality of this service are currently high, and older people can obtain it in a convenient and low-cost manner, thereby improving their life satisfaction. Furthermore, older people can benefit from the opportunity of interacting and communicating with others during meals which is likely to reduce their sense of loneliness (29). Medical and health services were found to have a positive impact on the psychological health of non-solitary older people, disabled older people, and wealthy older people. By comparison, the impact on other groups of older people was negative. This may be due to disabled older people having a stronger demand for medical services, requiring family support to a certain extent, and relatively high cost (30). Visiting and chat services have a differential impact on those who live alone and have a significant positive impact on their psychological health. This is most likely because solitary older people lack the care of their families and children and are eager to communicate with others (31). Emotional counseling services have a positive impact on the psychological health of different groups of older people, particularly the solitary and the disabled, who are more vulnerable to psychological problems due to life stresses. Emergency rescue services have a positive impact on the psychological health of the solitary and disabled older people, while their impact on other groups of older people was negative. This may be because solitary and disabled older people have impaired physical functioning and may be more likely to face the risk of sudden illness and/or accidental injury. Emergency rescue services provide protection for older people through remote monitoring, improving their sense of security. Cultural and sports activities services have a significant positive impact on the psychological health of all groups of older people, indicating that regardless of residential, physical, and economic status, older people have a clear need to participate in various social interaction activities to reduce loneliness, improve life satisfaction, social skills, and maintain their cognitive abilities. Legal aid service has a negative impact on the psychological health of non-solitary and poor older people, while it has a positive impact on other groups of older people. This may be because, when faced with legal disputes, the non-solitary group typically has children to assist them, while those who are poor may have limited understanding of the legal system, and generally resolve disputes through other means. Overall, community care services significantly impact the psychological health of the solitary, disabled and poor older people, proving that these vulnerable groups have a greater need for community care services. At the same time, different types of services have different effects on different groups of older people. Those who are solitary need more spiritual services, the disabled need more life support services, and the poor need more cultural Frontiers in Public Health 07 frontiersin.org and entertainment services. This analysis result is consistent with other Chinese scholars' research reports. For example, Wu (21) reported that solitary older people lack the support and care of their families regarding life and spirituality and need social interventions to improve their psychological health. Huang (32) also reported that disabled older people need customized and specific services based on daily living and care services, but the current supply is still significantly insufficient. Wang (33) reported that poor older people have fewer opportunities to participate in social activities and created the phenomenon of "conceptual poverty" and "spiritual poverty. " It is necessary to encourage and support poor older people to participate in community activities and improve their quality of life. --- Discussion This study examined the impact of community care services on older people's psychological health in Liaoning Province. Overall, community care services have a significant positive impact on older people's psychological health, which is consistent with previous research (34). This study further indicates that only five of the eight service items have a positive impact, while the remaining three have a negative impact, including medical and health services, emergency rescue services, and legal aid services. Previous studies have not found this phenomenon, but some Chinese scholars have pointed out that these three services have obvious shortcomings such as insufficient supply and poor professionalism, and thus cannot meet the needs of older people (35), which confirms the conclusion of this study. In addition, this study indicates that living conditions, disability, and older people's economic circumstances have a differential impact on the results. Community care services have a greater impact on the psychological health of older people living alone, the disabled, and the poor. This indicates that vulnerable groups need more customized and high-quality community services which has also been confirmed by other relevant studies. In China, the "Healthy Aging" policy is committed to improving the comprehensive health of older people, especially their psychological health and social adaptability. This study is based on the policy background of "healthy aging, " and differs from other studies that only focus on physical health. Instead, it focuses on multidimensional psychological health, to test the effectiveness of different community care services. This study fills the research gap in this field and suggests directions for Chinese community care services to address issues associated with population aging. At the same time, however, this study is limited to Liaoning Province, which has a very large aging population, and does not cover the whole country, which may result in the research results not being widely representative. In the future, with further research, more comprehensive conclusions may be drawn. --- Conclusion This study was based on survey data from pilot programs of community care services in Liaoning Province, to explore the impact of community care services on the psychological health of older people. The empirical results show that, of the eight service elements stipulated by policy, the implementation of five types of services had a significant positive impact on older people's psychological health, namely, cultural and sports activities services, visiting and chat services, emotional counseling services, meal and cleaning services, and living care services, while the other three service elements -medical and health services, emergency rescue services and legal aid services were found to have no significant effects, and may even have a negative impact on older people's mental health. This is mainly due to the relatively short development time of the three services, which are still in the initial stage of development, insufficient experience of practitioners, high service costs, and insufficient coverage, resulting in a lower overall service level. In addition, there are differences in the results of community care services under different living conditions, disability conditions, and economic conditions. It is necessary to focus on the psychological health of older people living alone, the disabled, and the poor, and provide specific community care services for these groups. In response to these conclusions and to the policy spirit of "healthy aging" and "active aging" in China, the following suggestions are proposed to improve the community-based elder care model, and improve the psychological health status of the older population: First, provide comprehensive and high-quality community care services for older people, focusing on improving the quality and coverage of spiritual services. Research shows that with the development of China's economy and society, older people's needs are characterized by diversity. After basic physiological and survival needs are met, emotional satisfaction, the acquisition of dignity, and the establishment of social interaction are equally important. From the perspective of the supply side of care services, existing care services for older people are characterized by "emphasizing physiology but neglecting psychology. " The coverage of spiritual services is relatively low, and the service level is relatively poor. It is recommended that the development of spiritual services for older people in the community be strengthened through fund allocation, talent cultivation, facility construction, team cooperation, and project introduction. Second, build a social activity platform for older people, and organize diverse cultural, sports, and other entertainment activities for them. Of the eight community care service elements, cultural and sports activities had the most significant impact on older people's psychological health, which requires lower capital costs in comparison to other services. In the case of limited community resources, priority could be given to the implementation of recreational activities and services. Communities should establish activity centers for older people, fully mobilize their human resources, and regularly organize activities for them, thereby connecting them to the community and enriching their lives. Third, it is important to focus on the psychological health of the most vulnerable and disadvantaged groups of older people, i.e., those who are solitary, disabled and/or poor, and give them more respect and care. Generally speaking, while impairments in physical functioning are obvious and receive attention, loneliness and lack of social interaction are more difficult to detect but are likely to result in loneliness, depression and an increased risk of mental illness. Therefore, when providing care services for older people, it is important to provide targeted services based on the particular circumstances of those who are particularly vulnerable, and develop corresponding service guidelines. All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. --- Data availability statement The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author. --- Ethics statement The studies involving human participants were reviewed and approved by Northeastern University Ethics Committee. The patients/ participants provided their written informed consent to participate in this study. --- Author contributions All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication. --- Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's note

Background: Rapid population aging in China means it is imperative to establish a comprehensive care service system for older people. Currently, China is vigorously promoting the development of community care services for older people which should, ideally, focus on psychological health in addition to physical health. This study examined the impact of community care services on older people's psychological health. Methods: Survey data (n = 741) were collected from people aged 60 years and older in Liaoning Province, China, in which various community care services for older people were provided. Information was collected regarding the types of services provided (e.g., meal services, medical and social care), participants' demographic details (age, gender, economic circumstances, etc.), and their psychological health (e.g., loneliness, life satisfaction). The impact of the various care services on older people's psychological health was subsequently examined through the construction of a structural equation model. Results: Community care services for older people had a significant positive impact on their psychological health, with the most significant positive impact on cultural and sports activities, visiting and chat services, and emotional counseling. The impact of community care services on sub-groups of older people (e.g., those who were disabled, socially isolated and/or poor) was different.It is necessary to provide comprehensive and high-quality community care services, organize diverse cultural, sports, and recreational activities, provide differentiated and specific services for older people, and formulate corresponding service guidelines.

Parent training is one of the most widely studied, cost-effective interventions known for reducing or preventing child mental health disparities through the promotion of effective parenting strategies (Brestan & Eyberg, 1998;Webster-Stratton & Taylor, 2001). However, the Incredible Years Parenting Program, a standardized parent training program developed for European American parents, did not generalize well to KA parents due to cultural and linguistic differences (Kim, Cain, & Webster-Stratton, 2008;Kim, Choe, & Webster-Stratton, 2010). In response, researchers developed the Korean Parent Training Program (KPTP). This program addresses parenting factors related to child mental health problems and can be delivered in the context of Korean culture and faith. Faith-based, health promotion program can be effective for KAs because 70 to 85% of them regularly attend KA churches (Kim, Han, & McCubbin, 2007;Kwon, 2004). The purpose of this study was to pilot-test the preliminary efficacy of the KPTP. --- Protective and Risk Factors Related to Korean American Child Mental Health KAs are a hard to reach population, making them medically underserved and understudied by researchers. Existing data clearly indicate child mental health disparities exist and there is an urgent need for early intervention in this population. Studies found 41% of KA children had behavior problems while 35% of U.S. children experienced similar problems (Kim, Guo, Koh, & Cain, 2010;Webster-Stratton & Hammond, 1998). KA children also scored significantly higher on depressive symptomatology than European American children (Nahm, 2006). Recent studies found 60% of KA adolescents had depression symptoms compared to 24% of European American adolescents (Duggal, Carlson, Sroufe, & Egeland, 2001;Nam, 2013). The most important protective and risk factors related to KA child mental health were parenting practices. Protective parenting factors relating to fewer behavior problems and more social competence among children were parental warmth, reasoning, expressions of affection through hugging/kissing and saying "I love you," and correcting misbehaviors (Kim, Guo, et al., 2010;Kim, et al., 2007). In contrast, risk parenting factors included parental rejection, physical punishment, intergenerational acculturation conflicts, and parental depression (Kim, 2005(Kim,, 2008a(Kim,, 2008b(Kim,, 2011(Kim,, 2012;;Kim, Cain, & McCubbin, 2006;Kim & Cain, 2008;Kim, Guo, et al., 2010;Kim, et al., 2007). Parental rejection was related to children's behavior problems and social competence, and adolescents' psychological adjustment and depression (Kim, 2008b(Kim,, 2012;;Kim & Cain, 2008;Kim, Guo, et al., 2010;Kim, et al., 2007), while physical punishment was linked to behavior problems (Kim, Guo, et al., 2010). For KAs, parental cultural adaptation to the U.S. (i.e., acculturation) moderated the relationship between parenting and child mental health. Acculturation is cultural and psychological adaptation into the mainstream culture (Berry, 2006). When mothers adopted fewer U.S. cultural practices, maternal rejection was negatively related to children's social competence. In comparison, when mothers adopted more U.S. cultural practices it was found to be unrelated to children's social competence (Kim, et al., 2007). Similar results were also found among adolescents; those adolescents whose mothers adopted more U.S culture had better psychological adjustment than those whose mothers adopted less (Kim, et al., 2006). In addition, mothers who used strict parental control and maintained their Korean culture while adapting to U.S. culture had adolescents who were better adjusted psychologically than adolescents whose mothers maintained less Korean culture while adapting to U.S. culture. These findings indicate effective and positive parenting practices can help to decrease KA child mental health problems and intergenerational acculturation conflict. It also indicates that adapting to U.S. culture functions as a protective sociocultural factor. Likewise, maintaining Korean culture while adapting to U.S. culture also functions as a protective factor. --- Standardized Parent Training Programs Parent training is a focused, short-term intervention aimed at improving effective parenting practices and decreasing child mental health problems (Barlow, Smailagic, Ferriter, Bennett, & Jones, 2010;Dretzke et al., 2009). According to the American Academy of Pediatrics (AAP, 1998), effective parenting practices include: (1) positive discipline to build intimacy (e.g., hug/kiss); (2) appropriate discipline to promote desirable behaviors (e.g., sticker charts) and reduce undesirable behaviors (e.g., timeouts); and (3) the avoidance of harsh discipline (e.g., verbal and physical punishment). A meta-analysis of 57 randomized controlled trials in the Cochrane Review supports the use of parent training, with a Heges's g standardized mean difference between -0.62 to -0.67, indicating that parent training leads to more decreases behavior problems of children than control group (Dretzke, et al., 2009). In another meta-analysis of 63 randomized controlled trials, Cohen's d effect size was 0.42 for improving child problem behaviors and 0.47 for improving effective parenting practices (Lundahl, Risser, & Lovejoy, 2005). However, most standardized parent trainings have been developed for U.S.-born parents familiar with U.S.-based parenting practices. They may not reflect the concerns or backgrounds of ethnic minority parents and may require significant adaptation. When adaptations were made, such as reducing the length of intervention period or eliminating critical core content, retention rates increase but effect size decreases (Gorman & Balter, 1997;Kumpfer, Alvarado, Smith, & Bellamy, 2002). Recently, a few culturally tailored parent training programs were developed and tested using more rigorous research methods targeting Chinese Americans (Lau, Fung, Ho, & Liu, 2011), African Americans (Coard, Wallace, Stevenson Jr, & Brotman, 2004), and Mexican Americans (Martinez & Eddy, 2005), and impacts were found to be positive and quite promising (Coard, et al., 2004;Gross et al., 2009;Lau, et al., 2011;Martinez & Eddy, 2005). The lead author pilot tested the Incredible Years Parent Training with a KA audience. She used the program without adaptations to keep the fidelity to the curriculum. The Incredible Years program, an exemplar version of parent training, uses recommendations from the American Academy of Pediatrics to teach effective parenting practices. Program outcomes have been examined repeatedly in a variety of rigorous trials (Brestan & Eyberg, 1998). Results from the lead author's pilot study showed intervention group mothers increased positive discipline in comparison to control group mothers but intervention group children did not decrease behavior problems (Kim, et al., 2008). Within the intervention group, mothers with less acculturation to the U.S. decreased harsh discipline whereas mothers with high acculturation increased appropriate discipline (Kim, et al., 2008). Parents noticed cultural differences in the skills taught and reported negative feelings about watching program videotapes that featured mostly English speaking European Americans characters. Parents advocated for a KA adaptation of parent training that reflected their culture, language, and parenting norms (Kim, Choe, et al., 2010). This led to the development of the Korean Parent Training Program (KPTP). --- Korean Parent Training Program As shown in Table 1, the KPTP uses multiple theories to bring together traditional Korean cultural elements (e.g., Confucianism, Korean parenting virtues), more recent cultural elements (e.g., Christianity), and effective parenting practices. First, the KPTP helps KA parents understand how living in two cultures impacts both their parenting and their child's mental health. Then, it reviews Christian parenting principles, such as unconditional parental love, and traditional Confucianism and Korean parenting virtues, such as role-modeling behavior that is respectful, warm, patient, and generous, as well as having high standards on children. Second, within the context of the parental acceptance-rejection theory, it shows parents how to use positive discipline strategies to build intimacy with their children. Third, the KPTP supports the use of effective discipline strategies, including how to help children with emotion regulation. Another useful parenting strategy taught in the KPTP is helping parents learn how to use operant conditioning procedures involving the application of consequences following a child's behavior to influence future behavior (Gilmel & Holland, 2003). The strategies are adapted from the Bright Futures in Practice: Mental Health, a nationwide mental health promotion program (Jellinek, Patel, & Froehle, 2002). For example, parents role-play a common physical punishment method among Koreans-raising their arms up in the air while kneeling for five minutes (Kim & Hong, 2007). After this roleplay, some parents have the insight that the method may not be as effective as they thought because it tends to make children angry rather than helping them reflect on their misbehaviors. It also emphasize a biblical parenting principle that says, 'do not provoke your children, lest they brome discouraged (Colossians 3:21).' This realization can motivate parents to find an alternative, non-punitive method, such as a brief timeout (Kim, Choe, et al., 2010). The KPTP manual is available in both Korean and English. --- Hypothesis It is hypothesized that: 1) upon completing the KPTP, intervention group (IG) mothers would demonstrate an increase in effective discipline strategies, parental warmth, and parent self-efficacy, as compared to waiting-control group (WCG) mothers; and 2) upon completing the KPTP, IG children would demonstrate a decrease in emotional and behavioral problems and intergenerational acculturation conflict in comparison to WCG children. --- Methods --- Study Design This study used a partial group randomized controlled experimental study design to test the effectiveness of the KPTP. --- Sample The study sample included 48 KA mothers, their children, and their children's Sunday School teachers. All participants were recruited from six, partnered KA churches in the Pacific Northwest. Criteria for inclusion of mothers were: (1) having a child between three to eight years of age; (2) being a first generation KA originally born in Korea; (3) being willing to take the initial twelve-week KPTP class followed by three, monthly booster sessions as a member of the intervention group (IG) or the waiting-control group (WCG); and (4) being willing to complete pre-(T1), post-(T2), and 3-month follow-up (T3) assessments following the KPTP intervention. Children between three and eight who were born to the participating mothers and teachers who interacted with participating children each week during Sunday School were also included. --- Instruments Discipline-The Korean Parent Discipline Interview (KPDI) (Kim, Guo, et al., 2010) was adapted from the Oregon Social Learning Center's Discipline Questionnaire to assess discipline strategies that the KPTP attempted to increase or decrease among participants. The KPDI is a 65-item, 7-point, Likert-type scale instrument with four subscales that measure harsh discipline (e.g., yelling, spanking), positive discipline (i.e., hug, praise), and appropriate discipline (e.g., timeouts, consequences). This instrument is available in English and Korean. Higher mean scores indicates higher tendency of using each discipline technique. Webster-Stratton and colleagues established predictive validity of the instrument by finding the instrument to be sensitive to the intervention (Webster-Stratton, 1998;Webster-Stratton, Reid, & Hammonde, 2001). Cronbach's alpha for the current KA sample was.68 for harsh discipline, 0.72 for positive discipline, and 0.60 for appropriate discipline. (Gottman, Declaire, & Goleman, 1998) is an instrument that has 81 yes/no items with four subscales that measure four parenting styles in emotion coaching including dismissing (e.g., 'Children really have very little to be sad about'), disapproving (e.g., 'A children's anger deserves time-out'), laissez-faire (e.g., 'You should express the anger you feel'), and emotion coaching (e.g., 'Anger is an emotion worth exploring').. The scale was translated forward and backward using Brislin's method (Brislin, 1970). One positive emotion coaching style score was computed by calculating score of emotion coaching score minus mean score of dismissing, disapproving, and laissez-faire Cronbach's alpha for the current KA sample was 0.74 for dismissing, 0.78 for disapproving, 0.55 for laissez-faire, and 0.60 for emotion coaching. --- Emotion coaching-The Emotion Coaching Parenting Style Parental warmth-rejection-The Parental Acceptance-Rejection Questionnaire (PARQ) is a 24-item, 4-point, Likert-type scale instrument that includes eight questions measuring warmth and sixteen measuring rejection (Rohner, 1991). Warmth subscale scores range from 8 to 32 with higher scores indicating higher parental warmth. Rejection subscale scores range from 16 to 64 with higher scores indicating higher parental rejection. Rohner (1991) reported evidence for convergent, discriminant, and construct validity. This instrument is available in English and Korean. Cronbach's alpha was 0.81 for warmth, 0.58 for rejection, and 0.85 for total scale. Parenting self-efficacy-The Parenting Self-Efficacy Scale (Choe & Chung, 2010) is a 37 item, 5-point, Likert-type scale instrument that assesses self-efficacy and parental ability to communicate with, discipline, and raise healthy children. Scores range from 37 to 185 with a higher score indicating a higher parenting self-efficacy. It was developed in Korean and was translated into English using the forward and backward method (Brislin, 1970). Choe and Chung reported construct validity and Cronbach's alpha of 0.95 for Korean mothers in Korea. Cronbach's alpha for current sample was 0.96. Child emotional and behavioral problems-The Pediatric Symptoms Checklist (Jellinek & Murphy, 1990) is a 35-item, 3-point, Likert-type scale instrument that assesses cognitive, emotional, and behavioral problems in children ages four to sixteen. Scores range from 0 to 105 with a higher score indicating a higher level of emotional and behavioral problems. It has 66% sensitivity and 95% specificity (Jellinek & Murphy, 1990). Jellinek and Murphy (1990) reported evidence for convergent and construct validity. The scale was translated forward and backward using Brislin's method (Brislin, 1970). The Cronbach's alpha for the Korean version in the current study was 0.73. --- Intergenerational acculturation conflicts-The Child Intergenerational Acculturation Conflict Scale was adapted from the Asian American Family Conflict Scale (AAFCS) (Lee, Choe, Kim, & Ngo, 2000). It is a 10-item, a 5-point, Likert-type scale instrument that measures the frequency of Asian American parent-child acculturation conflicts over values and practices. Scores range from 10 to 50, with higher scores indicating a higher frequency of parent-adolescent conflict. The validity was demonstrated by statistically significant correlations between the AAFCS scores and the Social, Attitudinal, Familial, and Environmental Acculturation Stress Scale scores (Gamst, Liang, & Der-Karabetian, 2011). Cronbach's alpha for the current study sample was 0.89. Observational data-Videotapes (see Procedure for details) were analyzed using Dyadic parent-child interaction coding system (Eyberg & Robinson, 2000). Positive interactions included a parent validating a child's emotion, helping with problem solving, reflecting the child's statements, making descriptive comments of the child's behaviors, following the child's lead, or giving the child options. Researchers also tracked the number of times a parent praised a child for a specific action within a 15-minute period. The coder was not blinded to the group assignment or assessment interval. Intervention group mothers' satisfaction with the program-The Parent Satisfaction Questionnaire was developed specifically for this program. This 65-item, 5point Likert-type scale instrument consists of five subscales to measure intervention group mothers' general satisfaction with the KPTP program, the trainer, the usefulness of the program delivery method, and the usefulness/applicability of the parenting techniques taught in the program. Cronbach's alphas ranged from 0.72 to 0.88 for the current KA study sample. --- Procedure Participants were recruited from six participating Korean ethnic churches from August to September 2011. During that time, a total of 58 women agreed to be part of the study. After completion of pre-intervention assessments for all participants, churches were assigned to intervention group (IG) and waiting-control group (WCG) using partial group randomization. The second author suggested pairing the churches based on the number of volunteers from each. Two churches, each with large numbers of participants were paired; however one church ultimately asked to take the class in spring (i.e., be in a waiting-control group) due to their church schedule. The third church has a large number of participating mothers who spoke English more fluently than Korean; as a result, this church was assigned to the intervention group. The rest three small churches were paired and randomized. Mothers could chose a KPTP intervention provided in Korean or in English. By the end of the study, five groups had received KPTP in Korean and one group had received it in English. There were four IG composed of a total of 31 mothers and two WCG composed of 27 mothers. Each group had five to twelve mothers. Once the groups were formed, the KPTP was delivered in a small group setting to the IG participants from September 2011 to March 2012. Data were collected from IG and WCG at pre-intervention (T1), post-intervention (T2), and 3-month follow up (T3). After T3 data collection the KPTP was offered to the WCG. No additional data were collected after KPTP was offered to the WCG. The institutional review board approved the study and the researchers obtained informed consent from all mothers and targeted Sunday School teachers. They also obtained assent from all children whose mothers had agreed to be in the study. Among the 31 mothers in the IG, three mothers withdrew (reasons: class ended too late in the evening, heath problem, returned to Korea) and three did not complete the survey. Among the 27 mothers in the WCG, four withdrew (reasons: pregnancy, two had care for a sick family member, one had to visit Korea). The data, therefore, represent responses from 48 total mothers (IG n = 25, WCG n = 23) for the T1 and T2 comparison. At T3, two mothers in IG and two in WCG dropped out (reasons: families did not want to fill out the questionnaires). Thus, for the T1 and T3 comparison, the data represent responses from 44 total mothers (IG n = 23, WCG n = 21). Using ANOVA, no significant differences on the demographic variables were found between mothers who dropped out (n = 11) and those who remained in the study (n = 48). Each mother received a $10 gift certificate upon completion of each T1, T2, and T3 assessment. Sunday School teachers (IG n = 8, WCG n = 8) completed self-report surveys on child emotional and behavioral problems observed in church. Teachers were blinded to weather parents of children were in the experimental or waiting-control group and to the assessment interval. Teachers received a $5 gift certificate upon completion of each T1, T2, and T3 assessment. Children (IG n = 16, WCG n = 9) who could read and write filled out surveys on perceived acculturation conflicts that occurred between them and their mothers. The research team read the questionnaires to the non-literate children (ages 5 and older) and recorded their answers. Children received a $10 gift certificate upon completion of each T1, T2, and T3 assessment. Observational data were gathered in the form of videotapes filmed at T1, T2, and T3 assessment. The tapes showed participating mothers and children interacting using a Tinkertoy construction set or a Jenga game. Parents were instructed to play with their child as they would normally play and to ask their child to clean up the toys once fifteen minutes had passed. Intervention delivery-Two bilingual and bicultural interventionists (i.e., the lead author and one interventionist) delivered the KPTP separately in small group sessions. The KPTP was delivered over a twelve-week period using weekly, 3-hour classes followed by single booster sessions (also 3-hours in length) held monthly for a total of three months. Intervention integrity and dosage-To assure integrity and consistency of each class, the lead author trained the interventionist by reviewing the PowerPoint of the program protocol including parenting principles, role-plays, and homework at weekly supervision meetings. The lead author also observed several sessions delivered by the interventionist. IG mothers attended the program an average of 9.71 classes (81%, range 5-12 classes) and completed 80% of the assigned weekly homework. --- Data Analysis Baseline characteristics were compared using descriptive statistics, ANOVA, and Chi square tests. The primary aim of the study was to test for IG versus WCG differences in the primary outcome variables. Analyses used a generalized estimating equation (GEE) model in SPSS 17 to test for treatment group differences at follow up times while controlling for the baseline value of the outcome variable. GEE estimates the parameters of a linear model with a possible correlation between subjects within clusters (Diggle, Heagerty, Liang, & Zeger, 2001). For the purposes of this study, clusters are the class groups. The GEE uses robust standard error estimates based on the'sandwich estimator' to take into account withinsubjects correlations of the study variable scores during the follow up period. P values were 2-sided with a significance level of 0.05, and confidence intervals were at the 95% level. In this study, there are twelve outcome variables that were measured at two outcome time points (post-intervention and 3 months). Each outcome variable was analyzed separately at each time point. No formal adjustment was made for multiple comparisons in this pilot study, thus results should be interpreted with this in mind. --- Results --- Demographic Characteristics As shown in Table 2, no significant difference on all demographic variables was found between mothers, children, and Sunday School teachers between IG and WCG at T1, T2, and T3, except U.S. residency for T1 and T2. Overall, IG mothers lived in the U.S. significantly longer than WCG mothers, which can be attributed to the inclusion of the English-speaking group in the IG. This group consisted primarily of women who had come to the U.S. at a young age. However, when data were analyzed with and without controlling the length of stay in the U.S., the results were basically the same. The data presented in Table 3 did not control for the length of stay in the U.S. There were no significant differences on the demographic variables between children or Sunday School teachers in the IG or the WCG. --- Pre-Intervention Group Equivalence and Descriptive Data Means and standard deviations for all study variables by group and for T1, T2, and T3 are reported in Table 3. Oneway ANOVA indicated no significant differences on any of the study variables between IG and WCG at T1. --- Pre-and Post-Between Group Hypotheses Testing Using GEE As shown in Table 3, IG mothers, when compared to WCG mothers, increased effective and appropriate discipline, parental warmth, emotion coaching, and self-efficacy, and decreased parental rejection and harsh discipline (hypothesis 1). Observational data (i.e. videotapes) showed that IG mothers increased use of positive interactions and praise and decreased use of negative interactions more than WCG mothers. According to reports from mothers and Sunday School teachers, IG children decreased emotional and behavioral problems as compared with WCG children (hypothesis 2). IG children self-reported fewer intergenerational acculturation conflicts with their mothers as compared with WCG children (hypothesis 2). Therefore, both hypothesis 1 and 2 were fully supported using self-report data from parents, Sunday School teachers, and children, as well as videotape data of parentchild interactions. All significant treatment effects were maintained at T3 except teacher's report of child emotional and behavioral problems, child's report of intergenerational acculturation conflict, and videotaped data of mothers' positive and negative interactions with children. --- Intervention Group Mothers' Satisfaction with the Program The overall score of IG mother's general satisfaction with KPTP was 4.1 out of 5 (SD = 0.5), indicating that mothers approved of the program. Additionally, mothers said the program's delivery methods (M = 4.2 <unk> 0.4) and the parenting skills (M = 4.3 <unk> 0.3) covered were helpful. The parenting strategies most appreciated by mothers included: helping children with emotion regulation (20/42, 48%); learning what the Bible states about parenting (10/42, 31%); use of sticker charts (10/42, 24%) or specific praise (10/42, 24%) to promote good behavior; and how to express love to one's child (9/42, 21%). Overall, participants thought the strategies featured in the KPTP were easy to practice (M = 2.5 <unk> 0.5), with the easiest being expressing affection to one's child (M = 1.5 <unk>0.9), followed by using sticker charts (M = 2.1 <unk> 0.9), and giving specific praise (M = 2.1 <unk> 0.9). The hardest parenting strategies to use included: anger management (M = 3.3 <unk> 0.9); relaxation/ meditation (M = 3.1 <unk> 1.2); and emotion coaching (M = 2.9 <unk> 1.1). The most commonly cited reasons for attending the class were: to learn effective discipline strategies (31/42 74%); to learn emotion coaching (21/42, 50%); to better understand parental roles reflected in the Bible (14/42, 33%); to learn about parental stress and emotion regulation (13/42, 31%); and to practice ways to decrease intergenerational acculturation conflict (10/42, 24%). --- Discussion The major finding of this study was that the KPTP appears to be a promising way to promote effective KA parenting and increase positive mental health outcomes for KA children. Compared to WCG mothers, IG mothers were significantly more likely to employ effective/appropriate discipline strategies at both T2 and T3. These results are consistent with previous studies that demonstrated how parent training increases effective parenting strategies among ethnically diverse families (Gross et al., 2003;Martinez & Eddy, 2005;Tucker, Gross, Fogg, Delaney, & Lapporte, 1998). The findings of this study are particularly noteworthy given the results of the lead author's previous research testing the Incredible Years program with KA parents. In that study, all participants increased only their use of positive parenting strategies (Kim, et al., 2008). Furthermore, acculturation showed a significant impact on treatment effect; mothers with low acculturation significantly decreased harsh discipline, whereas those with high acculturation showed a trend of increased, appropriate discipline (Kim, et al., 2008). In the current study, the treatment seemed to impact all mothers as IG mothers reported less harsh discipline, more positive discipline, more appropriate discipline, and more emotion coaching. Videotape data of parent-child interactions also indicated that study participants significantly increased positive interaction and decreased negative interactions at T2. While IG mothers praised their children an average of 3 times/videotaped session at T1, this number jumped to 9-11 times/videotaped session at T2 and T3. This study also found KA mothers significantly increased parental warmth and decreased parental rejection at T2 and T3. These factors are significantly related to KA children's emotional and behavioral problems and social competence, as well as depressive symptoms among KA adolescents (Kim, 2002;Kim, et al., 2006;Kim & Cain, 2008;Kim, Guo, et al., 2010;Kim, et al., 2007). Compared to a standardized parenting program, the KPTP was developed to fit Korean culture and faith, which are important social determinants that affect health, functioning, and quality of life outcomes (US DHHS, 2012). The specialized curriculum allowed participating mothers to easily learn and practice strategies taught in the program, which resulted in a high degree of participant satisfaction with the curriculum. KA mothers felt more comfortable adopting and practicing new parenting styles because the KPTP linked these practices to KA culture and faith. For example, a previous study found that KA parents made decisions for their children and asked children to follow them (Kim, Im, Nahm, & Hong, 2012). This practice is strongly supported in collectivistic Confucian Korean society, where parents view their children as extensions of themselves, and as such, assume full responsibility for their children's good and bad behaviors and outcomes (Ahn, 1994). KA parents also perceived this strict parental control as a way of expressing their love for their children (Kim, 2005). The KPTP helped participating parents realize this practice may not fit well within the individualistic American social context where children are encouraged to be independent, nor did it fit with Christian faith principle, which views a child as a separate entity from his/her parents. The KPTP also helped parents to understand that this practice might hinder the development of a child's autonomy. Once learned, the parenting strategies in the KPTP helped participating mothers feel more competent, as evidenced by their significantly higher parenting self-efficacy scores recorded immediately and three months after taking the KPTP. Previous studies have also demonstrated that participants who take parent trainings become empowered through their experience (Gross, Fogg, & Tucker, 1995;Tucker, et al., 1998). In the late 1970s, about 60% of first generation KAs in the Midwest saw a need to alter their parenting practices because existing practices delayed cultural adaptation into American society (Hurh, Kim, & Kim, 1979). Recent studies found that KA parents reconstructed their parenting practices after evaluating the pros and cons of Korean parenting, constructed in Korea, and American parenting, observed in the U.S. (Kim & Hong, 2007;Kim, et al., 2012). Findings from the current study indicate the KPTP could help KA mothers make this transition more quickly and successfully. This finding is consistent with health promotion studies that found culturally and linguistically appropriate programs were effective in changing immigrants' health behaviors (Kelly, Huffman, Mendoza, & Robinson, 2003;Zuniga de Nuncio, Nader, Sawyer, & Guire, 2003). However, this pilot-study needs to be expanded using a larger sample. In this study, children's baseline emotional and behavioral problem score was low in both conditions. Nonetheless, IG mothers reported that their children significantly decreased emotional and behavioral problems at T2 and T3. Sunday School teachers also reported that participating IG children significantly decreased emotional and behavioral problems at T2. Previous studies have shown that parent training promotes effective parenting practices and can lead to a decrease of children's emotional and behavioral problems and the prevention of conduct problems in adolescence (Barlow, et al., 2010;Dretzke, et al., 2009). Children of IG mothers reported significantly fewer intergenerational acculturation conflicts at T2. Previous studies have shown intergenerational acculturation conflict was significantly related to KA adolescents' depressive symptoms (Kim & Cain, 2008). The KPTP teaches parents how living in two cultures impacts the parent-child relationship and children's mental health based on the acculturative family distancing theory (Hwang, 2006). This study shows that when parents understand and work with the root causes of intergenerational acculturation conflict, children will perceive less conflict. Since this effect was not maintained at T3, this needs to be reexamined using a bigger sample. Several limitations must be noted. Each outcome variable was analyzed separately at each time point and formal adjustment was made for multiple comparisons; therefore, results should be interpreted with this in mind. Since participants were recruited from KA churches, they may have been more involved in the KA community and more willing to learn new discipline strategies. Hence, results cannot be generalized to the total KA parent population. A partial group randomization assignment weakened the strength of a randomized controlled experimental study. The treatment fidelity was checked by the first author who also was one of the interventionists, which may have decreased the meaningful assessment. Finally, Cronbach's alpha reliabilities for harsh discipline, appropriate discipline, liaise-faire, emotion coaching, and parental rejection were low. A few strengths need to be stated. The KPTP was developed using a great degree of cultural sensitivity and this influenced the positive effects of the program. It combines well researched, effective parenting techniques with KA culture and faith to positively impact KA parenting and child mental health using the parental acceptance-rejection theory and the acculturative family distancing theory as theoretical frameworks. Not only have parental acceptance-rejection and intergenerational acculturation conflict been proven to impact child mental health, they carry particular significance for the KA population and are, therefore, appropriate and effective. This study also used multiple sources of data to test the research hypotheses. The approach used in this study can be applied other immigrant families who face similar challenges of raising children in two cultural parenting norms. --- Conclusion The KPTP was effective in changing KA mothers' parenting practices and improving their children's mental health outcomes because it taught effective discipline strategies using KA culture and faith as a foundation. Future research should be conducted with a larger sample size, should include an assessment of both mothers and fathers, and each site should have both an intervention and a waiting-control group. For example, small churches can be grouped to allow for 20-30 families in one research site. Furthermore, Korean language videotapes of the KPTP program would facilitate understanding of the content. Implications for healthcare practice must be viewed with caution since this was a pilot study utilizing a small sample. Nonetheless, this study shows that healthcare professionals can positively impact the discipline practices of KA mothers through culturally relevant parenting education training. With practice, specific discipline strategies can effectively increase positive parenting among KA mothers, reduce child/parent intergenerational conflict, and improve the mental health of KA children. (Gottman, et al., 1998): parents help children with emotion regulation by recognizing their emotions, taking deep breath to calm self, listening and validating emotions, and helping children to solve problems that triggered the emotion. <unk> Operant conditioning procedures (Gilmel & Holland, 2003): parents apply consequences following a child's behavior in order to increase or decrease the likelihood of that behavior recurring. <unk> Bright Futures in Practice: Mental Health (Jellinek, et al., 2002): a national health promotion and disease prevention initiative that addresses children's health needs in the context of family and community

Problems-Korean American (KA) children experience mental health problems due to difficulties in parenting dysfunction complicated by living in two cultures.Parent Training Program (KPTP) was pilot tested with 48 KA mothers of children (ages 3-8) using partial group randomized controlled experimental study design. Selfreport survey and observation data were gathered. Findings-Analyses using generalized estimating equation indicated the intervention group mothers increased effective parenting and their children decreased behavior problems and reported less acculturation conflict with mothers.The KPTP is a promising way to promote effective parenting and increase positive child mental health in KA families.Korean American; parent training; randomized controlled experimental study; culturally and linguistically appropriate intervention Korean American (KA) immigrant parents encounter two cultural parenting norms which make them feel incompetent in their parenting; firsthand experience with Korean parenting practices and an unfamiliarity with desired parenting practices in the U.S. (Kim & Hong, 2007). Living in two conflicting cultures can also create conflict between parents, who practice Korean style parenting, and children, who quickly adopt U.S. cultural practices. Parenting dysfunction and the complications that arise when immigrant families must bridge two cultures are related to mental health problems among KA children and adolescents. KA children and teens report more emotional and behavioral problems than European American children or other Asian American subgroups (Choi,

in one research site. Furthermore, Korean language videotapes of the KPTP program would facilitate understanding of the content. Implications for healthcare practice must be viewed with caution since this was a pilot study utilizing a small sample. Nonetheless, this study shows that healthcare professionals can positively impact the discipline practices of KA mothers through culturally relevant parenting education training. With practice, specific discipline strategies can effectively increase positive parenting among KA mothers, reduce child/parent intergenerational conflict, and improve the mental health of KA children. (Gottman, et al., 1998): parents help children with emotion regulation by recognizing their emotions, taking deep breath to calm self, listening and validating emotions, and helping children to solve problems that triggered the emotion. <unk> Operant conditioning procedures (Gilmel & Holland, 2003): parents apply consequences following a child's behavior in order to increase or decrease the likelihood of that behavior recurring. <unk> Bright Futures in Practice: Mental Health (Jellinek, et al., 2002): a national health promotion and disease prevention initiative that addresses children's health needs in the context of family and community. Promotes use of positive discipline (hugging/kissing) and appropriate discipline (sticker charts, timeouts). --- Author Manuscript Author Manuscript Kim et al. Page 17

Problems-Korean American (KA) children experience mental health problems due to difficulties in parenting dysfunction complicated by living in two cultures.Parent Training Program (KPTP) was pilot tested with 48 KA mothers of children (ages 3-8) using partial group randomized controlled experimental study design. Selfreport survey and observation data were gathered. Findings-Analyses using generalized estimating equation indicated the intervention group mothers increased effective parenting and their children decreased behavior problems and reported less acculturation conflict with mothers.The KPTP is a promising way to promote effective parenting and increase positive child mental health in KA families.Korean American; parent training; randomized controlled experimental study; culturally and linguistically appropriate intervention Korean American (KA) immigrant parents encounter two cultural parenting norms which make them feel incompetent in their parenting; firsthand experience with Korean parenting practices and an unfamiliarity with desired parenting practices in the U.S. (Kim & Hong, 2007). Living in two conflicting cultures can also create conflict between parents, who practice Korean style parenting, and children, who quickly adopt U.S. cultural practices. Parenting dysfunction and the complications that arise when immigrant families must bridge two cultures are related to mental health problems among KA children and adolescents. KA children and teens report more emotional and behavioral problems than European American children or other Asian American subgroups (Choi,

Introduction The interest in the exchange and consumption of news is one of the oldest human habits (Stephens 2006). What has changed over time is the medium to disseminate news: from word-of-mouth, to written, to printed, to broadcasted and finally to online publishing and social media sharing. The shift from print to online offers new opportunities to study people's access to news, which human scientists, but also journalists and news organizations can leverage to understand readers' interests and better adapt their palette of content. The literature is rife with data-driven studies of online news consumption (see Sec. 2). These studies characterize various aspects, for example the virality of news (Kourogi et al. 2015), news coverage in social media compared to traditional media (Olteanu et al. 2015), or user preferences compared to editors' suggestions (Boczkowski 2010;Abbar et al. ). However, as with any empirical work, these studies are limited by what their input data allows to observe. We identify four types of data sources used so far to characterize online news. (i) The first is published content. For example, projects like GDLET (Leetaru and Schrodt 2013) and EventRegistry (Leban et al. 2014) crawl online news articles worldwide and make them available to researchers. Although this approach faithfully captures the "supplier" part, such data does not convey what users really consume. (ii) The second source is what is publicly shared in popular social networks (Osborne and Dredze 2014). Although this approach nicely captures how users relay and react to news, it misses what happens outside the social network. (iii) The third source is collecting usage data statistics from an individual online newspaper (Boczkowski 2010;Abbar et al. ;Dezsö et al. 2006), a comprehensive view but limited to one news outlet. (iv) Finally, researchers and practitioners counter the partial visibility of these sources with more traditional sources like user surveys. Institutes like Pew Research and Reuters regularly issue survey-based reports (Mitchell and Page 2014;Newman, Levy, and Nielsen 2015) tracking various aspects of news consumption habits worldwide. Although they offer insights on larger aspects of human behavior and preferences, it is not clear how precise user explicit feedback is in capturing actual usage of online news. In this work, we uncover a different source of data and contrast it, to the extent possible, to the previously described four existing data sources. In particular, we extract and analyze online news visits made by a population of users. We define a news visit as a click to a web page containing a single news article. In contrast to prior work focusing on a single news outlet (e.g., Dezsö et al. (2006)), we observe all news visits of individuals in a population of users across any online news website they visit when connected to their Internet provider network. More precisely, we extract online news visits by passively observing traffic traversing a network link, from which we extract the pages containing news that users visited. We perform our analysis on data collected for about one year and a half at vantage points we installed in Italy: three are located in the network of a large Internet Service Provider (ISP) and the forth is located in a large university campus. During this period we observe 80 million visits to 5.4 million distinct articles overall. In this paper, we focus in particular on the popularity of news categories. We define a category as the general theme of a news article, as directly assigned by the publisher, as opposed to the specific story or event it covers. News categories correspond thus to the sections under which news editors often group their articles (e.g. Sport or International). We quantify the popularity of news categories in terms of number of visits and put it in perspective with their popularity considering various sources: (i) news discovered on Facebook, (ii) news visits to major news outlets, (3) published articles, and finally, (iii) user preferences expressed in surveys (Sec. 5). To estimate the view that social media and major news outlets have, we analyze the referral of the news visits, i.e., the previous page from which the visit came, to study how users discover news articles and quantify the importance of social networks and major publishers in the overall consumption volumes (Sec. 4). We regularly crawl all news outlets in Italy for a duration of three months to get a dataset of all published articles. Finally, to better assess the significance of the category popularity, we study their stability across locations (spatial) (Sec. 7) and across time of day and months (temporal) (Sec. 6). We summarize our findings as follows. • Usage data obtained from the network complements existing data sources well. At most 11% of news visits in our datasets come from social networks (Sec. 4). As such, studying only news shared on social media captures a small fraction of news visits-as already pointed by Flaxman, Goel, and Rao (2016). Popularity of categories as inferred from news discovered on Facebook does not match the popularity obtained from the analysis of all data. In particular, "Tragedies and Weird" news are over-represented on Facebook, while "Sport" is under-represented. Similarly, the popularity of categories we observe when considering all visits does not match that observed individually by a major national news outlet either. • "Tragedies and Weird" and Sport are the most popular categories of news in our datasets (around 16% of visits each). Politics is surprisingly less popular with 3.5% of visits. Overall, despite notable differences, the popularity of many categories follows the supply, or what is published by editors. Surprisingly, Reuters' survey estimates do not match our data. For instance, respondents report International and Politics news as some of the most popular categories, far beyond Sport and Weird news, suggesting a dissociation between what users prefer and what they actually consume or are attracted by (Sec. 5). • We discuss two possible causes to explain this dissociation. The first is social desirability bias whereby users would show a preference for categories that give a better image of themselves. The second, inspired by the incentive-sensitization theory of addiction (Berridge and Robinson 2016;1998) in neuroscience, postulates that what we observe is a legitimate difference between what people "want" (the motivational part of a given reward) and what they "like". Our discussions with two experts (the professor of political communication and editor of the Reuters survey-based report, and the biopsychologist who formulated the incentive-sensitization theory of addiction) indicates that the second cause is more likely (Sec. 8). • News-category popularity is overall stable across locations, despite demographic differences between the populations we study (Sec. 7). Albeit overall stable, the popularity of certain categories slightly varies depending on time of day or certain events of the year (e.g., vacations, beginning and end of the sports season and major events like the Paris terrorist attacks) (Sec. 6). --- Related Work Since the introduction of online newspapers, researchers in different communities have seen new opportunities to study news consumption. In this paper, we study online news with a focus on the difference between various data sources. Comparison of various news sources. Some of the related work compared various sources of news in terms of coverage. Zhao et al. (2011) use topic models to compare the topics in news shared in Twitter with those of the New York Times for a period of three months. The authors found that Twitter covers more personal life and pop culture and that users tweet less about world events, but retweet a lot, which causes the news to spread. Olteanu et al. (2015) compare the coverage of news about climate change in traditional news media and in Twitter. They show that the scope of the traditional news media they consider and Twitter are different. Chakraborty et al. (2016) also compare the topical coverage of the New York times, Twitter and Facebook, and highlight other differences. Kwak et al. (2018) compare the gap between "media attention"-measured as popular topics at a news aggregator -and "public attention" -measured with Google Trends -on international news coverage across hundreds of countries. Finally, Boczkowski (2010) In this work, we not only compare news categories discovered on social media and those published by news outlets, but we additionally compare these categories to (i) categories considering all news visits of a population, (ii) those considering only visits to major news outlets and, finally, (iii) those preferred by users in surveys. We similarly find differences between published articles and those visited from Facebook. In addition to prior work, we find that focusing on a single news outlet, although very popular, does not capture the overall behavior of users when including all outlets. Hence, what is learned from a large single outlet cannot be representative of all traditional news media. Unlike prior work, ours spans visits to all news outlets, for different populations of users and for a long period of time. This allows us to observe how habits change over time and regions. Usage of surveys. Surveys remain a valuable tool to understand user behavior online (Mitchell and Page 2014;Newman, Levy, and Nielsen 2015;Prior 2009;Lee and Chyi 2014). Research institutes like Reuters or Pew Research regularly issue survey-based reports tracking various aspects of user behavior. Here, we contrast popularity inferred from usage data with self-reported preferences in surveys (Newman, Levy, and Nielsen 2015). Similarly to our work, Lee and Chyi (2014) question the perceived noteworthiness of consumed news. The authors rely on a survey to show that only about one third of the content produced by the mainstream news media is perceived as noteworthy. Our work shows that what users value might not necessarily match what they will actually view. Surveys are also known to suffer various biases. For what concerns news, in a prominent work, Prior (2009) has confirmed anecdotal evidence suggesting the inability of surveys to accurately measure news media exposure. In particular, by directly comparing Nielson's audience measurement estimates to survey-based estimates, Prior showed that selfreports tend to overly inflate media exposure, by a factor of three on average and up to eight for certain demographics. Our work complements his by comparing survey estimates and usage data based on the finer granularity of what users value (news-category preferences) and not only exposure. We reveal indeed a major difference between self-reported preferences and actual views. News on social media. Less directly related to our work, in recent years, social media has risen as a means to share news. This has led to the explosion of social media datadriven studies of news. We report few examples. Kwak et al. (2010) show that 85% of the topics in Twitter are related to news headlines. Osborne and Dredze (2014) compare the performance of Twitter, Facebook and Google Plus in terms of news coverage and latency. They find that Twitter is faster while the other two offer more diversity. Saez-Trumper, Castillo, and Lalmas (2013) identify different kinds of biases in news shared on Twitter. Wang and Mark (2013) study the news consumption from social media in China. They characterize different types of audience for different types of news. Kourogi et al. (2015) extract features from the headlines and text of tweeted news and use an SVM ranker to infer features that can predict the virality of a news article. Morgan, Lampe, and Shafiq (2013) show that users sharing news on Twitter present no bias based on their perceived ideology of the news outlets, i.e., they see no bias due to selective exposure in their datasets. Flaxman, Goel, and Rao (2016) study whether social media helps increase exposure to diverse perspectives or lead to ideological segregation by creating filter bubbles. In their analysis, they find that most news consumption comes from people directly visiting mainstream newspapers web sites, with only a small fraction of news visits coming from Facebook. In this paper, we confirm that only a small fraction of news consumption in our datasets comes from social media and highlight a difference in the popularity of certain news categories when considering overall visits and those coming from Facebook. --- Usage data: a new observation point In this paper, we rely on datasets that gives us a privileged vantage point to study news categories' consumption habits. In this section, we present our datasets, the ethical issues around their use and how they complement existing sources. --- Datasets and methodology Our analysis relies on a record of news visits we extract from passive Internet traffic observation on four different networks with tens of thousands of users who access the Internet from their PCs, smartphones and tablets. We first describe our methodology to extract news visits from the raw traffic. Data collection. We collected data from four networks: at one university campus in Italy (Campus-City2) and in a large residential ISP in two different Italian cities (ISP-City1, ISP-City2-a, and ISP-City2-b). In the first scenario, we monitor the link that connects the campus LAN and WiFi networks to the Internet. In the residential scenarios, we monitor traffic from one district in one city (ISP-City1 ), and two different districts in another city (ISP-City2-a and ISP-City2-b ). Here, users connect to the Internet thanks to ADSL or FTTH access, which is shared through WiFi inside their houses. The traces were obtained by running Tstat (Finamore et al. 2011). Tstat processes network packets and extracts raw HTTP requests. More precisely, for each HTTP request, it extracts the requested URL, the referral (which captures the URL of the previous page visited by the user, if any), an anonymized IP address, the timestamp of the visit, and the user agent (which describes the browser or the app originating the request). All in all, we analyze tens of billions of HTTP connections spanning the period from January 1st 2015 to May 15th 2016. We rely on a fifth dataset, Published-articles obtained by crawling the 667 major news outlets in Italy for a period of 3 months. This allows us to contrast what is published with what is consumed. All in all, we collect 80k articles. From HTTP traces to visits. HTTP traces contain a large number of requests to images, scripts, or ads that compose webpages, but that do not correspond to the page users actually visited. Hence, we first need to extract visits to webpages. For this, we use a Hadoop cluster and adapt algorithms from our earlier work (Ben Houidi et al. 2014) to extract user visits all HTTP requests. Identifying news articles. From the set of visited URLs, we keep only those URLs that correspond to well-known online newspaper sites. Defining which webpage corresponds to a news article is subjective and tricky. In particular, lists of news outlets used in prior studies (Bakshy, Messing, and Adamic 2015;Ribeiro et al. 2018) focus mainly on Englishspeaking news outlets and hence are unfit to identify news articles in our Italian-based dataset. To overcome this issue, we rely on the Google News authority and consider all publishers that it indexes as potential news outlets. To build our list, we crawl Google News Italy once every 20 minutes during one month and build a list of unique outlets. The resulting list has 667 online outlets that we make available online (onl online April 2019). Google News does a thorough job in indexing news. Our manual inspection of outlets in the list shows that it covers all the popular online outlets in Italy including some influential blogs. Our analysis, however, may miss some foreign or unpopular niche newspapers. Table. 1 provides for each of the four locations the number of distinct news articles that we observe. Assigning categories to articles. Finally, we label news articles with categories. Assigning a category to a news article is also difficult and subjective. To counter this, we rely on the sections under which editors publish articles. So, similarly to prior work (dos Rieis et al. 2015), whenever available, we extract the category of news assigned by the news editor. Using this method, we identify first 167 unique section names in the four datasets. Since different editors may give different names to the same categories (e.g., use of plural and singular), we manually merge similar categories. Similarly, we group sub-categories into larger ones by hand. We hence obtain 40 distinct categories. This allows us to confidently label more than 70% of news articles. The remaining 30% are hard to automatically label. By manually inspecting them, we observed that most of them come from minor unpopular websites that do not adopt proper labeling of sections. To shed light on the composition of some ambiguous categories, we manually verify three categories that sometimes overlap with others: • International: Many articles are published under the International section, but it is difficult to further pinpoint the category of the news article beyond the fact that it happened abroad. We manually label 200 randomly chosen articles from this category and find that 60% is about "Tragedies/Weird" news and 16.5% of International politics. An additional 11% are a mixture between politics and tragedies (wars, fraud, terrorism attacks, etc.). • Ed. Columns: Many articles are grouped by editors under various names like "columns", "opinions" or (special) "editions". For the ease of the presentation, we group all these articles under the label "Ed. Columns". Similarly, labeling a 200-article sample, we find that such columns cover a large number of topics, with "Tragedies/Weird" having 25% of articles and Politics 17%. • Tragedies/Weird: One popular section referred in Italian as "cronaca" is notorious for reporting mainly Crime stories. We manually label 200 randomly chosen articles from this category and find that it contains 42% of Crime-related articles, 31% are about natural disasters and accidents, and around 10% are weird news. For the ease of the presentation, we refer to this category as "Tragedies/Weird", but it contains a majority of crime news. We manually review samples of the rest of the categories as well to confirm that articles are correctly labeled. We also publish the scripts we use to group editor sections into categories as well as the 80K articles in Published-articles dataset and their associated categories (onl online April 2019) to allow further verification of our methods. Identifying users: Households and surfers. One challenge for our analysis is that the traces have no per-user identification. A single user may appear multiple times in the traces, because she may connect from multiple devices at home. Conversely, multiple users may share the home gateway's IP address, e.g., members of a household connected to the Internet through the same ADSL/FTTH gateway. To overcome this, we study "users" at two granularities: surfers and households. We identify a household by the anonymized IP address. Because of the static assignment by the ISP of IP addresses to home gateways, this corresponds to a household in ISP traces. 1 We define a surfer as the concatenation of the anonymized IP address and the user agent. A surfer captures a particular user surfing the web from a particular browser or mobile app. As documented in our prior work (Vassio et al. 2018), we observe that smartphones and tablets account for more than 50% of devices in the datasets and generate around 20% of page visits. Note that with this definition, the same person might still appear as multiple surfers. This is particularly likely if we take into account the long duration of our traces and the fact that user agents change with software upgrades. Neither surfers nor households defines precisely a user, we will use them both with caution to approximate per-user behavior. Table. 1 shows the numbers of observed households and surfers in the traces.2 Finally, unless otherwise stated, we consider the number of visits to a particular article as the number of distinct surfers that clicked on the article. Note that counting the popularity of articles considering the number of distinct households instead of the number of distinct surfers does not considerably change the results: only very few articles are read by more than one surfer in a household. --- Ethical Issues Privacy protection mechanisms implemented in Tstat have been devised in close collaboration with ISPs' privacy officers and legals (Trammell et al. 2014). Tstat processes packets in real time and obfuscates any Personally Identifiable Information (PII). IP addresses are anonymized using consistent and irreversible hashing functions -so that it is impossible even for the ISP network administrators to link traffic to a customer identity. In this work, we instrumented Tstat to collect the minimum information required for our analysis. It only logs URLs and referrals and user agent strings. Moreover, we instrument Tstat to sanitize URL to avoid exposing users' identifiers such as email addresses, by, e.g., removing query parameters after the "?" character. Web cookies and locations are not logged. Note that Tstat has no visibility on encrypted traffic (HTTPS), where the sensitive information concentrates. 3 Our analysis ignores the traffic generated by users who activate the Do-Not-Track flag in their clients (observably less than 1%). Our traffic monitoring activity and data collection obtained the approval of the Security and Privacy Offices of the campus and the ISP in which we deploy our traffic monitoring probes. We discussed the opportunity to make customers aware of the data collection. Given the anonymity and the technical precautions we adopted to preserve users' privacy, this was considered to be not required. With the 2018 European General Data Protection Regulation (GDPR), however, this type of data collection does require explicit user consent. Thus, future data collection efforts of this type will become more challenging in Europe. Finally, the data is only available to the researchers for the sake of pursuing their research objectives, which involves only the study and publication of aggregated results. Hence, none of the analysis or the results we present in the paper can be linked back to a single user. The data analysis (which was conducted post-GDPR) was subject to a privacy impact assessment, conducted with the data protection officer of Inria. Given that the datasets were already available, that they do not contain any PII, and that the analysis builds on aggregated trends, the conclusion was that no additional formal ethical review was required for the analysis. --- Limits Our methodology and datasets have some limits that we summarize hereafter. • First, our perspective is spatially constrained: we can monitor news browsing behavior (on both mobile and laptop) only when users are connected to the networks we monitor, which correspond to either their home or work. As such, we miss online news consumption when users connect to the Internet using cellular networks, e.g., when they read the news from their mobile while travelling or commuting. • Second, our dataset focuses only on actual clicks to news articles and can not capture when users get exposed to online news without visiting the news article, for instance by seeing the preview as shared on a social network. • Third, we lack a precise way to identify a user. To this end, we mainly focus our analysis on properties exhibited by aggregated populations of users. Whenever needed, we carefully rely on the notions of surfer and household to approximate users. This limitation also prevents us by design from obtaining demographic information about users to be used in the analysis. • A fourth limitation is that this dataset comes from a single country. Therefore, our findings might not generalize to other regions and other cultures. • Finally, especially when later compared to surveys, our dataset lacks of provable guarantees of representativeness of users in the studied country. In general, whenever applicable, we tested the statistical significance of our results. Given the large scale of our dataset, when comparing empirical CDFs for instance, even barely visible differences between CDFs were significant up to 3% of significance level. So while we can assess how our results and data (which is considerably large both in time and space) are significant and representative for the populations we study, we have no guarantees that --- Weight of social media and online publishers Both social media and online publishers represent standard sources to study news consumption. To understand how usage data complements these sources and infer their importance, we study the news' referrals, i.e., the pathways from where the visit to the news article came. Using this approach, we can measure the share of visits in our datasets that comes from online social networks and online publishers. Tab. 2 presents the results for the three residential areas for two separate months, April 2015 and April 2016. The table shows a ranking of the most important sources and their respective shares. First, note that "Self referral" means that the visit comes from a webpage within the same website. The "Direct Browsing" class corresponds to visits with no referral. This may be due to direct visits (e.g., bookmarks) or often to links obtained via alternate channels like email or messaging apps that do not pass the referer, something sometimes called the "Dark social Web". 4 Note also that we verified that Google, Facebook and Twitter always passed the referral at the time of the data collection. Hence, we fortunately do not miss their referral traffic and can precisely estimate their share. Finally, it is worth noting that the impact of HTTPS on our referral analysis is limited. In the worst case, the referral share of other possible unknown HTTPS websites is upper bounded by 10-15%, which is the share of "Direct Browsing" (no referral). Overall, with the exception of the few variations that we discuss below, the results are stable across locations and especially across months. Not shown on the table, we verify that the results change only slightly between months. Regardless of the location, most of the visits come from the homepage of the online newspaper itself. Indeed, 57% to 69% of all news visits are self referrals. This result implies that users mostly rely on visiting the website of their favorite online newspaper to discover news. Far below, the next most frequent referrals are "Google search" 5 (around 10%), "Direct Browsing" (12-15%) and "Facebook" at around 10%, with the exception of ISP-City2a(around 5%), which we explain later. Twitter's share is much smaller (0.1-0.27%). Google News accounts for around 1% despite 22% of surveyed people in Italy claim to use it to read news (Newman, Levy, and Nielsen 2015). Two sources of data show visible variations across locations and across months. The first is the "Direct Browsing" which steadily increased in one year. We believe that this is due to the increase of HTTPS adoption, which implies that less referers are passed; more visits appear to us as "No referer". The second is Facebook, which has around 10% in ISP-City2-band ISP-City1 but only 4 to 5% in ISP-City2-a. We conjecture that this is due to the fact that ISP-City2-a aggregates traffic for an area which has a lot of office buildings. In a working environment, people may be less likely to use Facebook to discover news. This dataset, as opposed to the others, has indeed higher activity during the day and week days and lower activity at nights and weekends. Finally, the relatively small share of social networks (11% at most) is somewhat surprising. To verify that our datasets do not introduce a bias concerning social networks, we study the percentage of active Facebook and Twitter users. 6 We find that 65% of surfers in our datasets visited Facebook in 17 months, and only 7% visited Twitter. These percentages are similar to the available statistics of Facebook and Twitter usage in Italy (Manson online May 2016;Statista online May 2016). This result also confirms the study of Flaxman, Goel, and Rao (2016), which found a small fraction of news visits coming from Facebook. Notice that the immediate referral metric might underestimate the weight of social media. For instance, users coming from Facebook and landing on a news article may end up visiting a number of other articles on the site. We analyze the whole chain of referrals starting from Facebook. We find that the vast majority of visits leaving Facebook stops only one page away. This behavior is documented in more details in our prior work (Vassio et al. 2018). Since online newspapers are the most influential news pathways with 55% of referral traffic, we study also the percentage of visits that each of them drives individually. We 5 News visits from Google Search suggest that users either were intentionally looking for news about a particular event, or used the search engine to find the homepage of a newspaper. 6 Traffic towards these services is encrypted using HTTPS (thanks to the refer field, we can only see visits that come from these websites because they always pass the referrer when transitioning from HTTPS to HTTP). Hence, for this task, we use the Tstat TCP and DNS logs available with our datasets. find a large disparity between news outlets. While 80% of online newspapers are referral for less than 0.1% of all news visits, 31% of visits comes from the top 5 newspapers. To better understand this, we show in Fig. 1 the top 20 newspapers in terms of referral share in ISP-City1 (red bars), together with the referral shares of the same newspapers in ISP-City2b (blue bars). The top newspapers are similar, but present differences in their ranking due to regional preferences. For instance, outlet-city1-1 and outlet-city2 are very popular local online newspapers based in City1 and City2, respectively. We explore such a spatial effect in more details in Sec. 7. Takeaway. Although many studies of news consumption are based on news shared in social networks, direct access to online newspapers remains the main source for most users in our datasets. Only 16% of clicks to news articles in our dataset come from social networks. Understanding how and what news are shared is clearly important per se, but these news might only represent a small fraction of what is consumed. Depending on the geographical locations, certain news outlets account alone for up to 18% of all news referral traffic. We will further assess in Sec. 5 if data obtained from a single major news outlet or Facebook can be representative of news-category usage behavior at large. --- Popular News Categories We now measure the popularity of news categories in our datasets and put it in perspective with the popularity that can be inferred from various other sources. The first row of Table 3 presents the top 11 most popular categories by aggregating data from all our four locations. We define the popularity here by the percentage of visits that each category gets. The top 11 categories capture together around 85% of all visits. We see that the most visited categories are Tragedies/Weird, Sport, and Ed. Columns. Politics is surprisingly unpopular with less than 3.5% of the overall visits. We will show later that, despite notable variations, these results represent a behavior that is stable across time (Sec. 6) and space (Sec. 7). Table 4: Simulation of Top 11 most popular categories (percentage of interested surfers and households). --- Popularity according to various sources We now put in perspective the popularity we observe in usage data with the one that could be observed by other sources. For social media, we focus on Facebook and extract the subset of visits to news articles with Facebook as referral. We measure the popularity of various categories and report the results in the second row of Table 3. For usage data obtained from major news outlets, we similarly extract the subset of visits to such outlets. We present the results for Repubblica.it in the third row of the table. In the fourth row, we report the popularity of news categories as expressed by users in the Reuters survey in Italy (Newman, Levy, and Nielsen 2015). Note that with the exception of "Local" (somewhat included in Region in our case) and " Weird" (somewhat included in "Tragedies/Weird") news for which we do not have a perfect direct match in our editor-based categories, remaining categories have a suitable match in our dataset. Finally, we apply the same principle on the Published-articles dataset to measure category popularity among published articles in the last row. First, looking at news discovered on Facebook yields interesting differences with the overall usage data. For example, Sport news seem under-represented on Facebook with only 4% of visits originating from Facebook, whereas it represents 16% of visits in usage data. Tragedies/Weird on the other hand is over represented with 23% of popularity in Facebook against 16%. Other categories like Health (less than 1% when considering all visits) and People seem over represented on visits originated from Facebook. One possible explanation for these differences is that what people think valuable to share with others does not necessarily match what they themselves view. This explanation seems inline with a recent study in psychology that analyzed the neural correlates of message propagation (Falk et al. 2013). The study has found that people especially mentalized about what will be appealing to others when propagating a message. However, another explanation could be that the population that uses Facebook to discover news is different compared to the rest of the population. Further investigation is needed to understand these differences. Second, comparing with the popularity as seen by major news outlets, we also find noticeable differences. In the reported case of Repubblica.it, few categories like Ed. Column, International, Travel, Technology, Cars and Motos are particularly more popular compared to the entire data. This result suggests, at least, that relying on usage data from a single news editor is not representative of user consumption habits at large. Other outlets exhibit other differences. Our next comparison point is with surveys. At first sight, our usage data seems far from Reuters' (Newman, Levy, and Nielsen 2015) survey-based preference estimates for the same year in Italy. 7 Although 46% of users report being interested in Politics and 44% in Science, each of these categories accounts for less than 4% of the visits. The survey estimates are however expressed in terms of percentages of users. We perform various simulations to investigate in more details the differences between reported self-preferences and actual usage data in Sec. 5.2. Finally, to put these numbers further in perspective, we study the popularity of news categories in terms of number of published articles (and not visits). We remind that a direct comparison is thus not possible due to the difference of units. Interestingly, overall, the percentages of published articles seem to perfectly match the percentages of visited articles for many categories. There are, however, as in the previous cases, noticeable exceptions. Especially, Sport and to a lesser extent International and Science have higher popularity in terms of number of visits compared to published articles. Sport has indeed 16% of visited articles despite only 6.5% of published articles. International has also 7% of visits despite only accumulating 3.5% of published articles. Not shown in the table, only 1.9% of published articles are about Science but almost 4% of the views relate to Science. At the opposite side, the People category has a higher number of published articles (9%) than the actual fraction of visits it attracts (4.85%). Takeaway. First, usage data shows that Tragedies/Weird and Sport are by far the most accessed news categories. Despite being liked in surveys by almost one out of two users, Politics and Science attract less than 4% of visits. Second, none of the existing comparable sources of data can capture precisely the news category popularity at large when considering all visits from all users. Each of these sources yields a different popularity distribution that reflects the peculiarity of the data source. We next further investigate the difference between usage data and self-reported preferences in surveys. --- Usage Data versus Self Reports Although the Reuters survey was not conducted on the same exact set of users in our datasets, it was done on a small, yet representative sample of the population in Italy in the same year of our collection. To put both results in perspective and better contextualize our findings, we perform a more thorough comparison in this section. The Reuters survey counts the popularity in terms of number of users, we count it in terms of total visits. To make a fairer comparison, we run different experiments with the aim of estimating the percentages of users that are interested in each category based on our datasets. We consider surfers and households and assume that a surfer (household) "is interested" in a given category if

Most of our knowledge about online news consumption comes from survey-based news market reports, partial usage data from a single editor, or what people publicly share on social networks. This paper complements these sources by presenting the first holistic study of visits across online news outlets that a population uses to read news. We monitor the entire network traffic generated by Internet users in four locations in Italy. Together these users generated 80 million visits to 5.4 million news articles in about one year and a half. This unique view allows us to evaluate how usage data complements existing data sources. We find for instance that only 16% of news visits in our datasets came from online social networks. In addition, the popularity of news categories when considering all visits is quite different from the one when considering only news discovered on social media, or visits to a single major news outlet. Interestingly, a substantial mismatch emerges between self-reported news-category preferences (as measured by Reuters Institute in the same year and same country) and their actual popularity in terms of visits in our datasets. In particular, unlike self-reported preferences expressed by users in surveys that put "Politics", "Science" and "International" as the most appreciated categories, "Tragedies and Weird news"' and "Sport" are by far the most visited. We discuss two possible causes of this mismatch and conjecture that the most plausible reason is the disassociation that may occur between individuals' cognitive values and their cue-triggered attraction. * Work performed when the authors were at Nokia Bell Labs.

being liked in surveys by almost one out of two users, Politics and Science attract less than 4% of visits. Second, none of the existing comparable sources of data can capture precisely the news category popularity at large when considering all visits from all users. Each of these sources yields a different popularity distribution that reflects the peculiarity of the data source. We next further investigate the difference between usage data and self-reported preferences in surveys. --- Usage Data versus Self Reports Although the Reuters survey was not conducted on the same exact set of users in our datasets, it was done on a small, yet representative sample of the population in Italy in the same year of our collection. To put both results in perspective and better contextualize our findings, we perform a more thorough comparison in this section. The Reuters survey counts the popularity in terms of number of users, we count it in terms of total visits. To make a fairer comparison, we run different experiments with the aim of estimating the percentages of users that are interested in each category based on our datasets. We consider surfers and households and assume that a surfer (household) "is interested" in a given category if it has viewed at least k times a news article from that category. We vary the threshold from 2 to 20 visits (in the entire 17 months) and compute each time the fraction of interested surfers (households) in each category, separately for each location and aggregated across locations. 8 For a fair comparison with the survey and to consider a worst case estimation, we obtain the fraction by dividing this number, not by the total number of surfers (households), but by the volume of surfers (households) which have shown interest in at least one category. Intuitively, when the threshold is low, category popularity is inflated, which gives us an upper bound on the number of surfers and households interested in each category. We report the results for two thresholds, for the entire same year as the survey, in Table 4. With k = 2 for surfers and k = 5 for households, we find that at maximum 10% of surfers and 24.08% of households are "interested" in Politics. These drop to 8.55% and 13.7%, respectively, when considering higher thresholds. Overall, unlike the survey report, Tragedies/Weird and Sport are still by far the most "interesting" categories for both surfers and households. Despite our lack of precise user identification, results in Table 4 show a strong mismatch between what users report to prefer and what they actually consume. We further discuss possible reasons in Sec. 8. Finally, to complement the above simulation, we study the distribution of visits per household. Fig. 2 shows the empirical distributions of visits to each category across households. The figure shows that there is a large variation amongst households for all categories. Yet, Politics and Science span much less users compared to Sport and especially Tragedies/Weird news. Takeaway. Unlike their preferences in surveys, users seem to consume news articles from more "catchy" categories such as Tragedies/Weird, Sport and Entertainment, while Politics, Science and Economy are far less popular. This result could reveal a social desirability bias that survey institutes should better account for. Other explanations are also plausible: (i) users actually prefer certain categories but cannot resist the "urge" of clicking on other appealing categories, (ii) users consume what the supply provides. The latter is, however, less plausible. First, there exists categories for which the demand does not match the offer (e.g., People has more articles than visits). Second, editors try to publish what they estimate attractive for users. We will further discuss these possibilities in Sec.8. The percentages of category popularity in our data are representative of the populations we study. Sub-sampling both surfers and households from this population does not change the popularity of categories. However, one question for us is how stable are these results in both space and time. We will explore this question in the next sections. --- Temporal stability We now study how stable is the popularity of categories over different hours of the day and across months. --- Over time of day Fig. 3(a) shows the percentages of visits per news category at various hours of the day. Fig. 3(b) complements it by showing the absolute numbers of visits for each category. Both figures are based on aggregated data, across the 17 months and various locations. We notice that, for the periods with the highest activity, i.e., from 9:00am till 11:00pm, the percentages of each category seem to be overall stable. There are however few curious variations. "Tragedies/Weird" attracts more visits than Sport in both percentages and total number of visits around midday, before leaving the first place to Sport during the afternoon. This trend continues till late at night where Sport reaches its highest share compared to "Tragedies/Weird" (e.g. 19% vs 13% at midnight). This could be due to the fact that sport events usually happen at night. Another noticeable variation is Science which reaches a peak around 6% between 8:00am and 9:00am then lays around 3% the rest of the day. Interestingly, the last two trends are observed in each location separately, including in residential areas, and resist to sampling (across households). Finally, notice the low number of users between 1:00am and 8:00am. Statistics during this period need thus to be interpreted with caution. Nonetheless, various locations curiously exhibited similar trends during this time: a much higher proportion of Sport compared to Tragedies/Weird and a rise of Entertainment. --- Over months Fig. 3(c) shows the percentage of visits to each news category at each month of 2015 in one of the locations. Fig. 3(d) complements it by showing the absolute numbers of visits, as well as the mean number of visits per active surfer during the month, a metric that approximates the activity of users during the month. The major decrease in August is due to holidays and people moving out of the cities. Analyzing the figures, we find that, although being overall stable, the popularity of a couple of categories varies depending on cyclic or special events of the year. For instance, Sport, which has the most noticeable variation, peaks around May and September, which correspond to the end and the beginning of the sports season. Travel and, not shown, Photo Gallery exhibit a peak around the summer vacation. The latter, by manual inspection, is boosted by photos of celebrities during their summer vacation. Entertainment peaks in February, boosted by a famous country-wide festival. Finally, International shows a peak around November 2015, the period of the terrorist attacks in Paris. During this period, the number of visits per surfer slightly increased but the absolute number of visits for all the remaining categories has decreased, suggesting that the Paris attacks event impacted the other categories. Finally, performing the same analysis on a weekly basis yield no unexpected noticeable behavior. Most of the time, the rankings and the overall popularity are comparable to the yearly data. Interestingly, a category like Politics was remarkably stable. During our 17 months of analysis, however, we were not aware of any major or local political event. To further explore this, we split the category popularity on a daily basis looking for peaks in the popularity of politics. We found 3 days in which Politics peaked up to 7% of popularity in all locations (and up to 10%, one of the days, in the campus trace). When further investigating, we found that these days preceded an Italian referendum in 2016, to which almost 16 million voters nation-wide participated. However, these small daily peaks did not have a remarkable influence on the monthly popularity of Politics. Takeaway. Although the exact popularity percentages of certain news categories slightly vary according to the time of day and regular or exceptional events of the year, they are overall stable in terms of ranking. --- Spatial stability We now study the stability of category popularity across our locations and hence also across different demographics since the campus network contains mainly a young population of students. Fig. 4 presents the top-10 categories for each location in terms of percentage of visits they attract. We see that news categories exhibit only slight differences across locations with few exceptions. ISP-City2-a and Campus-City2 both have a slightly higher percentage of Tragedies/Weird. However, these two locations have, in contrast to the two others, a majority of active users during day time, a period of the day where Tragedies/Weird is often slightly more popular regardless of the location as we saw previously. Campus-City2 users read more Sport and less People compared to the rest. Notice that in Campus-City2's university the number of male students is much greater than the number of female ones because the university hosts mainly engineering courses. This fact can explain why Sport is more popular in Campus-City2 than in other datasets. Finally, ISP-City1 has a higher share of Regional news. Takeaway. Despite differences in geographical locations, category popularities seem to be rather stable across our four vantage points. --- Informal discussion The comparison of users' self-reported preferences and actual usage data presented in Sec. 5 shows, for comparable categories, that there is a mismatch between preferred categories and popular ones. This difference is visible for both percentages of visits as well as the penetration in terms of users. Furthermore, the category popularity that we observe is not circumstantial. Indeed, with the exception of few events that slightly change the popularity of some categories, the popularity remains stable over time. Additionally, the popularity of categories is also rather stable across the various locations we study. This important observation leads us to wonder why preferences expressed by users differ from usage data, i.e., users' actual behavior. Unfortunately, our dataset alone can not help answering this question. Hence, we build on prior work to informally discuss two possible causes, which we run by two experts. This section reports these informal discussions. Further research is required to evaluate which of these possible causes explains these differences. The first possible cause is the so called social desirability bias, i.e., the tendency of survey respondents to answer questions in a manner that will be viewed favorably by others. Indeed, people might be "ashamed" of saying that they like "weird" news or news about violence or might have a better image of themselves if they show they are interested in more intellectual categories like Science and Politics. The second possible cause is that this dissociation is a "legitimate" difference between what people want, and hence do (e.g., click on link), and what they actually prefer (e.g., an expression of interest in a given topic). This possible explanation is inspired by the incentive-sensitization theory of addiction (Berridge and Robinson 2016;1998) in neuroscience. Individuals are driven by various rewards (e.g., food, sex, information or surprise). According to this theory, a reward has two components: Like, the pleasure one obtains from it, and want, i.e., how much one desires it, or how much effort one is ready to spend for it. Individuals may develop addiction to a given reward: The addictive behavior or substance hijacks the want system. The want, the motivational part of the reward becomes thus disconnected and higher than the like, the pleasure that the individual gets from the reward.9 Assuming that users' expressed preference coincides with their perceived "pleasure", i.e., their like, then we conjecture that the dissociation we observe might be telling of the addictive nature of certain news categories. In other terms, it could be that people do not like (anymore), for instance sensational news, but ended up being "forced" to click on them in a cue-triggered way, like in addiction. However, this latter explanation could be too simplistic because what is subjectively perceived as preference is probably complex and cannot be reduced to the pleasure that individuals get from a "simple" reward like drugs. Other aspects like self-knowledge or self-image might enter also into play to guide people defining what is valuable. For instance, the need to maximize feelings of self-esteem can lead to the wish of reading noble or serious categories like Science and Politics. To shed more light on this, we contacted Prof. Rasmus Kleis Nielsen, who is one of the lead authors of the Reuters Institute report (Newman, Levy, and Nielsen 2015). We asked for his feedback about our findings and whether he thinks that the mismatch we observe could be due to a social desirability bias. He stated that he is not surprised by the difference between surveys and usage data results. Interestingly, he provided an interpretation of our finding that goes in the direction of the second possible cause, i.e., what people do must not necessarily follow what they prefer. Concerning the response bias, he replied that surveys are designed in such a way to minimize any kind of biases, so he would not support the first hypothesis. Indeed, prior to fielding surveys, a large effort is spent by Reuters and its partners all over the world to test for various issues, including social desirability bias. He also pointed us to work by Prior (2009), which demonstrates that self-reports immensely over-estimate news exposure. We additionally contacted Prof. Kent C. Berridge, one of the two biopsychologists/neuroscientists who formulated the incentive-sensitization theory of addiction. We asked for his feedback about the second possible cause. Prof. Berridge said he thinks our second conjecture is correct: the dissociation between individuals' cognitive values, on one hand, and cuetriggered attraction, on the other, is probably what causes the mismatch we observe. As such, for him, the expressed news preferences reflect people's cognitive judgments about value (one of which is probably influenced by self-image) whereas their actions reflect their motivational-driven choices. Finally, he further added that this is "probably a little different" from what happens in the mesolimbic dopamine system in the brain that causes a dissociation between want and like for the same thing, as it happens when engaging with addictive behavior. Finally, it is worth noting, that regardless of the reasons, the popularity of "Tragedies/Weird" news in our datasets seems in line with the history of news consumption where sensational news have been the most popular categories since ancient times (Stephens 2006). --- Conclusion In this paper, we analyzed 80 million news visits to 5.4 million news articles, extracted from a dataset of 17 month long anonymized HTTP traces. We focused on the popularity of news categories in this dataset and put it in perspective with the category popularity in four other sources: what is discovered on Facebook, what is observed by major news outlets, what is published, and finally user preferences expressed in surveys. The results of our analysis show that news consumed in social networks represent a small fraction of the overall news consumption and that none of the existing sources can faithfully capture news-category consumption at large. Interestingly, our results further demonstrate that the analysis of usage data complements survey-based data with new insights. Indeed, by putting survey-based results in perspective with usage data, we observed a clear mismatch between news categories which users claim to prefer and those they actually visit. We conjectured that this can be explained by a disassociation between individuals' cognitive values and their cue-triggered attraction.

Most of our knowledge about online news consumption comes from survey-based news market reports, partial usage data from a single editor, or what people publicly share on social networks. This paper complements these sources by presenting the first holistic study of visits across online news outlets that a population uses to read news. We monitor the entire network traffic generated by Internet users in four locations in Italy. Together these users generated 80 million visits to 5.4 million news articles in about one year and a half. This unique view allows us to evaluate how usage data complements existing data sources. We find for instance that only 16% of news visits in our datasets came from online social networks. In addition, the popularity of news categories when considering all visits is quite different from the one when considering only news discovered on social media, or visits to a single major news outlet. Interestingly, a substantial mismatch emerges between self-reported news-category preferences (as measured by Reuters Institute in the same year and same country) and their actual popularity in terms of visits in our datasets. In particular, unlike self-reported preferences expressed by users in surveys that put "Politics", "Science" and "International" as the most appreciated categories, "Tragedies and Weird news"' and "Sport" are by far the most visited. We discuss two possible causes of this mismatch and conjecture that the most plausible reason is the disassociation that may occur between individuals' cognitive values and their cue-triggered attraction. * Work performed when the authors were at Nokia Bell Labs.

Background Obesity and chronic disease behavior management can be especially challenging in rural areas in the Deep South. Nearly half of the rural population in the United States (U.S) resides in the South [1], and in states such as Mississippi where a large percentage (79%) of counties are classified as rural [1], the percentage of obese residents (37.3%) is among the highest in the region [2]. Further, the state of Mississippi has greater Black population density (37.8%) compared with other Deep South states including Alabama (26.8%), Louisiana (32.6%), and Georgia (32.2%) [3] as well as a stark difference between the proportion of Black (29%) and White (10%) populations living below poverty levels [4]. Mississippi has a history of being medically underserved with a majority of the state designated as a primary care provider health professional shortage area [5]. Therefore, it is difficult to reach rural and remote populations with few available resources to support health education and management. Community-and church-based programs have been effective in improving health outcomes and reaching minority and underserved populations. However, greater program attendance is key for better health outcomes as shown in previous church-based intervention research [6,7]. Additionally, various types of social support, including general, religious and church social and instrumental support, have been associated with improved diet and physical activity behaviors among minority and rural communities [8,9]. Social support may be crucial for engaging minorities in healthy behaviors compared to Caucasian counterparts [10] and technology may facilitate the support of health program participants. The supportive policies for the provision of telehealth/-medicine, including coverage and reimbursement, point toward the potential for Mississippi to be a model state for the use of technology in health behavior management programs [11]. Thus, the Church Bridge Project was developed and tested over a two-phase intervention among African Americans living in southern Mississippi [12]. This research was intended to support the need for adoption and implementation of weight management interventions, focusing on diet and physical activity education and behavior change, among minority communities in the South. The purpose of this study was to examine program acceptability, barriers, and facilitators of participation in a church-based weight management intervention among young to middle-aged adult African Americans in southern Mississippi. Additionally, this research explored perception about weight management programs and the use of technology in weight management programs. --- Methods Focus group research was selected as the methodology for this study to gain collective information and rich understanding from a population that was reached (and either participated in the program or not) to enroll in a church-based weight management program that was not captured by our quantitative outcomes [13]. Focus groups were conducted after conclusion of the Church Bridge Project weight management intervention that was implemented in two churches from February 2017 to September 2017 [14]. The weight management intervention program included 12, 1-h group-based sessions that entailed a 20-min education component, 20-min motivational interviewing component and 15-to 20-min snack and social component. Eligible focus group participants were those determined at intervention enrollment to be overweight/obese based on BMI and self-identified as being African American between the ages of 18 and 50 years of age. The participants were recruited from a referral and enrollment database of potential participants developed for the intervention study [14]. The two types of participants recruited for the focus groups were classified as intervention "completers" and "non-participants." "Completers" were defined as individuals who successfully completed the intervention program. "Nonparticipants" were defined as individuals eligible for the intervention, approached to participate and enrolled in the intervention but did not participate in the intervention sessions. Two focus groups were conducted for "completers" and for "non-participants" with a goal to recruit six to eight participants each. This sample size goal is in alignment with widely accepted sample size recommendations [15]. Potential participants for each focus group were contacted via telephone, invited to participate and given a verbal overview of the focus group purpose and processes. All study protocol and materials were approved by The University of Southern Mississippi Institutional Review Board. This study follows the guidelines set forth in the Belmont Report which require that human research subjects, a) will not participate in research unless and until they have given voluntary and informed consent, b) confidential information received from participants will be fully protected within the limits of the law, both during and after research is conducted, d) subjects may withdraw their participation at any time without penalty or loss of benefits to which they would otherwise be entitled, e) burdens put on research participants must be reasonable relative to anticipated benefits to themselves and to society as a whole, f) the selection of research participants must be equitable and defensible in terms of both the goals of the research and general considerations of fairness. Upon arrival, focus group participants were assigned a random number to maintain confidentiality. Research staff reviewed the study consent information and obtained consent to participate after all questions were answered. The focus groups were conducted by two trained research assistants with one serving as a moderator and the second as a note-taker. A script was provided that served as a guide for the focus groups, which included study overview and procedures related to how the focus group was to be conducted and recorded. After providing informed consent, participants were informed that a digital voice recorder would record the conversation until the end of the session. Each focus group lasted approximately 1 h. The "completer" focus group included open-ended questions related to barriers and facilitators to participation in a weight management intervention and suggestions for program improvement, whereas the "non-participant" focus group included questions related to weight management program perceptions as well as barriers and facilitators to participation and the use of technology in a weight management intervention. Focus group participants were given a $25 gift certificate to a superstore at the completion of the session. Our qualitative research process was guided by the four tenets of trustworthiness: credibility, confirmability, transferability, and dependability [16]. After the focus groups were completed, recordings were transcribed by one research assistant and reviewed for accuracy by a second. Data analysis was conducted using a series of steps [17] based on Braun and Clarke's six-step framework for thematic content analysis [18]. The initial transcription and review of the data by two staff was considered to be first step to "become familiar with the data." All notes or initial thoughts regarding the data were noted and recorded by each research staff. The next step involved the generation of initial codes which followed a theoretical and open coding approach. This included developed and modified codes assigned to pieces of text from the data that were related to the research question and interesting. The third step was to examine the initial codes and search for common themes across the codes that identified something significant about the data. These initial steps described were conducted independently by two staff. In the next step, the two staff met to compare and review the themes each one developed, discuss common and divergent themes, and review whether data supported the proposed themes and if there was any overlap between themes. Staff developed a final, refined single list of themes which were subsequently reviewed by the larger team of staff and investigators. Research staff and investigators together reviewed and discussed the draft themes and developed the final list of themes to include in the final results. As data were analyzed, we also considered saturation of the data and defined saturation based on theoretical saturation or the emergence of new themes in the data [19]. Theoretical saturation for this project was "the point where no new codes are emerging in the data [19]." In alignment with grounded theory development guidance [19,20], we also considered saturation as the justification for termination of analysis. In consideration of data saturation, we also considered prior theoretical work [14]. --- Results A total of 12 individuals (n = 6 "completers" and n = 6 "nonparticipants") participated in two focus groups. Themes were identified based on barriers and facilitators to and suggestions for improvement of the weight management program as well as diet values and weight management program perceptions. Table 1 provides integrated results of main themes identified. --- Barriers identified by "completers" Two themes for weight management program barriers were identified by "completers": 1) technology learning curve, and 2) competing priorities. Participants found it difficult to locate specific food items or activities in the mobile app to log their nutrition and physical activity behaviors. However, the technological barrier was viewed as something that could be overcome with time with one person stating "... but I got into how to do it and stuff, and I learned how to do it so it wasn't too bad..." It was agreed by all participants that the main barriers When asked what would motivate them to participate in a nutrition and physical activity program, limited cost and convenience were identified. Additional motivators identified were health benefits to include pre-existing health conditions and overall health status. Pre-existing conditions were chronic diseases, including high blood pressure and diabetes. Overall health status included both mental and physical health. One participant explained: "Right now, what's motivating me is my health. The second thing motivating me is loss of self, because I look in the mirror and I go, 'Who is that person?' I don't recognize her because I've always been smaller my entire life. It's just the last 5 or 10 years that I've slowly continued to gain weight." "Non-participants" expressed that if they were able to participate in a program, they would hope to gain program support as accountability to achieve health goals. "Non-participants" also voiced technology could promote participation by providing support via being motivational, assisting with coaching, and providing reminders. Others expressed concerns with feeling burdened by needing to be consistent with using technology. Table 2 includes themes and sample coded text. --- Discussion This study examined perceptions about a weight management program delivered in a church-based setting that were useful for identifying factors related to the weight management intervention design and implementation. The themes gleaned from the focus groups suggested that the participants of the intervention, the "completers", had positive perceptions about the program. However, use of technology seemed to be both a barrier and facilitator of program success. Competing priorities were discussed as barriers to weight management program participation among both "completers" and "non-participants." Another theme involved the negative connotation associated with the "weight management program" terminology. Suggestions included framing the wording as a nutrition and physical activity program. Social support and personal health were also identified as key factors for the success of a weight management program. The results of the focus groups indicate that careful considerations are needed when incorporating technology into health intervention research. Mobile technology was viewed as a way to facilitate motivation, session attendance, and social support; however, a predominant theme among "completers" was that the food log function of the mobile application was difficult to use. A review of articles about dietary assessment using mobile phones found that six of the seven included studies reviewed had high participant satisfaction regarding use of mobile phones for dietary assessment [21]. The results of two additional studies also support the general acceptance of mobile food record methods among adults, including a community sample [22,23]. Prior research found that higher user satisfaction was associated with more accurate dietary intake reporting [24]; however, it --- Longevity and Improved Health "The better you eat, the better you feel, and the longer you'll be around for your kids. I have high blood pressure, but I'm on two different blood pressure medications and that was a wake-up call for me. I'm gonna get fit, I'm gonna eat right, or I'm gonna die. So, having nutrition and diet is everything." Perceptions of "Weight Management Program" --- Overwhelm and Defeat "Well for me, when I hear weight management, I think about all those negatives. I can no longer have, I feel deprived. I go to the negative part, instead of thinking about the positive partwhat it's going to do for me, the health and the finding myself again and all those different things. So, I don't think about the positive. It's all negative." Perceptions of "Nutrition and Physical Activity Program" --- Positive and Approachable "It's a little better than weight management because to me that sounds like exercise and eating healthy, the two... So when I look at the portions and eating healthy and exercise, the two, I think of that term." was not related to perceived participant burden to remember and log food intake [23]. Additionally, Krebs & Duncan [25] found that 44.5% of surveyed users discontinued using a health-related app because it took too long to enter data. Thus, while the use mobile applications for dietary assessment may be perceived as a satisfactory mode of data collection for participants, it may not reduce the perceived burden of data collection, which has data integrity implications. While mobile technology may not reduce data collection burden, the supplementation of behavior interventions with technology may assist with maintaining program intensity/contact and providing individualized participant support without increasing participant burden. A review indicated that technology-based weight loss interventions induced positive weight-related outcomes, enhanced social support and self-monitoring opportunities, and improvement in program adherence [26]. Our previous research as well as conclusions from another study [27] examining health-related virtual communities support the hypothesis that technology, by increasing convenience and access, can enhance perceived support, thus improving outcomes. As we found, simple strategies such as providing participants live-streamed access to sessions when they cannot physically attend may increase program support and maintain participant commitment to the program. Since previous churchbased intervention research studies have concluded that greater program attendance positively influences health outcomes [6,7], an important theme was that competing priorities serve as major barriers of program attendance by young to middle aged African American adults. Further research is needed to ascertain from the population what strategies might assist them with circumventing program participation barriers. Various types of social support (general, religious and church social and instrumental support) have been associated with improved diet and physical activity behaviors among minority and rural communities [8,9] and is crucial for engaging the population in healthy behaviors [10]. Preliminary work also identified that family support was associated with weight loss intentions among an African American population. A similar theme found in this study was that the "church family" was a motivating factor for program participants. However, research examining potentially beneficial health effects of perceived social support from church and group-based interventions is minimal. An additional theme, disease management and prevention objectives were identified as motivating factors of weight management program participation. It was interesting to note that although "non-participants" voiced a common theme of the importance of participating in weight management programs for health and longevity, they were not sufficiently motivated to actually participate in the program when offered. The relevant themes from the focus group data and previous literature support the importance of shifting the conversation about obesity from body ideals towards health in a culturally appropriate manner [12]. Challenges associated with this goal include perceptions among African Americans that obesity in itself is not an indicator of poor health [28], health is independent of obesity status [29], and "bigger is healthier" [30]. Other African American focus group participants have also viewed body mass index negatively [31] which corroborates our results of African Americans' preference to frame "weight management interventions" as "nutrition and physical activity interventions" to enhance healthy lifestyle determinants and diminish the focus on weight. This suggestion is supported by currently published concepts that individuals should focus on being healthy and not achieving a certain body weight ideal [31]. --- Limitations and strengths In alignment with the four tenets of trustworthiness of qualitive inquiry [16], our results were credible, transferable, dependable, and ultimately, confirmable. Credibility addresses the goodness of fit between the participants' ideas and the researchers' interpretation of those ideas. We used a multi-step approach that included research triangulation, or multiple observers, at various steps to define and confirm themes that represent the data and reduce potential biases. Code and theme development was also documented at every step and examined by multiple researchers during the refinement and finalization of themes. Transferability refers to the generalizability of the knowledge gained from the research. This study underpins the importance of addressing weight management from a disease prevention and preventive health perspective. Additionally, churchbased programs, compared to traditional medical settings, may be advantageously equipped with social support to facilitate diet, physical activity and other health behavior change goals. While our results are limited to an African American population in rural, Mississippi and a small sample size, these findings are not completely isolated in the literature and may have implications for other populations, especially across the Deep South. We also only conducted one focus group for each group type. Our research staff felt that they had saturated the pool of potential participants with recruitment efforts and did not feel further recruitment would yield enough participants for an additional focus group. Our sample does uniquely represent a predominantly younger to middle aged African American population, which will be a crucial target for health education and behavior programs to go beyond the management of disease toward the reduction of disease risk and ultimately, health disparities. Last, our research is both dependable and confirmable. Dependability refers to the research process in that it is clear and reproducible. Our steps are documented and in alignment with best practice guidelines for qualitative inquiry. Confirmability requires demonstration of how conclusions were reached which is met when credibility, transferability and dependability are achieved. Our study outlines a clear qualitative process for how themes were derived with multiple checkpoints and results have clear implications for weight management in a specified population, which may have implications for and guide research in other populations. --- Conclusion The findings from this study may help improve the development of lifestyle interventions intended for implementation among underserved, rural populations. The language used to describe weight management programs may deter participants and should consider a focus on the key factors identified by the population of interest (ie. improving nutrition and physical activity behaviors to achieve weight loss) versus the weight loss itself. Technology may serve to alleviate certain barriers and introduce others. Careful consideration of the population to be served and understanding technology related barriers is recommended to ensure that the pros outweigh the cons of use. --- Authors' contributions JL and PA were informed the design of the study and provided study oversight as well as were major contributors in forming the manuscript. RG, AB, and LB facilitated focus group implementation and analysis and were major contributors in informing the manuscript. LA and TG provided external review of the study and were major contributor in informing the manuscript. All authors read and approved the final manuscript. --- Funding Research reported in this publication was supported by the National Institute on Minority Health and Health Disparities of the National Institutes of Health (NIH) under Award Number R15MD010213 and Mississippi INBRE, funded by an Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the NIH under grant # P20GM103476. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. We would like to acknowledge other grant staff, intervention participants and community partners for facilitating this research. The authors declare no conflict of interest. --- Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. --- Declarations Ethics approval and consent to participate All study protocol and materials, including consent information, were approved by The University of Southern Mississippi Institutional Review Board. The USM IRB complies with federal regulations, including Federal Policy for the Protection of Human Subjects and operates under a Federalwide Assurance (FWA) on file with the Office for Human Research Protection (OHRP, Registration number: FWA00002393) in the U.S. Department of Health and Human Services. Human subjects research conducted at or supported by the University honors the three ethical principles enshrined in the "Belmont Report": respect for persons, beneficence, and justice. These general principles require that human research subjects, a) will not participate in research unless and until they have given voluntary and informed consent, b) persons under the age of 18, legal incompetents, and those unable to consent may participate in research only with consent from a parent or legal guardian and, when possible, their assent, c) confidential information received from participants will be fully protected within the limits of the law, both during and after research is conducted, d) subjects may withdraw their participation at any time without penalty or loss of benefits to which they would otherwise be entitled, e) burdens put on research participants must be reasonable relative to anticipated benefits to themselves and to society as a whole, f) the selection of research participants must be equitable and defensible in terms of both the goals of the research and general considerations of fairness. --- Consent for publication Not applicable. --- Competing interests Authors report no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The prevalence of obesity is disproportionately high among African Americans in the Southern US. More information is needed about factors that influence participation in nutrition and physical activity programs to promote healthy weight. Objective: The purpose of this study is to explore the weight management perceptions of young to middle aged adult African Americans.The Church Bridge Project intervention participants were recruited for two focus groups. Qualitative data were recorded, transcribed and a thematic content analysis was conducted to identify major themes. Results: Barriers included technology learning curve/burden and competing priorities. Facilitators included support, limited cost, convenience, and health. Participants perceived the term "weight management" program as overwhelming and defeating.The Church Bridge Project model confirmed social support and disease prevention as key factors for weight management. Further work should substantiate social support as a key factor to guide minority health efforts.

Introduction The relationship between low socioeconomic status (SES) and poor health outcomes has been extensively documented through empirical research [1]. The nature of this relationship is, however, highly complex and the precise mechanisms through which SES translates into health disadvantages (for low SES persons) or advantages (for higher SES persons) are not fully understood [2,3]. One problem is that one's adult health status may reflect SES-related risk or protective factors which are cumulative in terms of their effects [4][5][6]. Thus point-intime indicators of adult SES such as income and educational attainment may not adequately reflect the influences shaped by one's SES during childhood, which either persisted or changed during the life course [7,8]. Other influences associated with SES may be difficult to quantify, including homeownership, indebtedness, family wealth, and dynamics operating at the community level such as one's network of social relationships [3,[9][10][11]. For the above-mentioned reasons, conventional indicators of SES (which typically refer to income, occupation, and/or educational attainment) may have poor construct validity, highlighting the need for a more critical and careful consideration of how SES is defined and measured in the context of health research [2,[12][13][14]. Some studies have demonstrated a significant relationship between SES and adult health when measures of SES are based on perceived social ranking, and that one's perceived status may provide a better conceptual/ measurement framework for exploring the associations between adult SES and health outcomes [9,10,15]. This may be, in part, because subjective measures are arguably 'global' in nature and account for one's understanding of the stability of their social status, as well as for the prospects of social mobility, while 'point-in-time' measures, for example, do not. [10,15] Consistent with findings from adult and child health research, low maternal SES has been shown to be persistently and pervasively related to adverse pregnancy and birth outcomes [16][17][18]. However, we are aware of only one published study that examined the relationship between SES and maternal health, defining SES in terms of a woman's perceived or subjective social status (SSS) [19]. In that study women who perceived their status as relatively low were found to be in relatively poor health, independent of their income and educational attainment. Moreover, among some ethnic groups, perceived status was reported to be more strongly related to maternal health than income and education. The study was, however, limited to only one measure of maternal health-namely, a woman's overall selfrated health status during pregnancy. The purpose of our investigation was [1] to assess the extent to which subjective social status is related to a wide variety of maternal health-related outcomes; and [2] to compare and contrast the magnitude of the relationship between subjective social status and maternal health outcomes to conventional measures of SES and maternal outcomes. --- Methods This research was part of a larger, prospective, community-based study examining maternal stress, birth outcomes, and maternal and infant health and health-related behaviors. The study involved recruitment of women who first enrolled for prenatal care at a consortium of six community health centers in Philadelphia, Pennsylvania, between February 2000 and November 2002. It was funded by the Centers for Disease Control and the National Institute for Child Health and Human Development and review for the ethical treatment and study of human subjects was approved by the institutional review boards (IRBs) at Thomas Jefferson University, Drexel University and the University of Pennsylvania. Subject consent was obtained through signed consent forms. Additional findings from, and more details about, the larger study are available elsewhere [20,21]. Women were recruited and interviewed at the time of prenatal care enrollment, and scheduled for three follow-up surveys in the subjects' homes, at approximately 3 months postpartum, 11 months postpartum, and at 24 months postpartum. Only information from the prenatal, 3 month, and 11 month postpartum interviews was used for the purposes of this study. Women with a singleton intrauterine pregnancy who spoke either English or Spanish were considered eligible for enrollment; among those, women with pregnancies ending in a live birth were designated for postpartum follow-up (n = 2,026). More than 98% (n = 1,984) agreed to participate in the follow-up phase of the study, and attempts to complete both the 3 month and 11 month postpartum surveys were successful for 67% (n = 1,322) of those who did agree. Data from the prenatal interviews indicated that women who did not complete the first two postpartum interviews (n = 562) did not differ from those who did with respect to income, education, marital status, or parity (data not shown). All interviews were based on structured surveys, conducted in English or Spanish by trained female interviewers. To minimize the chances of interviewer bias or misinformation each interviewer received approximately 30 h of training in use of this instrument, including instruction on when and when not to prompt women for responses, and how to explain the meaning of a question should any misunderstanding arise. The training included role playing and each interviewer was assessed and rated as competent by an experienced project supervisor before assignment to the field. For quality control purposes, a ten percent random sample consisting of 10% of all interviews from each interviewer were reviewed. Weekly meetings were held with the interviewers, project staff, and principal investigator (Dr. Jennifer Culhane) to discuss any issues related to the interviewing process. --- Study Variables Preliminary analyses of the data indicated that t he relationships between the health outcomes of interest and three primary measures of social status (SSS, income and education) were ordinal and non-linear in nature. For that reason women were ranked as either being of 'low'','medium' or 'high' for each variable, as described below. --- Subjective Social Status The surveys conducted at 11 months postpartum included the MacArthur Scale of Subjective Social Status, a measure which has been shown to be a valid and reliable indicator of one's perceived status in society [22]. Specifically women were presented with a "social ladder" with 10 rungs and were asked to rank themselves by placing an "X" on the step they felt represented their relative position in society at large. Each rung was numbered and the women were read the following statement: "In our society, at the top, are those who have the most money, education and best jobs; at the bottom are those that have the worse jobs or no jobs. Where do you think you fit in?" Their self-rank on this ladder, ranging from 1-10 was used as the basis for our measure of subjective social status (SSS). Women with a ladder score of 0 through 3 were classified as being in the 'low' SSS group; those with a score of 4 through 7 as being in the'medium' SSS group; and those with a score of 8 through 10 as being in the 'high' SSS group. The distribution of SSS was approximately normal with a mean value of 5.9 and standard deviation of 1.85. The abovementioned cutoffs for 'low' and 'high' groups represent scores that were approximately 1.5 standard deviations below the mean and 1.5 standard deviations above the mean of 5.9, respectively. All others between 1.5 standard deviations above and below the mean were classified as being in the'medium' SSS group. --- Objective Indicators of Social Status Several measures of the respondents' objective indicators of socioeconomic status were used in the analyses. The major indicators were educational attainment and personal income. Income was determined based on response to the question in the 11 month postpartum survey, which appeared as follows: "What is the total amount of money you received from all sources during a typical month?" Respondents were classified based on quartile rankings. Specifically, those in the first quartile of the income distribution for the sample (<unk>$500 per month) were categorized as being the low income group; those in the second or third quartiles were categorized as the'medium' income group ($500-$1200 per month), and those in the highest quartile ([>$1200 per month) were categorized in the 'high' income group. Educational attainment was measured in number of years of education, collected at the time of the initial interview along with other descriptive sociodemographic characteristics such as age, race/ethnicity, and parity. Women were grouped into three categories in terms of their education attainment: low (did not complete high school), medium (completed high school) and high (some college or more). Other variables representing measures of objective social standing that were included in the analyses were public assistance status (on food stamps or state cash welfare payments), homelessness (ever experiencing homelessness since childbirth) and ever having an involuntary utility (water, gas or electric) shut-off. --- Health Outcomes Several measures related to both the mental and physical health of respondents were available from one or both of the surveys and used as the dependent variables in the analyses. They included overall self rated physical and emotional health during pregnancy (reported by respondents at the 3 month postpartum interview) and overall physical and emotional health in the postpartum period (reported by respondents at the 11 month postpartum interview). All self reported measures of physical and emotional health were dichotomized, with respondents classified as being in relatively good (reporting 'excellent','very good', or 'good') as opposed to relatively poor (reporting only 'fair' or 'poor') health. Other measures of emotional health during the postpartum period used in the analyses included depressive symptomatology and perceived stress (reported at the 11 month interview). Depressive symptomatology was defined as a score of greater than or equal to 23 on the Center for Epidemiological Studies Depression Scale (CES-D). The CES-D is a pre-tested reliable and valid instrument used widely in studies of depression, including postpartum depression; scores of greater than or equal to 23 correspond to the 90th percentile in community samples and has been used by other researchers to define the presence of depressive symptoms in pregnant and postpartum women [23,24]. Perceived stress was based on the administration of the Cohen Perceived Stress Scale (PSS). The PSS possesses strong psychometric properties, including reliability across gender, socioeconomic status, age groups, race/ ethnicity, and other demographic characteristics [25]. This 14 item self report scale measures the degree to which a respondent appraises her life as being stressful. Women scoring below the 75th percentile (<unk>28) were defined as having relatively low stress while those equal to and above that were classified as the relatively "high stress" group. In addition, women were classified in terms of whether or not they had one or more major health problems (heart disease, asthma, diabetes, or high blood pressure) and whether or not they had ever been told by a health professional that their pregnancy was 'high risk.' Additional data describing the health outcomes used as the independent variables in the analyis are available from the authors upon request. --- Statistical Analyses We first used simple cross tabulation techniques to examine and compare the relationships between all health outcomes and SSS, education, and income. Goodman and Kruskal's gamma was used to determine statistical significance, magnitude and direction of all relationships. Gamma is based on the correspondence among the rank ordering of observations for any pair of variables ordinal in nature, and is appropriate whenever the relationships are likely or purported to be ordinal, as is the case here. Gamma values are themselves both measures of the degree and strength of the ordinal relationships and are normally distributed for large sample sizes, and thus the calculation of P values is straightforward [26]. Logistic regression (LR) was then used to explore the effects of SSS vs. income, education, and other measures of SES described above on the health outcome variables. Specifically, a series of LR models were created with SSS, and additional SES indicators were entered in a stepwise fashion. In this way, the strength and robustness of the effects of SSS on health, independent of and in contrast to other measures could be quantified by comparing adjusted odds ratios. Confidence intervals associated with all effect parameters were also calculated and are presented in the tables and figures that follow. All analyses were completed using Stata 10.0 [27]. --- Results The sociodemographic and health characteristics of the study population are shown in Table 1. African-Americans comprised 70.6%, Whites 9.8%, and Hispanics 16.5% of the 1,322 women for whom the three surveys were available. The age range for this cohort was 13-43 years, with a mean age of 23.9 years. Almost 40% (39.3%) of the women did not graduate from high school, while 17.7% received at least some post high school education. Almost 65% (64.8%) of the women reported a total income of less than $1200 per month. More than 90% (93.6%) of the women reported receiving public assistance. More than 5% (5.2%) indicated that they had experienced a utility shutoff and over 2% (2.6%) reported being homeless at some point since their last childbirth. The distribution of the Subjective Social Status scale indicated that the vast majority of respondents perceived themselves to be average, despite median incomes ($800/month) well below that for the U.S. as a whole. Specifically, more than 7 out of ten (71.9%) placed themselves close to the middle (rung scores from 4 to 7) of the subjective status ladder. The relationships between subjective social status, income, and education with each of the health outcome variables are provided in Table 2. In general the associations between SSS and health are strong and persistent. Relatively poor overall physical and emotional health during pregnancy, the presence of major health problems, perceived stress and depressive symptomatology are all significantly and inversely related to SSS. For example, compared to women who rated themselves high on the SSS scale, those who rated themselves in the lowest category were more than twice as likely to report their overall physical health during pregnancy (24.6 vs. 11.4%), and their emotional health in the postpartum period (31.4 vs. 14.1%) to be relatively poor. Consistent with these overall self-reported items, women in the lowest compared to highest SSS group were more likely to have indicated one or more major health problems (45.6 vs. 30.8%), far more likely to be in the high stress category (50.4 vs. 20.9%) and to have depressive symptomatology (32.9 vs. 15.4%). Income and education were inversely related to some of the health outcomes, although not as strongly or as consistently as was the case for SSS. Physical health during pregnancy was unrelated to either income or education, while overall physical health was related to education but not to income. Overall emotional health in the postpartum period was related to both income and education, but overall emotional health during pregnancy was related only to income. As was the case for SSS, perceived stress and depressive symptomatology were related to both income and education, although the relationships were somewhat stronger for SSS. The results from the logistic regression analyses are presented in Tables 3 and4. Models A, B and C show the unadjusted odds ratios for SSS, Income and Education, respectively. Model D presents the odds ratios adjusted for income and education; and in Model E the odds ratios are adjusted for all additional measures of social status. Presenting the results in this manner allows us to compare the relative strength and robustness of the relationship between SSS and high stress (Table 3) and self-rated poor physical health (Table 4), when all other measures of SES are taken into account. As we can see in Table 3, the relationship between SSS and high stress is strong and persistent even after accounting for objective SES measures. Women in the low SSS category were far more likely than women in the highest category to have high stress levels, even after controlling for income and education (OR 3.66; 95% CI 2.25, 5.95). Even after adjusting for all other SES measures those in the low SSS category were still far more likely than those in the highest category to have high stress levels (OR 3.24; 95% CI 1.98, 5.30). Note that after adjusting for SSS, education is unrelated to high stress and income is only weakly related. Women in the lowest income group were somewhat more likely to have high stress levels compared to women with the most income after adjusting for SSS (OR 1.87; 95% CI 1.31, 2.68), but the effect size is considerably smaller than that for SSS. The relationship between SSS and self-rated physical health was also persistent even after accounting for objective SES measures (Table 4). Specifically, women in the low SSS category were far more likely than women in the highest category to rate their overall postpartum physical health as poor, even after controlling education and income (OR 2.15; 95% CI 1.16, 3.99). After adjusting for all other SES measures those in the low SSS category were still far more likely than those in the highest category to have relatively poor postpartum physical health (OR 2.04; 95% CI 1.09, 3.82). Note that after adjusting for SSS, income and education were unrelated to poor physical health. A summary of results for additional health outcomes is reported in Fig. 1. Adjusted odds ratios and 95% confidence intervals associated with low and medium compared to high SSS are presented, in each case after adjusting for income, education and all three other objective SES measures. The results show that for every other outcome, women in the low SSS category were significantly more likely to have worse health outcomes, even after adjusting for SES measures. Women in the low SSS category, for example, were 1.95 times more likely than women in the high SSS category to have depressive symptomatology (OR 1.95; 95% CI 1.1.12, 3.38) (Fig. 1a), as well as two times more likely to report being in relatively poor physical health in pregnancy (OR 2.53; 95% CI 1.34, 4.80) (Fig. 1b). --- Discussion The analyses of the relationship between our measure of subjective social status (SSS) and several emotional and physical health outcomes revealed the following: [1] SSS was related to all health outcomes; [2] this relationship exists independently of the influence of more conventional or objective measures of socioeconomic status; and [3] the relationship between SSS and health appears to be more robust than that for objective measures of socioeconomic status. Our findings are consistent with those from other studies of adolescent, middle-aged, and older populations [28][29][30] demonstrating a significant relationship between measures of subjective social status and overall physical health, independent of measures of objective social status. They also lend support to the notion that, in some cases, measures of SES based on subjective indicators may in fact be more sensitive than those based on objective indicators [3]. Measuring subjective social status may be particularly important, as is the case here, where the study population is relatively poor. Income levels, for example, may be relatively restricted in terms of the range of values in low income populations. Similarly, subtle but important differences along other dimensions of status-such as savings/wealth, the quality of education received, and access to social and community-level resources-may be overlooked. SES-related influences on health behaviors or outcomes may, therefore, go undetected. This highlights the concern raised by other observers, of the need to carefully consider how SES is defined and measured in the context of health research, and how problems associated with concept validity may have important implications for the interpretation of empirical results [2,3]. A great deal of empirical health research in the U.S., for example, is designed to assess racial/ethnic differences in health behavior and outcomes ''independent'' of SES differences. To a large extent, however, the integrity and the interpretation of these findings hinges on an adequate and appropriate measurement of SES itself [2,31,32]. In summary, our findings suggest that subjective social status may be a critical assessment tool for the measurement of socioeconomic status in low income populations. --- Study Limitations The findings are of course limited with regard to both the characteristics of the sample and the study design. As noted earlier the study sample was derived from data collected as part of a larger, longitudinal investigation of the relationships between maternal stress, prenatal and postpartum health and infant birth outcomes. The protocol for that larger study involved the administration of the SSS Scale at 11 months postpartum. At least in some cases SSS was related here to health measures collected at earlier points in time. As a result the extent of the causal nature of relationships between socioeconomic status, including those as measured by SSS, was beyond the scope of this paper to address. Recruitment for the larger study occurred at prenatal sites in very low-income urban neighborhoods and yielded a study cohort with a medium income of less than $10,000 per year. The extent to which the findings reported here can be generalized to poorer, or nonimpoverished populations is unclear; as a result further research may be warranted in order to establish the generalizability of our findings. Finally, it is worth noting that the criteria used to categorize the original SSS ladder scores into groups resulted in a relatively small percentage (10.3%, n = 137) of women being categorized as 'low'. Since it was possible that the findings may, therefore, have been partly a statistical artifact of an 'outlier', low SSS group, we replicated the analyses which included women with a ladder score of '4' in the low as opposed to medium SSS group. The pattern of results did not change, and the magnitude of the differences between the low, medium and high SSS groups for all the health outcomes was virtually identical to that reported here. --- Research Box --- What is already known about this topic Conventional socio-economic status indicators such as income and education may have poor construct validity in health outcomes research, especially in homogenous lowincome populations. --- What this study adds This study has strengthened the evidence that subjective self-assessment of one's social standing should be considered in any empirical study that explores relationships between socioeconomic status and maternal health outcomes. The relationship of subjective social status on health outcomes. a. Adjusted odds ratios: summary of effects of SSS on measures of emotional health; b. Adjusted odds ratios: effects of SSS on measures of physical health Relationship between heath outcomes and subjective social status (SSS), income, and education (GAMMA = --- Table 3 Logistic regression results: relationships between SSS, income, education, selected measures of impoverishment and high stress levels a

Appropriate measurement of socioeconomic status (SES) in health research can be problematic. Conventional SES measures based on 'objective' indicators such as income, education, or occupation may have questionable validity in certain populations. The objective of this investigation was to determine if a relatively new measurement of SES, subjective social status (SSS), was more consistently and strongly associated with multiple health outcomes for low income mothers. Data available from a large scale community-based study examining maternal and infant health for a low income urban population were used to examine relationships between SSS and a wide range of postpartum physical and emotional health outcomes. Crosstabulations and multivariate analyses focused on the breadth and depth of these relationships; in addition, the relative strength of the relationships between SSS and the health outcomes was compared to that of conventional measures of SES, including both income and education. SSS was significantly related to all physical and emotional health outcomes examined. The overall pattern of findings indicated that these relationships were independent of, as well as more consistent and stronger than, those between conventional measures of SES and postpartum health outcomes. SSS represents an important dimension of the relationship between SES and postpartum physical and emotional health. In low income populations the failure to account for this dimension likely underestimates the influence of SES on postpartum health. This has important implications for the

Background High-risk injection practices, such as receptive injection equipment sharing (i.e., injecting with syringes, cookers, rinse water that were previously used by another person), play a central role in the transmission of infectious diseases (e.g., HIV, viral hepatitis) among people who inject drugs (PWID) [1][2][3][4]. In the United States (US), there are an estimated 750,000 people who injected drugs in the past year [5]. Studies have found that the prevalence of receptive injection equipment sharing among PWID varies across the United States and has been associated with infectious disease outbreaks [6][7][8][9][10][11][12][13][14][15][16][17][18][19]. For example, a study conducted in Baltimore City (Maryland) found that 16% of PWID reported having engaged in receptive syringe sharing in the past month [7]. Another study conducted among PWID in a rural county in West Virginia found that 43% reported engaging in receptive syringe sharing in the past 6 months [9]. Similarly, a study conducted in Kentucky found that 30.2% of a sample of PWID living with viral hepatitis reported having recently engaged in receptive syringe sharing [4]. These and other findings underscore the continued need for comprehensive interventions that increase access to sterile injection equipment. Several decades of research have been conducted to better understand unsafe injection practices among PWID. For example, prior studies have identified that these behaviors are driven by the intersections of individual-and structural-level factors, substance use, social context, and policy [7,9,13,16]. Inadequate access to sterile injection equipment has also been associated with syringe sharing [20][21][22]. Mitigating the consequences of high-risk injection practices (e.g., infectious disease acquisition) may be achieved through the implementation of interventions that aim to increase access to sterile injection equipment, including syringe services programs (SSPs) [23][24][25][26]. However, many communities lack SSPs due to restrictive policies, community-level opposition, and inaccurate fears that they may increase substance use, crime, or syringe litter [20,[26][27][28][29][30][31][32][33][34][35]. Stigma and discrimination against people who use drugs also negatively affect the implementation and utilization of SSPs and other evidence-based response strategies, such as medications for opioid use disorder (MOUD). The COVID-19 pandemic had far-reaching effects on public health, including among PWID. In some instances, SSPs closed or modified their operations to reduce COVID-19 transmission risks [36][37][38][39][40]. Some SSPs also had inadequate staffing during the pandemic which led to decreased service availability, such as onsite HIV and hepatitis testing [36]. Further, pandemic lockdowns also resulted in reductions in syringe distribution and infectious disease testing [41]. Mental health issues (e.g., depression, anxiety, and loneliness) worsened among people who use drugs during the pandemic [37,42,43]. In terms of substance use disorder treatment, a 2022 study found that there were substantial reductions in in-person services, but policy changes that provided flexibilities in treatment delivery (e.g., increased take-home medications, counseling by video/phone, and fewer urine drug screens) were well-received among people with histories of substance use [44]. Other COVID-19 era research has found that PWID struggled to get appointments with HIV counselors and physicians and that access to preexposure prophylaxis diminished during the pandemic [45,46]. Although existing research demonstrates several ways in which the COVID-19 pandemic affected PWID, limited research has been conducted to understand its impact on high-risk injection practices. One study found that syringe reuse was more common during the pandemic [43], but this was limited to a sample of PWID in New York City and may not be generalizable to other settings. Given that receptive injection equipment sharing is strongly associated with infectious disease transmission among PWID, better understanding this behavior in the context of COVID-19 may afford key insights about potential intervention opportunities in the ongoing pandemic and in ensuring sustainable access to sterile supplies in the future. This study utilizes data from a multistate survey conducted in late 2020 and early 2021 to examine factors associated with receptive injection equipment sharing among PWID. --- Methods --- Study context From August 2020 to January 2021, study participants were recruited from 22 substance use disorder treatment programs and harm reduction service providers in nine states (Maine, Maryland, Michigan, New Jersey, New Mexico, New York, Pennsylvania, Tennessee, and West Virginia) and the District of Columbia. Most participating drug treatment programs and harm reduction providers were engaged in the Bloomberg Opioid Initiative (a campaign supported by Bloomberg Philanthropies that aims to reduce overdose rates). Staff at collaborating organizations distributed study recruitment cards to clients. Each card featured the study logo, the study phone number, and a unique study identifier (to reduce duplicate and non-client participation). Persons who were interested in participating in the study contacted the data collection team via phone and were subsequently able to ask questions and be screened for eligibility. Eligibility criteria included being at least 18 years old, a current client of a collaborating organization, able to provide informed consent, and able to provide an unused unique study identifier. Participants received $40 compensation via a pre-paid gift card or Venmo payment. Overall, 587 responses were collected. Given our interest in receptive injection equipment sharing among PWID, we restricted the analytic sample to participants who had injected drugs in the past month (n = 266). We further removed a transgender participant to ensure their anonymity was protected. This research was approved by the Johns Hopkins School of Public Health Institutional Review Board. --- Measures --- Receptive injection equipment sharing in the past month Participants answered two questions about their receptive injection equipment sharing behaviors in the past month. Participants indicated if they had used a syringe or needle after someone else had used it and if they had used other injection equipment, like cookers or rinse water, after someone else. These two indicators had a high degree of overlap (85% of persons who shared syringes also shared other equipment); as a result, we created a binary indicator for receptive sharing of any injection equipment in the past month. --- Sociodemographic characteristics Participants reported their age (in years), gender (man/ woman), relationship status (single/in a relationship or married), sexual orientation (heterosexual or straight/ sexual minority), education level (less than high school, high school diploma or equivalent, or some college or more), and employment status (full time, part time, not working). Participants reported their race and ethnicity, which we dichotomized to non-Hispanic White and Racial/Ethnic Minority (e.g., Black, Hispanic, Multiracial/ Multiethnic) due to sample size constraints. Participants further reported if they were currently homeless (yes/ no), if they experienced hunger (defined as going to bed hungry due to lack of food) at least once a week since the COVID-19 pandemic (yes/no), if they had ever tested positive for HIV (yes/no), and if they traded sex for drugs or money since the pandemic started (yes/no). Based on the county participants reported living in, we created an urbanicity measure using the National Center for Health Statistics Rural Classification Scheme (codes range from 1-large central metro to 6 -non-core). We created a three-category measure of urbanicity: large metropolitan (codes 1 and 2), small metropolitan (codes 3 and 4), and non-metropolitan (codes 5 and 6). --- Injection drug use in the past month We created binary indicators of whether participants reported having injected each of the following drugs/ combinations of drugs in the past month: cocaine, heroin, fentanyl, heroin and fentanyl simultaneously, speedball (cocaine and heroin simultaneously), methamphetamine, methamphetamine and heroin simultaneously, prescription opioids, tranquilizers, and buprenorphine (e.g., Suboxone). We also created a variable that reflected the total number of drugs/combinations of drugs injected in the past month. --- COVID-related drug use behavior changes We included four measures of drug use-related behavior changes during COVID-19. First, we asked participants to indicate how often they injected drugs per day during COVID-19 relative to the pre-COVID era (less frequently, the same, more frequently). Participants indicated how often they used drugs with others during COVID-19 relative to before the pandemic (less frequently, the same, more frequently). Participants further indicated if they used mostly in private locations during COVID-19 (yes/no) and if they had avoided accessing syringe services programs due to COVID-19 fears (yes/ no). --- Service utilization We included three binary measures of drug treatment engagement. First, we created an indicator for any pastmonth drug treatment. We then created two indicators for the type of treatment received: any MOUD (buprenorphine, methadone, or naltrexone) and any non-MOUD treatment. The treatment types were not mutually exclusive. We also asked participants whether they had acquired sterile syringes from a syringe services program in the past month (yes/no). --- Analysis We first estimated the prevalence of past month receptive injection equipment sharing in our sample. We used Chi Square and t-tests, as appropriate, to assess bivariate relationships between variables and receptive injection equipment sharing. We used logistic regression to identify factors associated with PWID having recently engaged in receptive injection equipment sharing. We considered all correlates of receptive injection equipment sharing at the p <unk> 0.2 level for inclusion in multivariable logistic regression analyses. We elected to utilize the number of drugs injected instead of individual drug measures to achieve a more parsimonious model. We further excluded two variables (homelessness and MOUD treatment) from the multivariable model due to collinearity with other included variables (hunger and any drug treatment, respectively). In the multivariable logistic regression model, standard errors were clustered by the provider participants were recruited from to account for study design. Statistical analyses were performed using Stata 17 (StataCorp, College Station, TX). --- Results The average age of the sample was 39 years old (SD: 10.5). Half (50.2%) the participants were women and 62.9% identified as non-Hispanic White (Table 1). Fourteen percent identified as a sexual minority. Few (4.9%) reported having HIV. Over half (56.8%) of participants were in a relationship. Having a high school education was the most common education level (45.7%); the prevalence of having less than a high school education (27.2%) or some college or more (27.2%) were similar. Most (85.3%) participants were not working. About onequarter (27.7%) of participants were homeless and onethird (34.3%) reported weekly hunger. Urbanicity level varied (39.0% large metropolitan, 37.5% small metropolitan, 23.5% non-metropolitan). Approximately eleven percent (10.6%) reported engaging in transactional sex. On average, participants reported injecting three drugs in the past month. Most (85.6%) had accessed an SSP in the past month. One-third (32.1%) of participants reported more frequent drug injection during COVID-19. Just under half (46.0%) had received drug treatment in the past month. One in four participants reported having engaged in receptive injection equipment sharing in the past month. At the bivariate level (Table 1), participants who reported receptive injection equipment sharing were significantly younger than persons who did not (p = 0.04). Participants who identified as sexual minorities (p = 0.03), as non-Hispanic White (p = 0.004), experienced hunger at least weekly (p = 0.04), and who engaged in transactional sex (p = 0.02) were significantly more likely than their counterparts to report receptive injection equipment sharing. Participants with a high school education were more likely to report receptive injection equipment sharing than participants with other education levels (p = 0.01). Use of speedball (p = 0.03), methamphetamine (p = 0.003), and methamphetamine and heroin (p = 0.005) were all significantly associated with receptive injection equipment sharing. Participants who reported receptive injection equipment sharing, on average, used significantly more drugs than persons who did not (p = 0.006). Individuals who reported increased injection frequency during COVID-19 were significantly more likely to report receptive injection equipment sharing than persons who reported the same or less frequent injection (p = 0.02). In the multivariable model (Table 2), having a high school education or equivalent was associated with greater odds of receptive injection equipment sharing compared to having less than a high school education (adjusted odds ratio [aOR] = 2.14, 95% Confidence Interval [95% CI] 1.24, 3.69). Experiencing weekly hunger (aOR = 1.89, 95% CI 1.01, 3.56) and number of drugs injected (aOR = 1.15, 95% CI 1.02, 1.30) were also associated with greater odds of receptive injection equipment sharing. Older age (aOR = 0.97, 95% CI 0.94, 1.00) and living in a non-metropolitan area (aOR = 0.43, 95% CI 0.18, 1.02) were marginally associated with decreased odds of receptive injection equipment sharing. --- Discussion Using data from a geographically diverse sample of PWID during the early months of the COVID-19 pandemic, we found that approximately one in four participants reported having recently engaged in receptive injection equipment sharing. Factors associated with greater odds of recent receptive injection equipment sharing included experiencing hunger, number of drugs injected, and having a high school diploma. Our findings contribute to existing literature that examines receptive injection equipment sharing by demonstrating that this behavior was associated with factors identified in similar research that occurred before COVID-19 [7,9,47,48]. Eliminating infectious disease transmission among PWID will require novel, low-threshold interventions (e.g., peer-led SSPs, harm reduction vending machines, no-cost access to mail-order harm reduction supplies) that ensure PWID have access to sterile injection equipment during times of co-occurring crises. We found that 34% of our sample reported experiencing weekly hunger and that hunger was associated with greater odds of receptive injection equipment sharing. These findings parallel similar research conducted among PWID before the COVID-19 pandemic. For example, food insecurity has been associated with PWID engaging in high-risk behaviors (e.g., syringe sharing, condomless sex) for HIV/STI acquisition in prior research [9,[47][48][49]. For PWID with insufficient food access, obtaining food may compete with persons' engagement in health-promoting behaviors, such as always using sterile injection equipment. It is also plausible that hunger is a proxy for a mosaic of structural vulnerabilities (e.g., homelessness, unemployment) and having less agency to engage in risk minimizing behaviors. Among PWID living with HIV, research has also shown that inadequate food access increases severity of infectious diseases [50,51]. Communities should work to guarantee no person struggles with hunger. Strategies to mitigate hunger among PWID, and communities more broadly, should be holistic in nature given the overlapping nature of hunger with other structural vulnerabilities, including homelessness. Comprehensively addressing structural vulnerabilities among PWID may carry significant public health benefits via supporting reductions in high-risk injection behaviors. Future work should be conducted to identify exemplar models of care that integrate the provision of harm reduction services and food access. Notably, there are examples of service providers that integrate food provision and harm reduction [52][53][54]. Similar to research conducted before COVID-19, we found that the number of drugs PWID injected was positively associated with receptive injection equipment sharing [9]. This finding may be partially explained by associated needs for sterile injection equipment, i.e., persons who inject more types of drugs may require larger volumes of sterile injection equipment, including syringes. Given that the COVID-19 pandemic reduced access to SSPs, it is also plausible that PWID may have had challenges ensuring they had a sterile syringe and other supplies for each injection [55]. Further, many communities lack SSP access, potentially exacerbating risks for receptive injection equipment sharing [26]. Future work should be conducted to develop innovative strategies that afford PWID reliable and low threshold access to sterile injection equipment. Exemplar strategies to increase access to sterile injection equipment may include public health vending machines, mail order injection supplies, and distributing supplies at retail venues (e.g., pharmacies). Peer-based SSPs may also be particularly effective at reaching vulnerable PWID [56,57]. We found that living in a non-metropolitan area was marginally associated with decreased odds of recent injection equipment sharing. This finding warrants additional study given that many injection drug use-associated HIV outbreaks in rural communities have occurred in recent years [17][18][19]58]. Further, analyses that examined risks for injection drug use-associated HIV outbreaks identified many rural counties throughout the United States as vulnerable [59]. Though methodological differences limit comparability across studies (e.g., we recruited PWID who accessed services at drug treatment and harm reduction programs, which may be of limited availability in non-urban areas), receptive injection equipment sharing has been shown to be a relatively common phenomenon among rural PWID [3,13,15,60,61]. Our finding that non-metropolitan residence was associated with decreased odds of recent injection equipment sharing may also reflect both the considerable heterogeneity in where we recruited participants as well as how we operationalized urbanicity. Nevertheless, future studies should be conducted to more comprehensively understand factors associated with receptive injection equipment sharing among rural PWID and if these relationships are affected by the degree to which persons access drug treatment and harm reduction services. It is important to interpret the findings of this study relative to its limitations. Our outcome focused on PWID engaging in receptive injection equipment sharing in the past month. As such, we are only able to glean a snapshot of receptive injection equipment sharing among our participants rather than more comprehensive examinations of this behavior and how it may vary by context over time. Additionally, there is considerable variation in how high-risk injection practices are measured in the literature, limiting our ability to make direct comparisons. Due to sample size limitations, we trichotomized our measure of urbanicity. More robust sample sizes may afford nuanced analyses across the urban-rural continuum. In addition, we found that education was significantly associated with receptive injection equipment sharing; however, this finding should be interpreted with caution given both sample size constraints and our sampling strategy. Future lines of scientific inquiry should explore the role of educational attainment and engagement in high-risk injection practices. Efforts should also be undertaken to ensure PWID receive evidence-based education about the risks of sharing injection equipment. Another potential limitation relates to sampling bias given that we recruited persons from substance use disorder and harm reduction service providers in nine states and the District of Columbia. Our findings should not be considered representative of PWID across the US, nor reflective of the experiences of PWID who do not access substance use disorder treatment facilities or harm reduction services. Though our study is not without limitations, it contributes to the public health literature by examining factors associated with receptive injection equipment sharing among a sample of geographically diverse PWID during the early months of a global pandemic. In conclusion, we found that a quarter of PWID who were connected to drug treatment and harm reduction service providers reported receptive injection equipment sharing during the early months of the global COVID-19 pandemic, and that these behaviors varied according to education level, hunger, urbanicity and number of drugs injected. We also found that PWID residing in non-metropolitan communities had marginally decreased odds of receptive injection equipment sharing. Factors associated with receptive injection equipment sharing in our study had both similarities and differences to prior research. The COVID-19 pandemic affected risks for infectious disease acquisition among PWID throughout the world, and our results shed light on the high-risk injection practices among PWID that contributed to enduring infectious disease risks during the pandemic. --- Availability of data and materials Deidentified data that supported the findings of this study are available upon reasonable request. --- Abbreviations --- SSP Syringe services programs PWID People who inject drugs HIV Human immunodeficiency virus Author contributions STA, MM, SJH, BS, and SGS were involved in the conception of the study. STA and KES were involved in the analysis. All authors were involved in the interpretation of the findings. All authors were involved in drafting the manuscript. All authors reviewed and approved the final manuscript and agree to be held accountable for all aspects of the work. --- Declarations Ethics approval and consent to participate The study was approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board. --- Consent for publication Not applicable. --- Competing interests The authors report no conflicts of interest. • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research? Choose BMC and benefit from:? Choose BMC and benefit from: --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background Receptive injection equipment sharing (i.e., injecting with syringes, cookers, rinse water previously used by another person) plays a central role in the transmission of infectious diseases (e.g., HIV, viral hepatitis) among people who inject drugs. Better understanding these behaviors in the context of COVID-19 may afford insights about potential intervention opportunities in future health crises. Objective This study examines factors associated with receptive injection equipment sharing among people who inject drugs in the context of COVID-19.From August 2020 to January 2021, people who inject drugs were recruited from 22 substance use disorder treatment programs and harm reduction service providers in nine states and the District of Columbia to complete a survey that ascertained how the COVID-19 pandemic affected substance use behaviors. We used logistic regression to identify factors associated with people who inject drugs having recently engaged in receptive injection equipment sharing.One in four people who inject drugs in our sample reported having engaged in receptive injection equipment sharing in the past month. Factors associated with greater odds of receptive injection equipment sharing included: having a high school education or equivalent (adjusted odds ratio [aOR] = 2.14, 95% confidence interval [95% CI] 1.24, 3.69), experiencing hunger at least weekly (aOR = 1.89, 95% CI 1.01, 3.56), and number of drugs injected (aOR = 1.15, 95% CI 1.02, 1.30). Older age (aOR = 0.97, 95% CI 0.94, 1.00) and living in a non-metropolitan area (aOR = 0.43, 95% CI 0.18, 1.02) were marginally associated with decreased odds of receptive injection equipment sharing.Receptive injection equipment sharing was relatively common among our sample during the early months of the COVID-19 pandemic. Our findings contribute to existing literature that examines receptive injection equipment sharing by demonstrating that this behavior was associated with factors identified in similar research that occurred before COVID. Eliminating high-risk injection practices among people who inject drugs requires investments in low-threshold and evidence-based services that ensure persons have access to sterile injection equipment.

Introduction The Fulani peoples are the major pastoralist group across West Africa and have dominated cattle production in Nigeria for centuries [1]. Also known as Fulbe pastoralists, their population in Nigeria is estimated at 15.3 million [2]. In the late 1980s Fulani were estimated to manage 90% of Nigeria's ruminants [3]. A 1992 livestock survey found that Fulani pastoralists, the great majority of whom have now settled, grow crops and practice a form of limited seasonal transhumance, kept 83% of the cattle in Nigeria. Many arable farmers also practice animal husbandry. Traditional management in and around rural villages by non-Fulani accounted for 17% of cattle. Only 0.3% cattle were reared on commercial holdings in a peri-urban or urban settings [4]. Village and urban cattle keeping is increasing as business people invest in the current agricultural revolution in Nigeria and the local 'indigene' populations learn herd management skills from the Fulani. In 2014, the ruminant population of Nigeria was estimated at 19.4 million cattle, 40.6 million sheep and 71.0 million goats [5]. Traditionally, Fulani practiced year-round nomadism, partly in response to the need to migrate away from the high infection challenge presented by tsetse flies. Before the 1950s, herds from the northern savannah zone only grazed in the sub-humid zone further south during the dry season, when the risk from trypanosomiasis was lower. Since the 1950s there has been a southwards shift into the sub-humid zone for year-round grazing with Fulani pastoralists occupying 5% of the rural population of what was an inhabited zone. By 1988 it was estimated that the dry season cattle population decreased by approximately 40% in the wet season [6] indicating that an increasingly year-round population was present in this zone. Expansion of cultivation has reduced suitable tsetse habitat, making the area more hospitable to livestock keepers [7]. An increasing number of Fulani are giving up the wet season migration northwards, to engage in mixed crop/livestock farming and a more settled lifestyle [8]. Most Fulani now have permanent homesteads and practice only short-range dry and wet season transhumance, in part due to diminishing access to rangelands from farming pressure, increasing conflicts and insecurity [1]. Grazing reserves were established in Nigeria in the 1960s to encourage pastoralist sedentarisation [9]. The reserves were anticipated to increase productivity, providing critical resources for livestock keeping (water and land tenure) and access to markets, and to reduce clashes between pastoralists and crop farmers driven by competition for resources. The Kachia Grazing Reserve (KGR) was established by the Kaduna State Ministry of Animal and Forest Resources in 1967 to settle nomads in one location to improve their standard of living; to improve the quality of livestock produced; to reduce conflict between nomads and farmers and to provide an area for research [10]. The KGR is home to some 10,000 Fulani pastoralists and their 40,000 cattle. In May 2011, a month after the presidential election, KGR experienced a sudden influx of displaced families fleeing violent clashes in their areas of origin [11]. The study objectives are two-fold. Firstly, we describe and assess variation in KGR household characteristics in terms of the household head, wives and marriages, livelihood strategies, livestock keeping, crop farming, off-farm sources of income, mutual assistance and genderedwealth holdings. Secondly, we explore whether livestock and wealth are equally distributed among KGR households, and if not, what variables account for the variation seen across households. The KGR, the first grazing reserve to be established in Nigeria, is representative of Fulani livelihood diversification, wealth and household heterogeneity in a grazing reserve setting. Analysing the social and economic make-up of grazing reserve communities and their resilience to social change is pertinent because of the societal and political lobby for sendentarisation of pastoralist populations. Pastoral livelihoods are in a transitional state and understanding the household economy is crucial in achieving sustainable and effective development initiatives. --- Study site The KGR is situated in Kaduna State, north central Nigeria, and comprises 31,000 hectares between latitudes 10 <unk>03'-10 <unk>13'N and longitudes 7 <unk>55'-8 <unk>06'E. KGR lies within the sub-humid zone, 700-900 m above sea level and is fed by the Kaduna River. KGR exhibits northern Guinea Savannah woodland vegetation. The climate is tropical sub-humid, with a wet season running from June-October and dry season between November-May. The average temperature is 28 <unk>C (minimum of 19 <unk>C in January and maximum of 39 <unk>C at the start of the rains). KGR settlers are exclusively Fulani pastoralists. The KGR 'district' is called Ladduga or 'bush' in Fulfulde and the KGR headquarters and trading centre is called Tampol (after the tarpaulins that covered the first market stalls). Administratively, KGR is divided into 6 blocks. Block 2 is large and diverse, geographically, and is subcategorised into 2A and 2B (Fig 1). KGR has 9 Ardos or village heads, each representing a clan. The settlement areas within the blocks are named after the clan elder. --- Study design This mixed methods study comprised three comprehensive livelihoods surveys undertaken within KGR during 2011: March (mid-dry season), June (beginning of wet season) and October (end of wet season). This approach enabled themes to be covered not only through administration of questionnaires at different time points and to different cohorts of households/individuals but also through application of a range of different participatory research methods. Triangulation was employed to validate the repeatability of data obtained and ensure better reliability of evidence. This method also ensured that variations in characteristics, knowledge, perception and practices were captured. The household was the primary unit of assessment. In July 2010, a state census undertaken by the KGR Project Office recorded 581 households in KGR. For the survey undertaken in March 2011, 64 households were randomly selected from this total. In May 2011, one month after the presidential election, KGR experienced a sudden influx of displaced families fleeing violent clashes in their areas of origin. In total, 249 families (3,000 people) moved into the reserve with their livestock (20,000 cattle, 5,000 sheep and 1,500 goats). A census undertaken in June 2011 provided a revised figure of 777 households with a human, cattle, sheep and goat population of approximately 10,000, 40,000, 10,000 and 5,000 respectively [11]. Of the 752 households for which data on the year of settlement in KGR are available, 28.2% were established in the KGR before the period of inter-communal violence that began in the early 2000s. A further 38.7% settled in KGR between 2001 and 2010 and 33.1% of all households had moved into KGR in May 2011. Of these 249 households, all were inhabiting the reserve in October 2011 intending to settle permanently. In this study, households that moved into KGR during the mass immigration event of May 2011 are referred to as 'new immigrant' households and the remainder as 'old settlers'. For each of the June and October surveys, 40 households were randomly selected from across these 777 households by allocation and generation of random numbers using the Survey Toolbox1. Focus group discussions and key informant interviews. Focus group discussions [FGD] employing participatory rural appraisal techniques [12] were undertaken by the first author with 8 groups of 6-12 individuals of the same sex, with the assistance of a local translator. These discussions were supplemented by two key informant interviews. Topics of discussion and individuals/groups targeted are summarised in Table 1. Wealth and poverty were assessed using participatory wealth ranking, in which focus group discussants self-determined wealth reference points [13]. Questionnaires. A questionnaire was administered to each selected household. Interviews were undertaken by the first author with the assistance of a local translator. Respondents were household heads or, in a minority of cases, their sons or brothers. Not all selected households agreed to be interviewed. For the survey undertaken in June 2011 a single 'outlier' household was removed from analysis, having a household size of 277 and cattle herd of 1,500. Questionnaires focussed on four themes: household size and composition; the domestic animal population (species composition and holdings kept in and outside the KGR); household livelihoods strategies and sources of income. Livestock capital was used as the primary proxy for wealth. FGDs indicated that the number of animals was the most important parameter for ranking a household's wealth status in the KGR as previously reported [14,15]. To aggregate the livestock species maintained by a household, the total number of tropical livestock units (TLU, equivalent to 250 kg live-weight) were calculated. The following conversion factors were applied: cattle = 0.70, sheep and goats = 0.10, domestic fowl/poultry = 0.01 [16]. Other wealth indicators including number of buildings, hectares farmed, and educational level of the household head, were also examined. Our analyses are based on a total sample size of 133 households. Fifty-six households were interviewed in March 2011; 38 in June 2011 and 39 in October 2011. For the surveys undertaken in June and October 2011, approximately 30% of households were of new immigrants. Data pre-dating the immigration event have been analysed separately and these excluded households and livestock that were on dry season transhumance. For most analyses, responses from June and October 2011 are aggregated. Statistical analyses. A range of univariate analyses (t-tests, chi-square tests and simple linear regression) were carried out in R v3.1.1 [17]. A multi-variable general linear regression model to explore the key variables affecting the total livestock units within a household was created in R (regression modelling strategies 'rms' package) using a stepwise, forward-selection approach with Akaike information criterion (AIC) values. Multiple correspondence analysis (MCA) was performed using selected variables from the June and October 2011 survey data. The variables used were: household size, herd size, number of marriages of household head and sources of extra income (from the options: wages (casual labour), salary (salaried work) and some sort of business initiative-e.g. tea shop, motorcycle servicing, etc.). Wealth status (using TLU per capita as a proxy) and geographical location were included as supplementary variables, which does not affect the creation of the main dimensions but enables these variables to be projected onto the MCA plot. The MCA was performed in STATA v.13 (Statacorp LP, College Station, TX, USA). --- Ethics statement Ethical clearance for interviewing of human subjects was granted on 7th February 2011 by the Ministry of Health, Kaduna State (Nota MOH/HS/PER/VOL.I/234/70). Study participants were briefed on the purpose of the study and verbal informed consent was obtained. Written consent could not be obtained as the large majority of study participants were illiterate. Participant consent was documented directly in the questionnaires used to interview the study participants. The ethics committee of the Ministry of Health, Kaduna State, approved this consent procedure. --- Results --- Household characteristics Household head. The nucleus of all KGR Fulani households is its head (HHH) or jewuro, an adult male, who makes decisions on social, economic and political matters. FGD responses indicated that the main role of the HHH was to manage the herd or agricultural unit, being responsible for all aspects of herd security, maintenance and reproductive efficiency. HHH ages ranged from 23 to 87 years with a mean and median age across the surveys of 53 years. Over 50% of HHHs were aged between 45 and 64. There was a significant relationship between household size and age of the HHH (p<unk> 0.01), though linear regression indicated that this relationship accounted for only 8% of the variation in household size. Rates of formal education (primary, secondary or further), other than in Koranic schooling were low, at just over 10% of all HHHs. Wives and marriages. A Fulani man may take a maximum of four wives at any one time in accordance with Islamic rules. It is important to consider not just the current number of wives of HHHs, but also the number of marriages contracted. Following divorce or death of a spouse, children usually remain in their father's household. The majority of HHHs, 70%, had either one or two wives. Two was the modal number, when all of a HHH's marriages were considered, with 32% in the March survey and 44% in the June-October surveys having married twice. Three HHHs had married five times and one HHH interviewed had married 10 times, but was exceptional (Table 2). Household composition. The household, or wuro, is a group of agnatically related men, their wives and children. The FGDs with women revealed two wuro structures in the KGR: a three-generation household in which the HHH is elderly and his sons and their wives and children live under his directive and one in which the HHH has died and is replaced by his eldest son, who lives in the same household with his junior brothers, their wives and their children. The first wuro structure was more common: only 11% of households interviewed reported having a HHH living with his brothers and his brother's wives and children. 1 Number of times HHH married, inclusive of current, divorced and deceased wives 2 A HHH had one and another had 2 wives, but these wives died leaving both HHH with no wives 3 HHH divorced 5 wives 4 2 HHH had 5 wives overall and 1 HHH had 10 wives (same HHH as the one who divorced 5 wives) 5 HHH had 5 wives, 4 present and 1 that died doi:10.1371/journal.pone.0172866.t002 Respondents reported four phases of household expansion and division. (i) The household expands through the offspring of the HHH and his first wife, and may continue to expand to form a compound family if the HHH takes on more wives. (ii) The household expands when the sons of the HHH take their own wives and have children. (iii) Division occurs as sons and their wives separate from their father's household if the sons have built up large enough cattle herds, as illustrated by this statement from a young focus group discussant who decided to 'go it alone' and create a new wuro distinct from the one of his father on account of his large herd size and financial independence. A son from a poor household may also be driven to leave his father's wuro to improve his prospects by moving elsewhere: 'if someone does not have enough cows to give to all his sons then he will send his son to go and work for another herd so that he can work to earn a calf, the going rate is two years for a female and one year for a male'. (iv) Household dissolution occurs when the HHH dies and his herd is distributed amongst his sons and daughters in a 2:1 ratio. At this stage, each son may form his own household unit, although a household may continue to exist as a single unit even after the death of an elderly HHH. Certain factors such as death of a father, livestock wealth or poverty make household division more likely, but focus group discussions revealed that there is no typical threshold number of cattle or prescribed rule for an individual deciding to form his own household unit. The decision to divide is made by the household head, as illustrated by this statement from a focus group discussant: 'a son will only separate his animals and family if his father gives his approval'. FGDs in March 2011 showed that marriage occurs in individuals of 16 years or more. For the surveys undertaken in June/October 2011, 16 was considered the age of adulthood, accordingly 53% of the population were children (Table 3). Overall 51% of the household population was male; within the 5-15-year old age group, 60% were male. Marriage of young girls may have resulted in their being classified as older than they were. In the March 2011 survey, 16% of households reported hiring non-blood related 'cattle boys', classified as members of the households, accounting for 1.4% of the population. FGDs indicated these could be from non-Fulani ethnic groups. The mean household size was found to be higher (25.9) in June-October 2011 than in March 2011 (20.4). New immigrant households were significantly larger with a mean 3). Livelihood strategies. KGR is considered by government officials to be an 'agro-pastoralist' community, with the implication that 25-50% of income is derived from livestock and livestock-related activities [18]. At the time KGR was set up, it was stipulated that on settlement in KGR households should be allocated 10 hectares of land, with a proviso that 4 hectares should be dedicated to crop farming. FGD interviews showed a mismatch between the perceptions of the authorities and inhabitants. All households engaged in livestock keeping, with 97% ranking this activity as their primary source of income or subsistence. Households reported deriving more than 50% of their income from livestock which would categorise them as pastoralists. Households also engage in other livelihood strategies (cropping, mainly for subsistence, and off-farm activities). Ninety percent of KGR households grow crops, and 96% of the crop-growers ranked this activity second in terms of contribution to overall household income. Over half of KGR households engaged in off-farm activities, and ranked this activity third in terms of its contribution to the household economy. Remittances from family members living away from home and women's crafts also contributed to the income economy of some households, although these sources were typically ranked 3 or lower (Table 4). Livestock, milk and, to a lesser extent, crop sales meet the cash needs of the household. These include purchase of herbs, spices and condiments for cooking, clothes, school fees, human and veterinary drugs. Small ruminant sales cover most day-to-day cash needs whilst the sale of cattle is limited to major cash needs. There was no significant difference in household engagement in non-livestock related activities between new immigrant and old settler households (Table 5). Livestock keeping. Livestock species kept in KGR include cattle, sheep, goats and domestic fowl (chickens, turkeys and guinea fowl). Three households kept ducks and one kept pigeons). KGR households also keep small ruminants, dogs and cats: dogs for herding cattle and cats for population control of rodents that can devastate grain reserves. Cattle accounted for 96% of the overall TLUs. The contribution of each species to the overall livestock capital in terms of TLUs is shown in Table 6. The survey undertaken in March 2011 indicated many sub-herds being maintained by KGR households outside of KGR. Interviews conducted in June and October 2011 differentiated between livestock kept in the reserve and outside. Approximately 40% of the households sampled in June and October 2011, maintained cattle outside of the reserve. Despite these herds being smaller than those kept within the reserve (mean herd size outside 74 as opposed to 96 inside KGR), these sub-herds accounted for 23% overall TLUs. Some households kept goats, chickens, dogs and cats outside of the reserve, this suggests homesteads were maintained outside of the reserve as these species are not transhumant. Indeed 21% of interviewed households reported owning/hiring homesteads outside of the KGR, of whom half were old settlers. Some of the new immigrants reported that their old homesteads 'had burnt to ashes' in the post-election violence. These ''secondary" homesteads were mostly within Kaduna State (Kwoi, Birnin Gwari, Anchau, Kafanchan, Kagoro, Zangon-Kataf, Kachia, Fadan Kamantan). Some interviewees reported owning property in Bauchi, Plateau and Nassawara States. Household size was larger for new immigrants, who had correspondingly larger average TLU (Table 7). In June 2011, new immigrants and old settlers had 3.2 and 2.0 TLU per capita respectively; by October 2011, both groups had just over 2.5 TLU per capita. 1 Cattle kept within and outside KGR 2 5 HHs(households) kept sheep out of KGR, of which one had no sheep in KGR 3 1 HH kept goats out of KGR (this HH had no goats in KGR) 4 1 HH kept chickens both in and out of KGR 5 2 HHs kept guinea fowl both in and out of KGR Livestock contribute to household income and subsistence primarily through sale of cattle and small ruminants to generate cash and through the consumption of milk. Discussants reported rarely eating meat: "we do not have a taste for meat outside of slaughtering practiced as part of Islamic religious festivals (Eid el Kabir and Eid el Fitr), and even then we would rather sacrifice a sheep". The economic and sociocultural value of cattle in pastoral communities ranges from prestige-making, bartering potential or currency, sources of food and labour and asset saving or insurance against disasters. In KGR, milk from cattle is sold and or consumed. The small ruminants kept are not milk producing breeds. Among the 82% of KGR households that sell milk, half of the milk that is produced is sold, mostly within the KGR community. Some women will trek to non-Fulani villages and towns outside of KGR to sell milk and milk products. KGR inhabitants are cattle-keeping Fulani and so internal demand for purchase of milk is low. Most households take milk to the KGR central market area to sell directly to teashops. Women make nono (yogurt) and occasionally wara (cheese), sold on market days to supplement cash needs for cooking ingredients or school supplies and clothing for children. The lack of a milk market chain was described as a constraint by the community. Respondents recalled a company called 'Milkopal' which used to operate within the reserve, collecting milk directly from households and distributing to communities outside. Crop farming. Most households interviewed grew crops mostly for household consumption (Table 4). Crop farming detail was investigated during March 2011 and data refer to the old settlers in the reserve. The modal area of land farmed was 2 hectares, although some households reported farming up to 50 hectares. Half of crop farming households (51%) sold some of the crops produced and on average reported selling 40% of their produce. Less than 20% of households reported growing crops to feed livestock. Respondents ranked the importance of each crop grown in terms of subsistence and/or cash value. Almost all households engaged in crop farming grew maize and sorghum, which ranked as the two most important crops. Around 70% of households grew sweet potatoes and yams, while 40-55% of households grew cocoyam, soybean, beans, rice, cassava and groundnuts. Fewer than 30% of households grew millet. A few households cultivated ginger as a cash crop. The number of hectares farmed was not correlated with the year a household moved into the reserve (Pearson coefficient = 0.028). There was a moderate positive correlation between the number of hectares farmed and household size (Pearson's correlation coefficient = 0.396, p = 0.003) and between the number of hectares farmed and TLU per household (Pearson's coefficient = 0.431, p = 0.001). Households with the most livestock assets were found to farm the most crops. The number of livestock owned and household size are intrinsically linked, as the ability to look after large livestock herds also depends on the availability of manpower. There was a weak positive correlation between TLU/capita and hectares farmed (Pearson's coefficient = 0.150, p = 0.275). Off-farm income sources. Over half of KGR households have diversified their livelihoods through off-farm activities (Table 4), most citing 'business activities' as a source of additional income. Business activities included owning shops in Tampol, the trading centre of KGR (drug shops, teashops, a phone charging shop, a motorcycle repair shop, a general provision shop, maize grinding service and a tailor shop). Respondents also reported engagement in cattle trading or operating motorcycle taxi services. One respondent was a registered contractor of an agro-services company. One respondent had a house building and another a carpentry business. Salaried employment was also reported (Table 4). Employment included: teacher, bus driver, paramedic/ health worker, computer technician, policeman/other civil service roles. Fewer households cited engagement in casual waged labour but where this was reported it consisted of building and agriculture-related activities such as weeding, ridging, planting, sowing and ploughing. Mutual assistance. Approximately one third of households received money from family members who did not live within their homestead (Table 4). Gendered wealth holdings. Women can inherit cattle from their father. Upon the death of a household head, his cattle are distributed in a 2:1 ratio between his sons and daughters. A focus group discussant gave an example: 'if a HHH has 25 cows, 1 daughter and 2 sons, the daughter receives 5 and each son 10 cattle.' Women, however, do not hold on to this cattle wealth and will usually give these animals to her sons and husband. Transmission of cattle wealth to the next generation is also gender biased because a father will give one female calf to a newborn son but not a daughter. All subsequent calves and herd growth will usually come from this one animal, although relatives can sometimes give young boys a calf. A 28-year old discussant reported that the pregnant cow he received from his father on his second birthday enabled him to build up a herd 10 cows, 10 bulls and 5 calves. Focus group discussions revealed sheep are also owned and managed by men but that most goats and domestic fowl are reared and owned by women: 'if a woman has cash needs she can sell a goat or a chicken'. Women are also responsible for preparing and selling milk and milk products such as nono (yogurt), fura de nono (yogurt with millet), nebam (butter) wara (cheese), nyamri (porridge) and kindirmo (buttermilk). Focus group discussions with women revealed that half of milk goes to household consumption and the other half is sold. The cash generated from milk sales is managed by the household head. The only source of independent income for women is derived from women's crafts. Table 4 also shows that across 23% of households, women engaged in a range of activities including metalwork (flat pans), mats, soap, food products (bean cakes for sale on market days), sewing and dressmaking. A female FGD participant elaborated: "this enables us (women) to get some allowance for ourselves to spend on our homes and our children". --- Measures of household wealth status Association between KGR TLU and other household variables. The associations between KGR TLU and key household variables, were explored using linear regression models for all 133 households in the study (N = 133). An initial linear model was created with household size as primary predictor of household TLU. Household size was highly predictive for household TLU but accounted for only 28% of the variation seen across the households sampled. A series of additional variables were explored applying a stepwise, forward-selection approach using the adjusted R 2 and AIC values shown in Table 8. This analysis indicated, that in addition to household size, the total number of marriages ('wives') of the HHH was a significant predictor of household TLU. On average for each extra member of a household the value of its total TLU increased by 2.0 while for each additional marriage TLU increased by 15.7 (Table 9). Block membership, a variable linked to geographical location, showed marginal significance, but did not demonstrate a better fit (with a delta AIC 0.8 in the model based 121 households with complete data). A scatterplot of TLU values per household across the blocks (Fig 2) indicated differences across blocks and significant heterogeneity between households within the same block. Block 2B has the lowest median household TLU and is also the most homogeneous. Additional variables such as date of survey, old settlers versus new immigrants, years established in KGR, and number of buildings per household did not improve model fit. The introduction of various 'off-farm' activities: household engagement in business, salaried work and casual labour or receipt of money from family members living outside KGR also did not improve the model fit. Categorisation of KGR households in terms of per capita livestock holdings. The association between wealth in terms of household TLU for livestock kept in the KGR and other key variables indicated that household size was an important variable. TLU per capita were calculated and households were allocated into wealth categories based on TLU per capita as in [19] (Table 10), an approach to the estimation of wealth status that has also been widely adopted by other authors for categorising pastoralist and agropastoralists households [20,21]. There was a strong relationship between per capita and overall household TLU with a linear relationship explaining around 40% of the variance (Fig 3). Introducing a quadratic term (also highly significant) improved relationship fit, increasing the total amount of variance explained by <unk>10%. Households with a large livestock holding tended to have large livestock holdings per person. The nature of this relationship is however, more complex than that proposed by [20] who posited a maximum 'plateau' at 5 TLUs per capita. Here, many households exhibited a TLU per capita higher than this value (Fig 3) and 50% of the variation in this value could not be explained in terms of overall household TLU. A significant proportion (30%) of households in KGR change TLU-based wealth category when cattle outside of KGR are taken into consideration (Table 10). A general linear model yielded better predictions when 'all TLU' was considered as the outcome. Accordingly, for the MCA only data from 77 households interviewed in June and October 2011 were included in the analysis since the distinction between reported total and KGR cattle holdings was very clearly made in these interviews. Only two households fell in the lowest and highest wealth categories, these were put into the next nearest categories to generate a 4-way categorisation of wealth. Association between livestock holdings and other household variables. The MCA examined the association between household wealth status in terms of livestock holdings and a range of household variables (Fig 4). Specific components associated with the creation of weights on the first two dimensions are summarised in Table 11, including household size, total TLU at household level, number of marriages and off-farm income sources. While wealth status (based on TLU per capita) is shown on Fig 4, this is a possible consequence of the fact that it was entered as a'supplementary' variable-i.e. one that plays no part in the underlying analysis. The first two dimensions of the MCA plot, accounting for around 76% of the variability due to the variables included, are shown in Fig 4. Variables used in the construction of plot are detailed in Table 11. Wealth Category was included as a supplementary variable. The first dimension is highly dependent on the largest and smallest household size categories (HH++ and HH-) as well as these two extreme categories for Herd Size (Herd-and particularly Herd++). The differences between the categories of 'One' and 'Many' wives also contribute to this first dimension. Sources of additional income make almost no contribution. The second dimension is strongly influenced by the smaller household size category (HH-) and moderately sized herds (Herd+ and Herd-). The difference between the groups having one or two wives, again has an influence, while those having no extra sources of income, separate from those with some, or many sources of extra income. The 'wealth' category classes are well separated particularly on the second dimension. A cluster is observed in the top-left quadrant consisting of households which are the largest in size, have the most cattle and, also have many wives ('Cluster 1'). Conversely, in the topright quadrant we find households associated with very small household size, with only one wife and smaller herds ('Cluster 2'). Finally, households with smaller overall household size but with fairly large herds and two wives are clustered towards the bottom of the graph ('Cluster 3'). The moderately wealthy category is strongly aligned with Cluster 3, while the wealthiest class sits between all the clusters. The least wealthy households tend to align with Cluster 2. Households with smaller herd sizes and relatively large households were those most engaged in looking to sources of extra income. Geographical location was included as a supplementary variable. Block 2B aligned closely with Cluster 2, while Blocks 4, 5 and 6 tended to align with Cluster 3 and Block 1 showed some alignment to Cluster 1. The other blocks showed no clear orientation and in general the inertia accounted for by block location was limited, indicating a fair degree of spatial heterogeneity in terms of these categories as represented within the MCA. --- Discussion Fulani domestic units were traditionally comprised of agnatic lineages: primary kinship groups of 500-1,000 persons [22], whose common ancestor could be traced back to more than seven generations [23]. Interviews with the Fulani community in KGR show little has changed. Household members are entirely dependent on the head of household for economic, physical and moral support and for political representation in line with previous descriptions of Fulani households [23,24]. Members of the households acknowledged that the head is responsible for management of the herding and agricultural output, for the cattle herd's safety, maintenance and reproductive efficiency. Within the reserve, 50% of HHHs were aged between 45 and 64 years, showing no increase from observations made almost 40 years ago [23], although older than recorded in the 1950s [25,26]. Household demography was also little changed with 53% consisting of children, as compared to 48% reported the 1980s [23]. The average reported household size in KGR was larger than previously cited. A mean of 12 individuals was previously cited for semi-nomadic households on the Jos plateau [23] and average household size of 6 elsewhere in Nigeria [27]. In Senegalese pastoralist communities, an average household size of 11 was observed [28]. In this study, the KGR community defined a wuro (household) as the extended household, made up of multiple 'ruga' (homesteads), consisting of a collection of huts belonging to members of the same family. This is the unit representing a cattle-owning entity headed by the HHH even though individual cattle may in fact, belong to different family members. Differences may be attributed to interpretation of a wuro, as previous studies may have defined households as individual ruga. In pastoralist communities, the livestock holding (particularly cattle) is considered to be the node that ties different aspects of wealth and poverty [11]. Increasing wealth is more likely to be associated with accumulating livestock than increasing cropping [29] but there is a tendency to generalise pastoralists as poor, pursuing an out-dated livelihood strategy which generates impoverishment [30]. The acquisition of stock and ensuring its wellbeing has been described as a means in itself, rather than a means to an end [31]. Although livestock, especially cattle are seen as a source of prestige, this is tightly bound up with their economic function. They are the means of production, the source of both future livestock and of daily income from milk for consumption or sale and occasional income from sale or slaughter of stock. Several authors discuss the importance

A mixed methods study was undertaken in the Kachia Grazing Reserve of northern Nigeria. Surveys in March, June and October 2011 included focus group discussions, key informant and in-depth household interviews, concerning livelihood practices, animal health, ownership, and productivity. In May 2011, 249 Fulani families fleeing post-election violence entered the reserve with their livestock, increasing the number of households by one third.Despite being settled within a grazing reserve, over half of households sent all their cattle away on seasonal transhumance and another third sent some away. Cattle accounted for 96% of total tropical livestock units (TLU), of which 26% were cattle kept permanently outside the reserve. While all households cited livestock as their main source of income, 90% grew crops and 55% derived income from off-farm activities. A multiple correspondence analysis showed that for each extra member of a household its TLU value increased by 2.0 [95% CI, 1.4-2.7], while for each additional marriage its TLU increased by 15.7 [95% CI,. A strong association was also observed between small herds, small households with only one wife, alongside marked geographical wealth differences within the reserve. New immigrant families had larger household sizes (33) and livestock holdings (122 TLU) than old settlers (22 people and 67 TLU). Prior to the mass immigration, the distribution of TLU per person was unimodal: 41% of households were classified as 'poor' and 27% as 'medium', whereas post-immigration it was bi-modal, with 26% classified as 'very poor' and 28% as 'medium'.

consisting of a collection of huts belonging to members of the same family. This is the unit representing a cattle-owning entity headed by the HHH even though individual cattle may in fact, belong to different family members. Differences may be attributed to interpretation of a wuro, as previous studies may have defined households as individual ruga. In pastoralist communities, the livestock holding (particularly cattle) is considered to be the node that ties different aspects of wealth and poverty [11]. Increasing wealth is more likely to be associated with accumulating livestock than increasing cropping [29] but there is a tendency to generalise pastoralists as poor, pursuing an out-dated livelihood strategy which generates impoverishment [30]. The acquisition of stock and ensuring its wellbeing has been described as a means in itself, rather than a means to an end [31]. Although livestock, especially cattle are seen as a source of prestige, this is tightly bound up with their economic function. They are the means of production, the source of both future livestock and of daily income from milk for consumption or sale and occasional income from sale or slaughter of stock. Several authors discuss the importance of large herds for security in times of drought: 'a man who loses one-third of his stock is much better-off if he begins with 60 cows rather than with 6.' [32], a point underlined by the demographic modelling of the time taken recover from a drought event undertaken by [33]. Lastly, when a hardship strikes that is beyond remedying by the sale of smallstock, strikes, the so-called 'unproductive' or'surplus' animals are the ones that can be sold. These different functions are reflected in cattle management practices, as observed in the KGR [34]. An increasing body of evidence describes diversification in income sources of pastoral peoples [2,35,36,37,38]. A decrease in household livestock holdings or increase in demand for household inputs is considered to favour a transition from pastoralism to agro-pastoralism, with diversification of livelihood being regarded as a risk avoidance strategy, promoting resilience to the environmental and social conditions pushing pastoral communities into poverty. A study in Kenya reported that the poorest categories of pastoralist households (those with less than 1.0 TLU per capita) had the most diversified sources of income whereas those with more than 4.5 TLU per capita focussed heavily on pastoralist activities and avoided diversification [30]. Government policy in Nigeria continues to place an emphasis on reducing Fulani mobility and promoting sedentarisation. As discussed above, although the KGR is officially regarded as agro-pastoralist, this study clearly indicates that it is predominantly a pastoralist community. In this study, only 10% of the KGR Fulani households still relied on livestock as their sole source of income, significantly lower than the 30% observed in a recent study of Fulani households on the nearby Jos Plateau [2]. This study shows a clear association between household TLU, household size and the number of wives of the HHH. The association between family and herd size, and between prestige, polygyny and large families has been previously reported [39]. Households with more people raise more cattle, and larger cattle holdings can support more people through production of milk and cash generated by sales. While it is perceived that a cow-human equilibrium exists, at which the size of the household and herd functions as a viable unit [33], consideration of TLU/ capita shows that KGR pastoralists are heterogeneous. MCA analysis revealed three main household clusters: wealthy, with large numbers of people, many wives and big herds; poor, with small household size, smaller herds and only one wife and moderately wealthy, small households, moderately sized herds and two wives. Polygamy was regarded in male and female focus group discussants as a social marker for wealth, but marriage was described by the men as being 'costly', each bride coming with a 'bride price' (usually the transfer of animals from the groom's to the bride's family). Taking total TLU/capita as a proxy of wealth, 38% the households interviewed in KGR in June and October 2011 would be considered destitute, very poor or poor and a further 67% showed signs of moving into poverty. KGR households are diversifying their income sources, involving other livelihood strategies and deriving income from off-farm activities including: business activities, salaried work and casual labour. For households engaged in crop farming, the number of hectares farmed is dependent on household size, as most households rely on family labour for ploughing, seeding, weeding and harvesting. Crops are grown for subsistence so that farming brings in little additional cash to the household. The extent and pattern of diversification into crop farming and off-farm sources of income varied. Households at both ends of the wealth scale were more likely to engage in off-farm income generating activities and crop farming, than those in the middle. Households with smaller herd sizes and relatively large households were those most engaged in salaried work, casual labour and business activities. Thus, households with fewer livestock had often been largely divested of their pastoral livelihood. The push out of a nomadic pastoralism and pull towards sedentarisation and diversification was eloquently described by an elderly Ardo: "there is no future in sending animals into the wilderness. The future for nomadic pastoralism is bleak. If we do not learn how to grow crops for our own consumption and forage, the big farmers with big farms will remain only and nomads will be boxed out of their livelihoods". It is becoming increasingly difficult for Fulani men to practice transhumance in Nigeria. Younger Fulani were less sentimental towards nomadic life and more pragmatic concerning income generation. There is still prestige in having large cattle herds, but younger Fulani are open-minded about combining cattle herding with other sources of income, as one young man explained: "us youngsters are less motivated to have a very large herd, we are happy to get by growing crops for our families". The pattern of wealth, and income distribution, among African societies dependent on animal husbandry, is one of inequality [40]. Insufficient attention has been paid to the disparities in livestock ownership and wealth differentiation [13,41,42,43]. Economic inequality among pastoralists, arises from historical internal dynamics and unequal access [44]. It is important to distinguish between the distribution of livestock and wealth between households and the mechanisms which prevent permanent inequalities, such as transfer of assets and limitations on herd size imposed by family labour [45]. A geographical wealth bias was observed in KGR with one particular block of wealthier, long established settlers living near the central market and a poorer group of settlers living further away, with unfavourable access to transhumance routes and grazing reserve amenities. At inception, the reserve was divided into 6 blocks, sprayed with insecticide and declared tsetse-free to encourage pastoralists to settle. While the division of KGR into blocks is administrative, the KGR community regard the blocks as separate and distinct entities, referring to themselves as 'inhabitants of Block 1' or 'inhabitants of Block 2'. Early settlers established holdings in Blocks 1 and 2A, perceived to have the best land, the best access to transhumance routes and be best served in terms of infrastructure. Inhabitants settled according to clans and new families will settle close to relatives of the same clan. In this study, Blocks 1 and 2A contained many prosperous households and were inhabited by a large proportion of 'first settlers', referred to as the 'community elite'. They are regarded as the wealthiest members of the KGR, with large cattle herds and many wives; considered to have supremacy over the rest of the KGR community, and exercise power by living in the most advantageous location. Most community leaders, including the District Head, the chief or representative of KGR district, live in these blocks, as described by one respondent: "our fellow herdsmen who have been here the longest were from wealthy clans and were able to maintain or to build up their herds better than those that came after". These blocks now enjoy a prime location next to the main access road for KGR, proximity to the market, schools and other amenities (including health care), water access (boreholes and dams) and relatively large stretches of cleared woodland for crop farming. Block 2A is considered urban in character and the heart of the KGR community. In contrast, households in Block 2B are located along a poor road leading to the Kaduna River. Households in Block 2B are regarded as the poorest members of the KGR community typically having small cattle herds. These community members live in the most inaccessible, inhospitable and remote part of the reserve, in part due to members being of a 'poor' clan and by virtue of their poverty. This area is furthest away from the transhumance corridor, making it difficult to take cattle out of the reserve for grazing. Proximity to the river presents a higher risk of trypanosomiasis in cattle, confirmed by a recent epidemiological survey [46]. Keeping herds close to watering points or hydrological networks has previously been identified as a risk factor for trypanosomiasis in KGR [47]. The opinions and attitudes of individuals in the community reflected these differences: "if we go round the Fulani settlements in the KGR we will observe that not every Fulani household is endowed with a large cattle herd, as wealth of animals is something God gives to whom he will". Heterogeneity was observed between old settler and new immigrant households. Household and herd size for the new immigrants who sought sanctuary in KGR during the political clashes in May 2011, were significantly larger than observed for the old settlers. Almost all herds of over 300 cattle belonged to new immigrants. One such household owned 1,500 cattle, 80 sheep, 2,000 chickens and comprised 277 people (a TLU/capita of 3.89 placing it in a medium wealth category). The HHH had been married 5 times. The reason for new immigrants having larger herds and households was explained by KGR inhabitants as being a result of better grazing conditions in the areas in which these households had been living previously: "when the new immigrants fled from the violence they had to bring all their animals with them and their herds are larger than the ones we are used to here because of the abundant grasses in their places of origin as compared to the limited grazing available in KGR". In response to a lack of grazing in the reserve around 40% of households also still adhered to the traditional practice of taking their KGR-based herds on transhumance at least once a year, taking cattle north during the wet season and south during the dry season travelling between 40-80 km and 20% of households seek pasture within 20-30 km of the reserve. Some 40% of KGR households also maintained permanent sub-herds outside of the reserve [11,34]. The smaller herd sizes maintained by old settlers may be related to the level of sedentarisation. Herds that moved longer distances between seasons were much larger than herds that did not move or moved only short distances. Herd owners settled in the reserve for a shorter period had considerably larger herd sizes than those who had been in their present settlement for longer as previously reported [8]. Herd size slowly decreased over the first 10 years of settlement in the reserve, after which period herd sizes reduced sharply. By October 2011, households in the reserve reported high volumes of sales indicating that residents, in particular the new immigrants were divesting themselves of some of their animals. In the current climate of political instability, reduced opportunities for herd mobility and poor access to pasture, many KGR residents believe that a shift from a purely pastoral to an integrated cattle rearing and crop farming system is a way in which KGR households can become more resilient. The early settlers, regarded as the elite and wealthiest members of the community owe their success to their ability to embrace crop farming: 'when the early settlers decided to move here, they had to move with their dependents who did not have herds of their own, this stimulated them to go into crop farming as a way of reducing pressure on the needs of the household'. Poverty in KGR is attributed by community members to the decline of crop farming within KGR, the result of an increasing cattle population and the constraint of not having enough skilled labour: 'outside KGR where there are other settlers you have the privilege of hiring labour to work on your farm. The people in KGR are Fulani whose expertise is animals, not farming, and other tribes do not come in here'. The solution, according to the community, is not to co-habit with other communities but for the Fulani man to develop the technological know-how and skills to grow crops for his family and forage for his animals. The 'Lawol-Bote' dairy producers, a dairy cooperative in KGR, also had some very clear ideas about how the community needs to adapt to changing economic conditions. Their opinion is that fodder banking (cultivation of high yielding and drought resistant pasture crops which are then stored and fed to cattle during the dry season) is one of the ways to face the current challenges. This sentiment was shared by then current KGR Project Officer and State representative of the reserve: 'we need the enlighten pastoralists on constant movement with animals and teach them how to make fodder banks'. The president of the cooperative also reported owning bulls of Friesian breed, which are crossed with the local White Fulani females to produce a crossbreed which can produce more milk whilst remaining adapted to local conditions. Shifting to a system focused on milk production through genetic breed improvement was perceived as a potential route to dairy specialisation as a way of increasing resilience, but this needs to be implemented in parallel to organised milk collection schemes and infrastructure provision (refrigerated bulk milk tank, vehicles to transport milk to milk processing plant). The NLPD (National Livestock Development Project, Federal government) is focused on reducing the number of animals and increasing their productivity to reduce competition for limited resources and negative impacts on the ecosystem, thus it is promoting intensification and specialisation of cattle production. The federal government recognises that 'the state needs to do a lot more as the infrastructure of grazing reserves is decaying and cannot accommodate the increasing cattle population'. Governmental neglect has encouraged Fulani to establish cooperatives that campaign and seek funds to promote Fulani interests. One such cooperative in the KGR is a woman's milk cooperative called the 'rise of dawn', formed to improve milk marketing opportunities within the KGR. During a FGD, its female members mentioned that a company called 'Milkopal' from Kaduna used to collect and purchase the milk produced by the community. Unfortunately this scheme collapsed and after years of waiting for the state to replace it, the women took their fate in their own hands and secured funds as part of the Kaduna Agricultural Development Project. This project has built a refrigerated bulk milk tank in Tampol, to improve opportunities for milk marketing. In the absence of a complementary milk collection scheme to take the milk outside of the KGR to areas where demand is high, the impact of this scheme is uncertain. --- Conclusion While cattle remain the principal source of Fulani income and wealth in KGR, inhabitants of the reserve have diversified their livelihood strategies in response to their changing circumstances. There was a clear association between cattle holdings, number of marriages, household size, and overall wealth, with no tapering of livestock holdings per person to a threshold. A geographical wealth bias was also observed, with wealthier, long established settlers living near the central market and a poorer group of settlers living further away, with unfavourable access to transhumance routes and grazing reserve amenities. The limited availability of grazing within the reserve and continuing political insecurity outside of the reserve are stressing pastoralist communities, resulting in the maintenance of smaller livestock holdings, pushing households into poverty and increasing diversification. Proposed strategies for further adaptation to changing economic conditions have included fodder bank schemes and shifting to a more dairy based economy (increased milk production capacity through genetic improvement of cattle) and these have been explored to some extent in the KGR. The future of the KGR's established residents and its new immigrants will ultimately depend on their resilience and ability to adapt as already evidenced by their existing adoption of a mix of livelihood and cattle management strategies. --- All relevant data are within the paper and its Supporting Information file. All data are contained within Ducrotoy MJ. Livelihoods of Fulani pastoralists and burden of bacterial zoonoses in the Kachia grazing reserve, https://www.era.lib.ed.ac.uk/ handle/1842/15912. --- Supporting information S1

A mixed methods study was undertaken in the Kachia Grazing Reserve of northern Nigeria. Surveys in March, June and October 2011 included focus group discussions, key informant and in-depth household interviews, concerning livelihood practices, animal health, ownership, and productivity. In May 2011, 249 Fulani families fleeing post-election violence entered the reserve with their livestock, increasing the number of households by one third.Despite being settled within a grazing reserve, over half of households sent all their cattle away on seasonal transhumance and another third sent some away. Cattle accounted for 96% of total tropical livestock units (TLU), of which 26% were cattle kept permanently outside the reserve. While all households cited livestock as their main source of income, 90% grew crops and 55% derived income from off-farm activities. A multiple correspondence analysis showed that for each extra member of a household its TLU value increased by 2.0 [95% CI, 1.4-2.7], while for each additional marriage its TLU increased by 15.7 [95% CI,. A strong association was also observed between small herds, small households with only one wife, alongside marked geographical wealth differences within the reserve. New immigrant families had larger household sizes (33) and livestock holdings (122 TLU) than old settlers (22 people and 67 TLU). Prior to the mass immigration, the distribution of TLU per person was unimodal: 41% of households were classified as 'poor' and 27% as 'medium', whereas post-immigration it was bi-modal, with 26% classified as 'very poor' and 28% as 'medium'.

Background The coronavirus disease 2019 (COVID-19) has become a public health emergency worldwide. Among approximately one million laboratory confirmed COVID-19 cases in Sweden, over 57,000 have been hospitalised and more than 14,000 COVID-19 related deaths have been confirmed (up until May 15 th, 2021). Mechanisms explaining a higher vulnerability to severe COVID-19 have been linked to inflammation characterized by increased levels of several pro-inflammatory cytokines and the inflammasome [1]. In turn, this has resulted in an inter-individual variation in severity of COVID-19 infection, so that, for example, older age, male gender and one or more comorbidities have been associated with increased risk for hospitalization and mortality due to COVID-19 [2][3][4]. Also, lifestyle-related factors have been linked to COVID-19 severity. Overall/central obesity and hypertension were the first and most frequently reported factors found to be more prevalent in individuals who were hospitalized or died due to COVID-19 [5][6][7][8]. In later papers, physical inactivity has been linked to severe COVID-19 risk [9,10] and in a small sample of men and women, lower cardiorespiratory fitness (CRF) has also been associated with a higher risk of hospitalisation for COVID-19 [11]. Only a few studies have investigated the importance of lifestyle factors on severe COVID-19 risk, and it is plausible that, based on previous knowledge, a healthy lifestyle before infection may reduce the risk of severe COVID-19. A positive impact on inflammation and the immune system is one possible mechanistic pathway [12,13], as low-grade inflammation is considered to be a strong causal factor for chronic diseases such as cardiovascular disease and cancer [14]. Also, the possible impact of a healthy lifestyle on other risk factors, such as overweight/ obesity and hypertension [6,15], may induce protection against severe COVID-19, and regular physical activity (PA) has been suggested as a protective non-pharmacological tool against COVID-19 [12,16]. However, the research underpinning these assumptions is limited, as are previous studies looking at the importance of and the interaction between different lifestyle-related factors for COVID-19 severity. Apart from the above predictors, lower socioeconomic status (assessed as, for example, educational level, income or area of residence) has been related to more severe COVID-19 [2,17,18]. The subsequent severity of the COVID-19 infection may, however, not only be explained by structural socioeconomic factors, but also by more unfavourable lifestyle habits and poorer health status before infection in individuals with lower socioeconomic status [19,20]. If and how lifestyle mediates some of the associations seen between socioeconomic factors and severe COVID- 19 has not yet been investigated. The identified knowledge gaps above are addressed in the present study, with the main aim being to study a wide span of lifestyle-related and socioeconomic factors as potential predictors of severe COVID-19, and with special focus on CRF. Secondary aims are to study whether CRF may attenuate obesity-and hypertensionrelated risk of severe COVID-19, and whether lifestyle-related factors mediate the associations between socio-economic factors and severe COVID-19 risk. The hypotheses are that lifestyle-related (in particular CRF) and socioeconomic factors can predict severe COVID-19, and that variations in lifestyle-related factors mediate a large proportion of the risk of severe COVID-19 associated with socioeconomic factors. --- Methods The study is a nested case-control study based on data from the Health Profile Assessment (HPA) database (www. hpi. se). HPAs have been carried out in health services all around Sweden since the middle of the 1970s and is offered to all employees working for a company or an organization connected to occupational or healthrelated services. An HPA includes a questionnaire about lifestyle and health experiences, measurements of anthropometrics and blood pressure, estimations of maximal oxygen consumption (VO 2 max) from a submaximal cycle ergometer test, and a person-centred dialogue with a HPA coach. In February 2021, a total of 407,131 HPAs between 1992 and 2020 were available in the database, and the database was linked to national registries with data on severe COVID-19 (defined as hospitalization, intensive care or death due to COVID-19) using the unique Swedish personal identity number. A total of 857 (0.2%) confirmed cases with severe COVID-19 were identified, including COVID-19 hospitalization (n = 547, 0.1%), intensive care (n = 172, 0.04%) and death (n = 138, 0.03%). Controls were recruited from the same HPA database. All deceased controls before 2019-12-31 according to the national cause of death registry were excluded. To minimize internal drop-out, only participants without severe COVID-19 and with valid data on sex, age, educational level, CRF, body mass index (BMI), exercise and smoking were eligible as controls (a total of n = 278,598). Eligible controls were in general more often women, older, and had higher CRF and lower BMI (see overview of included and excluded participants in Additional file 1). The study was approved by the ethics board at the Stockholm Ethics Review Board (Dnr 2020-02,727). Informed consent was obtained from the participants prior to participation in the HPA. It was not possible to involve participants or the public in the design, conduct, reporting or dissemination plans of our research, due to its retrospective design. --- Estimation of VO 2 max Measurement of CRF as actual maximal oxygen uptake (VO 2 max), using a graded test to exhaustion, is limited in the general population for numerous reasons including health risks in non-athletic populations and dependence on laboratory equipment and expertise. Therefore, CRF was assessed as estimated VO 2 max (estVO 2 max) from the standardized submaximal <unk>strand cycle ergometer test in L•min -1 and also expressed in relative values (ml•min -1 •kg -1 ) [21]. Previous validation studies on adult population samples show small and non-significant mean differences on a group level (-0.07 L•min -1 95% CI -0.21 to 0.06) between estVO 2 max from the <unk>strand protocol and direct measured VO 2 max during treadmill running with an absolute error and coefficient of variance similar to other submaximal tests (SEE = 0.48 L min -1, CV = 18.1%) [22]. To minimize well-known errors with submaximal testing, participants were requested to refrain from vigorous activity the day before the test, consuming a heavy meal and smoking/using snuff three hours and one hour respectively before the test, as well as avoiding physiologic and emotional stress prior to the test. The participant cycled on a calibrated ergometer at an individually adapted submaximal work rate (aiming at a rate of perceived exertion of "Somewhat hard", 13-14, on the Borg RPE scale) for 6 min to achieve a steady-state pulse assessed during the last minute of cycling. Using the steady-state pulse and the work rate, VO 2 max was estimated from a sex-specific nomogram, with corresponding age-correction factors [21]. --- Other measurements Body mass was assessed in light-weight clothing using a calibrated scale and to the nearest 0.5 kg. Body height was assessed to the nearest 0.5 cm using a wall-mounted stadiometer. Body mass index (BMI) (kg•m -2 ) was subsequently calculated. Central obesity was assessed as waist circumference and measured to the nearest 0.5 cm with a tape measure at the midpoint between the top of the iliac crest and the lower margin of the last palpable rib in the mid axillary line after normal exhalation. Systolic and diastolic blood pressure (BP) were measured manually by the standard auscultation method in the seated position after 20 min of resting. --- Self-reported and register data Exercise, commute type, physical work situation, diet habits, alcohol habits, smoking, overall stress, and perceived health were self-reported (see Additional file 2). Highest educational attainment, occupation, income, civil status, and data on country of birth at the time for the HPA were obtained from Statistics Sweden by linking of the participants' personal identity numbers. Educational attainment was collected from the Swedish education nomenclature 2000 and was categorised into three categories: Elementary school, High school/Vocational education, and University. Each occupation is labelled and defined by a four-digit code according to the Swedish Standard Classification of Occupation [23]. In the present study, occupations were further aggregated according to the first digit into white-collar high-skilled (Major group 1-3), white-collar low-skilled (Major group 4-5), blue-collar high skilled (Major group 6-7) and blue-collar low-skilled (Major group 8-9). For a more detailed description see publication [24] by Väisänen et al. Total income from employment for the specific year was derived from the Income and Taxation Register. Cases of chronic disease before 2020 were ascertained through the Swedish national patient registry using ICD-coding; C00-D48 to define tumour disease, E10 to E14 for diabetes, I10 to I15 for hypertension, I26 to I28 + J44 to J45 for lung disease and I20 to I25 + I30 to I52 + I60 to I69 for cardiovascular disease. --- Severe COVID-19 surveillance The main outcome in the present study is severe COVID-19, which is defined as either hospital admission, admission to ICU and/or death due to COVID-19. Participants' personal identity numbers were linked to national registers. Data on hospitalization was obtained from the Swedish National Patient Register, data on intensive care was obtained from the Swedish Intensive Care Registry, and data on death was obtained from the National Cause of Death Register. --- Statistics Main analyses include cases with sex-and age-matched controls, as the risk of severe COVID-19 is strongly associated with male gender and higher age. Sensitivity analyses using unmatched controls are included in supplementary material (see Additional file 3). For matched analyses, each case was matched to four eligible controls out of the 278,598 eligible controls in the HPA database, with no tolerance in variation between sex or age (per year) between the case and the controls and without replacement of controls in the sampling. For unmatched analyses all eligible controls were used. To compare descriptive characteristics of cases and controls, paired t-test (continuous data), Cochran's Q test (nominal data) and Related-Samples Friedman's Two-Way Analysis of Variance by Ranks (categorical data) were used, and effect sizes as Cohen <unk>s d is presented for continuous data. To compare descriptive characteristics between cases of severe COVID-19, chi-square test of independence with false discovery rate correction for multiple testing (categorical data) and ANCOVA (continuous data) were used. Logistic regression modelling was used to calculate odds ratio (OR) with 95% confidence intervals (95% CI) for different predictors of serve COVID-19. Three models were used and adjusted for an increasing number of variables (see under each table/figure). Model 1 included sex, age and performed year of HPA. Model 2 additionally adjusted for educational level, civil status and country of birth, and Model 3 also adjusted for CRF, BMI, number of chronic diseases, exercise habits, smoking and overall stress. Due to internal drop-out for variables included in Model 2 and 3, two Model 1's are presented; one Model 1 with OR and 95% CI including all available individuals (labelled Non-complete data in the tables, referred to as Model 1-nc), and a second Model 1 including only individuals with complete data for all adjusting variables (labelled Complete data for all adjusting variables in the tables, referred to as Model 1-c). Further, BMI, waist circumference, blood pressure, estVO 2 max and income from employment were analyzed as continuous variables as well as after aggregation; BMI was aggregated into normal weight <unk> 25, overweight 25-29.9, obesity 30-34.9 and severe obesity <unk> 35 kg•m -2 ; large waist circumference was defined as <unk> 88 cm for women and <unk> 102 cm for men, (both BMI and waist circumference were categorized according to recommendations by the world health organization [25]); high systolic and diastolic BP were defined as <unk> 140 mmHg and <unk> 90 mmHg, respectively; estVO 2 max was arbitrarily categorised based on multiples of one METs (3.5 ml•min -1 •kg -1 ) into <unk> 25 (very low), 25-<unk> 32 (low), 32-<unk> 46 (moderate) and <unk> 46 (high) ml•min -1 •kg -1 ; and total income from employment into quartiles, percentile 25 = 281,143, percentile 50 = 362,718, percentile 75 = 479,764 Swedish crowns. All above analyses were performed using IBM SPSS (V.26.0.0.1) and Jamovi (The jamovi project (2021) Version 1.6. retrieved from https:// www. jamovi. org). Marginal effects plots for severe COVID were calculated by setting the covariates at a mean (for continuous variables) or average (for factor variables) level while varying the focal variables, using R (R Core Team, 2021) and the packages Tidyverse [26] and ggeffects [27]. Mplus version 8.6 [28] was used to estimate Bayesian parallel mediation models linking socioeconomic indicators to severe COVID-19 via multiple mediators (Fig. 2). Separate models were estimated for each socioeconomic indicator. The highest socioeconomic category was used as the reference category in each model. CRF, BMI, exercise, and stress were treated as continuous variables whereas smoking was dichotomized into daily smoker or never/seldom smoking. We used the proportion of the total effect that is mediated as an effect size measure [29]. The proportion mediated by each mediator was calculated by dividing the specific indirect effect by the total effect (a 1 b 1 ) / (a 1 b 1 + a 2 b 2 + a 3 b 3 + a 4 b 4 + a 5 b 5 + c'). The total proportion mediated was calculated by dividing the sum of the indirect effects by the total effect (a 1 b 1 + a 2 b 2 + a 3 b 3 + a 4 b 4 + a 5 b 5 ) / (a 1 b 1 + a 2 b 2 + a 3 b 3 + a 4 b 4 + a 5 b 5 + c'). Models were estimated using four Markov chain Monte Carlo chains and a minimum of 50,000 iterations. The first half of the iterations were discarded as burn-in and the remaining iterations were used to estimate the posterior distribution of the parameters. A stable potential scale reduction factor (PSFR) close to 1 was considered as evidence of chain convergence alongside inspection of trace plots and autocorrelation plots. Indirect effects were evaluated using 95% highest posterior density (HPD) credibility intervals [30]. The credibility interval indicates the probability that the parameter lies between the lower and upper bound of the interval [31]. If an interval did not include zero, the indirect effect was credible. The default non-informative prior specification in Mplus was used. --- Results --- Characteristics of cases and controls In the matched analyses, 857 cases of severe COVID-19 and 3426 matched controls were included (for one case, only two exact matched controls were identified). Mean age was 49.9 years (SD 10.7) and 70.4% (n = 603 cases and n = 2 412 controls) were men. The median year that the HPA was performed was 2012 (Q1 2008, Q3 2016) for controls and 2011 (Q1 2006, Q3 2016) for cases. In the unmatched analyses, the mean age for all eligible controls was significantly lower compared to cases (43.7 years (SD 11.6), p <unk> 0.001), and with a significantly lower proportion of men (53.8%, p <unk> 0.001) compared to the matched analyses. The median year that the HPI was performed for all eligible controls was similar to the matched controls 2012 (Q1 2007, Q3 2016). There were several differences between cases and matched controls for established COVID-19 risk factors, such as cases having higher BMI, blood pressure and presence of comorbidities as well as greater waist circumferences (Table 1). Cases also demonstrated significantly lower estVO 2 max and more unfavourable exercise patterns. There were also several differences in terms of educational level, country of birth, dietary habits and self-rated health. Moreover, cases with more severe complications from COVID-19 (death vs intensive care or hospitalization, and intensive care vs hospitalization) had significantly lower estVO 2 max, higher BMI, greater presence of comorbidities and were more often daily smokers (see Table 2). --- Impact of lifestyle-related characteristics Four models were used to quantify independent associations between potential lifestyle-related predictors and severe COVID-19 in the matched analyses, where two Model 1's (Model 1-nc and Model 1-c) enabled comparative analyses with non-complete and complete data for all adjusting variables in Model 2 and 3, see Table 3. In terms of CRF, there was a graded increase in odds with lower compared to high levels, OR = 1.62 (95% CI, 1.00 to 2.62) for moderate CRF (32 to <unk> 46 ml•min -1 •kg -1 ) and an approximately two-fold increased odds for low (OR = 2.02, 1.22 to 3.35) and very low fitness (OR = 1.91, 1.09 to 3.34), respectively (Table 3). Similarly, being overweight was associated with two-fold increased odds compared to normal weight (OR = 1.98, 1.53 to 2.56), and obesity and severe obesity was associated with three-fold increased odds (OR = 2.94, 2.13 to 4.07 and OR = 2.98, 1.80 to 4.94 respectively). A larger WC was associated with higher odds in the fully adjusted model, OR = 1.75, 1.20 to 2.55. Presence of chronic disease had a graded increase for every additional diagnosis, OR = 1.88 (95% CI: 1.44 to 2.45) for one chronic disease, and OR = 4.55 (1.83 to 11.33) for 4 to 5 chronic diseases. Neither high systolic nor diastolic blood pressure remained significantly associated with severe COVID-19 after multiadjustment. Reporting daily smoking (OR = 0.60, 0.41 to 0.89) as well as a high level of stress (OR = 1.36, 1.001 to 1.84) were significantly associated with severe COVID-19 in the fully adjusted model. Figure 1 presents the predicted probability of severe COVID-19 according to continuous levels of CRF, overall and central obesity (BMI and WC), and systolic and diastolic blood pressure. All obesity and blood pressure measures were attenuated by adjustment for lifestyle variables and CRF, however, these associations with severe COVID-19 remained significant. --- Impact of sociodemographic factors Low education (elementary vs. university) predicted severe COVID-19 (OR = 1.81, 1.28 to 2.54) after multiadjustment, as did being born outside Sweden vs. being born in Sweden (OR = 2.58, 1.97 to 3.38) (Table 4). No significant associations were seen for civil status, occupational groups, or income quartiles. In additional sensitivity analyses using unmatched controls (see Additional file 3, Supplement Tables 1 and2) the odds for severe COVID-19 were higher in men (Model 3, OR = 1.97, 1.62 to 2.40) and with increasing age (Model 3, per year OR = 1.02, 1.01 to 1.03). Individuals <unk> 70 years and 60 to 69 years had higher odds compared to those <unk> 60 years. However, the odds were attenuated by additional adjustments for lifestyle-related factors (Model 2 and 3 adjustment). Further, lifestylerelated and sociodemographic predictors showed similar associations as in the matched analyses, with only marginal variations in both OR and CI, which did not alter the results or conclusions of the unmatched analyses. --- Mediation analyses The mediation analyses are summarized in in the highest socioeconomic category, lower socioeconomic status was related to an increased risk of severe COVID-19 through BMI and CRF, and a lower risk for severe COVID-19 through smoking. The findings were similar across all three socioeconomic indicators. The total proportion mediated across all five mediators ranged from 49 to 86%. Because the mediation models were inconsistent (i.e., they included both positive and negative effects on the dependent variable), we calculated the proportion mediated based on absolute values [29]. Thus, these values represent the proportion of the absolute total effect that was mediated. Mediation analyses using an unmatched sample (N = 279,455) showed similar results (see Additional file 3, Supplement Table 3). --- Discussion The main results of the present study include strong associations of several lifestyle-related risk factors, including CRF, overweight/obesity, perceived stress, and high blood pressure, with severe COVID-19, even after adjustments for sociodemographic factors and previous diseases. Among patients with severe COVID-19, those with more severe COVID-19 (death vs intensive care or hospitalization, and intensive care vs hospitalization) had lower CRF. In mutually adjusted analyses, higher CRF attenuated some of the risks related to both obesity and hypertension. Low educational level, low income as well as blue collar/low skilled occupations were associated with increased risk of severe COVID-19. However, these associations were, to a relatively large proportion, mediated by CRF, BMI and smoking. The results were consistent when using matched or unmatched controls. --- Comparison with other studies This is, to our knowledge, the first study investigating the association between a wide variation of lifestyle-related risk factors, including CRF, and severe COVID-19. The results are consistent with the few previous existing studies that have found evidence of associations between PA, overweight/obesity and smoking with severe COVID-19 [9,10]. In 387,109 middle-aged men and women from the UK Biobank, there were 760 cases of hospitalization for COVID-19 [9]. After multi-adjustment, participants reporting no regular PA had a 32% higher risk of hospitalization for COVID-19 compared to those reporting some PA (active but below guidelines) or meeting activity guidelines (<unk> 150 min per week of moderate PA or 75 min per week of vigorous PA). Smoking, compared to not smoking, was associated with a 42% higher risk. Further, a lifestyle score was derived including both physical inactivity, smoking, heavy alcohol consumption, and overweight/obesity, which showed a dose-dependent increased risk of hospitalization for COVID-19 partly explained by C-reactive protein levels. Moreover, in 48,440 adult patients with a COVID-19 diagnosis, those who had been consistently inactive (0-10 min of PA per week) in the two years preceding COVID-19 infection, had a significantly higher odds of hospital admission (OR 2.26), admission to intensive care (OR 1.73) and death (OR 2.49) due to COVID-19 compared to patients reporting being inconsistently (10 to <unk> 150 min per week) or consistently (<unk> 150 min per week) moderately to vigorously physically active [10]. Although the present study did not find a significant association between PA levels and severe COVID-19, the strong and consistent association of CRF and COVID-19 may be even more important. All previous studies have relied on self-reported PA, which is a subjective measure of recent PA levels containing wellknown errors (recall-bias) that permit valid analyses on mainly aggregated PA levels [32]. In the present analyses, CRF was included as a more objective measure of recent PA as well as an indicator of the status of the cardiorespiratory system. This showed a lower risk of severe COVID-19 per ml•min -1 •kg -1 with a doubling of risk between the two lowest and the highest categories (<unk> 32 ml•min -1 •kg -1 and <unk> 46 ml•min -1 •kg -1 respectively). This is similar to a previous report on all-cause mortality and CVD morbidity risk, where decreases of 2.3% and 2.6% per ml increase in estVO 2 max were seen [33]. Only one previous study has studied the association between recent CRF and COVID-19. In a small sample of patients (n = 246) with positive tests for COVID-19, men (but not women) with lower CRF were more likely to be hospitalized than those with higher CRF [11]. A study using data from military conscript (<unk>18 years of age) between 1968 and 2005 showed that high CRF at conscript was associated with lower odds of severe COVID-19 later in life [34]. In the fully adjusted analyses, both perceived stress and smoking remained significantly associated with severe COVID-19. Reporting high overall stress was associated with significantly higher OR (1.36) compared to low stress. This is partly supported by findings from the UK Biobank participants [17] where a 58% increased risk of hospitalization due to COVID-19 was found among individuals reporting high psychological distress. In contrast to the present results, the association did not remain after full adjustment with comorbidities, other lifestyle variables and socioeconomics. More surprisingly in the present study, smokers had a significantly lower OR (0.60) compared to non-smokers, which adds to equivocal results in the current literature [4,9]. A hypothesis has been raised that nicotine may have beneficial effects on COVID-19 due to its interaction with the renin-angiotensin and effects on the immunomodulatory system [35], but further investigation of the mechanisms associated with these findings remains to be elucidated by better controlled studies. Consistent with other publications [5,6], both overweight and obesity were associated with a higher risk of severe COVID-19. This could partly be explained by a higher prevalence of metabolic risk factors and lowgrade inflammation in overweight/obese individuals, as these have been identified as central mechanisms for a higher vulnerability to severe COVID-19 [36]. Interestingly, a recent paper including over 17 million individuals found similar associations between commonly accepted risk factors (age, male sex, deprivation, obesity, and some comorbidities) for non-COVID (including CVD, cancer, dementia etc.) deaths and for COVID-19 deaths, suggesting that COVID-19 largely mirrors existing risks faced by patients [37]. However in the present study, obesity-risk was at least partly attenuated by CRF. Attenuation by CRF were also seen for central obesity (waist circumference) and high systolic and diastolic blood pressure related risks. These findings are highly clinically relevant and in line with previous studies on cardiovascular disease risk and premature death, where "fat but fit" individuals had significantly better prognoses for cardiovascular outcomes and mortality compared to obese but unfit individuals [38,39]. There are several suggested mechanisms for the beneficial effects of regular PA and higher CRF levels on both COVID-19 severity per se, as well as attenuation of the obesity-and hypertension-related risks [12,13,16,40]. One is the lower prevalence of obesity and hypertension in more active individuals [6,15]. Moreover, regular exercise induces a marked increase in several anti-inflammatory cytokines, counteracting the low-grade inflammatory state present in many chronic metabolic diseases (such as obesity and type 2 diabetes) [12,13]. It also induces a natural immune-protection against more severe COVID-19 by reducing the so-called "cytokine storm" (peaking of pro-inflammatory cytokines including interleukin-6 and tumour necrosis factor-alpha) that ICU-patients with severe COVID-19 experience [41,42]. Also, regular PA has shown a direct and positive effect on lung function, and the antibody concentration after vaccination is higher among regularly physically active individuals [40]. There were differences in sociodemographic factors between cases and controls in the present study, which is consistent with previous studies. Among 431,051 British adults, low levels of education, income and area deprivation doubled the risk of hospitalization due to COVID-19, with a 39% higher risk for those with occupations including personal service and sales compared to managers [17]. Across 3135 US counties, the counties with a higher percentage of households with poor housing had a higher incidence of COVID-19, as well as mortality due to COVID-19 [18]. These findings are supported by a large Swedish study, indicating that an educational level only up to elementary school, compared to higher educational levels, was associated with a higher risk for both intensive care and non-intensive care hospitalisation due to COVID-19 [2]. Also, bluecollar workers were significantly less likely to work from home or to change commuting habits in relation to the COVID-19 pandemic, compared to white-collar workers [43]. However, as health status prior to infection seems to heavily impact the severity of COVID-19, we hypothesized that the variation in health lifestyle factors would mediate some of the risk associated with socioeconomic factors. In the mediation analyses, lower socioeconomic status (indicated by education, income, and occupation) was related to an increased risk of severe COVID-19 through higher BMI and lower CRF, whereas lower socioeconomic status was related to a lower risk of severe COVID-19 through smoking. Similar mediation analyses have been performed for cardiovascular disease [19] and cancer morbidity and mortality [20], where modifiable factors including BMI and smoking explained between 42 and 46% of the association between low socioeconomic position and the outcomes. The proportion mediated in the current study ranged from 49 to 86%, indicating that the mediators accounted for a relatively large proportion of the association between socioeconomic factors and risk of severe COVID-19. The present indirect effects on severe COVID-19 risk through BMI and CRF highlight factors that could be targeted in interventions to strengthen the resilience for future severe infections. --- Strengths and weaknesses of the study A case-control study is not as powerful as other types of studies in confirming a causal relationship [44]. However, the strengths of this study are the large cohort of different aged women and men with variations in socioeconomic gradients, and the available data on several lifestyle-related factors assessed by standardised methods. Another strength is the highly corresponding results obtained using either the sex-and age-matched controls or all eligible controls in the analyses. The mediation analyses are also a strength, as they highlight processes through which socioeconomic inequalities may influence disease risk. In Sweden, patients from both low and high socioeconomic status have similar access to healthcare, which strengthens the argument for the role of lifestyle factors, including CRF, in preventing severe COVID-19. Limitations of the study include self-reported data regarding lifestyle habits, which risks recall bias [45]. However, questionnaires with categorical answer modes as used in the present study have been reported to provide superior validity compared to open answer modes for PA level [46]. The study design explores associations over time, but does not give information about causality, in this case between lifestyle related and socioeconomic risk factors and severe COVID-19. Moreover, the clinical status of the cases and controls between the time of their HPA and the follow-up period were not monitored. There is a risk of reversed causality due to individuals with a better health status possibly having higher CRF, lower BMI and lower blood pressure. However, the size of the study population made it possible to identify the effect of low CRF, obesity and elevated blood pressure by adjusting for multiple potential confounders and thereby reducing the risk of reverse causality. --- Conclusions Higher CRF was associated with better resilience for severe COVID-19, which is of great clinical value, particularly for high-risk individuals with obesity and/or hypertension. Further, the mediation analyses included in the present paper add important initial evidence of modifiable factors mediating the associations between socioeconomic variables and severe COVID-19. This should shift the focus from structural factors, such as educational level or income per se, having direct effects on disease risk, to instead highlighting and targeting modifiable factors, including CRF and BMI, to increase resilience. This is particularly important as a decrease by 10% (4.2 ml•min -1 •kg -1 ) in CRF has been reported over the last two decades in the Swedish working population [47]. This has been confirmed in international data [48]. During the same time period, the prevalence of obesity and severe obesity has increased by 153% and 86%, respectively [49]. This calls for an urgent need to implement interventions, such as PA on prescription, to increase CRF, preferably specifically targeting high-risk individuals. Further analyses on how sex and age moderate the association between CRF and severe COVID-19 are needed, as are studies including objective measures (e.g. accelerometers) for assessment of PA patterns. Although the mediation analyses highlight processes through which socioeconomic inequalities may influence disease risk, given the correlational nature of the data, these findings need to be replicated in future studies using designs that allow for stronger causal conclusions. --- Availability of data and materials The datasets generated and/or analysed during the current study are not publicly available due to them being the property of the HPI Health Profile Institute, but are available from the corresponding author, [email protected]. Additional information regarding technical details, statistical code, and derived data are also available from the corresponding author. --- Abbreviations --- Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s12966-021-01198-5. Additional file 1: Supplement figure 1. Flow chart of included and excluded participants. Contains a flow chart of included and excluded cases and controls. Additional file 2. Questions from Health Profile Assessment. Contains the questions from the Health Profile Assessment for self-reported data in the study. --- Additional file 3. Supplement Tables. Contains supplement table 1 to 3. --- Authors' information Not applicable. Authors' contributions EEB, AL, VB, LVK, BE, GA, PW contributed to the conception or design of the work. EEB, AL, DV, EH, JSE, ML, AS, TH contributed to the acquisition, analysis, or interpretation of data for the work. EEB, AL, VB, LVK, BE, EH drafted the manuscript. DV, GA, PW, JSE, ML, AS, TH revised it critically for important intellectual content. All authors critically revised the manuscript, gave final approval, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The corresponding author, EEB, is the manuscripts guarantor and attests that all listed authors meet authorship criteria, that no others meeting the criteria have been omitted and takes the full responsibility for the overall content. --- Declarations Ethics approval and consent to participate The protocols used were approved by the institutional review boards of the institutions involved in this study and all participants provided informed consent prior to data collection. The study was approved by the Stockholm Ethics Review Board (Dnr 2020-02727), and adhered to the Declaration of Helsinki. --- Consent for publication Not applicable. --- Competing interests Author GA (responsible

Background: The impact of cardiorespiratory fitness (CRF) and other lifestyle-related factors on severe COVID-19 risk is understudied. The present study aims to investigate lifestyle-related and socioeconomic factors as possible predictors of COVID-19, with special focus on CRF, and to further study whether these factors may attenuate obesity-and hypertension-related risks, as well as mediate associations between socioeconomic factors and severe COVID-19 risk. Methods: Out of initially 407,131 participants who participated in nationwide occupational health service screening between 1992 and 2020, n = 857 cases (70% men, mean age 49.9 years) of severe COVID-19 were identified. CRF was estimated using a sub-maximum cycle test, and other lifestyle variables were self-reported. Analyses were performed including both unmatched, n = 278,598, and sex-and age-matched, n = 3426, controls. Severe COVID-19 included hospitalization, intensive care or death due to COVID-19. Results: Patients with more severe COVID-19 had significantly lower CRF, higher BMI, a greater presence of comorbidities and were more often daily smokers. In matched analyses, there was a graded decrease in odds for severe COVID-19 with each ml in CRF (OR = 0.98, 95% CI 0.970 to 0.998), and a two-fold increase in odds between the lowest and highest (< 32 vs. ≥ 46 ml•min -1 •kg -1 ) CRF group. Higher BMI (per unit increase, OR = 1.09, 1.06 to 1.12), larger waist circumference (per cm, OR = 1.04, 1.02 to 1.06), daily smoking (OR = 0.60, 0.41 to 0.89) and high overall stress (OR = 1.36, 1.001 to 1.84) also remained significantly associated with severe COVID-19 risk. Obesity-and blood pressure-related risks were attenuated by adjustment for CRF and lifestyle variables. Mediation through CRF, BMI and smoking accounted for 9% to 54% of the associations between low education, low income and blue collar/low skilled occupations and severe COVID-19 risk. The results were consistent using either matched or unmatched controls. Conclusions: Both lifestyle-related and socioeconomic factors were associated with risk of severe COVID-19. However, higher CRF attenuated the risk associated with obesity and high blood pressure, and mediated the risk associated with various socioeconomic factors. This emphasises the importance of interventions to maintain or increase CRF in the general population to strengthen the resilience to severe COVID-19, especially in high-risk individuals.

SE, ML, AS, TH contributed to the acquisition, analysis, or interpretation of data for the work. EEB, AL, VB, LVK, BE, EH drafted the manuscript. DV, GA, PW, JSE, ML, AS, TH revised it critically for important intellectual content. All authors critically revised the manuscript, gave final approval, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The corresponding author, EEB, is the manuscripts guarantor and attests that all listed authors meet authorship criteria, that no others meeting the criteria have been omitted and takes the full responsibility for the overall content. --- Declarations Ethics approval and consent to participate The protocols used were approved by the institutional review boards of the institutions involved in this study and all participants provided informed consent prior to data collection. The study was approved by the Stockholm Ethics Review Board (Dnr 2020-02727), and adhered to the Declaration of Helsinki. --- Consent for publication Not applicable. --- Competing interests Author GA (responsible for research and method) and PW (CEO and responsible for research and method) are employed by HPI Health Profile Institute. The remaining authors declare that they have no competing interests. • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year --- • At BMC, research is always in progress. --- Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research? Choose BMC and benefit from:? Choose BMC and benefit from: --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: The impact of cardiorespiratory fitness (CRF) and other lifestyle-related factors on severe COVID-19 risk is understudied. The present study aims to investigate lifestyle-related and socioeconomic factors as possible predictors of COVID-19, with special focus on CRF, and to further study whether these factors may attenuate obesity-and hypertension-related risks, as well as mediate associations between socioeconomic factors and severe COVID-19 risk. Methods: Out of initially 407,131 participants who participated in nationwide occupational health service screening between 1992 and 2020, n = 857 cases (70% men, mean age 49.9 years) of severe COVID-19 were identified. CRF was estimated using a sub-maximum cycle test, and other lifestyle variables were self-reported. Analyses were performed including both unmatched, n = 278,598, and sex-and age-matched, n = 3426, controls. Severe COVID-19 included hospitalization, intensive care or death due to COVID-19. Results: Patients with more severe COVID-19 had significantly lower CRF, higher BMI, a greater presence of comorbidities and were more often daily smokers. In matched analyses, there was a graded decrease in odds for severe COVID-19 with each ml in CRF (OR = 0.98, 95% CI 0.970 to 0.998), and a two-fold increase in odds between the lowest and highest (< 32 vs. ≥ 46 ml•min -1 •kg -1 ) CRF group. Higher BMI (per unit increase, OR = 1.09, 1.06 to 1.12), larger waist circumference (per cm, OR = 1.04, 1.02 to 1.06), daily smoking (OR = 0.60, 0.41 to 0.89) and high overall stress (OR = 1.36, 1.001 to 1.84) also remained significantly associated with severe COVID-19 risk. Obesity-and blood pressure-related risks were attenuated by adjustment for CRF and lifestyle variables. Mediation through CRF, BMI and smoking accounted for 9% to 54% of the associations between low education, low income and blue collar/low skilled occupations and severe COVID-19 risk. The results were consistent using either matched or unmatched controls. Conclusions: Both lifestyle-related and socioeconomic factors were associated with risk of severe COVID-19. However, higher CRF attenuated the risk associated with obesity and high blood pressure, and mediated the risk associated with various socioeconomic factors. This emphasises the importance of interventions to maintain or increase CRF in the general population to strengthen the resilience to severe COVID-19, especially in high-risk individuals.

Introduction The complex issues faced by people with HIV-AIDS prioritizes the provision of comprehensive nursing care [1]. In Indonesia, the number of people living with HIV-AIDS remains relatively high, with a Case Fatality Rate of 1.03% as of 2018 [2], and is often accompanied by comorbidities, ranging from infectious diseases such as pulmonary TB, accounting for 21.65% and non-communicable diseases, such as hypertension (39.17%) [3], and Kaposi's sarcoma (2.6%) [4]. This is the main cause of hospitalization in patients [3], with a prevalence of 59% of the population dominated by individuals in productive age. In addition, the cost of care provided to these patients was almost six times higher than that of the controls and four times higher after taking antiretroviral drugs [5]. HIV-AIDS, which is characterized by a progressive decline in the immune system, has a serious impact on well-being. This is evidenced by the fact that patients with other diseases, such as pulmonary TB, are still willing to be open to appearing in public [6]; however, people with HIV-AIDS are very protective in their expression [7]. This difference is driven by the general stigma attached to people with HIV-AIDS in the community, which is far more discriminatory than other diseases [8], and the belief that HIV-AIDS is a punishment from God and the sins caused by their actions can add to the burden of suffering [9]. Therefore, it is important to understand the emotional responses of people living with HIV-AIDS during therapy programs. The severity experienced by people living with HIV-AIDS does not correspond with clinical practice because nurses fail to respond adequately to their demands, thus affecting their overall well-being [10]. This discrepancy between expectations and reality experienced by these patients, as well as worsening health as a result of their illness, emerged as a very agonizing experience, as reported in this study [11]. Meanwhile, the dominance of optimizing the anticipation of loss to follow-up and strengthening self-care has eliminated problems from other psychosocial aspects, so that it can change the healthcare system, which includes bio, psycho, social, cultural, and spiritual aspects [12]. It is clear that attention to physical problems is a serious concern without considering psychosocial problems or traumatic experiences when people living with HIV-AIDS interact in the community. For a wider interest in nursing practice, provision of nursing care to people living with HIV-AIDS should focus on psychosocial problems besides the physical aspects within the framework of holistic nursing care [13]. This is because, while providing nursing care to people with HIV-AIDS, the approach taken is concentrated and centered on clinical practice [14]. Meanwhile, the "Health" paradigm toward HIV-AIDS has shifted to a view of disease as a result of human behavior, and this must be integrated into the beliefs of people living with HIV-AIDS to support their well-being, which requires knowledge and an understanding of the patient's personal experiences and the meaning of these experiences in a patient's life, as these are much more important than just clinical pharmacological monitoring [15]. There is a relationship between well-being and increased immunity [13]. Therefore, in order to improve quality of life and understand their life experiences, it is possible to develop nursing interventions that can improve their well-being [16]. People living with HIV-AIDS frequently experience stigma and discrimination, especially in the surrounding community. This triggers psychological problems, including anxiety, depression, and the risk of suicide [17]. In this approach, the focus is on the experiences of people living with HIV-AIDS, and nursing care is critical because nurses as caregivers play an essential role in every nursing action and follow-up in continuing nursing services [18]. Therefore, nurses must view people living with HIV-AIDS as a combination of bio, psycho, sociocultural, and spiritual features, without distinguishing ethnicity, nationality, race, and religion. The relationship between patients and nurses is professional [12]. It is important for nurses to develop nursing care by enhancing interpersonal relationships between them and the patients as a mutually beneficial relationship; this approach encourages people-centered holistic nursing practice [19]. In addition, in the process of caring, the depth of the patient's experience is easy to understand, and nurses obtain new possibilities and opportunities to provide nursing services according to the needs and expectations of the patient [20]. Therefore, understanding the feelings of people living with HIV-AIDS is important, and this is used for the construction of future nursing interventions to improve patients' health status and happiness. Although several qualitative studies have reported events experienced by people suffering from HIV-AIDS [21], they did not specify the relationship between nurses and clients from the client's perspective and only focused on physical and functional disorders. Therefore, this study focuses on studies on the life experiences of people living with HIV-AIDS when interacting with nurses, and the meaning of these experiences. --- Material and methods --- Participants Overall, the study involved 22 participants, and factors included age, gender, type of work, and comorbidities. Further details can be found in Table 1. --- Procedure The life experiences of people living with HIV-AIDS that are overlooked by nurses during the interaction process as meaningful experiences have been mutually agreed upon in this study. In addition, the researchers used the COREQ strategy to determine consolidation criteria in conducting qualitative searches for this study [22]. The study process followed the Consolidated Criteria for Qualitative Research Reporting checklist. Individual and in-depth face-toface interviews with people living with HIV-AIDS were conducted in a semi-structured manner to obtain complete data on life experiences when interacting with nurses [23]. In order to meet the participants, mediation was obtained from nurses who organized the HIV-AIDS Disease Prevention and Eradication program at the Health Office, and were placed in the Community Health Center to monitor and provide treatment and counseling services. The nurse assessed whether the participants (people living with HIV-AIDS) were ready to be interviewed, initially asking potential participants if they agreed to be approached by the researcher to be invited to participate in the study. The researcher offered the participant information sheet to be read to the participants; if they were willing to participate, they provided written consent. Participants were encouraged to reflect on generally-accepted healthcare services and discuss their situations when interacting with nurses. Interview guides were used to remind researchers of the topics covered and ensure that all major topics were covered, including discussions of life experiences during their interactions. The participants were interviewed in a separate quiet room with treatment services that had been prepared in advance at the Community Health Centers. They had the option to stop the interview at any stage, if they wished. The results of individual interviews were recorded using cell phones, carefully written, and confronted with nonverbal responses through field notes for data analysis, and then reviewed to improve data accuracy [23]. --- Study design A phenomenological qualitative approach was used in this study [24]. The sample was selected through purposive sampling of patients with HIV-AIDS. To gain meaningful perceptions and experiences, the researchers recruited a diverse and representative sample that reflected the population of people living with HIV-AIDS. The inclusion criteria in this study were as follows: (1) people living with HIV-AIDS who were hospitalized or outpatients, (2) people living with HIV-AIDS who have had HIV-AIDS for more than two years, and (3) people living with HIV-AIDS routinely seeking treatment. Meanwhile, the exclusion criteria were people living with HIV-AIDS who were seriously ill; therefore, interviews were not possible. Data saturation was the final limit for determining the number of samples. Data saturation was reached at the 21 st interview because no new information emerged [23]. To strengthen the data, an additional participant was interviewed to ensure that no new information was obtained. Therefore, for the 22 st participant, data saturation was achieved. Furthermore, the researchers conducted a thematic analysis using a deductive method approach [23] by exploring important themes that described the phenomena that occurred in people living with HIV-AIDS when interacting with nurses. --- Data analysis All interview recordings were rewritten verbatim, coded, and labeled, and transcription of all data was carried out to be consistent with data reflection activities and used to determine new ideas. Interpretative Phenomenological Analysis was used for data analysis [25]. Researchers have also developed an Interview Guide (Box 1) based on the research objectives and the existing literature. Interview transcripts and field notes were read carefully and repeatedly to determine emerging themes by reading sentences in detail and then categorizing important terms that were related to each other through a selective approach. The researchers read the text collectively, tried to understand the overall meaning, and developed keywords and concepts through dialogue with the text. Additionally, the researchers maintained openness by reflecting on various interpretations to monitor assumptions and biases through the triangulation process, namely, linking interview data with field notes to clarify what is meant by clarifying to participants [26]. --- Box 1. Interview guide. The interview guide was developed for individual interviews based on a literature review. The interview guide was piloted with two HIV-AIDS sufferers. The content of the guide was relevant with no amendments required and included open-ended questions such as: 1. What has changed since you suffered from HIV-AIDS? Each sentence was analyzed by the researchers and confronted with data in the field notes. These themes were then reconstructed into a description of the life experiences of people living with HIV-AIDS that nurses overlooked when interacting [26], The researchers then connected categories based on the events experienced, and always paid attention to the balance of research themes by looking at each part as a whole. --- Ethical considerations The research procedure was performed in accordance with the principles of the Declaration of Helsinki and was approved by the Research Ethics Committee of the University of Muhammadiyah Lamongan (number:085/EC/KEPK-S2/05/2021. All participants provided informed consent and were told that they could withdraw from the study at any time. Informed consent was obtained from each study participant for publication of their responses while maintaining anonymity, and the place and time for the interview were arranged to maintain privacy and confidentiality. The respondents' identities were anonymized to maintain confidentiality. --- Results Broadly speaking, the theme that emerged was that people living with HIV-AIDS received limited attention regarding their psychosocial problems when interacting with nurses. Six sub-themes were found to support the overall theme:1) Difficulty in social access, 2) Forced to accept their situation and suppress their will, 3) Wanting to be recognized as others in general, and 4) Social stigma and self-stigmatization that affects the surroundings, 5) lack of enthusiasm for life expectancy, 6) always imagining being in the shadow "when death picks up.". --- The difficulty of social access HIV-AIDS affects all personal aspects of people infected, including social access, which makes them disconnected and lose their future. As a participant said: "I used to work, make friends, and hang out with other friends and family like normal people before I got this disease (HIV-AIDS)... but right now it's really hard for me to do these things. I feel forced to concentrate on a strict routine of treatment and self-care programs for my disease." (P-16). This situation causes people living with HIV-AIDS to lose control of their lives, lose enthusiasm, become disappointed, and regretful. According to a woman: "My condition prevented me from doing what I wanted. There's nothing I'm proud of myself right now, so I'm so sad... I'm so traumatized that I have to erase all my dreams. (P-03). --- Forcing to accept their situation and suppressing the will Participants often cited self-acceptance as a very positive attitude in their lives, and they were able to calm down even though they realized that it would take some time. One of the participants said: "I have to accept my condition. I don't force myself to do things that are difficult for me to do... I try not to be confused and not to worry, because I have to condition this situation." (P-10), The situation experienced by HIV-AIDS patients forces them to accept their situation, even though there is a desire to rebel. "I let my illness damage my body, I accept it even though I have to lose my beauty... it's already done... unless I surrender to God, maybe this is a way of life that I have to accept" (P-09). --- Wanting to be recognized like other people in general The perception of "pressure-free" in daily life makes participants feel that their lives are more relaxed, comfortable, and free, and their emotions more controlled. All desires can be carried out, such as getting along with everyone, and everyone can see them as normal people on an equal footing. Similar to what this lady said: "I want to be like him (a volunteer HIV-AIDS assistant who is also an HIV-AIDS patient)... it turns out that he can be like normal people in general... he is very relaxed. Having this disease (HIV-AIDS) can lead to a life where you do not have this disease. Finally, I believe that I can live a life like others because what he can do I can also do, even though I have to struggle." (P-08). They wanted others to treat their illness like any other infectious disease. They also did not want to be seen as "different" from their surroundings. Another participant pleaded: "I realized that I had this disease (HIV-AIDS)... but I didn't want the Health Officers to ask me what disease it was... he should have known what disease I had. He looked like he didn't know and didn't want to know my feelings. And should I answer that question (about HIV-AIDS) when many people ask me?" (P-05). --- Social stigma and self-stigmatization can affect surroundings This theme shows that people living with HIV-AIDS are aware that they are limiting their family life. Concurrently, they feel the need for their family. Participants believed that their condition had a negative impact on others. They believed that they were limiting their family life and that they were a burden to their family, but also desperately needed constant attention. One of the men expressed his feelings: "I feel very sorry for you (wife), because you carry a very heavy burden... I am being treated; you are also willing to wait for days when I have to go to the hospital, take me to control treatment at the hospital, and always accompany me... Activities in the surrounding community, as well as daily work, are always left to you, and makes it difficult for you" (P-06). This makes people living with HIV-AIDS feel tired, unmotivated, and hopeless. To avoid bothering other family members, they did not want to express their grievances. However, they felt ambiguous between wanting to be helped because they were still weak and needed help, and a desire not to bother their families. Some patients said that family support was important, but they felt frustrated that the support provided was useless and only inconvenienced their family. "My brother plays a significant role in my life; he watches me all day and helps me with everything. Without him, I don't know what will happen to me. However, I know that he only thinks about me and doesn't care about my needs. I feel bad... I am very sad because it always bothers my family, even though the effort is not worth the results obtained... until now, I have not shown any significant changes." (P-15). --- Lacking enthusiasm for life expectancy Having hope is the main weapon for people living with HIV-AIDS to survive and fight the disease, especially when the disease begins to progress. This hope helps them optimistically look to the future; however, it is that they never get. One mother said: "I always hope that I could live this life well, that this life deserved to be enjoyed... but my body is always controlled by this disease, it seems there is nothing to wait for (HIV-AIDS)' (P-02). The participants hoped that their illness would improve, but when they waited for hope, there was only despair, because what they wanted was never achieved, but when they never hoped, they wanted to get better quickly, as shown by the following quote: "The support of the people closest to me has helped me a lot... however, when they are beside me, I feel that the support cannot change my condition... I am still like this and always filled with suffering" (P-07) --- Always imagining the shadow "when death picks up" This theme describes the recurring thoughts of the patient about death. HIV-AIDS is frightening, as if death is in sight. Meanwhile, on other occasions, they are resigned and ready to face death. Some participants expressed that death was coming soon. A man expressed his feelings in the following way: "I know that the medicine I take is only to survive... only a few people survive, and in the end they also face death. Especially when I experience chills, I'm so scared, and I always think that my end has come" (P-18) Other participants also revealed: "I just surrender, and I'm ready to be picked up by death at any time" (P-17) Meanwhile, there were participants who wanted to die. In this case, death is seen as a solution to end their suffering, so there are those who think about planning suicide. One teenager confessed the following: "I am very tormented by this disease (HIV-AIDS), I don't think I can stand this suffering, and at some point in the day, I even think about ending my life, so I don't have to endure this suffering for too long." (P-21) --- Discussion Overall, this study emphasizes the physical impact of HIV-AIDS that affects feelings, purpose in life, and relationship with their immediate environment and social life, and several previous studies have reported social restrictions due to the negative impact of HIV-AIDS [27]. Uncertainty about their health condition is an additional feature, as a result of suffering from HIV-AIDS, as in this study [28]. However, previous studies have focused on behavioral followup of strict treatment regimens [29], medication adherence and routine control behaviors [30], and others have reported beneficial effects [31]. Despite these benefits, they are forced to make changes from routine activities that must be undertaken into new patterns of social life, and this requires major changes and readjustment into a personal life that can cause considerable discomfort to them. Regarding events experienced by people living with HIV-AIDS that make sufferers feel depressed, nurses must be physically present to discuss what to do, examine previous experiences, and help them assess and decide how to maintain their lifestyle to meet the self-care needs of people living with HIV-AIDS and/or modify their physical, psychological, and social environments according to their current situation, conditions, and demands [32]. Furthermore, other studies have also reported that having HIV-AIDS can lead people to see a change in their identity, and ultimately, they are able to judge and realize that they will never be who they were before [33]. However, this study has reported that this perception has a very negative impact on nurses, leading to prolonged sadness, stress, low motivation, and withdrawal, and it is increasingly difficult for nurses to understand if they are introverted and unwilling to express their feelings. This situation is a major problem that must be considered, and several researchers have reported positive consequences of efforts to increase self-esteem [34]. This is an important study by nurses to build effective communication that focuses on problems related to their ideas, thoughts, feelings, and hopes, through which nurses can contribute to altering their self-concept through changes in the construction of new identities [35]. The findings that nurses should be wary of are that participants conceal their problems and are unwilling to open up with nurses, even with their families, and they secretly make plans beyond common sense without the family's knowledge. This behavior is detrimental to them because they do not have time to share their experiences and stop their routine activities. However, this is not the case with the results reported in studies related to other chronic diseases where patients always express their feelings [36]. Family members had a positive influence on the disease process analyzed, especially those related to daily activities and psycho-emotional support. Other studies also support this finding, which explores the perspective of families in supporting people living with HIV-AIDS regarding their quality of life related to the burden borne [37]; many families are forced to change their lifestyle to provide support to those they care about [38]. Therefore, it is important for families to learn patterns of providing support and combining meaning in life for the benefits provided. This study also highlights that people living with HIV-AIDS have recurring thoughts about death. Another study also reported the same situation, especially when it came to the vital threat felt by patients with terminal cases so that they had difficulty expressing their feelings [39]. The results of this study highlight the concerns and suffering as part of their life experiences, especially those who are extremely scared of the possibility of dying, and some patients like this, as reported in this study [40], are not prepared for the possibility of death. However, despite the results of this study, qualitative studies have reported that some patients face this process calmly, there is no burden in their lives, and they see their illness as part of a life cycle that must be accepted [41] because all humans return to God. The results of this study also showed that some participants had a desire to end their lives and might view death as the best way to end their illness, as reported in this study [42]. Furthermore, this study also found a desire for a patient to end his life to end his suffering. Regarding the consequences of this problem, other studies have also reported that the roles of family and health workers worsen their psychological condition because they seem to let these patients die as if they avoid their duties and responsibilities [39]. However, there is a glimmer of hope for a terminal patient, such as a person living with HIV-AIDS, to enjoy his life, and if he has to die because of his condition, the hope that arises is to die peacefully and happily. In all these circumstances, it is important for nurses to establish effective communication with patients in a "physically present" effort, which is a starting point for understanding the possibility of death [43]. In the context of interactions in nursing, patients feel that the nurse is always there. This meeting makes them willing to discuss their panic so that it can increase prosperity and peace [44], as patients and their families wish nurses to be present in their situation. A strong desire to be able to enjoy life as a person who does not have a burden and hopes that the environment also accepts people with HIV-AIDS is an interesting finding to be discussed in this study. This finding is relevant to the goals and expectations of a prosperous life in their environment that they have experienced before suffering their illness. This is obtained when the environment around them does not make them feel depressed, so that they can live without psychological burdens. However, there have not been many reports in other studies that have been found to campaign for "free of stigma and discrimination" for people living with HIV-AIDS. However, several studies on other chronic diseases have discussed the importance of an atmosphere free from stigma and discrimination, and to this end, have discussed the importance of a stress-free life as a key to the psychological well-being of many of these patients [34] Several studies have acknowledged that people with chronic diseases need to recover their perception of a state without mental stress and adapt to a new environment [45]. In line with this, researchers have also found that people living with HIV-AIDS want to be identified as people suffering from the disease in general and there is no "emphasis of the HIV-AIDS accent word" for their disease in their environment. In this case, the stigma experienced, both social stigma and self-stigmatization, by people living with HIV-AIDS, is also experienced by those suffering from other chronic diseases [36]; therefore, as the findings of this study show, some people in their environment try to isolate them. In this case, as also reported in this study [46], living with an HIV-AIDS person is felt very differently and is considered psychologically stressful, and these feelings can lower self-esteem and increase insecurity. Another interesting finding of this study is their expectations. The participants agreed that there was a desire to live like they did before the illness, and this was to improve mood, quality of life, and a prosperous life [47]. On the contrary, as the researchers have seen in this study, when hope is far from the mind, despair and sadness become part of one's life, which is in line with this study [48]. This negative emotional response has been discussed in depth among people living with HIV-AIDS. In addition, this study also found that there was no hope for patients in crisis and unstable situations, thus giving the impression of being helpless in their fight against the illness. Studies conducted on other populations show that hope is an important mediator of effective coping strategies, although this is very difficult to be shown by patients with terminal cases [49]. It is important for nurses to find the relevance between expectations and problem-solving strategies faced by people living with HIV-AIDS, that the presence of high expectations from people living with HIV-AIDS allows nurses to motivate patients to decide on appropriate actions and overcome obstacles, including their experiences as a result of the disease. In this context, the nurse plays the role of a motivator and initiator in seeking a meaningful future in their lives [50]. --- Limitations The participants in this study were comfortable expressing their experiences while interacting with nurses during the ARV treatment program, but with different characteristics and emotional responses, their perceptions may also be different. They were very careful about expressing their feelings. In addition, the way they expressed their feelings also differed. In addition, they also did not want to be open to expressing all their wishes and hopes for the services provided by nurses because the focus of service was centered on definitive treatment. Although both pose potential limitations, they are not a barrier to obtaining natural and complete data. Armed with a therapeutic communication approach, all physical, psychological, and social problems can be expressed comprehensively. The analysis is based on the data that have been found, but triangulation with experts and HIV-AIDS program holders may contribute much more to their knowledge and perceptions through a professional approach, including with health care workers. --- Conclusion The interaction framework between nurses and clients in this study has made it possible to gain a greater knowledge and understanding of the meaning of the life experiences of people living with HIV-AIDS and to encourage new changes in providing nursing care for these patients. Furthermore, nurses innovate and reshape the nursing service system, which is centered on an individual approach. Therefore, nurses can position themselves optimally to become familiar with the patient's experience, and emphasize and maintain interpersonal relationships with patients and their families. This strategy will result in better nursing services and greater satisfaction for patients and professionals who care for them. vol. 11, no. 10, pp. 306-312, 2020, [Online]. https://www.sysrevpharm.org/articles/adaptive-strategyof-womens-leprosy-in-indonesia-psychic-experience-of-women-with-leprosy-in-living-a-community-life. pdf. --- All relevant data are within the manuscript and its Supporting information files. --- Supporting information --- S1 File. (DOCX) --- Author Contributions Conceptualization: Abd Nasir, Ah Yusuf, Fanni Okviasanti. --- Data curation: Makhfudli, Yanis Kartini. Formal analysis: Abd Nasir, Ah Yusuf, Susilo Harianto, Fanni Okviasanti. Methodology: Ah Yusuf, Fanni Okviasanti. Supervision: Abd Nasir. --- Validation: Makhfudli. Writing -original draft: Abd Nasir, Ah Yusuf, Makhfudli, Yanis Kartini. Writing -review & editing: Abd Nasir, Ah Yusuf, Susilo Harianto, Fanni Okviasanti.

Nurse-client interaction when providing nursing services is limited to optimizing treatment and self-care, with limited focus on the psychological problems of people living with HIV-AIDS. However, psychological problems manifest more often than the health risks of the disease itself. This study aimed to determine the emotional response of people living with HIV-AIDS who received limited attention from nurses from the perspective of nurse-client relationship.A phenomenological qualitative design was used through in-depth face-to-face interviews in a semi-structured manner, in an effort to obtain complete data. This research used purposive sampling with Participatory Interpretative Phenomenology analysis, involving 22 participants (14 males and 8 females).This research produces several themes, with six subcategories: 1) Difficulty of social access, 2) Forcing to accept their situation and suppressing their will, 3) Wanting to be recognized like other people in general, 4) Social stigma and self-stigmatization affecting surroundings, 5) Lacking enthusiasm for life expectancy, 6) Always lingering under the shadow "when death picks up."The results showed that mental stress was experienced more than physical problems by people living with HIV-AIDS, thus prompting new changes to nursing services for HIV-AIDS

Introduction Many international students travel to Western countries to pursue higher studies. According to UNESCO (2019), the international student population increased from 2 million in 2000 to over 5.3 million in 2017. More than 50% of these students are enrolled in the U.S., Europe, and Australia. After China, the second most common international students come --- Impact of Transitional Stress on Students The process of leaving their home country to study overseas makes international students homesick and alienated. They encounter culture shock due to the new academic environment, food, and unfamiliar climatic conditions. This transition process causes transitional stress among international students (Arthur, 2003;Johnson & Sandhu, 2007), negatively impacting their psychological, emotional, sociocultural, and academic well-being (Arthur, 2003). The transitional stress is manifested in various ways, such as anger, anxiety, cognitive impairment, confusion, exhaustion, defensiveness, depression, disorientation, exhaustion, fatigue, fear, gastrointestinal problems, headaches, homesickness, inferiority, insecurity, insomnia, irritability, lack of energy, loneliness, loss of appetite, loss of control, mood swings, muscle tension, overeating, resentment, sadness, sense of loss, unfamiliar body pain, and vague bodily sensations (Arthur, 2003). These findings were recently confirmed by other research studies, for instance, by Ching et al. (2017) and Kornienko et al. (2018). There has been a concern that international students might have low graduation rates because of the academic difficulties they encounter, such as being weak in the English language (Fass-Holmes & Vaughn, 2014), not understanding American teaching style (Ota, 2013;Roy, 2013), difficulty in maneuvering through new campus environment (Ota, 2013), and difficulty in interacting with fellow American peers and faculty (Roy, 2013). Though international students encounter academic challenges, they have been found to succeed academically (Fass-Holmes, 2016). However, their self-esteem decreased because of academic difficulties (Constantine et al., 2004;Lyken-Segosebe, 2017), reducing their work and personal life satisfaction (Shupe, 2007). --- Coping Strategies According to Lazarus and Folkman (1984), coping is a process where people evaluate their situation and consciously choose an effective coping strategy to solve their problems. Coping strategies are classified into emotion-focused and problem-focused (Lazarus, 2006). An emotion-focused coping strategy is where people try to distract themselves from stressful situations or reduce the negative emotions caused by such situations, for instance, by exercising, meditating, praying, or using relaxation techniques. A problem-focused coping strategy is where people solve stressful situations, for example, by identifying the source of the stress and eliminating it, seeking help to manage the situation, or detaching oneself from it (Folkman, 1997(Folkman,, 2013)). In most stressful situations, particularly in transitional stress, using emotion and problemfocused coping strategies are found to have influenced with effective outcomes (Lazarus, 2006). Prior research has highlighted that international students use diverse emotion-focused and problem-focused coping strategies to overcome transitional challenges (Alazzi & Chiodo, 2006;Vasileiou et al., 2019). For instance, Alazzi and Chiodo (2006) conducted a qualitative study with eight Middle Eastern students studying at a U.S. university. They found that the participants identified the cause of a problem and understood the consequences that could occur if the problem was not solved. In addition, the participants who considered themselves capable of solving their challenges had superior coping skills. Further, participants coped with loneliness and homesickness by engaging in religious activities, keeping themselves occupied at work, and maintaining good relations with family and friends. Such coping strategies were found to increase the confidence and happiness of the students (Alazzi & Chiodo 2006). Similarly, Poulakis et al. (2017) conducted a qualitative study to explore transitional stress among eight Greek international students studying in the U.S. They found that the participants developed strong relations with their families and peers to overcome their homesickness that helped them in gaining emotional support. In another study, engaging in positive self-talk, listening to music, and diverting oneself from thinking about home served as helpful coping strategies and promoted the overall well-being of fifteen international university students studying in the U.K. (Vasileiou et al., 2019). Concurring with this finding, a qualitative study conducted at a Malaysian university found that positive thinking and engaging in extracurricular activities such as exercising, reading, and praying increased students<unk> self-esteem (Saravanan et al., 2019). Additionally, taking the responsibility to understand the academic system of the host country, and honing their English language and social skills, helped five Latin American students overcome academic fears at a U.K. university (James, 2018). Also, focusing on developing communication skills, working hard, and engaging in community activities overcame transitional stress among 413 international students studying at a U.S. university (Mahmood & Burke, 2018). --- The Goal of the Study Literature indicates that transition is a process where regardless of the country international students transit to, they initially encounter transitional stress. Additionally, the more the difference between the home and the host countries in terms of cultures, attitudes, and academic systems, the more the transitional challenges and stress (Pham & Tran, 2015). Students transiting from India to the U.S. encounter significant transitional challenges because of the vast difference in the cultures and academic systems of India and the U.S. Therefore, to help students from India adjust to the U.S., it is vital to know the coping strategies used by previous Indian international students to overcome their transitional challenges. However, minimal research has focused on the coping strategies employed by Indian international students in the host countries. Thus, this study aims to (1) explore the coping strategies used by students from India to overcome the challenges in the U.S.; (2) promote awareness about various coping strategies; and (3) provide information to international student counselors to enhance the coping of Indian international students. --- Method Design A qualitative phenomenological methodology was employed to conduct this study. Considering that Creswell (1998) suggested a sample size ranging from five to twenty-five participants, while Morse (1994) recommended a minimum of six participants, a convenience sample of six Indian international students enrolled in a research-intensive public university in the southern United States was selected. The participants for this study were selected using the purposive sampling criterion, wherein participants are chosen based on their ability to provide comprehensive information concerning the phenomenon under investigation (Patton, 1990). To conduct purposive sampling, it was imperative to specify parameters for selecting study participants (Merriam, 1998). Accordingly, the following criteria were used: Participants must have been born and studied in India; they must not have visited the U.S. or any other country before enrolling as graduate students in the U.S; they must have lived in the U.S. for over a year; and they must be graduate students. --- Participants Six international students from India took part in the study. Of these, five were doctoral students, and one was a master<unk>s level student. At the time of data collection, the participants had lived in the U.S. for 2-11 years. All the participants were given the following pseudonyms to protect their identities: Krishna, Vishnu, Pragya, Shreya, Rishi, and Arjun. --- Procedure Upon obtaining the ethical clearance for this study from the Institutional Review Board (IRB), participants were invited for an interview. The interview protocol consisted of seven probes that explored the coping strategies used by the participants to adjust well in the U.S. The duration of the interviews ranged from 1-2 hours. The first author conducted interviews in the university<unk>s library, labs, and cafeteria. We informed participants about the confidentiality procedures before the interview. We took permission from them to record and transcribe the interviews. They were also told that they could withdraw from the interview at any time. The interviews were recorded and transcribed verbatim. The transcribed data were analyzed using the phenomenological methodology. --- Data Analysis Based on Moustakas<unk> (1994) modified version of Stevick-Colaizzi-Keen<unk>s method, the data analysis was performed using the following ten steps: (1) Epoche: When applying Epoche, Moustakas emphasized that the researcher must avoid preconceived notions about the phenomenon under study and instead concentrate on the data provided by the participants to understand the phenomenon from a newer perspective. Accordingly, the authors refrained from their thoughts, feelings, and assumptions regarding the coping strategies to obtain a newer understanding of the various coping strategies that can be employed for successful adjustment in the host country; (2) Obtaining an understanding of the data: To understand the data, all six recorded interviews were listened to multiple times to ensure the accuracy of the transcripts. Also, the transcribed interviews were read multiple times to better understand the participants<unk> viewpoints; (3) Horizonalization: This is a process of identifying verbatim parts of the transcribed interviews that explain the phenomenon (Moustakas, 1994). Accordingly, all interviews were thoroughly analyzed to determine significant verbatim parts describing the phenomenon of coping strategies; (4) Identifying the invariant constituent: Invariant constituents were identified by abstracting and labeling the statements essential in understanding the coping strategies. The statements that did not meet the criteria were repetitive, overlapping, or vague were deleted; (5) Identifying themes: The identified invariant constituents were placed under each heading, forming the fundamental themes of coping strategies; (6) Individual textural descriptions: In this step, descriptive narratives of each participant were developed based on the horizonalized statements, invariant constituents, and the themes identified in the previous step; (7) Individual structural descriptions: Here "Imaginative Variation" was applied (p. 33) "to arrive at structural descriptions of an experience, the underlying and precipitating factors that account for what is being experienced." (Moustakas 1994, p. 98); (8) Composite textural description: Inductive reasoning and analysis were used to combine the individual textural descriptions of all the participants formulated in step 6 to present a portrayal of the descriptive narratives of all the participants as a group; (9) Composite structural description: The individual structural descriptions were combined to formulate a composite structural description using imaginative variation, to describe what the participants, as a group, experienced and how they felt after employing those strategies; and (10) Textural-structural synthesis-essence of the phenomenon of coping strategies: This step involved combining textural and structural descriptions to understand the essence of the coping strategies from the participants<unk> perspective, which is the goal of this study. Lastly, member checking was performed to establish the trustworthiness of the transcribed data. The participants were contacted a second time after the primary interview to peruse their transcribed interviews and make any changes to ensure the accuracy of their responses. --- Findings The phenomenological data analysis resulted in the following four coping strategy themes: (1) open-mindedness, (2) goal-orientedness, (3) independent attitude, and (4) showing gratitude. The participants utilized these strategies to solve the transitional challenges encountered in their personal and academic lives in the U.S. The findings also indicated specific thoughts and actions that helped the participants acquire the identified four coping strategies, as shown in Figure 1. --- Open-Mindedness When the participants transited to the U.S., they struggled to form friendships and missed their home country<unk>s culture and atmosphere. For instance, Pragya stated: If there was a word that could multiply miserable times by thousands, that<unk>s how I felt. I did not like it here at first. I didn<unk>t want to meet anybody. I just wanted to go back home. I feel lonely, not being able to express my true feelings to somebody. Having to suppress my feelings and not being able to talk about certain things certainly makes me feel lonely. One coping strategy that helped participants form social bonds was being open-minded. According to the participants, open-mindedness is a mindset where students accept diverse people, ideas, and information. They demonstrated open-mindedness in various ways. For example, Shreya became more flexible, discouraged herself from judging people, initiated contact, and learned to be more expressive. Explaining, she commented: I had to keep an open mind and let things go. You cannot have everything according to the way you want, as you have it in India. Do not judge anyone based on their actions. Keep in contact and good contact with as many professors and administrative people over here in the university. And if problems arise, tell them immediately. Do not wait or think it will not be done. Just ask them. Arjun gave up generalizing and stereotyping. He expressed: I learned that many perceptions are out there; they are all different, not right or wrong. They are just different. Also, when you are in India, you have some stereotypes, but when you start interacting with people from different countries, you start realizing that, <unk>I need to discard a lot of these stereotypes that I carry.<unk> I have broadened my thinking and my mindset. In terms of my worldview, I have broadened. Rishi became more accommodating. He opined: One needs to learn how to share an apartment and how to share life with others. You have to be really flexible and open to adjusting to the setup here. One can<unk>t be firm and say, <unk>No, I like to do this; I won<unk>t do that.<unk> You must be open and adjust; that<unk>s the only way to survive. Rishi also invested an effort to understand the host culture: I understood that if you are in a society, you must understand it. For that, you need help from Americans. You need to open up more with Americans; ask them general questions about their background or education, but avoid personal questions. Once you start approaching, maybe others will also begin approaching from their side. That way, communication develops. Further, emphasizing on being more versatile, Rishi said: Most Indian students are very academically oriented. Outside of studies, they don<unk>t know anything. But they can change all those perspectives here. For example, they can join a hiking club. They can learn salsa, develop an all-round personality, and not just confine themselves to their department and apartment. Vishnu learned to be forthright. He explained: You need to be a lot franker over here --- Goal-Orientedness Participants got absorbed into the problems they encountered in the U.S. For instance, Shreya felt unhappy and angry when she saw a difference in her professor<unk>s treatment of her and her colleagues. She explained: When you see a difference in the treatment from anybody, you really get very angry and frustrated because you cannot speak up. Because you know that if you speak out, people will not like it and the relationship with them would be affected. You cannot go to your seniors, because for them it<unk>s a trivial matter and you think as to what they would think about me if I go every time and complain every time this happens. To overcome such difficulties, participants emphasized being focused and goal-oriented. Shreya explained, "you have come here to achieve a goal. Keep your goal in mind and keep doing that." Pragya believed that "the incoming students should imagine themselves five years from now. They should have a goal in mind, which could be educational, professional, personal, economical, or social, and try to reach it." Rishi said: "Success rate in graduate school determines the student<unk>s career and, accordingly, his/her happiness. Therefore, students should be earnest in setting goals and work diligently to accomplish them." Vishnu<unk>s goal was to excel in academics. He explained: Having a goal in mind in very essential. Enrolling in courses with a <unk>whatever works<unk> attitude is a serious mistake. Instead, courses should be selected based on interest and passion, so that we continue to excel in them. My goal was also to make my life in U.S. smooth. I had already met people; spoke to people right in India; we had gotten to know each other; we came here as a group; I already knew my roommate, fixed over there; and we had an apartment. I came to know of the India Association at Nurture University, who helped us. Arjun explained the importance of being goal-oriented and said, "If you are goal-oriented and focused, you can come out with good outcomes in the U.S. You have an opportunity to be recognized and excel if you want to. So, one must focus." --- Independent Attitude Another challenge that participants had to overcome was being dependent. Pragya stated: "my parents would protect me a lot, as a result of which I was much more dependent on them for decisions. Here, I am on my own, which is difficult." Similarly, Shruti stated: In India, you are not working when you are studying. So, when you are studying, you are concentrating a lot on your studies and your parents are very supportive throughout those 3 years. So it<unk>s basically just studying and you are not expected to do much else. But, here it<unk>s very difficult, like when you are just 18, and you have to adjust to cooking, cleaning and living on your own, navigating through classes and making all the decisions on your own. It<unk>s difficult. Back in India, parents take care of a lot of things, which would not be the case here. So, you would have to adjust to many new things after coming here. Participants emphasized cultivating an independent attitude to overcome this barrier, where they depended on their capabilities to accomplish tasks instead of relying on others. Participants had to completely change their beliefs about becoming self-reliant and independent. Explaining, Vishnu stated: In the United States, you are not just a student; you need to do everything by yourself in terms of cooking, laundry and so on. In India, you can just sit and study and your mom brings you tea or coffee. Here, if you want coffee, go make your coffee. You have experienced being pampered and living a happy life. Now go out to the real world and do the same things without all that support. Further, Rishi commented: Developing independent thinking and taking charge of your own life is very important. Rediff.com has a section for prospective students coming from India to the United States. India Association has a nice FAQ list, things to do, what to expect. Reading this information helps. Reading and talking to students already in the United States helps. Krishna took the initiative to solve problems on his own: I believed that I will get over the situation. I asked for help from my friends or relatives. I asked myself as to what makes me happy in U.S., instead of dwelling in negative thinking. And once I did this, I knew what exactly I needed to do. Re-emphasizing taking one<unk>s responsibility, Shreya stated: I learned not to keep too many expectations. I prepared mentally that I am alone here, I have to stand on my feet and take care of myself; there is not going to be anybody to take care of you. There will be problems, and it will take time to adjust. Similarly, Vishnu said, "I feel that you are 23, you are grown up and are responsible. Start living your life. So, I think having at least that confidence level is fair. Don<unk>t expect anyone to pamper you at this point." On similar lines, Pragya stated: We must be independent here. Life here is not as rosy as you see on television. It<unk>s shockingly different. Therefore, its important to go out and meet different people, but never forget your goal, which is to study and be responsible. Reading, keeping in touch with current affairs of United States, asking as many people as possible for their opinions are all important. Being independent helped the participants increase their self-confidence and inner strength. They became their own support system. Vishnu commented, "being independent helps you grow emotionally and mentally. That<unk>s how you learn things and solve problems on your own." --- Showing Gratitude Difficulties in the U.S. made participants delve into negative thinking, making them sad and stressed. For instance, Krishna described his initial days in the U.S. as "hell." Explaining, he stated: The daytime was split between loads of academic formalities, facing fierce competition for funding and course enrollment, meeting professors and prospective employers, sorting out living situation (lease, grocery, account, purchases), and staying in touch with family and friends back in India. In whatever little that was left of nighttime, it was very common to think of the reasons for coming here; occasional crying, self-consoling and wondering if all of it was worth the efforts. To overcome this challenge, participants began identifying the reasons they were thankful for. Vishnu stated: I am thankful that this country has accepted me. You came here to study. Do what you have to do and continue with your life. We are in a foreign land, and it is important to express our gratitude to the country that is providing us the opportunity to pursue our academic dreams and aspirations. Yes, if you cannot get into NASA, there is nothing wrong, because they ask for citizenship, and you are not a citizen. Sometimes people would say, <unk>I am not feeling good,<unk> and I would say <unk>think of any other international who were in India and think about what they would be going through.<unk> I mean, socially, we may also not be perfect, you know, in terms of treating an outsider in our community. So you need to think of it in that aspect and say, <unk>ok, this is not something really bad that happened to me, this is something that people sometimes undergo,<unk> not a big deal. Further, showing gratitude to his professor, Vishnu stated: My advisor has been great as far as mentoring is concerned. I like his style because he is an advisor and that<unk>s all he does, advise. He says, <unk>I am your advisor, and you are my student. We should always keep arguing and when you start winning more arguments than I do, it<unk>s time for you to graduate; that's a very cool concept, and I love that idea.' He<unk>s been a very good mentor and a role model that I would like to be like. He sends us out to conferences and says, <unk>Go talk to the professors; find out what they are doing, what their students are doing and try and see if you can collaborate with them.' So, things like this really help in the long run. I would have to give a lot of credit to my advisor for allowing us to do all this. Shreya expressed her gratitude by returning favors. Explaining, she commented: Friends help you a lot in adjusting. If someone helps me, I make sure that I return the favor or at least stay in touch with them. Just making use of a person and going is not good. That starts affecting other people too. House some people in your apartment, because they really need it, and the people who stay there should take care of people living in the house. Just don<unk>t be a problem for the person. In terms of expressing gratitude to the host society, Rishi stated: Whenever I go and talk to Americans, I know I am an outsider. I am not a part of this society, be it in terms of language or culture or anything. Usually, I talk with a sense of gratitude. I ask myself this question -suppose the same situation was reversed; let<unk>s say, this is India, you are an Indian, and they are Americans; it<unk>s like role reversal, will you be so acceptable? Will you be so welcoming to other cultures? Will you be so forthcoming to help others? If you see everything in that light, then what you see is definitely a service to you. So, in that sense I don<unk>t have any questions. Participants expressed that showing gratitude developed a positive frame of mind, which helped them integrate into the academic and social lives of the host country successfully. Thus, showing gratitude served as an effective coping strategy to adjust to life in the host country. --- Figure 1 Thoughts and Actions That Led to the Identified Coping Strategies Thoughts and Actions Coping Strategies --- Discussion The present qualitative study examined the coping strategies of Indian international students in U.S. higher education. Four specific coping strategies were identified: open-mindedness, goal-orientedness, an independent attitude, and showing gratitude. The participants reported that being open-minded helped them develop a more positive approach to dealing with challenges in the host country. Establishing clear goals for oneself helped them adjust to their academic and social life. Furthermore, a belief system that one must be independent and not rely on others increased participants<unk> self-esteem and happiness. Finally, developing the trait of having gratitude aided the participants in valuing the opportunities they received. Additionally, this study identified ways in which the above-mentioned four coping strategies can be acquired, as delineated in Figure 1. The results of this study indicated that the coping strategy of being open-minded aided participants in becoming more flexible in integrating and making friends from diverse cultures instead of limiting themselves to their own culture. This prevented them from experiencing major culture shock and transitional stress in the host country. These findings concur with previous studies wherein open-minded students understood other cultural viewpoints better and adjusted effectively to the host country (Alsahafi & Shin, 2016;Mahmood & Burke, 2018;Zhou et al., 2018). Participants adopted open-mindedness by improving their communication skills, avoiding rigid thinking, and accepting people from diverse backgrounds and cultures. The second coping strategy is goal-orientedness. Focusing on achieving the set goals increased participants<unk> confidence, they excelled in academics, and consequently, their stress levels were reduced. This finding adds to the previous research, where the results indicated that goal-oriented students are more competent, less focused on problems, and less challenged due to transitional stress (Mahmood & Burke, 2018). Further, the present study<unk>s findings are consistent with Park et al. (2016), who emphasized that setting goals related to developing a support system in the host country made the students feel more accepted by the host community. Participants achieved goal-orientedness by thinking in advance about the goals they wanted to achieve, setting achievable goals related to their academics, health, finances, and social skills, prioritizing their goals, and maintaining consistency in pursuing them. The third coping mechanism is to develop an independent attitude, which helped the students grow emotionally and psychologically, aiding them in solving problems. Independent thinking increased participants<unk> self-esteem. Vasileiou et al. ( 2019) viewed an independent attitude as a self-reliant coping strategy where the students recognized and accepted the problems and made a conscious effort to comfort themselves. According to Alsahafi and Shin (2016), independent students are responsible for improving their lives by developing healthy habits, such as avoiding substance abuse, encouraging more positive thoughts, honing their social skills, listening to music, and engaging in writing. Concurring with this finding, participants in this study took the responsibility to overcome their transitional challenges by becoming more independent. Participants achieved an independent attitude by realizing the need to be independent, taking the lead to better their lives, engaging in extracurricular activities, encouraging positive thoughts, honing their social skills, and taking the responsibility to solve their problems. The final coping strategy is showing gratitude. Being thankful helped the participants see the positive aspects of life and value the opportunities they received. This increased their satisfaction with the university, social interactions, academic system, and life, in general, in the host country. These findings were consistent with the previous literature that indicated a high correlation between expressing gratitude and contentment with the university experiences (Froh et al., 2008), developing healthy, long-lasting relationships (Emmons & McCullough, 2004), and enhancing student participation in the classroom (Flinchbaugh et al., 2012). Participants showed gratitude by returning favors and reciprocating to the kindness of others. --- Implications The participants in the present study employed four coping strategies that helped them successfully transit to the U.S. The most common resource materials provided by international student support services include online reading materials on their website, newsletter articles, and brochures (Madden-Dent et al., 2019). Therefore, the coping strategies identified in this study can be valuable to include in such resource materials to encourage coping skills needed for successful cross-cultural transition and adjustment. --- Limitations and Future Direction Although the sample in the present study met the requirements of saturation and phenomenological qualitative study (Lincoln & Guba, 1985;Strauss & Corbin, 1998), future studies should select larger samples. India is a large population where people differ based on their socioeconomic, linguistic, and cultural backgrounds. A larger sample of international students from India may allow students from different backgrounds to express their experiences. This information may be more beneficial for the diverse Indians who study overseas. --- Conclusion This study has highlighted the positive coping strategies that Indian international students use to combat transitional challenges encountered in the U.S. The findings indicate that the students can cope with such challenges by changing their thought processes and behaviors. Further, the findings add to the literature by discovering the role of openmindedness, goal-orientedness and showing gratitude for the first time in the international student population. The findings have implications for the stakeholders. The suggested coping strategies can be considered by higher education and could be recommended to the students during international student orientation and counseling sessions to help them manage their cross-cultural transitional challenges.

technologies.

INTRODUCTION The American Academy of Medical Colleges (AAMC)'s 2020 report demonstrates that the United States (US) physician workforce does not represent the racial or ethnic diversity of the population it serves. 1 However, efforts to diversify the physician workforce have been an active area of focus for decades. These efforts are partly in response to literature demonstrating health disparities among racial and ethnic minorities, increased health care access for underserved communities, and increased patient compliance when patients are cared for by physicians that connect with their social context. 2,3 Other works have shown improvement in patient care delivery and plan adherence outcomes with the inclusion of broader social perspectives in the culture of care delivery. 2,4 At a more fundamental level, independent of the patient benefits of increasing Black and Hispanic physician representation, governing bodies, national organizations, and member institutions recognize the importance of, and need for, a more diverse workforce. 1,3,5 However, many have found that efforts are inadequate to meet needs. The inadequacy of these efforts is in the setting of a growing body of work demonstrating ongoing barriers to a more diverse workforce; barriers that reflect persistent biases, racism, and discrimination within healthcare. [6][7][8] Despite this, and a growing focus on the topic, there is limited work that comprehensively explores trends in racial and ethnic representation among different medical specialties. 5 In an effort to understand our current state, we use the most comprehensive database on resident physicians to examine trends in the racial and ethnic diversity of Black and Hispanic physician trainees across the twenty largest medical specialties over the last 11 years. In addition, we present a conceptual illustration of what our future state could look like if the current trends continue. --- METHODS We reviewed 11 academic years (AY) of publicly available data on Accreditation Council for Graduate Medical Education (ACGME) training programs from the National GME Census (AY 2007-2018). 4 The National GME Census is a comprehensive database of demographic information on trainees in ACGME-accredited programs. Using AY 2007 data, we examined the proportion of physician trainees of Black race and Hispanic ethnicity among the twenty largest specialties; these account for the majority (81.4% in AY 2007 to 79.3% in AY 2017) of all ACGME trainees. To assess the presence of trends over the 11-year period, we used logistic regression modeling with the number of trainees aggregated by each of the 11 specialty-years, AY year as the predictor, and annual proportion of Black and Hispanic trainees as 2 outcomes. Odds ratios (OR) and confidence intervals (CI) were used to assess for direction and significance of trend; OR are based on statistical trend modeling for the 11-year period. To account for multiple testing, we used an alpha-level significance of 0.0025. We then focused on specialties with a significant annual increase in the proportion of Black or Hispanic residents and estimated the year in which these specialties would achieve representation proportional to the US population. To estimate projected annual increase (PAN) in population representation, we fit ordinary least squares regression models for the proportion of Black and Hispanic trainees using year as the predictor and Newey-West standard errors to account for autocorrelation. We used marginal estimation methods to generate out-of-sample predictions for proportions of Black and Hispanic trainees for each year (from 2018 onward) to calculate the number of years required for physician trainees in each specialty to reach the current racial-ethnic proportions of the US population. We used an alpha-level significance of 0.01 to account for multiple testing. The benchmark for proportional representation was the recent 2018 US Census population estimates for the US. We used Stata version 15 (College Station, TX) for all analyses. The institutional review board at Brigham and Women's Hospital by Partners Healthcare in Boston, MA, deemed this work exempt. --- RESULTS In 2018, 13.4% of the US population identified as Black and 18.3% as Hispanic. We found the total ACGME trainee population in AY 2017 included 5.5% who identified as Black and 7.8% as Hispanic. Obstetrics and Gynecology (10.38 to 8.62%) had the highest proportion of residents who identified as Black; Otolaryngology (2.56 to 2.34%) had the lowest proportion (Table 1). Among the 20 largest specialty training programs, Radiology was the only specialty with a statistically significant increase in the proportion of Black trainees. At the current pace, Radiology could take 77 years (PAN 0.13% (99% CI 0.07-0.19%)) to reach levels of Black representation comparable to that of the US (Supplemental Figure ). Obstetrics and Gynecology (7.36 to 10.14%) had the highest proportion of residents who identified as Hispanic; Dermatology (5.13 to 4.9%) had the lowest proportion (Table 2). Emergency Medicine, Internal Medicine/Pediatrics, Obstetrics/Gynecology, and Orthopedic Surgery demonstrated a statistically significant increase in the proportion of Hispanic trainees (Table 2). To achieve Hispanic representation comparable to that of the US population respectively, Orthopedic Surgery (PAN 0.14% (99% CI 0.06-0.22%)) could require 93 years, Internal Medicine/Pediatrics (PAN 0.20% (99% CI 0.03-0.36%)) could require 61 years, Emergency Medicine (PAN 0.20% (99% CI 0.12-0.27%)) could require 54 years, and Obstetrics/Gynecology (PAN 0.23% (99% CI 0.12-0.35%)) could require 35 years (Supplemental Figure ). --- DISCUSSION We found that Obstetrics and Gynecology had the highest proportion of residents who identified as Black or Hispanic, but no specialty represented either Black or Hispanic trainees in proportions comparable to the overall US population. In contrast to the 13.4% of the US population who identified as Black and 18.3% who identified as Hispanic, the majority of clinical specialties were found to have single-digit proportions of residents who identify as either Black or Hispanic. This absence of appropriate diversity is not limited to a particular discipline and is present in other primary care specialties (e.g., Family Medicine, Internal Medicine, and Internal Medicine/ Pediatrics). Furthermore, only a small number of specialties demonstrated statistically significant increases in the representation of Black or Hispanic trainees over time and, equally important, the majority of specialties showed no significant increases in representation over the period analyzed. The current pace for establishing proportional representation of Black and Hispanic trainees among the specialties we studied is absent for improvement in most, and in those with notable increases they are a generation (or more) away. A limitation of this conceptual approach is projections assume racial-ethnic proportions of the population will remain constant, at 2018 levels. Actual time-to-parity will inherently vary depending on how the US population changes over time. Our goal, however, was not to identify the exact year at which a certain specialty would achieve appropriate representation. Rather, it was to illustrate how the targeted change in representation is not being achieved with current efforts. We anticipate that this current illustration likely underestimates actual time to appropriate representation; the expected rate of change in diversity of the physician trainee population will likely lag behind that of the actual rate of change in the US population. We also anticipate that the US population will become less homogenous and more complex. 9 It should also be noted that the current study focuses on resident physicians of Black race and Hispanic ethnicity; we are unable to comment on other minority populations (e.g., American Indians/Alaska Native and Native Hawaiian/Pacific Islander). 5 These findings are concerning given that the resident population defines the future physician workforce. Even when resident diversity matches that of the US population, it will likely take decades still for this to translate into changes within the US physician workforce. Notably so given current AAMC estimates demonstrate that most US physicians are white (68.2%), while only a minority identify as either Black (2.6%) or of Hispanic ethnicity (3.8%), and that US physician demand is projected to grow disproportionately within minority populations. 1 A precursor to the resident physician population is medical students; an absence of diversity in medical school matriculants and medical students subsequently translates into an absence of diversity in the graduate medical education workforce. Although the AAMC requires medical schools to have pathway programs in place, and despite increases in the proportion of Black and Hispanic medical students, barriers and biases (e.g., implicit white race preference in medical school admissions, 10 insufficient recruitment and retention of individuals from racial and ethnic minorities, 11 and financial barriers to medical school for disadvantaged populations 12 ) within medicine make it less likely for Black or Hispanic candidates to succeed. Even among faculty, non-whites have disproportionately lower rates of promotion. 13 Consequently, non-white medical students and residents are faced with a low number of physicians with similar socio-cultural experiences that often facilitate mentor-mentee relationships. --- CONCLUSION We identify an absence of change over time in Black and Hispanic representation among resident physicians in nearly all medical specialties. In the minority of specialties with some degree of representation improvement over time, we find the time to proportional representation is decades (or a generation) away. Efforts to improve parity in representation to date have not resulted in an appropriate racial and ethnic representation among resident physicians. More direct action is needed to remove barriers limiting entry and success. To effectively address the persistence of underrepresentation within our workforce, we must acknowledge the systemic and structured biases that have shaped our profession's current demography and perpetuated the underrepresentation of Black and Hispanic physicians within medicine. Author Contribution: All authors had full access to the data; they take responsibility for the integrity of the data and accuracy of the data analysis. Corresponding Author: Christopher L. Bennett, MD, MA; Department of Emergency Medicine, Stanford University School of Medicine, Stanford, USA (e-mail: [email protected]). --- Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s11606-021-06650-7. --- Declarations: Conflict of Interest: The authors declare that they do not have a conflict of interest. Publisher's Note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The US physician workforce does not represent the racial or ethnic diversity of the population it serves. OBJECTIVES: To assess whether the proportion of US physician trainees of Black race and Hispanic ethnicity has changed over time and then provide a conceptual projection of future trends. DESIGN: Cross-sectional, retrospective, analysis based on 11 years of publicly available data paired with recent US census population estimates.

Background Persistent and marked inequity is observed for M<unk>ori at all levels of health, [1] education (Ministry of Health, Ministry of Education: M<unk>ori participation and attainment in science subjects, unpublished) and justice [2] in Aotearoa New Zealand. Non-M<unk>ori have life expectancies approximately 7 years longer than M<unk>ori, attain higher educational achievement in secondary and tertiary education, and are incarcerated at markedly lower levels. In primary and secondary health care, non-M<unk>ori patients are prescribed more effective medications [3], are referred more often to specialist services [4] and experience higher quality hospital care [5]. Non-M<unk>ori patients experience persistently lower rates of preventable diseases that lead to avoidable hospitalisation and unmet need in primary care [6]. M<unk>ori experience inequitable access to health services throughout the life course leading to higher rates of disability and multiple morbidity [7]. M<unk>ori are more likely than non-M<unk>ori to cite cost as a barrier to accessing primary care. In addition, nationwide quality improvement programs in Aotearoa New Zealand worsen inequity by differentially improving access to services for non-M<unk>ori [8][9][10]. Patient-centred research can facilitate the understanding of consumer experiences, perceptions and expectations of health services to generate insights and knowledge that guide improvements in healthcare acceptability and quality [11]. In the last two decades, there has been an increase in qualitative research to explore patient viewpoints to inform public policy and align health service development with consumer preferences and expectations [12]. Despite this shift toward greater inclusion of patient voices in clinical health research, including with M<unk>ori consumers, health outcomes remain inequitable across numerous clinical settings in Aotearoa New Zealand and for indigenous and tribal peoples worldwide [13]. In addition to seeking patient perspectives, qualitative studies can offer critical insights into the ways that researchers view and conceptualise the patient experience and how those experiences are problematised as a basis for interventions to improve health outcomes [14,15]. This study aimed to explore how M<unk>ori consumer experiences of health services and programs in Aotearoa New Zealand are conceptualised within qualitative research, to characterise how recommended strategies to improve M<unk>ori consumer experiences can inform policy directions to address health inequities and to identify gaps in the existing evidence base. --- Methods We did a systematic review and evidence synthesis of qualitative studies reporting M<unk>ori consumer experiences of health services and programs in Aotearoa New Zealand. We used the theoretical framework of the World Health Organization (WHO) Commission of Social Determinants in Health (CSDH) to categorise the factors reported to be associated with M<unk>ori consumer experiences of health and to evaluate how recommended strategies arising from the findings might inform strategies to address health inequities [16]. The Enhancing Transparency in Reporting the Synthesis of Qualitative Research (ENTREQ) framework was followed for this study [17]. --- Literature searching Electronic searches were conducted in AMED, CINAHL, EMBASE, MEDLINE, Google Scholar, PROQUEST, and PsycINFO for qualitative studies reporting M<unk>ori consumer experiences of health services and programs in Aotearoa New Zealand from each database inception up to week 2, February 2018. We used search terms using keywords relating to M<unk>ori (M<unk>ori, First nation, Oceanic ancestry, Native), and qualitative research (Content analysis, Descriptive, Discourse, Exploratory, Grounded theory, Interpretive, Interview, Mixed method, Multi method, Narrative, Phenomenology, Qualitative, Thematic, Theme) and experiences (Attitude, Belief, Experience, Perception, Perspective, Satisfaction, Value, View). Studies that included M<unk>ori participants but that did not provide separate data analysis for M<unk>ori and non-M<unk>ori participants were not eligible. --- Data extraction and quality assessment The following basic characteristics were extracted from each study: publication year, methods for ethnicity determination, number of M<unk>ori participants, gender, cohort characteristics, health setting, topic, study methodological framework, whether kaupapa M<unk>ori methodologies were used and funding source(s). The text of each paper including tables and figures was reviewed in full text by at least two of three authors (SCP, HG, SP) to extract the following from each study: stated purpose of research, determinants of M<unk>ori participant experiences, and research responses and recommendations arising from the findings. Two authors independently assessed the transparency of reporting using the Consolidated Criteria for Reporting Qualitative Research (COREQ), which assessed study methodological reporting of the research team, methodologies, context, analysis, and interpretation [18]. --- Data coding and analysis The extracted data underwent two cycles of coding by at least two of three authors (SCP, HG, SP). These authors (SCP and SP) have expertise in M<unk>ori health research or (HG) are training in M<unk>ori health research. Three authors who provided intellectual feedback on the coding also have expertise in M<unk>ori health research (TH, CL, LB). In the first coding cycle, descriptive coding was used to identify the basic topic for each passage of reported data in the results and discussion section of each included study [19]. In the second cycle of coding, the determinants of M<unk>ori consumer experiences of the health system (services and programs) were mapped against the World Health Organization Commission for Social Determinants of Health (CSDH) conceptual framework of the determinants, processes and pathways that generate health inequities [16]. The CSDH conceptual framework includes the socioeconomic and political context in which people live (governance, macroeconomic policies, social and public policies and culture and societal values), the socioeconomic positioning of people (social class, gender, racism, education, occupation and income) and the intermediary factors (material circumstances, behaviours and psychosocial factors) which, mediated through health services and programs, determine inequity in health and wellbeing. The actions recommended by researchers arising from the determinants of M<unk>ori consumer experiences were mapped against the CSDH framework for strategies tackling social determinants of health inequities. This framework describes a hierarchy of dimensions and directions for: 1) policies to reduce unequal consequences of illness in social, economic and health terms (individual interactions); 2) policies to reduce risk of exposure of disadvantaged people to health-damaging factors (community); 3) policies to reduce exposures of disadvantaged people to health-damaging factors (public policies); and 4) policies on stratification to reduce inequalities and mitigate effects of stratification (environment). --- Results Electronic searching yielded 4182 citations of which 293 were examined in full text (Fig. 1). Fifty-four qualitative studies were included [12,. The comprehensiveness of study reporting is shown in Additional file 1 and in the Additional file 2. Studies reported between 4 and 26 of the 32 Consolidated Criteria for Reporting Qualitative health Research (COREQ) criteria. M<unk>ori consumer experiences were reported within a range of health settings and programs including primary care, public health and screening initiatives, community health programs, disability and hospice services and hospital-based care (Table 1). The number of M<unk>ori participants in each study ranged between 4 and 130. Twenty-eight studies reported funding from government sources. Eleven studies reported kaupapa M<unk>ori (M<unk>oriled) methodologies [12,24,35,39,56,59,61,[65][66][67]70]. In 24 studies, ethnicity was reported as self-identified by participants [26, 27, 31, 34, 35, 37, 40, 45, 47-50, 55, 57, 59, 61, 63, 65-70, 72]. In two studies, ethnicity was identified through name or records, [29,56]. Most (38 studies) studies aimed to evaluate consumer experiences and perceptions of healthcare and health service delivery. Fewer studies evaluated patient experiences as they related to healthcare implementation and policy (8 studies), [46, 52-54, 56, 67, 68, 70] cultural appropriateness of healthcare (6 studies) [39,43,47,50,66,69] or to identify the causes and impact of health inequality (1 study) [36]. ---... not data available or none reported M<unk>ori consumer experiences of health that were mapped to social determinants of health inequity were most frequently identified as direct interactions with the health system (Fig. 2). These included patient-clinician communication and relationships, clinician cultural competency, tikanga (cultural mores) in health services, the physical clinical environment, wh<unk>nau (extended family) involvement in care, patient support, clinical respect for patient and clinician availability. Reported M<unk>ori consumer experiences mapped to intermediary factors that lead to differential exposures to healthdamaging factors included health beliefs, cultural (dis) connectedness, costs of clinical care and internalised blame. M<unk>ori consumer experiences in the available studies were less commonly mapped to socioeconomic positioning (previous health experiences, clinician ethnicity, racism, health literacy and socioeconomic factors) and the socioeconomic or political context (public health promotion and colonisation). Based on M<unk>ori consumer experiences, the most frequently recommended actions to improve M<unk>ori experiences of healthcare were aligned with reducing risk of exposure to health-damaging factors (such as integration of tikanga (cultural mores) in health services, health literacy interventions, increasing M<unk>ori workforce capacity and involvement in health service development, resources for cultural competency, accessibility of health services and clinician responsiveness to M<unk>ori consumers) (Fig. 3). Recommended actions to reduce the unequal consequences of illness in social, economic and health terms included culturally relevant interventions, support for wh<unk>nau (extended family)-based care and involvement in the health system, holistic models of care and reflexive clinical practices. Strategies aimed at reducing exposures to health damaging factors included improved referral practices, reducing clinician bias, increased awareness of health determinants and provision of cultural competency frameworks and strategy. Proposed strategies aligned with mitigating the effects of socioeconomic and political stratification included funding of health services including increasing specialist services, socioeconomic policy actions and preventative health care and programs. Researchers also suggested mechanisms to monitor and follow-up on health equity based on M<unk>ori consumer experiences. These included data collection processes, information sharing and empirical health research (Fig. 3). --- Discussion This paper summarises the determinants of M<unk>ori consumer experiences in health services and programs reported in qualitative studies and the actions recommended by researchers based on their findings. The findings have been mapped to the CSDH frameworks of health inequities to synthesise a range of potential strategies that might Fig. 2 Reported M<unk>ori consumer experiences of Aotearoa/New Zealand health services and programs mapped to the Commission for Social Determinants of Health (CSDH) conceptual framework of health inequities [16] address M<unk>ori health inequities informed by consumer experiences. Based on qualitative data from several health settings, direct consumer interactions with the health system and programs are important informants of the determinants of health inequity in Aotearoa New Zealand, suggesting this as a priority area for quality improvement. Aspects of care that were particularly noted were patient-clinician relationships and communication, including clinician cultural competencies. Clinical services lacked alignment with tikanga (cultural protocols and processes) and involvement of the wh<unk>nau in healthcare. Other intermediary factors that contributed to health inequity included costs of clinical care to consumers and internalised blame as a consequence of racism, leading to altered (reduced) care seeking actions. Systemic factors identified from exploration of M<unk>ori consumer experiences included socioeconomic and political factors such as colonisation, public health policies, institutional racism, power imbalances between clinicians and patients and health literacy. Mapping of the strategies that researchers recommended to improve M<unk>ori consumer experiences to the CSDH framework for tackling health inequities enabled the CDSH framework to be aligned to M<unk>ori patient, wh<unk>nau and consumer perspectives of the health system. Most recommendations provided strategies to reduce the risks of exposures to health-damaging factors experienced by M<unk>ori as the direct consequences of colonisation and racism. These responses included expanded use of tikanga and culturally competent practice in health services, capacity-building to support M<unk>ori participation in the health workforce and health service development, and greater access for M<unk>ori to clinical services including health system responsiveness to M<unk>ori consumer expressed needs and expectations. Health system level strategies based on consumer perspectives included greater funding of health services, expanding specialist services for M<unk>ori and increasing preventative health care and programs. Policies to reduce the unequal consequences of illness that further drive structural inequity included the development of culturally relevant interventions and wh<unk>nau (extended family) rather than individual-centred care. The actions identified by researchers that might reduce health inequities for Maori consumers in this synthesis are concordant with evidence within other national settings including Canada and Australia [73]. These include strengthening community-governed health services, addressing power imbalances during Indigenous patient interactions with health services through trust, reciprocity, and shared decision-making, as well as avoiding a deficit model of non-adherence by Indigenous patients as an explanation for health outcomes. Similarly, in a critical interpretive synthesis of healthcare in the United Kingdom among patients with socioeconomic disadvantage, equity of access was conceptualised as a complex interplay between social context and features of the health service Fig. 3 Researcher recommended actions to improve M<unk>ori experiences of Aotearoa/New Zealand health services and programs mapped to the Commission for Social Determinants of Health (CSDH) conceptual framework for tackling social determinants of health inequities [16] such as patient navigation and the permeability of health services to specific patient communities [74]. These findings are consistent with the present analysis that identified potential strategies to address inequities and improve services for M<unk>ori consumers informed by experiences include modifying referral structures, and increasing service and clinician accessibility and responsiveness. These findings suggest that, in addition to the role of qualitative research to evaluate individual and community-level M<unk>ori consumer experiences within specific health settings and encounters, there is an untapped potential for qualitative and participatory research with M<unk>ori consumers to inform the development and implementation of effective policies and interventions that reduce inequities and exposure to health damaging factors at a broader macro-level [75]. Aotearoa New Zealand has a governance system with the capacity to address health inequity as required by the Treaty of Waitangi. This system includes robust quantitative data collection and reporting on social determinants of health, legislative structures that enable intersectoral action on equity, a governmental framework linked to budget, and a strong public health system. While still relatively infrequent, qualitative research that is co-designed with M<unk>ori consumers has provided examples of how participatory research actions can inform system and policy-based change to address inequity. For example, research informed by M<unk>ori consumer and clinician experiences has led to consumer-designed health service improvements including increased specialist assessment and diagnostic services in rural Tai Tokerau/Northland [12] and a culturallyacceptable intervention to support safe bed-sharing (the wahakura sleeping pod), [21] that has subsequently been adopted as policy and evaluated in a randomised controlled trial [76,77]. Clinical quality improvement programs in Aotearoa New Zealand do not always improve health service experiences and outcomes for M<unk>ori and may exacerbate rather than address inequities [8][9][10]. Recent examples include inequities in childhood immunisations that were nearly eliminated in 2014 but have re-emerged through progressive loss of initial gains in immunisation rates among M<unk>ori children [7]. Similarly, progressive improvements in diabetes monitoring and kidney disease screening for non-M<unk>ori have not occurred for M<unk>ori patients. Knowledge of M<unk>ori consumer experiences of health care to inform understanding of structural determinants and intermediaries of health inequities may support healthcare quality improvements that sustainably reduce the unequal distribution of quality health care. Empirical analysis has demonstrated the necessity of critical reflections of power and institutional culture in the sustainable delivery of programs that are aimed to impact on health inequity [78,79]. The strengths of this paper include the a priori use of a conceptual framework that considers the structural determinants of health (CSDH), a broad literature search within multiple electronic databases and two levels of coding using established methodological processes. There are also limitations of this study that need to be considered when interpreting the findings. First, we may have not identified all the relevant qualitative studies available in the literature due to challenges in retrieval of qualitative research and a lack of searching of grey literature including governmental and non-governmental reports [80]. Second, we did not include qualitative studies exploring perspectives of health providers in the review which may have captured additional methodological approaches and theoretical frameworks in existing qualitative research of M<unk>ori consumer experiences. Third, the nature of qualitative research of patient experiences is frequently focused on the individual and their immediate family, and accordingly the type of study we evaluated would be most likely to examine the direct health consequences of unequal exposure to risk factors for disease and illness. --- Conclusions M<unk>ori consumer experiences of health services and programs are an important informer of variables that impact health inequity. Strategies to tackle health inequities informed by M<unk>ori consumer experiences can be drawn from existing empirical research. Future qualitative exploration of how socioeconomic, political and public policies influence M<unk>ori consumer experiences of health services and programs could inform a broader range of structural policies to address health inequities. --- Availability of data and materials The dataset generated and analysed during the current study is available in the Mendeley Data repository, https://doi.org/10.17632/wgmwjscsn9.1. --- Supplementary information Supplementary information accompanies this paper at https://doi.org/10. 1186/s12939-019-1057-4. Authors' contributions SCP conceived the study, screened the literature for eligible studies, extracted data, conducted the data coding and analysis, and drafted the manuscript. HG screened the literature for eligible studies, extracted data, conducted the data coding and analysis, and provided intellectual input into the manuscript. TH, CL, and LB were major contributors in writing the manuscript. SP conceived the study with SCP, conducted the data coding and analysis, and was a coequal contributor with SCP in writing the manuscript. All authors read and approved the final manuscript. --- Additional file Ethics approval and consent to participate Not applicable. The research is a meta-analysis of publicly available studies. --- Consent for publication Not applicable. --- Competing interests The authors declare they have no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Persistent inequities in health experiences and outcomes are observed for Māori compared to non-Māori in Aotearoa New Zealand. We conceptualised factors associated with Māori consumer experiences of health programs and services and characterise how the recommendations arising from qualitative research inform strategies to address inequities. Methods: In this systematic review, electronic literature searching was conducted in February 2018. Qualitative studies reporting Māori consumer experiences of health services and programs in Aotearoa New Zealand were eligible. Māori consumer experiences of health services were mapped to the WHO Commission of Social Determinants of Health (CSDH) conceptual framework on health inequities as related to: (i) the socioeconomic and political context; (ii) socioeconomic positioning; or (iii) intermediary factors that increase exposure to health-compromising conditions. Recommendations to improve consumer experiences were mapped to the CSDH framework for tackling social determinants of health inequities as policy directions on: (i) unequal consequences of illness (individual interaction); (ii) risks of exposure to health-damaging factors (community); (iii) exposures to health-damaging factors (public policies); and (iv) mitigating effects of socioeconomic and political stratification (environment). Results: Fifty-four studies were included. Māori consumer experiences mapped to social determinants of health inequities were most frequently related to direct interactions with health services and programs, particularly patient-clinician interactions (communication, relationships) and cultural competencies of clinicians and the system. Key recommendations by researchers mapped to potential strategies to address inequity were identified at all levels of the political, social and health system from individual interactions, community change, and broader public and system-level strategies. Recommendations were predominantly focused on actions to reduce risks of exposure to health-damaging factors including health literacy interventions, increased resources in cultural competencies and Māori capacity in health service development and workforce. Conclusions: Māori consumer experiences of health services and programs are an important informer of variables that impact health inequity. Strategies to tackle health inequities informed by Māori consumer experiences can be drawn from existing empirical research. Future qualitative exploration of how socioeconomic, political and public policies influence Māori consumer experiences of health services and programs could inform a broader range of structural policies to address health inequities.

Introduction Missed appointments and no-shows are a challenge to public service systems (J. A. DeFife, Conklin, Smith, & Poole, 2010;Pillay, Wassenaar, & Kramers, 2004), and in psychiatry, the rate of non-attendance i is double that of other specialties (Killaspy, Banerjee, King, & Lloyd, 2000;Stewart, 2013). Young people are especially vulnerable in regard to mental health services (Munford & Sanders, 2016), and due to a gap between child/adolescent and adult mental health services, many youth risk not getting the treatment they need in time or at all; only about half of children and adolescents with mental health difficulties receive treatment at all (Weitkamp, Klein, & Midgley, 2016). Furthermore, youth who do receive treatment have the highest risk of missing appointments (Chariatte, Michaud, Berchtold, Akre, & Suris, 2007). Missed appointments prevent people with mental health difficulties from getting professional help, increase risk of hospital admission (Chariatte et al., 2007), delay access for other patients (Kenwright & Marks, 2003) and increase the risk of suicide attempts (Pillay et al., 2004). Hence, non-attendance is not only a problem for service users, but it is also a problem in regard to inefficient use of resources as well as for staff morale (J. A. DeFife et al., 2010;Kenwright & Marks, 2003;Wiseman & McBride, 1998). Because of high levels of non-attendance throughout mental health services, there has been great attention in the literature on understanding factors contributing to service (dis)engagement and what might enhance attendance rates (Anderson et al., 2006;Block & Greeno, 2011;Jared A. DeFife, Smith, & Conklin, 2013;Edlund et al., 2002;Gibson, Cartwright, Kerrisk, Campbell, & Seymour, 2016;Henzen, Moeglin, Giannakopoulos, & Sentissi, 2016;Iachini, Hock, Thomas, & Clone, 2015;Munford & Sanders, 2016;Munson et al., 2016;Sheridan & McArdle, 2015;Thompson, Bender, Lantry, & Flynn, 2007). In two public service sectors in Denmark, the RENEW (Rehabilitation for Empowerment, Natural Supports, Education and Work) (Malloy, 2013) model was introduced to develop a targeted intervention to help young adults with mental health difficulties become more engaged in services as well as move closer toward attaining education. The initial focus of the project was not to investigate non-attendance, but to develop and adjust RENEW to the Danish setting through a participatory research approach. However, throughout 2015, it became clear that non-attendance was a prominent focus point in the public service sector, and hence, it was an inevitable topic in professionals' discussions as well as the occasional source of great frustration. There were many tape-recorded discussions about how RENEW could be qualified to try to reduce non-attendance; therefore, it made sense to investigate how this predominant focus on non-attendance has influenced professional's understanding of their work and of youth. In this article, we investigate professionals' discussions about non-attendance. By using discourse analysis, we will be able to better understand professional frustration by drawing attention to inherent and perhaps inappropriate practices which are based on naturalised and common-sense assumptions (Fairclough, 1992c, p. 90). Moreover, understanding discourses about non-attendance will help to reveal established normative organisational logics and identify ongoing or future cultural changes within the organisations (J<unk>rgensen & Phillips, 2002). This will enable professionals to be critical towards their own, perhaps inappropriate practices. Crawford et al. has described how compassionate language seems to be reduced due to time pressure in an environment of process-focused as opposed to person-focused care (Crawford, Gilbert, Gilbert, Gale, & Harvey, 2013). Kenwright and Marks describe how non-attending patients are referred to as 'unmotivated' or 'unreliable' (Kenwright & Marks, 2003), which might have unfortunate consequences for patient-staff relations. Based on empirical data from the participatory work with the RENEW model in Denmark, the purpose of the article is to investigate common-sense assumptions and inherent organisational logics by identifying discourses in professionals' negotiations about youth's non-attendance. Furthermore, we will discuss how potential tensions between discourses can be understood in the context of a broader social context of public service sectors targeting youth with mental health difficulties. --- The present study This article is based on empirical data from a study with a participatory approach in which a researcher and professionals have been adjusting and further developing the RENEW model (Rehabilitation for Empowerment, Natural Supports, Education and Work) (Malloy, 2013). In Denmark, RENEW targets young adults aged 18-30 with mental health difficulties, and the project was implemented in two public health service sectors: out-patient treatment in mental health services in the Capital Region of Denmark, and in an occupational centre in the City of Copenhagen ii. Secondary mental health services are free once patients have been referred through their general practitioner, and patients have a right to get a psychiatric assessment within one month, and a right to receive treatment within one to two months, depending on the severity of the disease. Treatment in out-patient services for non-psychotic disorders is restricted through so-called 'packages' comprising a certain number of different services dependent on diagnosis, e.g. patients with anxiety disorder receive a standardised package of 15 hours of clinical treatment altogether (Danish Regions, 2014). Treatment is typically interdisciplinary and managed by a psychiatrist, a psychiatric nurse and/or a trained psychologist and, if needed, a social worker and a physiotherapist. There are different types of occupational centres in Denmark, with the purpose of getting citizens on public welfare services into either education or employment, depending on age. The centre in this study targets young adults on social welfare benefits aged between 18 and 30 years. Most young adults connected to the centre have mental or social difficulties, and some have psychiatric diagnoses. The centre's main purpose of getting youth into education is aligned with the political agenda in Denmark, where it is decided that minimum of 95% of a youth generation should complete a formal education after finishing compulsory school levels (Cederberg & Lingärde, 2008;Danish Ministry of Education, 2015). Once connected to the centre, it is obligatory to attend meetings, and if youth fail to appear, their social benefits are reduced dependent upon the number of missed appointments. RENEW was developed in an American setting, and therefore, the purpose of this project was to further develop it in collaboration with staff, in order for it to make sense in the two Danish public service organisations. In December 2014, professionals were trained to work with youth in accordance with the RENEW model, and throughout 2015, professionals and a researcher held regular meetings to discuss the content of the model and implementation challenges. Throughout this process, professionals worked with youth while also participating in developing and adjusting the model. Hence, they could use practical experiences in the development process and test the changes made along the way. When working with the model, the researcher introduced the common term 'youth' to describe the target group instead of public sector service conventionally used terms like 'patient' for psychiatry or 'citizen' for the occupational sector. This allowed professionals to speak a more common language. The term 'youth' will therefore also be preferred throughout the remainder of this article. --- Method Data collection Throughout the development process, the Michaela Hoej had offices both in the occupational centre and in the mental health clinic. Hence, she was part of the research field. In addition to her, five interdisciplinary professionals from the occupational centre and six from the mental health centre participated in the empirical project activities. The mental health services project team consisted of three nurses, one social worker, one psychologist and one psychiatrist, while the team in the occupational centre consisted of two nursing aids, one psychologist, one nurse and one team member with a Master's degree in social science iii. Recruitment for RENEW was voluntary in both the occupational centre and the mental health clinic, and professionals were informed of the research character of the project from the beginning. In the outpatient clinic, team members consisted of approximately 15 to 20 professionals, and 4 volunteered to work with RENEW, whereas in the occupational centre, recruitment took place through a listing of a vacant positions on the intranet, where all 100 staff members could apply; however selected staff members were encouraged to apply. Four people were chosen out of the applications received. The qualitative empirical data in this article were collected throughout 2015, see Table 1. Summaries as well as transcriptions were then read thoroughly through while simultaneously listening to recordings, and discussions surrounding absence, no-shows or non-attendance was coded as such in NVivo. "To achieve the fullest possible understanding of the material, a rough and exploratory thematic analysis on discussions about non-attendance was then made. --- Type of data Context This did not result in selections or exclusion of the material, but rather it ensured a preliminary overview. This preliminary analysis was then followed by a more thorough critical discourse analysis based on selected analytical tools from Norman Fairclough's framework (Fairclough, 1992c)." This will be further elaborated below. Data were collected and analysed in Danish, and the chosen excerpts for the article were then translated into English. The primary writer did the initial translation and this as well as original citations in Danish were presented to co-authors, so all translations could be discussed to reach a mutual decision on the best possible translation. --- Ethics The project was approved by The Danish Data Protection Agency (journal number 03610 and ID-number: RHP-2015-006). The Regional Committee on Research Ethics was also contacted for approval (Protocol number: H-7-2014-FSP15), but the project was not liable to notification, because no biological material was included in the research. Hence, no approval was necessary. Furthermore, the Danish National Board of Health was contacted (Case number 2014111813), but the project was not liable for notification here either. When observations and interviews were conducted, informed verbal consent was collected from both professionals and youth. Participation in the research was not a condition for working with RENEW, but all professionals agreed to participate anyway. We chose not to collect written consent for several reasons. Firstly, data collection was ongoing throughout 2015, and it seemed more appropriate to start every meeting and workshop with an oral reminder that the meeting was being recorded for research purposes than to collect written consent only once in the start-up phase of the project. Furthermore, oral consent seemed less formal, and the aim was to create a relaxing and informal atmosphere between the researcher and professionals. Whether to involve youth more in the development process was considered at the beginning of the project, but since this was a process lasting more than a year, which would be longer than some youth participated in RENEW, it was decided to primarily involve staff. Youth was still involved in the development of RENEW through focus groups. However, the empirical data from these are not used in this article, since the focus is on which discourses professionals use when discussing non-attendance. Assumptions, theoretical framework and analytical strategy As described above, discourse analysis can help shed light on inappropriate and unacknowledged organisational common-sense practices (Fairclough, 1992c;J<unk>rgensen & Phillips, 2002), and this is what we wish to do in this article. We wish to further our understanding of inherent organisational logics by analysing discourses about nonattendance. We have identified discourses by using concepts from Fairclough's critical discourse analysis framework, because there is specific focus on linking the linguistic tradition to social practice, and by doing so, understanding microsocial as well as macrosocial processes (Fairclough, 1992c). Methodologically, this article has a social constructivist point of view, understanding reality as not in itself immediately accessible, but shaped and framed through discourses. The relationship between discourse and social practice is dialectical in the sense that society produces discourse and discourses also produce and reproduce society. Hence, discourse is a reflection of reality as well as a constructer of social identities, social relations, as well as knowledge and meaning structures (Fairclough, 1992c(Fairclough,, 2001a)). In our analysis, we have sought to investigate this dialectical relationship primarily through looking at the concept of intertextuality which comprises the fact that texts vi absorb and are created by drawing on fragments from earlier texts (Fairclough, 1992b(Fairclough,, 1992c)). Hence, texts constitute history and facilitate change by responding to, accentuating and reworking past texts. Therefore looking at intertextuality of a text can clarify earlier assumptions and provide insight into occurring changes. Intertextuality is a useful concept in analysing how concrete discursive practices can reflect wider macrosocial practices (Fairclough, 1995). As will be shown later in the article, in our analysis, we have utilised it to assist our understanding of how professionals' discourses about non-attendance reflect organisational logics. We have analysed intertextuality by identifying where the texts explicitly refer to an earlier text (manifest intertextuality) or where they draw on specific orders of discourses. An order of discourse is "the way in which diverse genres and discourses and styles are networked together. An order of discourse is a social structuring of semiotic difference in a particular social ordering of relationships amongst different ways of making meaning, i.e. different discourse and genres and styles." (Fairclough, 2001b, p. 2). Hence, orders of discourses can be understood to be the sum of different discursive types that exist within a given domain (J<unk>rgensen & Phillips, 2002). Inspired by the critical discourse analysis, we have identified discourses by analysing the textual dimension descriptively. Focus has been placed on the linguistic characteristics of the text, such as vocabulary. Specifically, we examine the ambivalence vii in professionals' discourse representations viii to describe the entanglement and complexity of non-attendance, and we pay especial attention to whether they are sentimental or distant (Fairclough, 1992b) to analyse which discourses the chosen wordings might represent. We have also been particularly interested in presuppositions, which is the 'taken-forgrantedness' the text producer communicates as an already established fact (Fairclough, 1992b). Presuppositions are especially valuable in identifying inherent organisational logics in professionals' discourse representations. Finally, we have linked texts to context by examining the social practice (societal context) with focus on ideology and discursive battles. Fairclough argues that discourses can be understood to be ideological, since ideologies are 'built into various dimensions of the forms/meanings of discursive practices and [they] contribute to the production, reproduction or transformation of relations of domination." (Fairclough, 1992c, p. 87). The most efficient ideologies are embedded in discourses and become naturalised or common-sense. When this occurs, hegemony is achieved. Hegemony should not, however, be understood exclusively as a sign of dominance, but rather as a form of leadership in which negotiation and constructing alliances through arguments takes place (Fairclough, 1992a, p. 91ff). Therefore, some ideologies hold more power in practice, and understanding the discursive battles can help us understand the professionals' social practice. An example of this can be found in the introduction, where it is described how time pressure effects compassionate language in care (Crawford et al., 2013). Therefore, our focus has been to look for ideologies behind discourses and to determine whether discourses reveal common-sense practices in order to determine power balances. --- Analysis In the following analysis, we present discourse representations about non-attendance as well as described and observed explanations and strategies. The purpose is not to identify which particular explanations or strategies are used, but rather how they are represented discursively and what underlying assumptions they represent. We will commence by introducing the complexity of understanding non-attendance. Then, we will describe the discourses derived in our analysis. We will analyse how these discourses can be understood based on intertextuality, and the focus will be on wording, ambivalence and presupposition. We will then look at the social practice by discussing which orders the discourses are part of and how some orders of discourse might hold more power than others. Ambivalence in discourse representations due to entanglement and complexity of discourses This example shows the complexity by illustrating the ambivalence in the discourse representations: Looking at vocabulary, Professional 3 argues that youth do not attend because they are not able (line 11) to do so, due to their anxiety (line 9). She says that youth has problems, indicating that their mental health problems are the main reason for their not attending. Although in line 12-14, the second professional understands and expresses agreement, she, as does the first in line 20, still uses the word obliged in line 8. This presupposes that youth stay at home when they do not feel obliged to come. This presupposition is also visible in line 16, in which the third professional highlights that youth have said that they liked RENEW, which indicates that not liking RENEW would be enough to make them not attend. The concept of ambivalence, also described earlier, that appears in this dialogue was present throughout the data, and the ambivalence was not dependent on membership of professional group or on professional background. We will now describe the three discourses identified in our analysis: the responsibility discourse, the solicitude discourse and the youth culture discourse. As exemplified in the discussion above, the discourses are not delimited entities, but rather discourses appear as a mixture of explanations and strategies to deal with non-attendance. Each discourse offers a different perspective on the perception of youth as a result of which words are chosen in the particular discourse representation, and each discourse emanates one way of understanding the societal development, or one order of discourse. In the following section, we will analyse how these three discourses can be identified and discuss what orders of discourse they draw on. --- Solicitude discourse The solicitude discourse is constructed with many sentimental utterances, as Fairclough might label them (Fairclough, 1992b, p. 272), and discourse representations entail descriptions of the target group as vulnerable patients in need of help. For example, in the above discussion when Professional 3 focuses on problems and anxiety as primary explanations as to why youth do not attend, this is done though a solicitude discourse. This discourse is highly salient throughout the data, and in addition to anxiety, it draws on other symptom explanations, such as depression, personality disorders, etc. Hence, the primary focus is on mental illness and vulnerability, and typical examples of discourse representations explain how it can be difficult for youth to get out of their bed, that they are anxious to come or that it can be hard for youth to attend a new service at a new place and meet new people due to their illness. Youth are explicated as fragile and in need of care, and, naturally, the primary roles of professionals in these discourse representations are the roles of caregivers and helpers. Professionals describe how it can be necessary for them to aid in making it easier for youth to attend, including, for example a) helping with transportation (on occasion, young adults from the mental health centre would have the possibility of being picked up in a taxi); b) sending them a friendly reminder of the next session via text; or c) trying to be flexible with the planning of sessions, e.g. not scheduling sessions in the morning, when it might be difficult for youth to attend due to irregular sleeping patterns. Other explicated strategies include family engagement, e.g. arranging parents to drive youth to RENEW, or making sure their boyfriend encourages them to attend the appointment when the youth finds it difficult to leave home. We can attempt to understand this discourse through looking at intertextuality and relating to a wider social practice. By doing this, we can see that focusing on vulnerability and solicitude is related to a traditional understanding of the nursing role. According to the American Nursing Association nursing can be defined as: "the protection, promotion, and optimization of health and abilities, prevention of illness and injury, facilitation of healing, alleviation of suffering through the diagnosis and treatment of human response, and advocacy in the care of individuals, families, groups, communities, and populations" (The American Nursing Association, 2017). In patients' perspective organizational factors also effect quality of nursing (Irurita, 1999), and according to nurses themselves high quality nursing care entails interactions with family and should address 'all needs' of patients (Williams, 1998). The role of the nurse have been described like that of a mother, involving physical touch, and it is connected to being intimate, engaged and empathically caring (Helman, 2007, p. 162-163). Thus, the solicitude discourse draws on an order of discourse, which might be called the nursing order of discourse. It entails an understanding of the youth as ill, and it focuses on the illness as a breakdown in the body (Helman, 2007, p. 32). Therefore, in order to help the body and mind recover, professionals will need to assist with activities which would normally be categorised as 'private', e.g. help with sleeping, eating, grooming, etc. (Nettleton, 2013, p. 111). This type of work has also been characterised by Strauss et al. as sentimental work (Strauss, Fagerhaugh, Suczek, & Wiener, 1982), which further stipulates the sentimentality of the discourse representations within this order of discourse. --- Responsibility discourse The responsibility discourse is detectable when professionals accentuate that the youths themselves have a responsibility for attending services, and one strategy to try to make youth feel more responsible for attending is trying to create a sense of obligation. This strategy is represented in the positions taken by the two professionals in the discourse representation above, in which they focus on the importance of making youth feel obligated to come. The responsibility discourse can also be seen in the following discourse representation from a dialogue meeting in which a social worker from mental health services, talks about nonattendance with a trained nursing aid from the occupational service. underlines youth agency and the fact that professionals can only do so much to try to make them attend. Therefore, trying to make youth attend becomes a negotiation process. This is also detectable in lines 3-4 where business-inspired linguistics are used; professionals have to do what they can to sell the service and make it interesting. This wording highlights the fact that negotiating the terms of attending is a bilateral process in which the youth is the consumer of the service and the professionals are the salespeople. Youths, therefore, have to take responsibility for participating in this negotiation. In this discourse representation, the focus is not on what Fairclough describes as sentimentalities, such as illness or problems, but rather, more distantly or prosaically on negotiating terms and making youths take responsibility. Other negotiation strategies employed in the field to deal with non-attendance are outlined at the beginning of the RENEW process: using a RENEW contract, which is signed by both the youth and the facilitator, or explicating expectations that youth attend and, if not able to do so, at least give notice that they are not attending. Professionals also frequently discuss youth motivation, e.g. how can professionals work on making youth reflect on reasons for participating or help them see their own progress since starting. When youths are the ones who are responsible for coming, they must obviously be motivated to come; hence, focusing on motivation also speaks into a responsibility framework. This discourse therefore places a great amount of responsibility on youth, making it not only their responsibility to attend, but also their responsibility to make the most of their own RENEW process. Placing responsibility on service-users is recognised as a general developmental tendency within the professional-patient relationship (Nettleton, 2013, p. 145), and it is part of a development within healthcare where prevention and health promotion are becoming increasingly important (Nettleton, 2013, p. 230ff). With this development, focus changes from hospitals and patients to people and their everyday lives. This results in focusing more on what measures, both preventative and curative, that individuals need to take themselves to live healthier lives. Therefore, strategies that attempt to enhance a sense of empowerment -encouraging decision-making skills and facilitating action -become increasingly important for professionals. As the responsibility discourse can be seen as part for a more general developmental tendency, we argue that this discourse is constructed through what we might call an empowerment order of discourse (Nettleton, 2013, p. 232). Focusing on empowerment might also have a 'flipside', however, when individuals are not able to take on this responsibility. This will be discussed further in the section The empowerment order of discourse holds more power than the nursing order of discourse. --- Youth discourse The final discourse we have identified is the youth discourse. Interestingly, here the professional focuses on age as an explanation of lack of focus rather than, for example, illness. The discourse representation makes a presupposed connection between age and ability to focus and assumes that youths are not able to handle too many things at once. The wording 'youth' about the participants in RENEW is also explicated frequently in the data, and as opposed to e.g. 'patients', 'citizens', 'clients' or'students', this draws the attention toward age and the generation as a whole. Professionals adapted this term from the participating researcher, because she used this term about the target group in trying to speak a common language between professionals in the two sectors. Normally, though, professionals in psychiatry would use the term patients (which is suited to the solicitude discourse), whereas professionals in the occupational centre would frequently refer to them as course participants xi. Thereby, the researcher played a central role, not only in the analysis, but also as a co-creator in regard to focusing on youth. In the following inquiry, the youth discourse is also represented. The discourse representation is from a dialogue meeting with professionals from both the occupational and mental health sector discussing how to deal with non-attendance. In this discussion a mental health nurse, expresses a wish to better understand the youths' generation to be able to come up with new strategies: Though she is educated and employed as a nurse in a psychiatric context that strongly accentuates solicitude and care, in the above discourse representation, Professional 6 still seeks more knowledge about youth nowadays and the culture (lines 7-8) as opposed to mental illnesses. In the above discussion, manifest intertextuality can also be identified; The professional explicitly refers to both an unnamed professor and a Noemi (Katznelson), who is a Danish professor researching youth culture and how to understand youth. Hence, in this discourse, the target group is perceived as a generation of youth with certain characteristics, which professionals need to be able to understand in order to be able to do their job, and nonattendance is described as a generation trait rather than a vulnerability. P6: "What I'm The youth discourse draws on what we could call a youth culture order of discourse. This order of discourse is present in society in regard to how we understand youth culture as more disengaged and liberated, or as Anthony Giddens describes it, we are increasingly free to choose what we want to do and who we want to be (Giddens, 1991). Youths are described as the ''me, me, me' generation' who, compared to earlier generations, have a reduced sense of empathy and show little consideration for others (Karkov, 2012;Konrath, O'Brien, & Hsing, 2011).Youths are also described as demanding, as having a hard time concentrating and finishing things, and even as justifying their choice (and right) to live on social benefits (Pultz & M<unk>rch, 2015). In Denmark, the term zapper-culture is frequently used to describe youth as unfocused and moving quickly from one thing to another, and this term also appeared more than once in data. A similar English term NEET (Not in Education, Employment or Training) was first used to describe needs of 16-18 year olds not eligible for benefits, but later, the meaning has expanded both in age range and due to the fact that it is generally used to describe unemployed young people (Hutchinson & Kettlewell, 2015). With the expansion, certain characteristics have also been attached to NEET, e.g. low engagement and disruptive behaviour. This order of discourse is present in public spheres, but has also found its way into the psychiatric and occupational sectors according to this empirical data. However, in the above discourse representation, the youth discourse is not the only discourse that can be detected. The ambivalence described earlier can also be found. In line 9, the professional describes how she sometimes wishes youths would pull themselves together. This wording represents the responsibility discourse, in which youths actively decide not to show up, and if so, it is because they do not wish to participate. Furthermore, the solicitude discourse is represented in lines 14-15, in which the professional expresses that she feels she sometimes needs to get a little perspective to understand what youth are faced with and what they bring with them. This presupposes youth being faced with difficulties, e.g. symptoms, bringing along some sort of difficult baggage that needs to be taken into consideration when dealing with them. --- Discussion We will now discuss the above findings by looking first at how the discourses are interrelated, with a focus on the tensions and similarities between them. We will then connect them to social practice by looking at inherent organisational logic and by analysing discursive battles between the different discourses. Furthermore, we will discuss how tensions between discourses and inherent logic leads to frustration for professionals, and how the youth discourse can be understood as a venting mechanism to cope with this frustration. To underline our analytical points further, data will also be introduced in this section. --- Discourse interrelation: tensions and similarities Looking at social practices, discourses about non-attendance are ambivalent, interwoven, tension-filled at times and even contradictory. The solicitude discourse and the responsibility discourse are, for example, inherently tensionfilled. The responsibility discourse emanates from an empowerment order of discourse that places agency and accountability on youths, whereas the nursing order of discourse places agency and responsibility on the professionals. Contrarily, the youth culture and the responsibility discourse have similarities, and they even coincide with regard to some of the strategies proposed to reduce non-attendance. For example, giving youths structuring tools, such as week or day planners, helping them to manage their time more constructively or praising them for regular attendance. Professionals also discussed the possibility of using 'time-outs' from the service, meaning youths should be banned for a period if they failed to give notice when not attending (this was never effectuated, though). More 'punishment' style strategies included different forms of sanctioning, e.g. (for the municipal part of RENEW) reducing youth social benefits if they neglected to attend without cancelling. A different 'punishment' style strategy discussed was saying names of non-attendees out loud in the group (which professionals knew youth would not like) or telephoning them. Interestingly, however, telephoning youth was explicated as both a punishment for not attending and a strategy to get them to come on other occasions. Hence, telephoning youth was both understood to be caring and punishing, dependent on context. Therefore, telephoning as a strategy is not immediately possible to ascribe to one discourse rather than another. The empowerment order of discourse holds more power than the nursing order of discourse So, how can we understand the interrelated and complex discourses and the ambivalent discourse representations? To answer this question, Fairclough suggests analysing discursive battles and power relations by looking further at inherent organisational logic and social practices. We claim that the empowerment order of discourse holds more power than e.g. the nursing order of discourse and build on this with two arguments: 1) As described above, the solicitude discourse is related to a nursing order of discourse, and the nursing order of discourse is an established discourse within healthcare systems. However emotional labour and nursing tasks are carried out by those in lower positions in the medical hierarchy (Helman, 2007, p. 94-95;Nettleton, 2013, p. 142-143), which means that not much power is associated with this discourse. 2) The empowerment order of discourse can be understood in the light of the governance regime, New Public Management (NPM), used for decades in public service systems. The empowerment order of discourse builds on the idea that individuals can and should take responsibility for themselves and their own lives. Working on making youth take responsibility then, in essence, is an indirect form of governing, since the purpose is for youth to self-regulate, or conduct their own conduct, rather than trying to manage them through authoritarian or disciplinary actions (Dean, 1999). Since empowerment can be understood as a strategy for governing individuals with the aim of reducing non-attendance, it can be understood in the light of NPM. NPM has been widely used to manage health systems through strategies such as decentralisation, visible control systems, goal setting through means of quantitative indicators of success, performance measurements and trying to raise efficiency levels by 'doing more with less' (Hood, 1991). One strategy to raise efficiency in an organisation can be to raise levels of empowerment among workers (Argyris, 1998;Kaymakç<unk> & Babacan, 2013). Similarly, in this project, one strategy to raise efficiency is empowering youth through increasing their sense of responsibility. Other strategies include registering attendance and formalising procedures through a LEAN xii -regime (Lawal et al., 2014), e.g. implementing an 'attendance-sheet' to be filled out by youths in order to contain possible future attendance challenges. These strategies all contribute to raising efficiency through reducing non-attendance, and they also match NPM strategies. Since the empowerment order of discourse can be understood in an NPM perspective, it becomes evident that this order of discourse carries more power than does the nursing order of discourse. The empowerment order of discourse reflects a governing style with great political and managerial support, whereas the nursing order of discourse, though persistent in public service sectors, is less influential. Since the empowerment order of discourse is more powerful than the nursing order of discourse, a related and highly relevant discussion becomes about how the empowerment order of discourse might influence social practice: Though empowerment in the above-described governing understanding might seem somewhat calculated, it is clear from the discourse representations in this material that professionals' aspirations to help youths take responsibility and become more empowered comes from a place of care. However, the question is: is it possible to 'give' someone empowerment? Yeich and Levine argue that it is not: "Empowerment seems to be a process that one must do for oneself -not something that someone can do for or to another" (Yeich & Levine, 1992). Furthermore, if empowerment is to be understood as an individual process, there will be individuals who cannot live up to this and therefore do not become empowered or able to take on this responsibility. This group therefore risks further marginalisation as a result of the empowerment order of discourse. --- Professional frustration It has been established that

Non-attendance constitutes a profound challenge in public sector services targeting young adults with mental health difficulties. Therefore, researchers and practitioners are occupied with trying to resolve this. In order for clinicians to be aware of their own naturalised and perhaps inappropriate communicative practices, we investigated the established normative organisational logics behind explanations and strategies related to non-attendance. We performed a critical discourse analysis on material collected through participatory research throughout 2015. Three discourses were identified: solicitude, responsibility and youth discourse. Although the discourses were complex and entangled, they were used by all practitioners. Furthermore, some of the discourses, especially the responsibility and the solicitude discourses were inherently tension-filled, and practitioners experienced frustration in dealing with these tensions. The youth discourse can be understood as a coping mechanism to deal with these tensions because it distributes responsibility for non-attendance to general social and cultural processes.

reflects a governing style with great political and managerial support, whereas the nursing order of discourse, though persistent in public service sectors, is less influential. Since the empowerment order of discourse is more powerful than the nursing order of discourse, a related and highly relevant discussion becomes about how the empowerment order of discourse might influence social practice: Though empowerment in the above-described governing understanding might seem somewhat calculated, it is clear from the discourse representations in this material that professionals' aspirations to help youths take responsibility and become more empowered comes from a place of care. However, the question is: is it possible to 'give' someone empowerment? Yeich and Levine argue that it is not: "Empowerment seems to be a process that one must do for oneself -not something that someone can do for or to another" (Yeich & Levine, 1992). Furthermore, if empowerment is to be understood as an individual process, there will be individuals who cannot live up to this and therefore do not become empowered or able to take on this responsibility. This group therefore risks further marginalisation as a result of the empowerment order of discourse. --- Professional frustration It has been established that the empowerment order of discourse holds more power than the nursing order of discourse, and that there is a predominant organisational focus on raising efficiency. The tensions between the two discourses and the organisational focus on efficiency also cause frustration and exhaustion among professionals because they perceive it as a pressure: P6: "We feel a pressure when youth don't show up. We don't have enough confrontation hours with youth, and then we feel pressured to come up with solutions to solve this problem. (...) It's a defence against system pressure, right." In the discourse representation, the professional expresses how they feel obliged to act, but at the same time, they lack meaningful and efficient options available for action. Furthermore, professionals also express the feeling of being 'caught between a rock and a hard place' (direct quotation). In the organisations, there is, as described, a profound focus on attendance, and professionals explicate how it would be nice to sometimes focus on different aspects in their work. As expressed in the above quote, professionals feel pressured and caught in the middle, and they express, how they feel that many of the strategies available to deal with non-attendance seems inappropriate and lacking effect. For example, at one methodological meeting, Barbara explained that management assigned her the task of trying to get her colleagues to integrate an exercise focusing on reducing non-attendance in their youth group work. She therefore presented the idea of integrating the exercise in the youth group to her colleagues. However, in a discussion about whether to use it, it became evident that Barbara herself did not find the exercise valuable and that she would have found other initiatives more helpful, e.g. installing a system with automatic text-message reminders to youth before sessions. Hence, professionals feel that some tools introduced by the organisation to try to reduce nonattendance are inappropriate and add pressure to both them and the youths; e.g. the above exercise or registration of non-attendance that might lead to a reduction in social benefits. Furthermore, they do not believe these tools will work in practice. At the same time, they feel that meaningful tools, e.g. a text messaging system, might actually work, but that the organisation does not acknowledge these tools or their potential benefits. This underlines the tension between the nursing order of discourse and the empowerment order of discourse, which leads to great frustration for professionals. --- A venting mechanism In this section, we will answer the following question: How does the youth culture order of --- Implications for practice We have described how the inherent tensions between the empowerment and the nursing order of discourse might result in professional frustrations leading to integration of the youth culture order of discourse in practice as a coping mechanism. Therefore, one important implication for practitioners is being aware of how focusing on youth responsibility for attending is part of an empowerment order of discourse, which can be understood as being connected to NPM. This is especially important in light of the newest literature on the ineffectiveness of NPM (Hood & Dixon, 2015). Furthermore, the professionals' experiences of pressure due to organisational attempts to streamline health care services are also described elsewhere (Allen, 2015;Martin, Armstrong, Aveling, Herbert, & Dixon-Woods, 2015). Therefore, this article adds to the existing discussion dealing with how health care professionalism might be challenged by management initiatives focusing on optimal utilization of services. A different but equally important practical implication is the risk of further marginalisation of the most vulnerable youth group when choosing an empowerment-based approach in trying to reduce non-attendance. Finally, this article calls for reflections on the implications of the use of the youth culture order of discourse as a coping mechanism for professionals. By introducing more general cultural and social explanatory processes for youth non-attendance, this also alleviates professionals of their responsibility to deal with the issue. Professionals and policy-makers in countries such as Denmark, that pride themselves on having strong social safety net in the form of social welfare systems for the most vulnerable groups, must ask themselves if this would be a desirable outcome If not, we need to assist professionals in dealing with the frustration they experience due to the tensions between the empowerment and nursing orders of discourse -perhaps simply by acknowledging them and reflecting on them collectively. representations mutually afterwards. It is most likely unavoidable that modest nuances of meaning might be perceived slightly different. However, we argue that the important meaning structures are preserved because of a rigorous translation process. Furthermore, to ensure transferability, we chose to describe original Danish meanings of words in the text and in notes, when translations were debated between authors. The critical discourse analysis was chosen for two reasons. The first reason is that we have not been able to find critical studies on discourses of non-attendance. The literature regarding non-attendance is action-oriented and deals with prevalence, as well as reasons and strategies for professionals. However, we strongly advocate the importance of knowing the contingencies of having a profound focus on non-attendance, which is the reason why this approach was chosen. A second reason was to try to create some analytical distance, since the primary author was part of the field of investigation and development process. For example, as earlier described, it was discovered in the analytical phase that the primary author might have contributed to a particular focus on youth. Hence, it was necessary to try to create both physical and analytical distance from the object of investigation. Fairclough argues that discourses constitute the social organisation of institutions, reflecting (and shaping) the norms and conventions embedded in them (Fairclough, 1992c). In the occupational sector, a predominant control-mechanism is reducing the social benefits for youths if they do not live up to their responsibilities (e.g. coming to services), and in the mental health sector, the overall aim is to heal people. Therefore, we might have expected that the professionals from the occupational sector would have been more inclined to use the responsibility discourse and that professionals from the mental health sector would have been more inclined to use the solicitude discourse. However, as described earlier, we did not see any clear distinctions between how discourses were used in the two areas. We have identified two possible explanations for this: 1) The professionals in the occupational sector were a selected, and perhaps not representative, group of occupational professionals, who might have been more inclined to use the solicitude discourse in their discourse representations than their coworkers. They did after all volunteer to participate in a project working with youths in cooperation with mental health professionals, and some were even trained health professionals who had earlier worked in mental health settings. These combined factors will most likely have influenced their perspective. 2) It could also be, however, that both empirical settings were part of the same public service sector, in which discourses were more influenced by the pressure of implemented performance measurements inspired by NPM (e.g. registering attendance), than they were by the overall organisational aims. Whether both, none or only one explanation applies, this will be worth investigating in similar public sector services. In the Introduction, we described how patients are described by Kenwright and Marks to be 'unreliable' and 'unmotivated' (2003). This might be a different example of a distancing process taking place in discourses about non-attendance. Therefore, it might be relevant to investigate whether this practice of placing responsibility for nonattendance in social and cultural processes takes place in other contexts than public social services targeting youth with mental health vulnerabilities, and, if so, what discourses might replace the youth discourse. --- Concluding comments In this article, we have investigated the complexity of understanding non-attendance in services provided for young adults with mental health difficulties. We uncovered three discourses used when discussing non-attendance: a solicitude discourse, which was based on understanding, with a focus on caring and providing the necessary help to assist what is believed to be a vulnerable target group in attending a needed service; a discourse drawing on ideas from NPW, in which negotiation of terms for attending is the focus, and the target group is primarily responsible for their own process; and finally, a discourse drawing on the understanding of the youth generation as disengaged, demanding and with low engagement as an explanation for their lack of attendance. We described how these discourses are often used simultaneously when discussing nonattendance, although there are inherent tensions between them. It is, for instance, contradictory to both provide unconditional assistance to aid youths in attending while also expecting youths to take responsibility for coming themselves. The tensions between discourses create frustration for professionals because they feel caught in what they feel is an inextricable situation. They experience pressure to act in accordance with the NPM understanding while also attending to their role as an empathetic and understanding caregiver. In order to cope with this situation, professionals search for new explanations that provide a strategy to place responsibility for dealing with youth non-attendance elsewhere, such as on societal and cultural processes. Explaining and understanding non-attendance through discourses from spheres which are normally not considered part of the mental health public service sectors might, however, have unforeseen consequences. Discourse representations regarding non-attendance in public mental health services that entail certain generationdependent presuppositions about youth might construct a different framework for the work. For instance, attributing the characteristics of the youth generation to the individual youth in the service takes away responsibility from professionals to deal with non-attendance, while also constructing new explanatory models for understanding youth with mental health difficulties. --- Notes i Consensus about the terminology on patients not attending services is not established in the literature. In this article, we have chosen the terminology 'non-attendance', unless referring to literature using a different terminology. ii Mental health difficulties constituted non-psychotic diagnoses, such as depression, personality disorders, anxiety and stress in mental health services, but young adults attached to the occupational service constituted a broader mental health profile, ranging from people in treatment for psychotic as well as non-psychotic disorders to people without diagnoses. iii The professional with a master's degree in social science did not participate in the workshops, because she joined the RENEW team in September 2015. She did, however, participate in methodological meetings. iv Furthermore a student assistant participated at occasions to help with practicalities such as video recordings etc. v There was one meeting which was not audio recorded, but detailed written minutes of this meeting were taken vi Text is to be understood in a broad sense, including both written and spoken language as well as photos and non-spoken language (Fairclough, 1995) vii Fairclough on ambivalance:'Intertextuality is the source of much of the ambivalence of texts. If the surface of a text may be multiply determined by the various other texts that go into its composition, then elements of that textual surface may not be clearly or unambiguously placed in relation to the text's intertextual network, and their meaning may be ambivalent-different meanings may coexist, and it may not be possible to determine "the" meaning.' (Fairclough, 1992b) viii Instead of the term quotation, the term discourse representation is used when describing a text because it accentuates that particular words are chosen as opposed to other words, and hence, that the words, tone, etc. is a (more or less) conscious choice (Fairclough, 1992b). ix All names in this article have been changed to protect informants' anonymity. x A youth group where youth meet to discuss their RENEW process and share experiences is part of the RENEW process in Denmark. xi The target group was actually 'kursist' in Danish, which directly translates to participant in course. xii There are different definitions of LEAN, but generally LEAN is described as a set of tools and techniques for improving practice through eliminating waste and adding value (Poksinska, 2010). --- 2017 --- Document Version Early version, also known as pre-print --- Data The empirical material generated in this project will not be publicly available due to the rules of the Danish Data Protection Agency. It will, however, be available from the corresponding author, after publication, upon reasonable request and following a signed confidentiality agreement with the Danish Data Protection Agency in the Capital Region of Denmark. --- discourse fit into the order of discourse puzzle? The tensions between the responsibility and the solicitude discourse lead to great amounts of frustration among professionals because they are the ones expected to act in accordance with both discourses (and it is debateable whether this is possible at all), while at the same time reducing non-attendance. Therefore, there is an inherent risk that professionals will be criticised for not fulfilling their duties to reduce non-attendance while also caring for the target group. Hence, professionals, being placed in this inextricable situation, seek new and different ways to understand non-attendance through a distinct and more general youth culture perspective, as exemplified by the request for more knowledge about youth culture. When professionals relate non-attendance to the youth generation and the youth culture, nonattendance is interpreted as a result of more general cultural and societal processes. Therefore, in this order of discourse, the problem of non-attendance is a general cultural trait, which must be handled at a general level, rather than trying to change the behaviour of the individual youth. Therefore, responsibility for reducing non-attendance becomes detached from both professionals and the individual youths. Hence, the youth discourse can be understood as a coping mechanism by which professionals allocate the responsibility for high rates of non-attendance to larger cultural and social processes. Furthermore, this framing entails a distancing of oneself from youth, since the problem of non-attendance is no longer related to the individual youth sitting in front of the professional, but rather the youth generation as a whole. This can be related to our earlier description of Crawford et al.'s conclusion that time pressure and an environment of processfocused care seem to reduce compassionate language (Crawford et al., 2013). Using less compassionate language due to time pressure and process-focused environments can be interpreted as a coping strategy as well, and it might be an expression of the same distancing process we have identified in our material regarding youth discourse. --- Limitations and recommendations for future research A majority of qualitative research uses interview data, but in this article, we chose to use empirical materials from'real-life' situations. Hence, we did not ask professionals to directly reflect and comment on non-attendance in an interview, but our empirical material on nonattendance comes from professionals' naturally occurring discussions on how RENEW could most appropriately be adjusted to a different context. We argue that this is a methodological strength, since the material then represents the professionals' discourse representations from actual situations, rather than 'artificial' interview situations, in which they were merely asked to describe or comment on these situations (Silverman, 1993). Furthermore, in interview situations, interviews might also be affected by how the interviewee and interviewer perceive each other and the situation (Silverman, 1993). By using'real-life' material, we believe that the material is less 'edited' or 'processed' by the individual professional and that the language in our discourse representations is a representation of how professionals would speak in natural settings. We chose to do a critical discourse analysis. This approach is very extensive, and therefore, we singled out specific analytical tools. It might be argued that we have neglected to use key concepts, e.g. consummation of the text. This would have entailed a more thorough analysis of how the receiver of the text employs earlier discourses in understanding and interpreting the text (J<unk>rgensen & Phillips, 2002).We argue, however, that it would not be possible to do a thorough analysis of every concept from Fairclough's terminology in one article. Hence, this can potentially be investigated elsewhere. Nuances in language might be lost in translation processes, and this is obviously important when analysing texts discursively. Therefore data was collected in Danish in order for participants to be able to use the mother-tongue when discussing non-attendance. Furthermore we chose to do the analysis in Danish, and then translate the selected discourse --- Declaration of conflicts of interest The Authors declare that they have no conflicts of interest.

Non-attendance constitutes a profound challenge in public sector services targeting young adults with mental health difficulties. Therefore, researchers and practitioners are occupied with trying to resolve this. In order for clinicians to be aware of their own naturalised and perhaps inappropriate communicative practices, we investigated the established normative organisational logics behind explanations and strategies related to non-attendance. We performed a critical discourse analysis on material collected through participatory research throughout 2015. Three discourses were identified: solicitude, responsibility and youth discourse. Although the discourses were complex and entangled, they were used by all practitioners. Furthermore, some of the discourses, especially the responsibility and the solicitude discourses were inherently tension-filled, and practitioners experienced frustration in dealing with these tensions. The youth discourse can be understood as a coping mechanism to deal with these tensions because it distributes responsibility for non-attendance to general social and cultural processes.

IntrODuCtIOn As life expectancy has increased in high-income countries, there has been a global debate about whether additional years of life are free from ill-health/disability. It is now largely accepted that old-age disability has declined in the USA (although varying by age/method), 1 2 although chronic illness increased, 3 and the picture beyond the USA is more mixed. [4][5][6] Yet, this research agenda has not been matched by similar attention to changes over time in morbidity in the working-age population. In the absence of direct evidence, policy-makers have often made claims based on self-reports of general health [6][7][8] which we know are unreliable. 9 10 The lack of evidence is even more problematic within social security, where many policy-makers have assumed that working-age morbidity must have improved in recent decades given improvements in mortality (despite the potential for declining mortality to coexist with rising morbidity) 6 -and that therefore high/rising levels of claims are not 'genuine'. 11 12 Almost the only direct evidence on changes over time in working-age morbidity in highincome countries comes from the USA. Contrary to policy-maker expectations, these studies have generally found deteriorating Open access morbidity since the mid-1990s, particularly activities of daily living and physical functioning. [13][14][15][16] Other studies have focused on the older working-age population with similar results. 2 17 Again, not all measures show deteriorations, and not all studies come to identical conclusions, 18 but there is little sign of any improvement in morbidity among working-age Americans-despite a 23% fall in working-age mortality 1993-2013 (online supplementary appendix 1). Outside of the USA, there is a paucity of evidence, but from the limited evidence that exists, there is again little sign of improving morbidity. [19][20][21][22] This study therefore asks: is there empirical support for the hypothesis that working-age morbidity in England has declined? (H 1 ). Or does the evidence support alternative hypotheses of stable (H2) or even declining (H3) morbidity? We answer this using the Health Survey for England (HSE), a high-quality Government survey with a combined sample of 140 000 individuals. We examine 39 specific aspects of morbidity rather than reducing morbidity to a single measure, partly because these produce more reliable trends, and partly to capture the multidimensional nature of morbidity. 23 However, we conclude by examining the broad picture of morbidity change, and how far this supports the competing hypotheses. This analysis makes two contributions. First, we provide one of the few systematic analyses of changes over time in working-age morbidity in any high-income country outside the USA. Second, we supplement self-report measures with 10 'biomarkers' which are particularly valuable for showing genuine changes over time (rather than merely changes in how people describe their health), but which have rarely been examined alongside self-reported working-age morbidity trends (Martin et al 24 being an exception). --- Data and methods This section follows the Strengthening the Reporting of Observational Studies in Epidemiology cross sectional reporting guidelines. 25 Data source Robust evidence of change over time requires consistently collected, high-quality data. We use the HSE, an annual government-sponsored cross-sectional survey of 3000-11 000 adults with no proxy responses. A particular advantage is that the interview is followed by a nurse visit which in selected years also includes a blood sample. Nevertheless, there are challenges in analysing change in HSE: <unk> First, HSE was run by the Government Office of Population Censuses and Surveys in 1991-93, before changing to NatCen in 1994. We focus on 1994-2014 given evidence of a discontinuity at this point. <unk> Second, topic coverage of HSE varies year-to-year, accompanied by changes in question wording/ filtering. Based on a systematic search of HSE questions, we have included every morbidity measure that is comparable over a significant duration. Even for measures that have been previously been analysed (eg, body mass index), 48 this new analysis uncovered further discontinuities (online supplementary appendices 2 and 3). <unk> Third, HSE excludes those in communal establishments. While a smaller problem for the working-age population than older ages, 2 we minimise the impact of rising university attendance by focusing on those aged 25+ (online supplementary appendix 3). The upper limit of the working-age population is set to 59 (women) and 64 (men) to match state pension ages at the start of the period. <unk> Fourth, HSE supplies non-response weights from 2003. However, there had been a substantial decline in response rates prior to the introduction of weights, particularly for blood samples (from 53.3% 1994 to 39.9% 2003; online supplementary appendix 3). We therefore reduce non-response biases by creating new non-response weights, described in online supplementary appendix 3. The resulting sample sizes for the various stages of data collection are shown in online supplementary appendix 3. Our dataset substantially extends an existing HSE time-series dataset (UK Data Archive SN7025); the code enabling other researchers to assemble this extended time-series dataset are freely available. 49 --- Patient involvement As this is a health monitoring study using secondary data, patients were not directly involved. However, from previous discussions we are aware that the study will be of interest to patient/disability advocacy groups, who will receive jargon-free summaries of the research. --- Measures We cannot interpret changes over time correctly without understanding different ways of operationalising'morbidity'. 1 General health/disability measures-for example, 'How is your health in general?'-are a simple way of measuring morbidity with a single indicator, and clearly do capture something meaningful. 50 However, their generality means that despite consistent question wording, different people may interpret questions or response options differently (eg, what 'good' health refers to). 51 p218-24 This can even occur within individuals, if they change their internal standards of measurement over time (contributing to'response shift'). 52 Numerous causal factors contribute to variable comprehension/ reporting, ranging from the experience of ill-health itself 52 to non-health factors such as social security incentives, 53 gendered-related and age-related expectations, and medicalisation. 54 These inconsistencies mean that general health/ disability measures are inadequate for answering our question: trends in such measures can differ wildly between different surveys covering nominally the same concept and population, for example, for disability in England 9 or self-rated health in the USA. 10 --- Open access general health' at the same time as activity limitations have fallen (changes over time in seven general HSE health/ disability measures are available in online supplementary appendix 4). Moreover, single indicator measures are potentially misleading in that they gloss over the multidimensional nature of morbidity. 1 To robustly answer our research question, we must instead focus on more specific morbidity measures that capture multiple aspects of morbidity. Our systematic search found 39 such measures that are comparable over time: these are summarised in table 1, with further details in online supplementary appendix 5. (A further 29 measures are also included in online supplementary appendix 6; this includes eight sub-components of measures in the main text, 16 reports of ever having a condition even if this not recent, and five other categories of longstanding illness (LSI).) These specific morbidity measures can be grouped into three types which have different strengths and weaknesses with respect to our question: 1. Medical labels: some measures are based on medical labels, either diagnosed chronic diseases or self-reported types of LSI. (Those reporting a LSI were asked, 'what is the matter with you?'; up to six responses were then coded by the interviewer based on the International Classification of Diseases (ICD)). These are imperfect measures of morbidity 55 as they partly reflect healthcare systems and medicalisation more broadly, both of which change over time. Nevertheless, they are an important element of morbidity as they have real consequences via increasing awareness/labelling of people's experiences. 2. Symptom-based: some measures are based on self-reports of ill-health symptoms or specific domains of activity limitations. These measures are either single items (eg, pain, anxiety/depression) or validated symptom scales (eg, the Rose angina scale, 56 57 General Health Questionnaire (GHQ) psychiatric distress). 58 The more specific and concrete nature of these measures prima facie makes them more likely to be interpreted consistently over time than medical labels and general measures. Others have reached a similar conclusion for comparisons across place, 55 particularly for disability measurement, 59 60 where the Washington Group on Disability Statistics-a UN agency founded in 2001have brokered a consensus that cross-country disability comparisons should be based on multiple measures of specific activity limitations. 61 62 We should nevertheless note that there is no guarantee that a given symptom/ impairment-based question will be interpreted identically over time. 63 64 3. Biomarkers-that is, objective measures of biological or physiological measures-have considerable strengths in analysing change, as they largely avoiding reporting biases that are likely to vary between socioeconomic groups and over time. 65 They do this at the price of an indirect and sometimes still-debated relationship to morbidity (see online supplementary appendix 5), and do not cover several important morbidity domains (eg, we lack good biomarkers for mental distress, pain and fatigue). These three types of measures are therefore complementary in understanding changing morbidity: biomarkers are least likely to be affected by changing respondent interpretations over time, but do not capture morbidity well; symptom-based measures capture morbidity well and are reasonably (if still imperfectly) reliable; and label-based measures are flawed in capturing symptoms/ limitations but do enable us to capture whether people consider themselves to have a medical condition. --- Analysis In the first instance, we look at unadjusted changes over time in each morbidity indicator, showing the actual levels of morbidity found in the population. However, we primarily focus on changes after adjustment for sex and age (following others), 66 67 akin to standardising for the agesex composition of the population. Given that our aim is to describe changes rather than to explain them, we do not further adjust for potential causal influences on morbidity that are likely to vary over the period, such as employment over economic cycles. This is a task for future research, but we should note that such analysis is possible using our publicly-available time-series dataset that includes inter alia employment status, education and region. We chose to examine discrete changes from the start to the end of available data for each measure, rather than using linear or non-linear trend terms. Given our aims of informing policy debates, this has three advantages: a discrete change is simple to interpret; it is compatible with the different start/end years available for different measures; and it does not require any assumptions about the functional form of trends (linear trends are particularly unlikely given the role of non-linear economic cycles). Individual survey years are grouped into 3-4year periods to increase sample size and precision, but single-year prevalence is given in online supplementary appendix 7. Given our binary outcome measures, we use logistic regression models with the following form: y i = logit[<unk> 1 period i + <unk> 2 age i + <unk> 3 male i + <unk>(age i * male i )]...where periodi refers to a vector of period dummy variables (covering all periods in which there were any observations: 1994-1996, 1997-2000, 2001-2003, 2004-2007, 2008-2010 and 2011-2014); <unk>1 is a vector of our primary outcome coefficients showing change between each period and the earliest available period; agei refers to a vector of age dummy variables; malei refers to a binary gender dummy variable and <unk>2, <unk>3 and <unk>4 refer to the coefficients on age, gender and their interaction, respectively. We present average marginal effects rather than odds ratios, partly because these are simple to understand-odds ratios have no easy real-world interpretation for policy-makersbut primarily because odds ratios are not fully comparable across different models, and cannot therefore underpin our comparison of changes over time between indicators. 68 on February 18, 2024 by guest. Protected by copyright. To avoid a binary cut-off of statistical significance, 69 95% CIs are used to convey precision. All analyses use weights, exclude boost samples that use different sampling methods, and adjust for the multistage clustered sample design and the stratification of the sample across survey years using the SVYSET command in Stata (although standard errors will be slightly underestimated as it is not possible to consistently adjust for sample stratification within years). For reasons of space, we are unable to discuss previous HSE studies of specific morbidity trends in the main text; these are instead described in online supplementary appendix 8. --- reSultS --- Conditions with sharply declining mortality We start by focussing on cardiovascular disease (CVD) and respiratory illness which have both seen large falls in mortality (by >50% and >25%, respectively, among 0-64 years old 1994-2013; online supplementary appendix 1). Changes over time in morbidity, however, are shown in table 2. Looking first at high blood pressure, biomarkermeasured high blood pressure has halved over two decades (similar improvements are found for the biomarkers for total and HDL cholesterol). Yet, when we look at self-reports (either people reporting this as an LSI, or in response to a direct question about having recent diagnosed high blood pressure), we see large rises over time. There has been an increasing diagnosis of high blood pressure and increasing prescriptions of blood pressure-lowering drugs; these may have helped reduce the underlying incidence of high blood pressure while simultaneously raising people's awareness of morbidity. Table 2 further shows declines in several key types of CVD (heart attack, mini-stroke, angina), whether measured through people's reports of the disease itself or their reports of its symptoms. Nevertheless, the morbidity declines (8%-50%) are often not on the scale of the declines in mortality (>50%); this is likely to be because mortality declines are partly driven by improved treatment 70 which means each incident CVD case is likely to last longer. 71 72 More surprisingly, the measures of 'any reported CVD' show no improvement (with some, uncertain signs of rises). Looking at its sub-components (online supplementary appendix 6), this seems to be due to possible increases in diagnosed irregular heart rhythm and other heart trouble. Finally, table 2 shows that symptoms-based measures of respiratory morbidity have improved, particularly COPD symptoms (regular cough and phlegm) and breathlessness (at both levels), and more uncertainly for recent wheezing/ asthma and wheezing stopping sleep. Again, though, diagnosis-related measures of asthma-reported diagnoses, or self-reports of having asthma as a LSI-have risen, even while underlying symptomatology is improving. Overall, table 2 illustrates how changes over time in morbidity do not necessarily follow changes in mortality. There are definite improvements in CVD risk factors and respiratory symptomatology on the scale of improvements in mortality. But the prevalence of self-reported CVD conditions such as heart attacks have only declined by a smaller amount, and recent doctor-diagnosed hypertension, any CVD, and asthma diagnoses have either stayed stable or risen. --- Conditions with claims of increasing prevalence The previous section focused on conditions where there may be an a priori expectation that morbidity has improved (given declining mortality); in this section, we focus on three areas where there have been widespread claims of increasing prevalence-obesity, diabetes and mental health. Looking at table 3, we do indeed confirm a large rise in obesity in HSE (an 8.0%-9.7% rise from an obesity prevalence of 16.9% in 1994-1996). The rise in high waist-hip ratios-sometimes suggested to be a better measure of potential morbidity 73 -is even larger. This has come alongside little change in the prevalence of being underweight over this period. Table 3 also confirms a large rise in diabetes. This can be seen whether diabetes is measured through people reporting diabetes as an LSI, a specific question about people currently taking medication for diabetes or via a diabetes biomarker (glycated haemoglobin). This clear rise in diabetes has occurred despite declining age 0-64 death rates from diabetes, which fell by more than one-third 1994-2013 (online supplemetnary appendix 1)-indeed, rising prevalence is because of falling mortality 74 -again demonstrating the difference between changes in mortality and morbidity. Trends in mental health are more contentious in the wider literature (see online supplementary appendix 8), and the measures in HSE are not as strong as the more occasional Adult Psychiatric Morbidity Surveys. 75 Nevertheless, HSE offers a unique annual perspective on self-reported mental health. As we might expect from increasing treatment/diagnosis, we see a doubling in people reporting a mental health LSI. However, the symptoms-based measures show a more mixed picture: <unk> Neither of the measures that capture more moderate mental ill-health show rising ill-health (these are psychological distress symptoms and people reporting a feeling of anxiety/depression today, both with a relatively common prevalence of 15%-25% appendix 7), we can see moderate mental ill-health symptoms fell between the mid-1990s and the mid-2000s, before rising in 2009. <unk> In contrast, the single measure capturing a feeling of extreme anxiety/depression today does show rising morbidity. To see if there were similar signs of rising mental ill-health at extremes in our other measure (psychological distress), we looked at a much higher GHQ threshold of 10 negative responses out of 12 questions (compared to the conventional threshold of 4). Unlike the conventional GHQ measure, this also showed an increase over time (95% CI of a 0.4% to 1.4% rise; see online supplementary appendix 6). While the GHQ is not designed to capture severe psychological distress in this way, others have similarly looked at moderate and extreme psychological distress using GHQ-and indeed, have found that rises in distress over time 1991-2008 are concentrated in the more extreme measure. 76 Overall, while labelling of mental health conditions has undoubtedly risen, trends in mental health symptoms vary across measures. If we interpret higher GHQ thresholds as indicating more serious psychological distress, then we can see a consistent picture: moderate mental illhealth symptoms fell from the mid-1990s to the mid-2000s before rising around the time of the 2008 economic crisis (as we would expect), 77 whereas more extreme mental illhealth has more consistently risen. --- Activity limitations, musculoskeletal and pain Pain/musculoskeletal conditions are a major component of working-age morbidity, yet very few previous studies show changes over time in symptomatology, and even those that exist 78 sometimes have debatable comparability. 79 Table 4 shows a fall in some-but not all-HSE measures focused on pain and musculoskeletal morbidity. Arthritis as a LSI has declined (the precision of the estimates is greater when looking at 2008-2010 rather than 2011-2014, and shows a decline of 0.3%-1.2%). There are some (similarly uncertain) signs that other musculoskeletal LSIs have also fallen, and noticeably fewer people say that they have any pain/discomfort today, although there has been no change in people saying they have extreme pain/discomfort. The echoes a previous study that found different trends in low back pain of different levels of severity. 80 In contrast, there has been a rise in all four activity limitations measures in HSE-although the increases are sometimes uncertain, and are smaller after adjusting for changes in age/sex structure. Moreover, the timing of the rises differ between the measures: the trend in limitations lasting at least a year shows a rise in 1994-1996 to 2001-2003, but the two measures of 'limitations today' do not, instead showing a possible slight rise in the more recent period (see online supplementary appendix 7; this difference remains if we focus on the sub-components of year-long limitations that more closely match to the 'limitations today' questions, see online supplementary appendix 6). The measures can collectively be seen as offering some, although relatively weak, evidence for an increase in activity limitations. --- Other measures Changes over time in other measures are shown in table 5. This includes four biomarkers that are more difficult to compare directly to self-reports: <unk> Changes over time are available for two biomarkers of inflammation (C-reactive protein (CRP) and fibrinogen). These are associated with a number of conditions including heart disease, diabetes, cancer 81 and-in the case of CRP-even depression. 82 Open access the CI is wide and there is a non-negligible possibility that the change is negative). <unk> The two other biomarkers available in HSE are clearly focused on anaemia and iron deficiency. table 5 shows that both of these have declined, with particularly clear evidence for a decline in iron deficiency. Table 5 also shows changes over time in sensory and communication-related morbidity. This shows a fall in eye/ear conditions (1994-1996 to 2011-2014) as well as hearing limitations in the earlier period (1994-1996 to 2001-2003), but no change in people having difficulty communicating with others. --- DISCuSSIOn Despite considerable evidence on morbidity trends among older people, there are few published studies on changes in morbidity among the working-age population, particularly outside the USA. In this paper, we have analysed changes over time in working-age morbidity in England 1994-2014 using a high-quality repeated crosssectional study. We see improvements in cardiovascular morbidity, respiratory morbidity and anaemia, but deteriorating obesity, diabetes, some biomarkers (fibrinogen and possibly also CRP) and feelings of extreme anxiety/ depression. We see little systematic change over time in more common mental ill-health or musculoskeletal conditions, pain/mobility and self-care limitations. Symptomatology and chronic disease diagnoses also often go in different directions-chronic disease diagnoses have sometimes stayed stable or even risen at the same time that underlying symptomatology has declined (such as for mental health conditions, asthma, hypertension and CVD as a whole), mirroring findings at older ages. 3 Our analysis has several strengths. We include every morbidity measure for which consistent changes can be constructed, including chronic disease, functioning and symptomatology, and biomarkers. We use a single survey series collected by a single survey organisation; exclude under-25s for whom comparability of survey coverage is unlikely; and construct new non-response weights. Nevertheless, we must note three limitations. First, response rates for each stage of the HSE have declined over time (see online supplementary appendix 3), and while we create new non-response weights covering the entire period, it is still possible that socioeconomically disadvantaged people (within any age-sex-region group) have become less likely to respond-and as they tend to be in worse health, this could mask deteriorating morbidity. Second, even if non-response biases have not changed, it is possible that people respond differently over time even to identical questions. Third, there are several dimensions of morbidity for which there is little comparable data in HSE. This includes several areas in which morbidity among the working-age population seems to be rising, including inter alia cognitive complaints, 83 allergic disorders 84 and liver cirrhosis (see online supplementary appendix 1), as well as some areas in which morbidity seems likely to have fallen, such as chronic kidney disease. 85 It is clear that there are different trends in different dimensions of morbidity-but for policy-makers, this leaves the question of whether working-age morbidity as a whole is unchanged (H2), getting better (H1) or getting worse (H3), to the extent that it makes sense to place health on a unidimensional scale. While we cannot create a single morbidity index here, online supplementary appendix 9 shows the association of each measure with bad general self-rated health (net of age, gender and education). This shows little systematic trend for falling morbidity to be seen in the measures that predict health the most (indeed, the evidence weakly points in the other direction, towards rising morbidity). This provides greater support for H2 than H1 or H3, mirroring evidence from In conclusion, despite considerable falls in working-age mortality and gains in life expectancy-and the ensuing expectations of social security policy-makers for improving morbidity-there is no evidence of systematic improvement in overall working-age morbidity in England from 1994 to 2014. However, two pieces of further research could strengthen this evidence base. First, the ideal measures for analysing changes in morbidity are functional limitations measures which are included in the HSE from 1996. However, these were last asked to the working-age population in 2001, and it is a priority to repeat these measures in future years of HSE. Second, there is a surprising paucity of studies looking at the changing morbidity of the working-age population outside the USA. Given their importance in public debate-particularly in discussions of retirement ages and disability benefits-we hope that other authors will repeat and extend our analyses here, including disaggregating these changes across different regions and sociodemographic groups. Correction notice This article has been corrected since it was first published. Data in the table 2-5 has been corrected. twitter Ben Baumberg Geiger @BenBaumberg --- http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2019-032378 on 15 March 2020. Downloaded from on February 18, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2019-032378 on 15 March 2020. Downloaded from on February 18, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2019-032378 on --- Data availability statement Data are available in a public, open access repository. The Health Survey for England 1994-2014 are available for free to registered users at the UK Data Service -see https:// beta. ukdataservice. ac. uk/ datacatalogue/ series/ series? id= 2000021#!/ abstract. There are no conditions for re-use for non-commercial applications of the data. The statistical code enabling replication using publicly available data is available from OSF (Morbidity in England 1994England -2014 --- Competing interests No. Patient consent for publication Not required. Provenance and peer review Not commissioned; externally peer reviewed.

BB. Has working-age morbidity been declining? Changes over time in survey measures of general health, chronic diseases, symptoms and biomarkers in

Introduction Events happen in real-world and on social networks. In online shopping, an event can represent user behaviors, such as click, cart or purchase (Liu et al. 2015). In geophysics, an event can be an earthquake (Sakaki, Okazaki, and Matsuo 2010). In online social media, events can be user actions (e.g., like, comment and retweet) over time, which have some features like user influence, content, time and connectivity of the social network (Fu 2011;Rizoiu et al. 2017;Shi et al. 2017). Online events usually follow the hot topics that caused by some significant news. For example, Figure 1 shows the interest about "Apple" over time in the last year 1. Apparently, the peaks are related to special events of Apple company (e.g., product launch conference) or holidays like Christmas. --- Apple Special Event Figure 1: Google trends about "Apple". We add the relative events on the figure. The so-called event sequences contain a series of events of different types in the continuous time domain. In an event sequence, the past events and the next event are related (Chang et al. 2015). Take the previous three types of events as examples: buyers' historical behaviors can be considered as their decision making processes. The aftershocks can happen with a month, or within days, from the main shock. Retweets could be grouped with topics of interest and timeline. The pattern of events may help cause or prevent future events. Thus, techniques to discover patterns among events are urgently required, so that the future of an event sequence can be accurately predicted (Xu et al. 2018). In the literature, feature based methods extract relevant features and apply different machine learning algorithms to predict the type and arrival time of the future events (Naveed et al. 2011;Cheng et al. 2014;Bakshy et al. 2011;Zaman et al. 2010). However, these models heavily depend on manually selected features. It is a fatal flaw as designing features requires much expertise, especially for large-scale dataset with high dimensional features, which may severely limit its application. Recently, some other prior arts based on generative approaches are proposed, (Shen et al. 2014;Cao et al. 2017;Luo et al. 2015;Lukasik et al. 2016), in which historical events are modeled to have impact on future ones. However, the generative methods depend on the design of stochastic process and the information hidden in the sequences cannot be fully leveraged. Variational Auto-Encoder (VAE) (Kingma and Welling 2014) is a powerful class of probabilistic models and has the ability to model complex distributions. In recent years, VAE are used in time-series (Babaeizadeh et al. 2018;Denton and Fergus 2018;Hu et al. 2017;Li and Mandt 2018). These models integrate VAE with RNN/LSTM to build a bridge between high interpretability and high predictive power. Along this line, in this paper, we present a novel probabilistic generative model for event sequences which we call Variational Event Point Process or VEPP. Firstly, we use LSTM to embed the event sequences, so the features can be automatically extracted and utilized by the powerful neural network. Secondly, our model introduces variational auto-encoder to event sequence modeling that can use the latent information and capture the distribution over event sequences. Finally, on two real-world datasets, we find that VEPP has higher log-likelihood in the mission of predicting event type and lower error in the mission of predicting time intervals. The experiments demonstrate that VEPP can model the future of event sequences. --- Related Work In this section, we briefly summarize the related work to deal with the event prediction problem as two groups, i.e., feature-based and generative approaches. The first category is feature based methods, which first extract some relevant features, including content, user information, original posters, network structure, and temporal features (Cheng et al. 2014;Lian et al. 2015;Wang et al. 2015). Then different machine learning algorithms are applied to build a regression or classification model, such as content-based models (Naveed et al. 2011), simple regression models (Cheng et al. 2014), regression trees (Bakshy et al. 2011) and probabilistic collaborative filtering (Zaman et al. 2010). However, these methods require much laborious feature engineering with expertise, which is hard to design, and their performance is highly sensitive to the quality of features. Besides, such approaches also have limitation in practice because they cannot be used in real-time online settings, like real-time event detection on Twitter. Given the large amount of data being produced every second, it is practically impossible to extract all the necessary features so the application is severely limited. The second type is generative approaches which are usually based on temporal point process, like Poisson process and its complexer variants (e.g., Reinforced Poisson Processes, Hawkes Process and Self-Correcting Process). A temporal point process can be used to capture the interarrival times of event sequences (Daley and Vere-Jones 2007). It directly models complicated event sequences in which historical events have influences on current and future ones. Reinforced Poisson Processes (RPP) is employed to model the phenomena in social networks (Shen et al. 2014). Hawkes process, a variant of Poisson process, has been proven to be useful for describing real-world data in social network analysis (Cao et al. 2017). Furthermore, multiply variants of Hawkes process are proposed to solve the issues of event sequences. Luo et al. (2015) proposed multi-task multi-dimensional Hawkes processes for modeling event sequences. Lukasik et al. (2016) applied Hawkes processes for rumour stance classification on Twitter. However, in practice the true distribution of events is never known and the performance depends on the design of stochastic process. Besides, these methods generally are not directly optimized for future events. They cannot fully leverage the information implied in the sequences for prediction. There still remains a gap between the interpretability and predictability. --- Preliminaries In this section, we first give the problem definition, and then briefly introduce the two basic models for the temporal point process and Variational Auto-Encoders. --- Problem definition As shown in Figure 2, the input is a sequence of events x 1:n = (x 1, • • •, x n ), where x n is the n-th event. The event x n = (k n, <unk> n ) is represented by the event type k n <unk> <unk>1, 2, • • •, K<unk> (K discrete event classes) and the time inter- val <unk> n <unk> R +. The time interval <unk> n = t n -t n-1 is the difference between the starting time of event x n-1 and x n. Given a sequence of events x 1:n-1, the event sequence modeling task is to produce a distribution over the event type k n and the time interval <unk> n of the next happening event. We aim to develop probabilistic models to predict what and when the next event will happen. --- Temporal point process A temporal point process is a random process which is used to capture the time intervals of event sequences (Daley and Vere-Jones 2007). A temporal point process is characterized by the conditional intensity function <unk>(t n |x 1:n-1 ), which is conditioned on the past events x 1:n-1. The conditional intensity is the expected infinitesimal rate at which events are expected to occur around time t. Given the n-1 past events, the probability density function for the time interval of next event is: f (<unk> n |x 1:n-1 ) = <unk>(<unk> n |x 1:n-1 )e -<unk>n 0 <unk>(u|x1:n-1)du. (1) The Poisson process (Kingman 2005) is the simplest and most ubiquitous example of point process, which assumes that events occur independently of one another. The conditional intensity is <unk>(<unk> n |x 1:n-1 ) = <unk>, where <unk> is a positive constant. Furthermore, more complex point processes have been proposed, like Hawkes Process (Hawkes 1971) and Self-Correcting Process (Isham and Westcott 1979). All these processes try to model the dependency on the past events. For example, Hawkes process is a self-exciting process in which the arrival of an event causes the conditional intensity (k 1, t 1 ) (k 2, t 2 ) (k 3, t 3 ) p(z 4 ) k 4 0. <unk>(t n |x 1:n-1 ) = <unk> 0 (t n ) + n-1 i=1 <unk>(t -T i ),(2) where T i <unk> t are all the event time having occurred before current time t n, and which contribute to the event intensity at time t n. <unk> 0 (t n ) is a deterministic base intensity function, and <unk> is called the memory kernel. However, the true model of the dependencies is never known in practice (Mei and Eisner 2017) and the performance depends on the design of conditional intensity. In this work, we learn a variational model that fits the conditional intensity by the history of events. --- Variational Auto-Encoders A Variational Auto-Encoder is a generative model which can effectively model complex multimodal distributions over the data space. A VAE introduces a set of latent random variables z, designed to capture the variations in the observed variables x. The joint distribution is defined as: p <unk> (z|x) <unk> p <unk> (x|z)p <unk> (z). The simple prior p <unk> (z) is usually chosen to be a multivariate Gaussian. The parameters of complex likelihood p <unk> (x|z) are produced by neural networks. Approximating the intractable posterior p <unk> (z|x) with a recognition neural network q <unk> (z|x) the parameters of the generative model <unk> as well as the recognition model <unk> can be jointly optimized by maximizing the evidence lower bound (ELBO) L on the marginal likelihood p <unk> (x): log p <unk> (x) = KL(q <unk> ||p <unk> ) + L(<unk>, <unk>) <unk> L(<unk>, <unk>) = -E q <unk> log q <unk>(z|x) p <unk> (z, x). (3 ) Recent works apply VAEs to time-series data including video (Babaeizadeh et al. 2018;Denton and Fergus 2018;Mehrasa et al. 2019), text (Hu et al. 2017), and audio (Li and Mandt 2018;Chung et al. 2015). Such models usually integrate a time-step VAE with RNN/LSTM. The ELBO thus becomes a summation of time-step-wise variational lower bound: L <unk>,<unk> = N n=1 E q <unk> (z1:n|x1:n) [log p <unk> (x n |x 1:n-1, z 1:n )] -KL(q <unk> (z n |x 1:n )||p <unk> (z n |x 1:n-1 )). (4) --- Variational Event Point Process In this section, we give the details of our VEPP model. We propose a generative model for event sequence modeling by using the VAEs. Figure 3 shows the architecture of our model. Overall, the types of events and their time intervals are encoded using a recurrent VAE model. At each step, the model uses past events to create a distribution over latent codes z n, a sample of which is then decoded into two probability distributions: one over the possible event types and another over the time intervals for the next event. --- Event representing and embedding As shown in Figure 3(a), at time step n, the model takes the event x n as input, which is the prediction target, and also the past events x 1:n-1. These inputs are used to product a conditional distribution q <unk> (z n |x 1:n ) from which a latent code z n is sampled. The true distribution over latent variables z n is intractable. We rely on a time-dependent inference network q <unk> (z n |x 1:n ) that approximates it with a conditional Gaussian distribution N (<unk> <unk>n, <unk> 2 <unk>n ). To prevent z n from just copying x n, we force q <unk> (z n |x 1:n ) to be close to the prior distribution p(z n ) using a Kullback-Leibler divergence term. At each step during training, a latent variable z n is drawn from the posterior distribution q <unk> (z n |x 1:n ). The output event xn is then sampled from the distribution p <unk> (x n |z n ) of the conditional generative model which is parameterized by <unk>. For convenience, we assume the event type and time inter-vals are conditionally independent given the latent code z n : p <unk> (x n |z n ) = p <unk> (k n, <unk> n |z n ) = p k <unk> (k n |z n )p <unk> <unk> (<unk> n |z n ),(5) where p k <unk> (k n |z n ) and p <unk> <unk> (<unk> n |z n ) are the conditional generative model for event type and time interval, respectively. It is a standard assumption in event prediction (Du et al. 2016). The sequential model generates two probability distributions: a categorical distribution over the event types and a temporal point process over the time interval for the next event. The event types are modeled with a multinomial distribution in which case k n can only take a finite number of values: K i=1 p k <unk> (k n = i|z n ) = 1,(6) where p k <unk> (k n = i|z n ) is the probability that event type i will occur, and K is the total number of event types. The time interval follows an exponential distribution whose parameter is <unk>(z n ), similar to a standard temporal point process model: p <unk> <unk> (<unk> n |z n ) = <unk>(z n )e -<unk>(zn)<unk> n if <unk> n <unk> 0,(7) where p <unk> (<unk> n |z n ) is a probability density function over variable <unk> n and <unk>(z n ) is the intensity of the temporal point process, which depends on the latent variable sample z n. At step n, the current event x n is represented as a vector x emb n with a two-step embedding strategy. First, we compute a representation for the event type k n and the time interval <unk> n separately. Then, we concatenate these two representations and get a new representation x emb n of the event: k emb n = Emb k (k n ), <unk> emb n = Emb <unk> (<unk> n ),(8) x emb n = Emb k,<unk> ([k emb n, <unk> emb n ]). Here, Emb k, Emb <unk> and Emb k,<unk> represent the embedding functions. A one-hot encoding is used to represent the event type k n. --- Generation The VEPP contains a VAE at every time step. However, these VAEs are conditioned on the state variable h n-1 of an LSTM. It will help the VAE to take into account the temporal structure of the sequential data. Unlike a standard VAE, the prior on the latent random variable is no longer a standard Gaussian distribution, but follows the distribution: z n <unk> N (<unk> n, <unk> 2 n ), <unk> n, <unk> 2 n = f prior (h n-1 ),(9) where <unk> n and <unk> n are the parameters of the conditional prior distribution and f prior can be any highly flexible function such as neural networks. x n-1 LSTM Embedding Decoder z n n (a) At each time step, the model uses the history of event sequences and inter-arrival times to generate a distribution over latent codes. Firstly, we sample z n from the prior to generate an event at step n. The parameters of the prior distribution are calculated based on the past n -1 events x 1:n-1. Then, an event type kn and time interval <unk>n are generated as follows: kn <unk> p k <unk> (k n |z n ), <unk>n <unk> p <unk> <unk> (<unk> n |z n ).(10) The decoder network for event type f k <unk> (z n ) is a MLP with a softmax output to generate the probability distribution in Equation ( 6): p k <unk> (k n |z n ) = f k <unk> (z n ). (11) The decoder network for time interval f <unk> <unk> (z n ) is another MLP, producing the parameter of the point process model for temporal distribution in Equation ( 7): <unk>(z n ) = f <unk> <unk> (z n ).(12) The LSTM encodes the current event and the past events into a vector representation: h n = LST M <unk> (x emb n, z emb n, h n-1 ). (13 ) Recurrent networks turn variable length sequences into meaningful, fixed-sized representations. The parameterization of the generative model results in the factorization: p(x 1:N, z 1:N ) = N n=1 p(xn|z 1:n, x 1,n-1 )p(zn|x 1:n-1, z 1:n-1 ) (14) --- Inference The posterior is proportional to the product of the likelihood and the prior. So the approximate posterior will not only be z n |x n <unk> N (<unk> z,n, <unk> 2 z,n ), <unk> z,n, <unk> 2 z,n = Enc(x n, h n-1 ),(15) where <unk> z,n and <unk> z,n denote the parameters of the approximate posterior. The encoding of the approximate posterior and the decoding for generation are tied through the LSTM hidden state h n-1. This conditioning on h n-1 results in the factorization: q(z 1:N, x 1:N ) = N n=1 q(z n |x 1:n, z 1:n-1 ). (16 ) --- Learning We train the model by optimizing the variational lower bound over the entire sequence comprised of N steps: L <unk>,<unk> (x 1:N ) = N n=1 (E q <unk> (zn|x1:n) [log p <unk> (x n |z n )] -KL(q <unk> (z n |x 1:n )||p <unk> (z n |x 1:n-1 ))). (17 ) Given the latent code z n, the event type and time interval are conditionally independent, so the log-likelihood can be written as follows: E q <unk> (zn|x1:n) [log p <unk> (x n |z n )] = E q <unk> (zn|x1:n) [log p k <unk> (k n |z n )] + E q <unk> (zn|x1:n) [log p <unk> <unk> (<unk> n |z n )]. (18 ) Given the form of p k <unk>, the log-likelihood reduces to a cross entropy between the predicted event type p k <unk> (k n |z n ) and the ground truth k * n. Given the ground truth time interval <unk> * n, we calculate its log-likelihood over a small time interval <unk> <unk> under the predicted distribution. log <unk> * n +<unk> <unk> * n p <unk> <unk> (<unk> n |z n )d<unk> n = log(1 -e -<unk>(zn)<unk>t ) -<unk>(z n )<unk> * n. (19 ) --- Experiments In this section, we evaluate the performance of VEPP on two real-world datasets, i.e., Retweets Dataset (Zhao et al. 2015) and MemeTrack Dataset (Leskovec, Backstrom, and Kleinberg 2009). --- Datasets Retweets dataset The Retweets dataset includes 166, 076 retweet sequences, each corresponding to some original tweet. Each retweet event is labeled with the retweet time relative to the original tweet creation, so that the time of the original tweet is 0. Each retweet event is also marked with the number of followers of the retweeter. As usual, we assume that these 166, 076 streams are drawn independently from the same process, so that retweets in different streams do not affect one another. Unfortunately, the dataset does not specify the identity of each retweeter, only his or her popularity. To distinguish different kinds of events that might have different rates and different influences on the future, following previous study of Mei and Eisner (2017), we divide the events into K = 3 types: retweets by "small", "medium" and "large" users. Small users have fewer than 120 followers (50% of events), medium users have fewer than 1, 363 (45% of events), and the rest are large users (5% of events). Given the past retweet history, our model must learn to predict how soon it will be retweeted again and how popular the retweeter is (i.e., which of the three types). We randomly sampled disjoint train and test sets with 20, 000 and 2, 000 sequences respectively. We truncated sequences to a maximum length of 264, which affected 20% of them. For computing training and test likelihoods, we treated each sequence as the complete set of events observed on the interval [0, T ], where 0 denotes the time of the original tweet, which is not included in the sequence, and T denotes the time of the last tweet in the truncated sequence. --- MemeTrack dataset The MemeTrack Dataset considers the reuse of fixed phrases, or "memes", in online media. It contains time-stamped instances of meme use in articles and posts from 1.5 million different blogs and news sites, spanning 10 months from August 2008 till May 2009, with several hundred million documents. The K event types correspond to the different websites. Given one meme's past trajectory across websites, our model can learn to predict how soon and where it will be mentioned again. We followed the previous study of Gomez-Rodriguez et al. ( 2013) to process the dataset, which selected the top 5, 000 websites in terms of the number of memes they mentioned. We truncated sequences to a maximum length of 31 and selected the minimum length of 2. We randomly sampled disjoint train and test sets with 96, 391, and 2, 470 sequences respectively, treating them as before. Table 1 shows statistics of the two datasets. The two datasets have very different characteristics. --- Architecture details The VEPP model architecture is shown in Figure 3 --- Implementation details The models are implemented with TensorFlow (Abadi et al. 2016) and are trained using the Adam (Kingma and Ba 2015) optimizer for 1, 000 epochs with batch size 32 and learning rate 0.001. We split both datasets into training and test sets containing 70% and 30% of samples respectively. We select the best model during training based on the model loss (18) on the test set. --- Baselines Poisson Process The intensity function is a constant, which produces an estimate of the average inter-event gaps. Hawkes Process (HP) Hawkes process is a self-exciting point process, in which past events from the history conspire to raise the intensity of each type of events. Such excitation is positive, additive over the past events, and exponentially decaying with time. Self-Correcting Process (SCP) We fit a self-correcting process with the intensity function in the book of Daley and Vere-Jones (2007). --- Recurrent Marked Temporal Point Processes (RMTPP) RMTPP is proposed in the study of Du et al. (2016). It views the intensity function of a temporal point process as a nonlinear function of the history, and uses a recurrent neural --- Metrics We use log-likelihood (LL) of event type to compare our model with the HP and RMTPP. For Poisson Process and Self-Correcting Process, their performance on event type prediction are similar to Hawkes Process and not very satisfactory. LL = log n i=1 f (x i |<unk>) = n i=1 log f (x i |<unk>) = n i=1 l(<unk>|x i ). We also compare Root Mean Square Error (RMSE) of interarrival time prediction..08 -6.58 -6.58 -6.48 -6.68 MemeTrack -10.9 -10.7 -10.5 -10.2 -10.0 RMSE = 1 N N i=1 (<unk> i -<unk> i ) 2 0.51 --- Experiment Results Overall performance Table 2 shows experimental results that compare VEPP with HP and RMTPP. VEPP outperforms HP and RMTPP on both Retweets and MemeTrack datasets. We believe that this is because the VEPP model is better in modeling the complex distribution over future events. For Retweets dataset, three methods all have relatively good performance. The reason may be the events of Retweets dataset have less types, so the patterns of event sequences are easy to model. Correspondingly, for Meme-Track dataset, VEPP and RMTPP significantly outperform HP. It is may because the event types are nearly 5, 000 and much larger than Retweets datasets. It also proves that VAE can use the latent information of event sequences and VEPP has the ability to modeling the complex distribution over future events. The prediction error is high for MemeTrack dataset due to the large number of types. Figure 4 shows RMSE in predicting the time interval given the history of previous events. VEPP achieves the lowest error, i.e., outperforms the other methods under the metric. The three methods based on point process have relatively higher error, because the designed point process does not fit the real situation. RMTPP and VEPP achieve better results, since they can learn the complex distribution. While VEPP can also use the latent information over the event sequence, it performs even better. Sensitive of latent variable dimensionality of LSTM We next explore the architecture of our model by varying the sizes of the latent variable. Table 3 shows the log-likelihood of our model for different sizes of the latent variable. We see that as we increase the size of the latent variable, we can model a more complex latent distribution which results in better performance. --- Learning curves --- Conclusion We presented a novel probabilistic model for sequence data, a variational auto-encoder that captures uncertainty in event types and arrival time. As a generative model, it could produce event sequences by sampling from a prior distribution, the parameters of which were updated based on neural networks that control the distributions over the next event type and temporal occurrence. The model could also be used to analyze given input sequences of events to determine the likelihood of observing particular sequences. We demonstrated empirically that the model is effective for capturing the uncertainty inherent in event prediction. In future, we will take into account the structural and contextual information to model the event sequences.

Many events occur in real-world and social networks. Events are related to the past and there are patterns in the evolution of event sequences. Understanding the patterns can help us better predict the type and arriving time of the next event. In the literature, both feature-based approaches and generative approaches are utilized to model the event sequence. Feature-based approaches extract a variety of features, and train a regression or classification model to make a prediction. Yet, their performance is dependent on the experience-based feature exaction. Generative approaches usually assume the evolution of events follow a stochastic point process (e.g., Poisson process or its complexer variants). However, the true distribution of events is never known and the performance depends on the design of stochastic process in practice. To solve the above challenges, in this paper, we present a novel probabilistic generative model for event sequences. The model is termed Variational Event Point Process (VEPP). Our model introduces variational auto-encoder to event sequence modeling that can better use the latent information and capture the distribution over inter-arrival time and types of event sequences. Experiments on real-world datasets prove effectiveness of our proposed model.

Introduction Social inequalities in mortality have been extensively documented. 1,2 Until fairly recently, most studies of social inequalities in adult mortality have focused on the role of adult characteristics. However, as evidence of the associations between childhood living conditions and adult health has accumulated, increased attention is now being paid to the contribution of the entire life course. [3][4][5][6] The accumulated evidence in life course epidemiology suggests that childhood nutritional status, health, socioeconomic status, place of residence and other household characteristics predict adult health and mortality. [7][8][9][10][11][12][13] The mechanisms through which childhood environment is hypothesized to influence adult health include indirect mechanisms operating through attained adult social characteristics (e.g. education, other socioeconomic status and lifestyle) and direct effects of in utero programming and childhood health. 14 Most of the evidence on the contribution of childhood conditions to adult mortality is based on historical cohorts and cohorts born in the early part of the 20th century. 4,5,8 The conditions at the time when these cohorts grew up were characterized by high rates of infectious diseases, poor nutrition and hygiene and high rates of poverty and material hardship. The findings obtained from these earlier cohorts may no longer apply to more recent cohorts. In the latter half of the 20th century, Western countries experienced rapid economic growth, social changes and improvements in health conditions and overall standard of living. However, because of the simultaneously decreasing household size, postponement of marriage and childbearing, higher rate of union dissolution and overall diversification of family forms, more recent cohorts have experienced a more diversified childhood family structure than earlier cohorts. These changes may have also influenced the mechanisms or pathways through which childhood circumstances influence adult health and mortality. Changes in economic structure, educational expansion and changing family forms have gradually transformed the work and family lives of cohorts that entered adulthood in the second half of the 20th century. Similarly, changing patterns of lifestyle risk factors-for example, the emergence and the decline of the smoking epidemic and the arrival of an obesogenic food environment-have influenced the health of more recent cohorts. Currently, there are no studies that we know of that compare the influence of early life circumstances on adult mortality of cohorts born before and after mid-20th century. In this paper, we: (i) compare the associations between various childhood circumstances and adult total and cause-specific mortality in Finland across cohorts born in 1936-50 and 1961-75; (ii) examine whether the effects of childhood characteristics are operating through attained adult social characteristics; and (iii) examine whether these effects may have changed across the two cohorts. --- Methods --- Participants and mortality follow-up We used two 10% random samples of households drawn from the 1950 and 1975 Finnish censuses. The samples were linked with register-based follow-up of death records using personal identifiers available to all permanent residents of Finland (Statistics Finland permission TK-53-789-10 and TK-53-155-17). We included individuals aged 0-14 years, living in private households at the time of the 1950 (birth cohorts 1936-50, n 1<unk>4 117 061) and the 1975 census (birth cohorts 1961-75, n 1<unk>4 102 280). The census data contained identifiers for families and households making it possible to identify family members. We used these identifiers in linking children with information on their parents' characteristics. The 1950 sample has been linked to quinquennial census records in the years 1970-95 and to death records for the years 1970-2007. [15][16][17][18] The 1975 sample was designed to be identical in structure; with linkage to census records in 1995, 2000 and 2005 and to death records until the end of 2016. For both samples we measured childhood characteristics from the household census (i.e. at age 0-14) and the individual's own educational attainment from the first quinquennial end-of-year census after the subjects turned 30. Adult characteristics were thus measured at ages 30-34 depending on the birth year. Mortality follow-up begun at the time of the assessment of adult characteristics at age 30-34. Of the sampled children, we excluded those whose parents could not be identified and those who, due to death or emigration, were not in the census when adult characteristics were measured (1950: n 1<unk>4 19 184, 1975: n 1<unk>4 8331). To guarantee comparability in the mortality follow-up, we followed the 1975 cohort to the end of 2016 and the 1950 cohort to the end of 1991 until the cohort members were aged 41-55. Cause of death was classified according to the 8th, 9th or 10th revision of the International Classification of Diseases and Deaths (ICD) with coding made comparable across the ICD revisions by Statistics Finland. Our final analyses were based on 1 467 360 person-years and 3592 deaths in the 1950 sample and 1 527 929 person-years and 2491 deaths in the 1975 sample. The design of the study is shown in Figure 1. --- Childhood characteristics Childhood characteristics came from the 1950 or the 1975 censuses when the cohort members were aged 0-14. We measure both socioeconomic and housing-related characteristics. For highest parental occupational class we distinguished the following groups: (i) non-manual, (ii) manual, (iii) farmers, (iv) employers/self-employed, and (v) others. We coded highest parental education as (i) upper secondary or higher, (ii) lower secondary and (iii) primary school or less. Of the housing-related characteristics, home ownership included: (i) home owners, (ii) renters (including those for whom the employer provided housing) and (iii) others/unknown. We characterized childhood housing conditions by household crowding coded as the number of persons per heated room in the dwelling (<unk>2, 2.0-2.99, 3.0-3.99, 4<unk> and unknown). Childhood family structure distinguished: (i) two-parent and (ii) single-parent families. --- Adult characteristics Adult characteristics were measured in subsequent censuses when the participants were aged between 30 and 34 years (the first 5-year census after turning 30) when most had completed their education and begun occupational and family careers. Occupational class was divided into (i) non-manual, (ii) manual, (iii) farmers, (iv) employers/self-employed and (v) others. Unemployed and retired persons were classified according to their previous occupations; housewives were categorized according to the occupation of the head of household. We distinguished between four educational levels: (i) higher tertiary (16<unk> years) education, (ii) lower tertiary (13-15 years), (iii) secondary (11-12 years) and (iv) basic or primary (9 years). Home ownership included (i) home owners, (ii) renters (including those for whom the employer provided housing) and (iii) others and unknown. Marital status was coded as (i) married, (ii) divorced/separated/widowed and (iii) never married or unknown. --- Causes of death We examined all-cause mortality as well as alcohol1<unk>4related, accidental and violent causes of death (ICD-10 F10, G312, G4051, G621, G721, I426, K292, K70, K860, K852, 0354, P043, Q860, V01-Y89 as the underlying cause) separately, as these are common among Finnish people. --- Statistical methods We used mixed-effects survival models to estimate allcause and cause-specific mortality from the age of 30-34 until the end of follow-up, which for the oldest individuals was 55 years. The baseline hazard was assumed to follow the Weibull distribution stratified by sex (we tested for sex interactions but these were negligible). As the data consist of persons nested in families, we use a two-level random intercept specification to take account of family clustering and allow for the estimation of family-specific random intercepts. The results are presented as hazard ratios based on coefficients from the proportional hazard parameterization [hazard ratio (HR) 1<unk>4 e b ] and their 95% confidence intervals (CIs). To test for change in the associations over the two samples, we pooled the data and included interactions with birth cohort. In order to assess the similarity in survival between siblings, we used the accelerated failure-time metric of the same model to estimate the family-level variance h. Because the individual level error term in Weibull models follows the Gumbel distribution, the individual-level variance is p 2 6p 2 where p is the shape parameter of the Weibull distribution and is obtained from the model. The sibling correlation measuring the sibling similarity in mortality (intraclass correlation is the share of total variation in midlife mortality which was explained by between-family differences) and is calculated according the method suggested by Goldstein et al. 19 for each model: Intraclass correlation 1<unk>4 h h <unk> p 2 6p 2 We take the sibling intraclass correlation to be a summary measure of all early life influences, both observed and unobserved, that are shared by siblings, 6 thus reflecting shared genetic and social influences. The'mestreg' command in Stata 14.2 was used for the estimation and 'nlcom' command for obtaining the 95% confidence intervals for the sibling correlation. --- Results The later cohorts had more highly educated parents who were much more likely to be employed in professional occupations (Table 1). Similarly, they had somewhat higher levels of education and were somewhat more likely to be employed in professional occupations than the cohorts born before the middle of the 20th century. In the 1950 sample, less educated or manual worker parents and children living in crowded or single-parent households had higher all-cause mortality in early mid-life, whereas home ownership in childhood was unassociated with early mid-life mortality (Table 2). There was no consistent evidence that these mortality differentials would have declined for the more recent birth cohorts in the 1975 sample. Instead, some of these associations appear to have increased. In particular, increase in the strength of the associations was especially strong for housing-related circumstances, such as living in rented housing (HR 1<unk>4 1.03 to HR 1<unk>4 1.21, with a cohort interaction parameter 1.17 95% CI 1.05, 1.31) in the last column) and living in single-parent households and in households with three persons per room. However, differences in mortality by parental education and occupational class have remained stable. Tables 3 and4 replicate the analyses for two broad causes of death; internal causes (Table 3) and external (alcohol-related, accidental and violent) causes of death (Table 4). For internal causes in the 1950 sample, most childhood characteristics were associated with elevated mortality, and these differentials were mostly similar to those observed in the 1975 sample with the exception of living in a household with over three people per heated room. However, the increase in the all-cause mortality differentials by childhood housing-related circumstances is mainly driven by external causes. For these causes of death, Separately for each index category, tests whether the hazard ratios are equal between the two cohorts in the baseline model. most childhood characteristics were more strongly associated with the risk of death in the 1975 than the 1950 sample. For example, the hazard ratio for single parenthood had increased by 30%. The associations between parental characteristics and mid-life mortality were substantially attenuated after adjusting for the individual's own adult social characteristics. This occurred for both cause of death groups and allcause mortality. After these adjustments, the associations between childhood circumstances and mid-life mortality were modest, indicating strong mediation by achieved adult social characteristics. Further adjustment for all childhood characteristics simultaneously changed the model estimates only little. Our estimates of sibling correlations-a summary measure of all early life and familial influences shared by siblings, both observed and unobserved-have increased for total mortality from 0.17 (95% CI 0.10, 0.24) in the 1950 cohort to 0.29 (95% CI 0.20, 0.37) in the 1975 cohort (Table 5). Sibling correlations for external causes of death increased from 0.28 (95 % CI 0.17-0.39) to 0.45 (95 % CI 0.36-0.54), and this increase cannot be explained by adjustment for achieved adult social characteristics. --- Discussion We evaluated the association between childhood circumstances and mid-life mortality in two Finnish cohorts, born in 1936-50 and 1961-75. We showed persistent associations between early life socioeconomic and family circumstances and mid-life mortality which are almost fully mediated through achieved social characteristics at ages 30-34. Despite the extremely rapid modernization and increasing affluence of the Finnish society in the latter part of the 20th century, these associations have not decreased between the cohorts. On the contrary, similarly to increased mortality differentials by adult social characteristics, [20][21][22] differentials according to childhood circumstances have increased over time for external causes of death. These increases have been substantial enough for housing-related childhood conditions to have a corresponding effect on allcause mortality. For internal causes the associations of childhood characteristics have remained stable. Most of the evidence on the association between childhood socioeconomic characteristics and later life mortality is from cohorts born in the first decades of the 20th century. Comparing separate studies that have recruited participants from several birth cohorts seems to tentatively indicate that the associations have not disappeared in more recent cohorts. 9 However, this conclusion is based on qualitative comparison of distinct studies using different methodologies, statistical adjustments and social contexts. Using consistent approaches over birth cohorts we show that, despite consistent temporal change towards improving childhood material living conditions, the 'long arm of childhood' has not only remained robust but has grown stronger. This increase is particularly evident for childhood experiences of rented accommodation, single-parent household and, in the case of alcohol-related, accidental and violent causes, also with having a primary educated or manual parental background. The observation that alcohol-related, accidental and violent causes are particularly strongly associated with early life is not new. 23 It is likely that these associations are driven partially by early adoption of risky behaviours, in particular alcohol and other substance use, in the more disadvantaged childhood environments with both chronic and binge-drinking being responsible. Our particular contribution, however, is to show increasing associations between several measures of childhood disadvantage with these causes of death over time. The association between childhood disadvantage and adult problem drinking is not fully conclusive. 24,25 For the purposes of this study, reliable evidence on alcohol use is difficult to obtain, mainly because of survey non-response, recollection and desirability biases and lack of comparative information on childhood disadvantage over different cohorts. However, it is tempting to speculate that alcohol and other substance use is ever more strongly associated with childhood disadvantage and underlies the increasing association with external mortality found in this study. It is possible that the long-term relaxation of alcohol control policies in Finland-affecting the younger cohorts more strongly-may have exacerbated these associations. Overall, these results imply that despite declining mortality in mid-life over time, a socioeconomically adverse start in life is in relative terms increasingly more hazardous to health. Empirically disentangling the role of various life course models remains difficult. Our results and many previous analyses of childhood socioeconomic influences on mortality indicate that these associations can be best understood in a life course framework, in which the associations of childhood socioeconomic characteristics are strongly mediated through achieved later life social position, 5,26 with own education being particularly relevant. 9 Similar findings have also been obtained in studies on health outcomes other than mortality. [27][28][29][30] The results also indicate that the associations of childhood characteristics on mid-life external mortality, which are mediated through attained social characteristics, have grown over time. This may come about because of at least two processes; first, the emergence of an ever-stronger association between the mediators and mortality. This possibility is supported by the fact that we know that the associations of achieved social characteristics with mortality have grown over time in Finland. [31][32][33] Second, the rise of a stronger association between childhood and adult characteristics. This possibility is supported by complementary analyses presented in Supplementary Tables 1234, available as Supplementary data at IJE online. These show that the associations between childhood-in particular those between childhood family type and home ownership-and disadvantaged adult social characteristics have emerged or increased over time. This development has been particularly strong for those living in rented housing in childhood. More research to confirm and extend these findings in other settings, for other health outcomes and with other methodological approaches to disentangling between changes in direct and indirect effects, are needed. Our analyses of sibling intraclass correlations further confirm the conclusion that the associations of early lifecapturing shared genetic and social influences-with later life mortality has increased. The analyses further show that after adjustment for a wide range of observed childhood and achieved adult characteristics, the intraclass correlation remains mostly unchanged. Importantly, this finding indicates that the specific early life characteristics that we measure-characteristics that are typically also measured in earlier studies-do not fully capture the childhood characteristics that are important, nor do we fully understand the pathways though which attained adult characteristics mediate childhood influences on later life outcomes. Further research should make efforts to evaluate the contribution of unstudied childhood and adult social circumstances for these differentials, including hitherto poorly understood social and genetic influences. --- Methodological considerations From a methodological point of view this paper contributes in several ways. First, information on childhood and adult characteristics was drawn from census records and thus was not subject to recall bias (common problem with retrospective data) or loss to follow-up (common problem with prospective data). Second, our relatively large samples enabled us to study both total and cause-specific mortality and evaluate the changing contribution of adult characteristics on the mortality differentials by childhood conditions over time. Third, we had access to two samples representing two birth cohorts 25 years apart with harmonized measurement and design. The distributions of some of the childhood characteristics have changed between the two cohorts, making it challenging to compare point estimates between the samples. For example, in 1950, 90% of children had parents with primary education, whereas in 1975 this proportion was 46% (Table 1). Thus, our measures may differ across cohorts in their ability to differentiate between childhood circumstances. In the case of education, this distributional shift should have led to smaller differentials over time. Distributional change has been less or none for other social characteristics, and the results for these confirm our main conclusions. However, even if the distributions were similar across cohorts, their meaning may have changed. For example, around 10% of children lived in single-parent families in both cohorts, but in 1950 the single parents were more often widows whereas in 1975 they were divorced. To overcome these measurement concerns we also assessed the contribution of all observed and unobserved childhood circumstances to the variation in mid-life mortality. This was done by calculating a sibling correlation (intraclass correlation) and comparing this measure across cohorts. The observed increase in sibling correlation is in line with our main result that many of the observed childhood characteristics had stronger associations with mortality in the more recent cohort. In conclusion, regardless of rapid economic development in the latter part of the 20th century, we show that the associations of childhood socioeconomic characteristics and mid-life mortality are substantial. Whereas we show that these associations are mainly indirect, this does not undermine the importance of childhood circumstances as important early life course experiences that set individuals on a path of later life social achievement and health. We further show that the effects of childhood circumstances on mortality have grown for the more recent birth cohorts. To a large extent the increase in the contribution of childhood conditions on mid-life mortality is driven by ever stronger associations of childhood circumstances with alcohol-related, accidental and violent causes of death. --- Supplementary Data Supplementary data are available at IJE online. --- Conflict of Interest None declared.

Background: Life course epidemiology suggests that early life circumstances affect adult mortality, but most of the evidence is based on cohorts born in the beginning of the 20th century. It remains unclear whether and how the influences of early life circumstances on mortality have changed in later birth cohorts. Methods: Analyses rely on 10% register-based samples of households drawn from the 1950 and the 1975 Finnish censuses, with consistent follow-up of socioeconomic and housing-related characteristics and early mid-life mortality (at ages 30-55 years). We estimate survival models for the associations between childhood circumstances and all-cause, internal and external mortality for cohorts born in 1936-50 and 1961-75 adjusting for attained social characteristics. We estimate sibling intraclass correlations as summary measures of all early life and familial influences. Results: Adverse childhood social circumstances were typically associated with about 10-30% excess cause-specific mortality. These associations were almost fully attenuated by adjustment for achieved later life social characteristics. Early life influences have grown over time for mortality from external causes, particularly as related to home ownership and family type. Differentials have remained stable for internal causes. The intraclass correlations further confirmed the increasing association of early life circumstances on external-cause mortality. Conclusions: Our analyses show that the associations between childhood characteristics and mid-life mortality are substantial and almost fully mediated by achieved adult social

the leadership of urban social organisations pass to the workers as bourgeois cities became proletarian spheres. 13 Yet there has been little investigation of what happened to many of the locally based organisations which made up the self help world of the late nineteenth century city, and especially not of the workings of urban hospital contributory schemes. Their social, cultural and political make up has received limited attention from either historians of the working class or of the urban infrastructure and thus and we know very little about who joined, how the various types of schemes operated on a day to day basis and the extent to which they became politicised in the increasingly class ridden inter war years. 14 This is unusual as, by 1938, contributory scheme membership had reached about 11 million by 1942 making them a key element in working class strategies to minimise social risk. 15 However, the schemes have not gone unnoticed. Central to understanding their social, political and financial importance between the wars is the work of Martin Gorsky, John Mohan and Tim Willis, whose wide ranging monograph on the movement has provided a framework for their further study. Yet, while overall they accept that the contributory movement achieved a great deal in a relatively short space of time, did much to keep the voluntary hospitals afloat and facilitated some elementary rationalisation in local hospital systems, they are sceptical about the degree to which members really engaged with the activist opportunities of the schemes and the extent to which they built a pan-class base for an alternative to local or national state control. Indeed they suggest most members only joined for the promise of treatment and cared little about either the voluntary hospital system as a morally superior way to deliver hospital care or for the claims of men like Sheffield's Moses Humberstone who asserted that it was 'a great humanitarian effort', a movement and not an insurance scheme. 16 Broad and deep as their research is, it only scratches the surface of the schemes' everyday existence and activities. Indeed, they admit that 'the identities and motivations of... ordinary members are now impossible to retrieve' turning instead to the local leadership which must suffice for understanding the membership. 17 However, by drawing on a range of local records, including the annual reports, executive committee minutes and the pages of the members' own monthly magazine, this article will examine some of these themes through a case study of The Leeds and District Workpeople's Hospital Fund between 1887 and 1939. 18 It will pay particular attention to the development of the Fund, the growth of its membership, income and services to its members, the structure and activities of the grassroots organisations and the political positions adopted by various key figures associated with the organisation over the inter war period. 19 Overall, it will consider the extent to which the Fund continued to provide a space for a mutualist approach to social provision in which a pan-class alliance could exist to provide a community service separate from the state. Hospital provision in late nineteenth and early twentieth century England was divided between a public sector of poor law and municipal institutions and a smaller voluntary sector of general and specialist hospitals. 20 The public sector expanded from the 1860s as many more new workhouses included separate Infirmaries charged largely with the care of the elderly and the chronic sick while councils dealt with those who might pose a threat to the community, including the infectious and the mentally ill. Between the wars the public sector attempted to shake off its Victorian roots and to develop general services for the wider community. This process was facilitated by the 1929 Local Government Act which transferred responsibility for the poor law to local authorities and gave them the option to 'appropriate' the poor law hospitals as council run medical and surgical institutions. By the end of the 1930s the majority of big cities had followed this route and there is evidence that Doyle in some areas municipal provision was improving significantly. 21 Yet it still lacked the resources, skills and reputation of the voluntary hospitals. These acute general and specialist institutions, which provided just over one third of the beds, were at the forefront of the rapid developments in medicine which characterised the early twentieth century. However, they were also invariably short of money as their traditional sources of income -the philanthropic middle classes -were unable, on their own, to meet these new obligations. As the voluntary sector feared state funding and was wary of adopting a system of direct payment, they welcomed the development of contributory schemes which could provide large, regular payments while continuing to rely on the traditional givers to meet much of their capital requirements. 22 These contributory schemes emerged out of nineteenth century working class organisation which sought to minimise the risks of modern capitalist life through collective or mutual pooling of resources. They included, as noted, trade unions, especially those of the skilled workers, the co-operative movement, friendly societies, various forms of thrift organisations such as burial clubs, sick clubs and even Christmas clubs, often associated with churches and chapels. 23 Although the latter largely disappeared after the welfare legislation of the Edwardian era and the disruptions of the First World War, trades unions, the co-op movement and even friendly societies flourished between the wars. As with the contributory schemes, debates remain about whether these organisations retained the commitment of their membership or if the relationship had become largely functional or whether mutualism had become redundant with the advent of state welfare. 24 It is thus within these contexts that mutualist hospital contributory schemes emerged in the late nineteenth century and came to play such a prominent part in the hospital economics and culture of the inter-war period. --- The Leeds Workpeople's Hospital Fund The Leeds Workpeople's Hospital Fund came into operation at the end of 1886 on the initiative of Alderman Fred Spark, a mid-century Radical Liberal and editor of the radical Leeds Express. 25 Spark was a central figure in working class organisations in the city and had secured a position on the Board of Leeds General Infirmary representing workers' interests. The Fund aimed to develop an effective means of mobilising workers' contributions to the hospitals in a more organised way through a regular contribution of 1d a week. It would replace the existing ad hoc arrangements including Saturday and Sunday Funds and other forms of collection at works which provided just under £2000 a year to the Infirmary. Spark worked with the key mutualist organisations -such as the trade unions, friendly societies and Jewish representatives -to set up the organisation and develop a local management structure whilst the first meeting was held at the Cooperative Hall. 26 The link between these organisations is very important -not just through Spark's personal role but also those of a range of other activists -and it is a feature which has been underestimated in much of the discussion of contributory societies and workers' funds. Indeed, it would seem that this embedding within the mutualist sub-culture was the key to the success of the Leeds organisation both in its formation and continued growth between the wars. Within a year there were ward organisations, a central committee with Executive and hundreds of subscribers both in workshops and as individuals. The initial membership was based on male industries such as metal working, engineering and railways, yet from the beginning the Fund recognised the need to mobilise the greatest number of potential members and courted both women workers -of which there were thousands in north Leeds -and the tailoring trades more generally. 27 Indeed there was a conscious decision to incorporate the 'Jewish shops' from the very start and to overcome potential prejudices, the Executive complaining vociferously about discriminatory treatment received by Jewish patients at the Public Dispensary, leading to the dismissal of a porter. 28 Ward committees were formed to create democratic structures which would engage members with the Fund, promote localised fundraising, provide a means for independent and home workers to contribute to the fund and generally raise its profile in the city. Early members and activists were drawn heavily from the Lib-Lab and Radical Liberal communities, such as Henry Maundrill in East Ward, along with community leaders like Joseph Kavanagh, representative of the Irish and Sigismund Ash from the city's Jewish community. 29 In the pre Great War period the patterns of local fund raising -beyond the collection of subscriptions from the works -took shape and came to include the organisation of collections in pubs and clubs, incentivised by competitions for Emblems for the pubs and prizes for the collectors. 30 Local committees were also active in promoting concerts in public parks during the summer. These were so significant that the Executive decided to buy music stands and music for the local committees to ease the costs of these popular events. 31 A crisis arose in the early Edwardian period when the City Council made a bid to restrict or even ban concerts in the public parks after accusations of rowdiness and complaints from the churches that they were offering too much of an alternative attraction on a Sunday evening. Following representations from, amongst others, Spark who emphasised the important charitable purpose of the concerts, the Council backed down and the concert programme continued. 32 Less successful was the decision of the executive to buy automatic collecting boxes which attracted limited contributions but the continuous attention of thieves and vandals. Alongside these local activities -which included carnivals and other public events such as a football match featuring the local pantomime troupe in costume 33 -was the Leeds Hospital Gala, a large scale event which was held annually on August Bank Holiday Monday and Tuesday from 1887 to 1939 at Roundhay Park -a large public park provided free of charge by the Corporation. 34 It received extensive coverage in the local press, especially following the death of one of the performers, Professor Higgins, a balloon artist, whose act went fatally awry in 1891. 35 The Gala proved to be both an effective profile raising activity and a financial success, contributing almost 10% of income within a couple of years and continuing as an important money spinner throughout the Victorian and Edwardian eras. 36 Thus, by the outbreak of the First World War the sources and approximate distribution of income was as in Table 1, with around three quarters of income coming from the workshop collections, substantial contributions from pub collections and the Gala (though this could be very erratic due to the weather) and a range of smaller events bringing in around £1000 but serving an important purpose in engaging a wide range of people in income generating activities for the Fund. The income raised by the contributors and the committee, which rose from £1800 in the first year to almost £14,000 by the outbreak of the First World War, was assigned to a growing number of activities. Initially the key focus was to support the hospitals and this remained the main purpose until the formation of the NHS. Given the origin of the Fund in a discussion at the Leeds General Infirmary (LGI) of ways to mobilise workers' contributions, it received the lion's share of the income raised by the Fund. Thus, in the first few years following the removal of a small administration charge of around 5-6%, the income was divided 82% to LGI, 10% to the Dispensary and 8% to the Women and Children's Hospital, which by 1890 was equivalent to around £4,300, £500 and £400. 37 In recognition of their contributions, the new Fund was allocated 20 workman governors at LGI and 10 at the Leeds Public Dispensary and by 1914 the Fund had a representative on the Boards of the LGI, the Hospital for Women and Children, the Public Dispensary, the District Nurses Association, Leeds TB Association, and a place on the Board, the Election Committee and the General Committee of the Maternity Hospital. 38 This increased representation reflected changes in the direction of funding. In the 1890s the Fund developed a significant convalescent system which by the First World War included a home for men in Horsforth and one for women in Ilkley, Derbyshire. This was a major undertaking with over 2000 admissions in 1913 -1300 men and 770 women -with the committee recognising the need to expand the provision for women. 39 The convalescent homes were costing over £4000 a year to run by this time, a sizeable proportion of the income of £13,000-14,000 per annum, and given the Fund had employed a general secretary in 1897 and had further diversified the recipients of support from the Fund to 10 institutions, the proportion allocated to the LGI fell significantly. Thus, by 1913 it received just £5,250 of the £13,250 income, around 40% of the Fund's total income and not much more than the convalescent homes. 40 As this suggests the convalescent activity had become a central concern of the fund in the early twentieth century. However, this shift to investing heavily in convalescence was causing tension with the Hospitals. For example, when, in 1914, the Fund stated it was planning to spend a substantial sum on improving the convalescence provision for women the representative of the Infirmary attending the AGM, Charles Lupton, stated 'He noticed, with some sorrow, that one of their Homes was so popular that they were wishing to increase the accommodation. He had hoped that the Infirmary would have been regarded as so popular that they would have wished to increase their donation to that institution'. 41 Thus, it is clear that the LWHF was a mutualist organisation which owed much to the central tenets of the mid-nineteenth century radical liberalism and Lib-Labism of Leeds. 42 But it also owed much to a deep rooted view of independence and collective support bound up in Spark's attack on National Insurance at the 1913 AGM: When they came to consider that the working men of Leeds had to pay by forced payment 4d per week out of their incomes [for the new National Insurance cover], and, in addition, were voluntarily giving a penny per week to the Fund, they must all agree it is very creditable. He ventured to think that the penny paid to the Fund was worth more to the working man in personal benefit than the fourpence which he was taxed. 43 That such organisations existed in pre-war Britain is widely acknowledged -the question is to what extent did they survive operationally and ideologically between the wars? --- Inter war Income Not only did the Leeds Workpeople's Hospital Fund survive into the interwar period, it positively flourished as a source of income for the city's hospitals and as a means of providing access, free at the point of delivery, for Fund members. The income of the Fund witnessed an impressive increase of over fourfold between 1918 and 1938, from £22,000 to £98,000. However, this increase was not straightforward as income grew only slowly during the 1920s, largely due to very poor economic conditions in the city's traditional industries. 44 Yet despite the impact of the great depression, which was severe on the engineering and textile sectors, 45 the Fund's income increased rapidly in the 1930s, more than doubling between 1930 and 1938 to reach almost £100,000. This major improvement was secured without the introduction of any significant new category of income; the sources changing little between the wars (see Fig. 1). The main changes were in the workplace scheme where, in 1919, the base contribution was raised to 2d per week and in 1930 when a contributory scheme was commenced. 46 The contributory scheme protected income in the early 1930s and probably helped to incorporate new members, including large numbers of women in the clothing industry such as the 8000 who contributed from Montague Burton's, along with support from offices and to a lesser extent retailing, reflecting the changing character of the Leeds economy and demand for hospital treatment. 47 The key aim of the contributory scheme was to concentrate all workplace collecting in the LWHF scheme to make it more efficient and regular through 'the withdrawal of all casual collections and, by arrangement with the Hospitals, the discontinuance of all collections in works for individual hospitals'. 48 Its success was measured both in terms of the huge growth in income and in membership which had reached 150,000 by March 1932. 49 Part of this growth was also associated with the expansion of the Scheme into the suburbs, especially to the east and south east in places like Morley so that in 1933 the name was changed to the Leeds and District Workpeople's Fund. 50 The launch of the contributory scheme caused some problems with the local medical profession who were worried that, in trying to mobilise middle class supporters for the scheme, this would allow those who should really pay for private treatment to gain access to subsidised care in the hospitals. 51 When addressing the AGM, the Lord Mayor, the Liberal Dr Hawkyard, noted his disquiet with the new method of collection, stating 'there were one or two things in it which he did not quite agree with'. 52 However, others at the AGM supported the development, Charles Lupton of the LGI assuring Dr Hawkyard that he had questioned the Fund on 'whether free hospital treatment was going to be available for those contributors who could well afford to pay doctor's fees and nursing home charges' and confirmed that the Fund would only send those who were a 'proper subject for voluntary hospital treatment' whilst the Hon. Sec., James Tate, reassured the Mayor that the scheme was 'designed for the necessary purpose of increasing our income' and that he was 'quite confident it will not be abused'. 53 Although the works' contributions were always the dominant element in fundraising the proportion drawn from this source varied from 75-95% of income over the course of the interwar period. Workplace contributions were at their lowest in the first few years of the post war period when collections from pubs and clubs and income from the Gala were both significant (Fig. 2). Similarly, ward income (from additional collections and fundraising activities) peaked in the early twenties then fell sharply from 1929, largely due to the incorporation of part of their income in the new contributory scheme. The Gala was also strong in the early 1920s, although this was always dependent upon the weather -as were many of the ward fundraising activities -with one or two years completely washed out. 54 This notwithstanding, Gala income was fairly consistent throughout the 1920s but it then fell sharply in the 1930s. The proceeds from pubs and clubs were closely linked to male tertiary disposable income, seeing a peak in 1920 during an era of rapid wage inflation, but then falling very sharply due to the appearance of large scale unemployment in areas like engineering and as a result of the general strike and the related lengthy coal strike. 55 However, unlike the other forms of income, pub and club collections were rising again towards the end of the 1930s, largely as a result of a renewed period of wage inflation. Overall, from 1930 worker contributions were the main source of income for the Fund, suggesting that from this point the movement had become an insurance scheme -though conceivably one in which the members took a greater interest than, for example, their engagement with the Prudential. 56 --- Interwar Expenditure As the sources of income for the Fund became more concentrated and the relationship between member and Fund more 'commercial' the range of organisations assisted and 'products' offered increased substantially. Thus the key features of the interwar period were the expansion of the services offered by the Fund; the extension of the range of organisations and hospitals benefitting from grants; and the significant decline in the importance of the convalescent homes within the mix of services offered. The Leeds General Infirmary remained the main beneficiary, given its role in the foundation of the Fund, but other institutions also benefitted through a significant growth in cash grants. The Public Dispensary, the Hospital for Women and Children (HWC) and the Leeds Maternity Hospital each secured more of the income and together, the four other hospitals (including the Jewish Hospital) and the Leeds Nursing Association received over £20,000 by the outbreak of the Second World War (Fig. 3). This distribution reflected the greater diversity of the membership, especially the contribution made by women workers, with the result that Leeds Maternity Hospital (which drew most of its income from the Fund) was the largest voluntary maternity institution in the country, delivering 45% of the babies born in the city by 1938. 57 Increased grants to the Nursing associations similarly reflected the spatial extension of the Fund to the outer suburbs and satellite villages. 58 Moreover, services to members grew. Thus, although dental provision had been present before the First World War, 59 it expanded greatly, especially in the 1930s. Support consisted mainly of grants towards the purchase of dentures, the patient paying the first 30 shillings, with the result that grants rose from just £150 in 1918 to over £3,200 by 1938. PT Leigh, President of the Fund at the end of the 1930s and the initiator of the Dental Hospital in Leeds, joked that in the early days they had more staff than patients but that the interest in dentistry reflected growing health consciousness amongst the population. 60 Provision was further diversified with the development of an ambulance service introduced in 1926 to bring in-patients, to and from their homes. 61 As membership expanded outside the city, partnerships were formed with ambulance services in the townships, for example at Morley, Guisley, Otley and Kippax by 1935. 62 Both the cost and the popularity of the service began to rise rapidly in the early 1930s with the fleet of ambulances conveying over 7000 patients by the end of the decade. 63 In a similar vein to the dental benefit were schemes to provide massage therapy and access to the Royal Bath Hospital, Harrogate for rheumatism suffers. 64 Furthermore, by the end of the decade the Fund was operating reciprocal arrangements with 48 hospitals across the UK with almost 500 members treated outside Leeds. This was an important development for, as Gorsky, et al have shown, this proved to be a particularly complex issue for contributory schemes to manage in this period. 65 However, more important than these innovations was the growing relationship with the municipal hospitals which began in 1931 when a deal was agreed with the Public Assistance Committee (Leeds did not appropriate its hospitals until 1933) to treat Fund members and their dependents at St James' and St Mary's. 66 There was, however, some scepticism amongst the voluntary hospital leadership about the arrangement which involved 350 patients in the first year. The Tory Alderman George Martin -chairman of the Municipal Health Committee -noted that on average patients spent three times as long at the PAC hospitals than at LGI, suggesting that the deal might impose a considerable cost on the Fund. 67 However, this was dealt with by the PAC/Municipal Hospitals only admitting patients who would have been eligible for voluntary hospital treatment and excluding 'Tuberculosis, Mental and Chronic Care cases' from as early as 1931. 68 Yet, over 5000 members and dependents were being treated by the end of the decade, an equivalent number to those treated in the LGI. 69 The expansion of this element of the Fund's work was stimulated by the chronic waiting lists at the LGI which meant members often had to wait months for treatment. However, if they wanted to make use of the municipal hospital, St James's, they were required to pay, and rightly they objected to 'paying twice' and demanded the cost of their treatment was covered by the Fund. The significant expansion in the numbers treated also reflected the upgrading of services at St James's and the marked improvement in relations between the voluntary and municipal sectors after 1935. 70 On the other hand the contribution of the convalescent service -central to the Edwardian Fund -declined. From the mid-1920s the number of patients sent to the homes stagnated at around 1,000 per annum, with a slight surplus of women over men. 71 As a result, the cost remained flat and proportionately declined whilst the homes saw no significant investment in new buildings. This was a surprising development given that from the late 1890s up to the early 1920s it had been such a major part, if not the main work of the Fund. When the homes were first opened, Fred Spark noted that they signified the maturity of the Fund which rather than simply making grants 'wanted to be absolute managers of something'. 72 This tradition was maintained by Sir William Middlebrook, Liberal MP and President of the Fund during the 1920s, who stated on a number of occasions that the homes were a principle and even 'the primary object' of the Fund. 73 This view was not shared by the hospitals, with Lupton of the LGI noting in early 1932 that he was pleased that they had kept up grants to the hospitals, for though the convalescent homes did'splendid' work, he felt there 'could be no doubt that the first charge upon their generosity was the need of the hospitals of the city.' 74 Thus, between the wars patterns of Fund expenditure became more complex as they focused on securing access to a diverse range of hospital treatments for their members, rather than on creating and running their own services like the convalescent homes. Moreover, the growing breadth of hospital services was clearly reflected in the spread of the grants, whilst the more sophisticated demands of the members -through the take up of dental services for example -and the widening spread of the Fund -through the ambulance service -complimented the rising standard of and demand for medical attention most obviously seen in the growth of births at the maternity hospital. 75 Indeed, those associated with the Fund took the opportunity of the AGM to reflect on the growth in demand for treatment in hospital, Charles Lupton suggesting that it was 'testimony to the increased belief in hospitals that grew up during the War' whilst the Labour Mayor, John Arnott noted a similar change in attitude, contending that in the past people 'did not believe in hospitals, but to-day that was all changed, and the work done in the hospitals was honoured and admired.' 76 Moreover, the wider range of services provided reflected demand created by the changing membership body. No longer was the Fund dominated by male manual workers from the engineering shops south of the river, but increasingly from men and women from across the Leeds district who expected a wider range of benefits than simply care and recuperation. --- Membership and Activism The Leeds Workpeople's Hospital Fund was a very successful organisation which saw its membership on a continuous upward curve, reaching almost 200,000 by the outbreak of the Second World War. As has been argued, especially by Gorsky et al, the overwhelming majority of these members took no interest in the organisation of the Fund beyond paying their dues. 77 On the face of it the organisational work and the engagement with the political and cultural objectives of the LWHF were restricted to a relatively small group of activists whilst the rest treated it as another form of insurance. 78 Yet this view may not be entirely fair and there is evidence of a broader support for the movement in a limited and occasional form by a much wider body of people demonstrated through their attendance at Fund events or their casual contributions through their social and recreational clubs and associations (see Figs 1 and2 above). This association with the Fund resembled workers' other mutualist engagements whether as members of the Cooperative Society -which had over 100,000 members in the city by the 1930s, most of whom used it as a cheap sources of basic groceries -or even as trade unionists, the vast majority of whom played little or no part in the day to day activities of their lodge. However, it is possible to identify some of the organisations and people involved in supporting this vital plank in the working class support structures of Leeds. --- Doyle It is true that identifying the background of even a few of the local members is very difficult, however this research points to a wide range of interests participating in the grass roots of the Leeds Fund. In broad terms membership covered a diversity of work places from mining and iron working through textiles, engineering and clothing, to local government and administration. 79 As noted earlier, this included a significant proportion of the female working population, unusually for this type of scheme in the north where women tended only to belong as dependents. 80 This presence was manifested to some extent in women's involvement in the local leadership of the Fund and as we have seen in the kind of services provided. 81 However, despite many attempts to engage shop workers, retailing -with the exception of the Co-op -was not widely represented, though prior to the setting up of the contributory scheme shopkeepers in certain areas did club together to undertake collections, the North East Ward Shopkeepers donating £13 in the summer of 1928. Moreover, despite criticism to the contrary, some of the multiple stores, including Maypole Dairy and Woolworth were organising collections before the switch to the contributory scheme. 82 Although individual contributors subscribing through their place of work were the mainstay of the movement, a range of other organisations and activities were also involved in maintaining both the income and vitality of the Fund. For example, a great many clubs were represented in the subscriber lists until the 1930s, including the by then rather quaint North West Ward Tariff Reform Club which donated £2 in the summer of 1928. 83 Even in 1932 the North East Ward had significant collections from both the Conservative Club and the Labour Club. More surprisingly the list of collections from the East ward the same year included entries for the United Irish Democrats next to that of the East Ward Conservatives and the East Ward Liberals -though not on this occasion the Labour Club. 84 Whilst political clubs were important, they were supported by a range of working men's clubs, such as those at Meanwood and Beeston, British Legions like that in Middleton, the RAOB and the Forresters Institute or the East Leeds National Railway Union Club whilst various sporting clubs supported the events organised by the wards, such as Kirkstall's annual fishing competition. 85 Leadership varied but could be very wide ranging. In city centre, Brunswick ward the president in 1931 was the local Conservative Alderman, PT Leigh, who later served as president of the Fund, whilst the Vice Presidents included his wife, also a Conservative councillor in the ward. They were supported on the committee by Vice Chair Mrs Joe Cohen and George Kelly, a leading member of the Ancient Order of Forresters Friendly Society from an Irish background, whilst the secretary was JC Sanderson, a vocal proponent of the working class identity of the movement from a Liberal heritage. In the Labour heartland of North-East Ward, the Conservative Councillor, J.W. Hemmingway, was president and a prominent supporter of the movement, providing prizes of gold watches for the pub collectors. North-west Ward elected a woman as its representative worker governor. In socialist East Hunslet a prominent local Liberal was Vice President whilst in New Wortley the President was a Labour councillor. 86 It is possible that Liberals remained disproportionately active in the movement, whilst in some areas the organisation was increasingly run by Conservatives, especially in the city centre and the north and west of the borough. Yet partisanship seems to have been rare, with all mainstream political subcultures associating themselves with the Fund either at the level of activist involvement or pub and club collection competitions. It is conceivable that local politicians and parties even used the Fund as part of a banal form of political promotion which allowed them publicity all year round. Certainly for many it was an appropriate venue in which to undertake important community based activity to demonstrate their commitment to the interests of the population of the ward. --- Doyle Activism and engagement was also demonstrated by support for the income generating activities of the Fund at the local and city level. The main event in the calendar was the Gala which continued to be held annually in Roundhay Park over August Bank Holiday Monday and Tuesday with a programme which combined sporting events with spectacle. 87 The day could attract some prominent sports people, such as the Huddersfield based Olympic athlete Tom Matthewman who competed from 1924 onwards. These athletes competed for substantial cash prizes in the 1920s in both foot racing and cycling whilst attendees were also entertained by a range of comedic and spectacular acts including aerial acrobatics, tumblers and other circus acts, jugglers and western style riding. 88 However, the Gala gradually lost its appeal. Attendances in the early twenties were usually over 40,000 for the Monday and over 10,000 for the Tuesday, with 1920 seeing almost 70,000 pass through the turnstiles. By 1930, however, attendances had fallen to nearly half that figure, numbering around 25-30,000 for the Monday and just 5-7000 for the Tuesday. 89 Even the stock attraction of the hot air balloon with its parachute descents had paled by this time as local people increasingly left town on bank holidays for the coast or the country and fewer people were attracted to an increasingly unsophisticated day out. 90 The Gala was suspended for the Second World War but never reinstated. Yet despite that, over the course of the interwar period over three quarters of a million people attended an event organised for and promoting the Fund. Whether this was a positive affirmation of the actions of the Fund or just a day out cannot be established -but it does suggest something about both the depth of organisation the Fund was still able to draw upon until the end of the thirties and of the pool of good will which existed in the city. At ward level there were a range of small scale fundraising activities which collectively maintained the profile of the Fund on the ground and contributed a few thousand pounds to the coffers. Very much of the age were the whist drives and dances (including fancy dress events) which occupied the committees in the winter months, often held in the political clubs of the wards and bringing together a few dozen members. 91 These were aimed specifically at the active members and served to cement a collective social culture and reward the efforts of the collectors as well as possibly act as a way of recruiting new activists. In addition there were a number of community events, usually in the summer, such as the annual carnivals run in the suburban areas with varying degrees of success. In the case of Bramley ward this may account for their large women's committee which seemed to consist of the daughters of the leading activists. In Kirkstall there was an annual Angling contest -again with its own committee -whilst across the city there was a bowls cup held every summer. 92 On the other hand the later 1920s saw the decline of the public park band concerts, a mainstay of early fundraising but largely abandoned by most wards by 1930. Around a dozen concerts were still held over the summer months but takings were rarely significant and many were affected by the poor weather leading the Secretary to comment in 1932 that: Band performances in the Parks of the City appear to be no longer an attraction to the Public, and in consequence are a dwindling source of revenue to the Fund. Ward committee efforts in other directions have, however, met with favourable responses. 93 In some cases these were replaced with either sacred concerts or organ concerts indoors, especially in cinemas on Sunday evenings. However, this could prove problematic prior to 1932 as local by-laws prohibited the selling of tickets at the venue on the night. 94 Thus, the committees in the wards did attempt to keep up to date in their fundraising methods, though their effects were probably more propagandistic and morale boosting than particularly beneficial to the coffers of the Fund. Indeed in most cases the Pubs and Clubs were the mainstay of ward collections and the focus of much of their communal activity as special meetings were held to celebrate the most successful and award prizes to venues and Doyle Page 10 Labour Hist Rev. Author manuscript; available in PMC 2014 January 06. --- Europe PMC Funders Author Manuscripts Europe PMC Funders Author Manuscripts collectors. These were an opportunity for a bit of fun for members in contrast to the dry annual meeting or the hard work of fund raising. 95 Overall the local organisation demonstrated a broad range of interests and activities. There were few ethnic, religious, political or social divisions. Drawing on a wide range of aspects of working class popular culture from the thrift organisations of the nineteenth century; the premises of the free church community; and the pastimes of the working man (and woman) -whether uplifting, for example fishing or bowls; bucolic, like the May Carnival; or traditional, such as the pub and club -the local efforts of the Fund suggest a commitment which helped to hold back the march of class and state. --- Ideological Battles Yet for all that there were some ideological battles within the movement. For whilst it may have been capable of maintaining a broad church in terms of support, it does not mean everyone understood the movement in the same way. There were broadly three positions articulated in various forums. The first was the voluntarist position, held to most strongly by the large number of Conservative activists

In the immediate aftermath of the First World War, British voluntary hospitals faced a financial crisis which looked likely to jeopardise their independence from state control. 2 Inflation and the seeming inability of traditional sources of income to keep up with rising costs meant the hospitals needed a new and regular supply of funds -preferably one which kept local and national government out of their board rooms. The solution was found in the development of a range of contributory schemes designed to raise money from patients in return for more or less assured access to treatment, transforming the funding of hospitals in interwar Britain. 3 Based more or less on nineteenth century models of working class mutualism, the form of these schemes varied significantly from traditional Saturday and Sunday Fund collections 4 through workplace collections with representation on hospital committees; 5 city wide collection funds independent of the hospitals to contributory schemes based either on a hospital or group of hospitals or a locality. 6 Recent work by Steve Cherry and Gorsky, Mohan and Willis as well as the research by Doyle on Middlesbrough and Gosling on Bristol, 7 has helped to foreground the importance of these activities to the financial viability of interwar voluntary hospitals and has gone some way to open up the operation of these schemes to greater scrutiny and understanding. But their place in early twentieth century urban culture and the developing social systems of the working-class, especially the fate of mutualist organisations between the wars, remains under-researched. In general, mutualism has received less attention from labour historians of the inter war period than trade unionism, socialist organisations or the Labour party. 8 The mid to late nineteenth century is often posited as the high point of mutualism with the rise of craft unions, friendly societies, cooperative societies, building societies and a wide range of thrift organisations associated with mechanics institutes, penny banks and nonconformist churches. 9 However, historians have perceived a shift from the 1880s away from radical led self-help and mutualism towards more politically directed bodies targeting unskilled and excluded workers, underpinned by socialist ideals and looking increasingly to a statist response to social or economic problems. 10 As a result, the role of mutualism after the mid-1880s has been downplayed, squeezed by professionalised, bureaucratic organisations in education, housing or insurance 11 on the one hand and by calls for, and the implementation of, state responses on the other. While not characterised as an immediate or linear development, historians such as Martin Daunton and Pamela Graves have seen the interwar period as one in which workers abandoned their faith in mutualism and a fear of the state in their quest for universal, high quality welfare services. Indeed Daunton, has asserted that by the 1930s 'The proportions in the mixed economy of welfare had been reversed, and where there was a mismatch between voluntarism and the state, it was resolved decisively in favour of the latter'. 12 Similarly, the community, pan-class possibilities inherent in these mutualist organisations have been downplayed as historians, such as Mike Savage and Andy Miles, have emphasised increasing class segregation in the structure of civil society. In this model a combination of assertions of working class independence and middle class withdrawal saw

ers Author Manuscripts Europe PMC Funders Author Manuscripts collectors. These were an opportunity for a bit of fun for members in contrast to the dry annual meeting or the hard work of fund raising. 95 Overall the local organisation demonstrated a broad range of interests and activities. There were few ethnic, religious, political or social divisions. Drawing on a wide range of aspects of working class popular culture from the thrift organisations of the nineteenth century; the premises of the free church community; and the pastimes of the working man (and woman) -whether uplifting, for example fishing or bowls; bucolic, like the May Carnival; or traditional, such as the pub and club -the local efforts of the Fund suggest a commitment which helped to hold back the march of class and state. --- Ideological Battles Yet for all that there were some ideological battles within the movement. For whilst it may have been capable of maintaining a broad church in terms of support, it does not mean everyone understood the movement in the same way. There were broadly three positions articulated in various forums. The first was the voluntarist position, held to most strongly by the large number of Conservative activists involved in the Fund at various levels and who became increasingly prominent in the broader voluntary hospital culture of the city. People like Charles Lupton, PT Leigh, Noel Whittaker, and especially Sir George Martin, used the meetings of the Fund to champion the voluntary system, attack state control and link the Fund to the preservation of the voluntary hospital system. 96 Thus, in February 1930, in a characteristic assault prompted by the introduction of the Local Government Act, Martin asserted that: He regarded the Fund as a sort of bulwark against the municipal control of the voluntary hospitals of the city. However excellent municipal control might be, however painstaking, he was satisfied it would not be to the advantage of the hospitals to come under the municipality. It would be a very bad day for those who required the help and beneficent care which a hospital could give when such institutions were regulated and controlled by the municipality. (Applause). 97 Such views could be multiplied across a number of speakers both representing the Fund and other medical charities in the city. Admittedly many of them spoke favourably of the potential for cooperation -though even here Martin was sceptical in the early 1930s -but most saw this as an arms-length relationship rather than a positive development to be actively pursued. 98 Second, there were those who took the view that the Fund represented a democratic response to the crisis of the hospitals between the wars and that it was the working man who had saved voluntarism. In this context, the voluntary system had an obligation to work closely with the municipal sector to secure the best hospital care for the community. These views were usually put forward by visiting Labour Mayors, including the middle class figures of D B Foster 99 and John Badlay. The latter, who claimed to have attended meetings as an ordinary member, suggested that This Fund has had an effect upon our social life altogether apart from the financial side. I am convinced that it has led to the formation of independence of character on the part of the people who have helped to carry on this work. When men now go to the Infirmary they don't go timidly to the back door, they go there as a right, feeling that they have paid for what they are going to get and have assisted others to obtain the same services. (Applause). 100 But it was not just the visiting dignitaries who asserted this line. Towards the end of the period ordinary committee members became more vocal, and in particular, J C Sanderson took an independent line on a number of occasions, asserting in 1938 that: It was obvious that had it not been for the working people of the city the voluntary hospitals would have been in a very bad way in the last ten years... [as well as]... paying a warm tribute to the efficiency of the medical work and the nursing at St James's Hospital, he said that the Workpeople's Hospital Fund had done a great service for the people of Leeds when they made arrangements for the patients to be received and treated at that institution. 101 Third was the traditional radical Liberal view which encapsulated the original aims and ambitions of the movement but was increasingly marginalised by the later 1930s in the shadow boxing over voluntary or state provision. Given the large number of Liberals involved in the organisation at a local level, this position may have been more widely held than articulated but it did receive a full exposition from the Liberal Lord Mayor, Charles Boyle, at the AGM early in 1940. He mused on the early history of the Fund, noting: Some men in Leeds got together and said the position was not good enough for them, and that they desired, as Englishmen, to be independent of charity in their medical services. The result of such a feeling was the wonderful one of raising today something like £100,000 a year from the workers of the city in support of medical charities. But, although that was good, they had not got full independence yet. They still had to rely to some extent upon the State and upon charities, which was a thing Englishmen abominated. He hoped that some day, perhaps not far distant, they would pay every penny of the cost of medical services in the City, which would, indeed, be a great achievement. 102 --- Conclusion Although Boyle's clear articulation of the independent, mutualist position may have owed something to a war setting in which the essence of Englishness was often being evoked, it also summed up a view which has been largely written out of the inter war history of the movement. Mutualism rather than voluntarism, was still widespread in mid-twentieth century Britain -evident in Friendly Societies, the Co-op, Trade Unions, and hospital contributory schemes. 103 Although they were from the working class these ideas and organisations were not exclusive to them. They were pan class, communitarian, democratic and often still an essential support to daily life. As Labour Lord Mayor, John Arnott pointed out in 1925, they were 'in the nature of a mutual insurance, an effort to pool resources, so that a person who did suffer from misfortune should have not only his own personal provision to help him in need, but that of his fellows who cooperated to help him.' 104 All such organisations offered a form of insurance or pooling of risk which may have owed little to overtly political considerations but suggests more than simply a commercial transaction. They were born out of a need to use self-help to gain access to scarce or privileged resources and to prepare for the inevitable costs of sickness, accident and old age. As such, these institutions promoted independence from patronage as well as want. The Leeds Workpeople's Hospital Fund was phenomenally successful in increasing income, supporting the hospital system, promoting integration and efficiency and mobilising thousands of people to act as citizens sustaining the fabric of their city. However, it was able to do this because of its roots in nineteenth century radical liberalism and nascent working class organisation. It was temporally contingent, an artefact of its time. Certainly the slow decay of its fundraising wing suggests that its ability to mobilise may have been on the wane, whilst the resistance of the hospitals to an increase in their role on the boards may have put a block on further development. 105 Yet the Fund retained a strong financial base into the forties and may even have benefitted from the growth in tertiary expenditure which characterised the post war period. 106 The Leeds Workpeople's Hospital Fund continued to hold the respect and attract the pennies of hundreds of thousands of people. It therefore provides clear evidence that mutualism in health care provision was still very healthy up to the mid twentieth century and could have formed part of a mixed economy response to future provision as it did in France and Germany.

In the immediate aftermath of the First World War, British voluntary hospitals faced a financial crisis which looked likely to jeopardise their independence from state control. 2 Inflation and the seeming inability of traditional sources of income to keep up with rising costs meant the hospitals needed a new and regular supply of funds -preferably one which kept local and national government out of their board rooms. The solution was found in the development of a range of contributory schemes designed to raise money from patients in return for more or less assured access to treatment, transforming the funding of hospitals in interwar Britain. 3 Based more or less on nineteenth century models of working class mutualism, the form of these schemes varied significantly from traditional Saturday and Sunday Fund collections 4 through workplace collections with representation on hospital committees; 5 city wide collection funds independent of the hospitals to contributory schemes based either on a hospital or group of hospitals or a locality. 6 Recent work by Steve Cherry and Gorsky, Mohan and Willis as well as the research by Doyle on Middlesbrough and Gosling on Bristol, 7 has helped to foreground the importance of these activities to the financial viability of interwar voluntary hospitals and has gone some way to open up the operation of these schemes to greater scrutiny and understanding. But their place in early twentieth century urban culture and the developing social systems of the working-class, especially the fate of mutualist organisations between the wars, remains under-researched. In general, mutualism has received less attention from labour historians of the inter war period than trade unionism, socialist organisations or the Labour party. 8 The mid to late nineteenth century is often posited as the high point of mutualism with the rise of craft unions, friendly societies, cooperative societies, building societies and a wide range of thrift organisations associated with mechanics institutes, penny banks and nonconformist churches. 9 However, historians have perceived a shift from the 1880s away from radical led self-help and mutualism towards more politically directed bodies targeting unskilled and excluded workers, underpinned by socialist ideals and looking increasingly to a statist response to social or economic problems. 10 As a result, the role of mutualism after the mid-1880s has been downplayed, squeezed by professionalised, bureaucratic organisations in education, housing or insurance 11 on the one hand and by calls for, and the implementation of, state responses on the other. While not characterised as an immediate or linear development, historians such as Martin Daunton and Pamela Graves have seen the interwar period as one in which workers abandoned their faith in mutualism and a fear of the state in their quest for universal, high quality welfare services. Indeed Daunton, has asserted that by the 1930s 'The proportions in the mixed economy of welfare had been reversed, and where there was a mismatch between voluntarism and the state, it was resolved decisively in favour of the latter'. 12 Similarly, the community, pan-class possibilities inherent in these mutualist organisations have been downplayed as historians, such as Mike Savage and Andy Miles, have emphasised increasing class segregation in the structure of civil society. In this model a combination of assertions of working class independence and middle class withdrawal saw

bACkgrounD Although there has been a reduction in new infections of HIV of about 41% between 2000 and 2015, HIV infection continues to be a public health problem in Sub-Saharan Africa, where 25.6 million people lived with HIV. Women of reproductive age account for about two-thirds of them. 1 The decline in the incidence of HIV infection is the effect of multiple interventions implemented to prevent HIV transmission among heterosexual and especially from mother-to-child HIV transmission. About 66 000 children under 14 years were reported to be newly infected by HIV in Western and Central Africa at the end of 2015. 2 Most of these infections could have been prevented by antiretroviral therapy (ART) during pregnancy and more by contraceptive use among HIV-infected women, which is the fourth component of preventing mother-to-child transmission (PMTCT). In 2005, Reynolds et al 3 estimated that the level of contraceptive use could prevent over 173 000 unintended HIV-infected births each year in Sub-Saharan Africa. Indeed, family planning showed numerous benefits for HIV-infected women, such as the reduction of morbidity and mortality due to pregnancy and also the improvement of health of HIV-infected women by reducing unintended pregnancies. Therefore family planning contributes to PMTCT of HIV. 4 The strengths and limitations of this study <unk> The use of contraceptives such as condoms could provide dual protection against acquisition or transmission of sexually transmitted infections, including HIV. <unk> This study is one of the first studies in Togo focusing its interest on the contraceptive use among HIVpositive women. <unk> Education level, marital status, AIDS clinical stage, type of health centre and having children had influenced the use of contraceptive methods among HIVinfected women. <unk> It is suggested that caregivers should integrate family planning services in HIV care during follow-up visits. <unk> Finally, we relied on self-reported reproductive health behaviour through a face-to-face interview, which may introduce information bias. Open Access use of contraceptives, mainly male and female condoms, could provide dual protection. It could protect against acquisition or transmission of sexually transmitted infections, including HIV. 5 6 However, in Sub-Saharan Africa, given the progress in PMTCT with the expansion of ART, most of the HIV-infected women still have desire to have children, 7 8 and the prevalence of contraceptive use varies according to the studies. In Western Africa, the use of contraceptive methods is not effective in HIV-infected women. [9][10][11] In Ghana, a study showed that only 42.6% of HIV-infected women used modern contraceptive methods at Komfo Anokye Teaching Hospital in 2012, 11 while in eastern Nigeria it was reported that 73.1% of HIV-infected women were using modern contraceptives in the same period. 10 In many studies conducted in Sub-Saharan Africa, the reported factors influencing the use of contraceptive among HIV-positive women were sociodemographic characteristics (age, education, marital status), reproductive characteristics (desire to have children, number of children), being on ART and knowledge of the partner's HIV status. [9][10][11][12][13][14] In 2015, Togo had about 110 000 people living with HIV and AIDS (PLWHA), including 59 000 women aged more than 15 years old and 9000 children aged 0-14 years. 15 In addition, the prevalence of HIV infection is almost twice higher among women than men (3.1% vs 1.7%). 16 HIV prevalence among women attending antenatal consultation in Togo was 2% in 2015. 17 In women of reproductive age, the prevalence of contraceptive use was estimated at 22% for all methods (16% for modern methods) in 2013-2014. 16 However, contraceptive use among HIV-infected women in Togo is poorly documented. This study aims at assessing the prevalence of contraceptive use, as well as individual and structural factors that influenced the use of contraceptives among HIV-infected women enrolled in HIV care centres in two health regions in Togo. --- MethoDs study design This was a cross-sectional study conducted in two health regions in Togo (Centrale and Kara regions) over a period of 4 months from May to August 2016. The study targeted HIV-positive women of reproductive age (15-49 years) and who were sexually active. setting and study population Centrale and Kara regions are two of the six health regions in Togo located, respectively, about 350 km and 420 km from the capital Lomé. These two regions are characterised by important seasonal arrival of populations (from the other health regions, and sometimes from outside of the country (Togo)) during cultural and traditional events, increasing the risks of HIV transmission. In 2013, the prevalence of HIV infection was estimated at 2.2% in the Centrale region and 1.8% in the Kara region. 16 There were 30 accredited medical centres in the two regions to provide healthcare to about 10 361 PLWHA, including 616 children, enrolled and followed up in these centres. 17 Women were eligible to participate in the study if they (1) were aged between 15 and 49 years old, (2) had a sexual partner within the last 6 months and (3) were enrolled in the selected centres for active ART. sampling First, a random probability sampling proportional to the number of patients in the active file of the 30 HIV care centres in the two regions was carried out to ensure representative centres with a high number of PLWHA. In a minimalist scenario, it is assumed that a sample of 15% of HIV care centres should be representative of all the centres. This led to the random selection of five HIV care centres for the implementation of this study. Then, a non-probabilistic, convenience sampling was carried out. Indeed, in these five selected HIV care centres, it was proposed to enrol any PLWHA who consulted for a follow-up from May to August 2016, who met the inclusion criteria and who consented to participate in the study. The prevalence of the use of at least one contraceptive method was assumed to be 50% with the precision of 5% and 20% refusal or incomplete data. Based on this assumption, the sample size was estimated at 461 HIV-infected women. --- Data collection Data collection was done using a standardised questionnaire in French, explained in the local language for participants if needed. The questionnaires were filled by the health workers. The questionnaire included sociodemographic information, clinical features, information on ART, sexual activity status, and knowledge on contraceptives and their use. Data on HIV status disclosure to the sexual partner were also collected among PLWHA. We defined modern contraception as the use of one of the following methods: sterilisation, condoms, contraceptive pill/oral contraceptives, intrauterine device, diaphragm, injection, emergency contraception or implant. statistical analysis of data Data entry was performed using EpiData V.3.1 software. Data were then exported for statistical analyses using SPSS V.17.0. For continuous variables, mean and SD were calculated, while for categorical variables we calculated the proportions. Our main outcome variable was HIV-infected women who were currently using any modern contraceptive methods. Pearson's <unk> 2 test or Fisher's exact test was used when appropriate in bivariate analysis. Multivariate backwards stepwise logistic regression analysis was performed to identify independent risk factors for the dichotomous outcome contraceptive use or not. All variables significant during bivariate analysis at a p value <unk>0.20 were introduced in a logistic regression model to estimate the adjusted OR (aOR). Interactions between the independent variables were tested. All these analyses Open Access were performed with 95% CI. To verify the robustness of the results, a sensitivity analysis was conducted to confirm the results found. We have performed the analysis by including women who have never heard about contraceptive methods, and then we excluded them to appreciate the variability of the statistical associations. We did not observe significant variability and logically we excluded those women (15/461) from the model. Our final model included only HIV-infected women who had heard about contraceptives methods. --- ethical issues We obtained consent from patients who participated in the study. For each of the person surveyed, the objectives, benefits to participate in the survey and progress of the investigation were clearly stated, as well as their right to interrupt the interview without justification. An informed consent form signed after the verbal explanation was made by the investigating officer in the language understood by the participant. --- results sociodemographic and clinical characteristics Table 1 shows the sociodemographic and clinical characteristics of the participants. In total 461 HIV-infected women of reproductive age and who were sexually active were enrolled into this study, of whom 267 (57.9%) were in the Centrale region and 194 (42.1%) in the Kara region. The mean (<unk>SD) age of the participants was 34.3<unk>7.1 years, ranging from 16 to 49 years. Among them, 40.6% had primary education or higher education level (45.3%), 54.7% were living in couple and 62.3% were living in an urban area. Out of the 461 women interviewed, 415 (90.0%) were on ART and for more than 2 years for 71% of them. The mean (<unk>SD) duration on ART was 4.1<unk>2.8 years. A quarter (25.4%) of the participants had a CD4 cell count of <unk>0.35 <unk> 10 9 /L at the last visit. Patients were classified as stage I (55.0%) or stage II (27.3%) based on the WHO clinical classification at the time of the survey. The partner's HIV status was unknown for 56.4% (260/461) of the participants. Most of the patients were followed up in a public hospital centre (61.8%), in a centre with available doctor (86.3%) or with a psychologist (55.7%). --- reproductive health outcome The majority of women (85.3%) reported having at least one child, while six out of ten participants expressed desire of having kids. Four hundred and forty six (96.7%) women reported that they heard about contraceptives methods. Among them, 70.2% (313/446) were informed by healthcare providers. At the time of the survey, the proportion of women who were using a contraceptive method was 74.7% (333/446) (table 1), and among them 74.5% (248/333) reported using condom, 8.4% (28/333) were using hormonal contraceptive and 16.9% (56/333) reported using dual methods combining condom and hormonal contraceptive. For 15 (3.3%) women who have never heard about contraceptives methods, we considered they had missing data for the variable contraceptive use. Furthermore the proportion of women using contraceptive was significantly higher among participants with higher level of education, ranging from 57.1% in women with no education to 82.8% in those having reached the secondary education level or higher (p<unk>0.001). However, this proportion was significantly higher (p<unk>0.001) among women at stage I (82.1%) than those at stage II (70.0%) or those at stages III and IV (58.4%). The prevalence of contraceptive use was significantly lower in women living in couple (p=0.002), higher among those who had children (p=0.032) and higher in women on ART for 2 years or more (p=0.025). Women followed up in a private centre (p<unk>0.001) or a centre with a psychologist (p=0.011) were more susceptible to use contraceptive methods (table 1). --- Factors associated with contraceptive use In multivariate analysis, only five factors remained associated with contraceptive use: education level, marital status, clinical stage, type of health centre and having children. Women with primary education level or those with secondary education level or higher were, respectively, two times more likely (aOR=1.99, 95% CI (1.05 to 3.76)) and four times more likely (aOR=3.95, 95% CI (2.03 to 7.67)) to use a contraceptive method than those who did not have any education, whereas participants living in couple were 55% less likely (aOR=0.45, 95% CI (0.28 to 0.74)) to use a contraceptive method. Women who reported having children and those followed up in a private structure were almost three times more likely (respectively, aOR=2.51, 95% CI (1.41 to 4.5) and aOR=2.54, 95% CI (1.22 to 5.29)) to report current use of contraceptive. Patients at AIDS clinical stage III or IV were 53% less likely (aOR=0.47, 95% CI (0.23 to 0.87)) to use a contraceptive method than those at clinical stage I. In addition this likelihood of using contraceptive is even lower for HIV-infected women at AIDS clinical stage III or IV followed up in private centres (aOR=0.13, 95% CI (0.03 to 0.64)) (table 2). --- DIsCussIon In this study, conducted in two regions in north of Togo, 96.7% of HIV-infected women had heard about contraceptive methods and mainly from health providers. Education level, marital status, AIDS clinical stage, type of health centre and having children had influenced the use of contraceptive methods among HIV-infected women. A high proportion of knowledge on contraceptive methods was also reported among women aged 15-49 years in the general population in Togo (96.3%) in 2013-2014 16 and among HIV-infected women on ART in Ethiopia (97.0%) in 2013. 18 This showed that seropositive women were as exposed to reproductive health Open Access Open Access information as the rest of the general population, which is a major step towards the use of contraceptive methods. We also found that almost three-quarters (74.7%) of the participants were using a contraceptive method. This proportion was higher than that of 19.9% reported in the general population in 2013 in the Demographic and Health Surveys (DHS) in Togo. 16 This was evidenced in prior studies conducted in Sub-Saharan Africa, 12 19 which reported that HIV-infected women seemed to be more likely to use contraceptives than HIV-negative women in public health facilities. This could be explained by the fact that HIV-infected women are frequently in contact with health workers and received more information on reproductive health during their follow-up. HIV-positive women are strongly encouraged to use contraceptive methods to prevent mother-to-child transmission of HIV. Similar results were found in South-East Nigeria (73.1% 10 ) or in Ethiopia (71.0% 20 ), where a high proportion of HIV-infected women used contraceptives. However, lower rates of contraceptive use were reported among HIV-positive mothers in some previous studies in Ethiopia (45.7% 18 ), Uganda (45% 14 ) or Ghana (42.6% 11 ). These differences might be due to temporal and regional variabilities, including policy guidelines on reproductive health. In fact cultural characteristics and beliefs, including the perception of the value of a child in the community, may vary through African regions, and this could strongly influence the decision of using a contraceptive method or not. In addition, in Togo, healthcare providers in HIV clinics have recently benefited from a programme of strengthening their skills on family planning. Non-governmental organisations have also been established to promote reproductive health through the distribution of contraceptive methods. This would have encouraged HIV-infected women to use these methods. As reported in previous studies in Western Africa, 9-11 condom was the most common contraceptive method used by HIV-infected women in this study. Fearing possible pharmacokinetic interactions between hormonal contraceptive and antiretroviral in people living with HIV, 21 22 healthcare providers more often advised HIV-infected women to avoid hormonal contraceptive method in favour of condoms. Condoms, in addition to being accessible and very effective in preventing unwanted pregnancies, contribute to the secondary prevention of HIV, especially in this context where more than four out of ten HIV-infected women did not know the HIV serostatus of their sexual partner. It is therefore important to know the influence of the HIV-infected woman in the decision-making process of condom use or not in a couple. In this study, the decision to use contraceptive methods in HIV-infected women was influenced by several factors, including education level, marital status, AIDS clinical stage, type of health centre and having children. Our findings revealed that women who attended school regardless of level (primary, secondary or higher education) were more likely to use contraceptives than those who could not read and write. Women education, which --- Open Access is key to societal development, remains a strong determinant of contraceptive use in Sub-Saharan African. This result was consistent with different studies conducted in Ethiopia 23 24 and in Malawi. 25 Women who attended school were able to better analyse reproductive health information, and to understand the mechanisms of action, and the advantages and disadvantages of different contraceptive methods. This could reassure them about the use of these methods and allow them to make a judicious choice. Our study showed that HIV-infected women who were living with a partner at the time of the survey were less likely to use contraceptive methods. In fact women in couple were more likely to express their desire to have children. This desire is increased by social pressure, which underlines a great importance of having children in the well-being of the couple. In addition, unmarried women would use contraceptive methods more than married women to avoid unwanted pregnancies, which could deteriorate their social image. Unlike our finding, Worke et al 24 in Ethiopia reported that married women were more likely to use contraceptives than those who were divorced or separated. In our study, HIV-infected women followed up in private centres were more likely to use a contraceptive method. Although very few studies compared the quality of healthcare between the public and private sectors, particularly in family planning services delivery, it is generally evidenced that healthcare is better in private health centres. 26 27 In private centres, even though they are non-profit-making facilities, it is reported that care providers are motivated to maximise patients' satisfaction in order to retain them, and this through the provision of good-quality healthcare. However, prior studies 28 29 reported that public health facilities were more likely to offer family planning services. Therefore, it is important to improve the quality of family planning counselling and services delivery in public health facilities in order to encourage patients, including HIV-infected women, to use these services. Among HIV-infected women, we found that the odds of contraceptive use were lowest at high AIDS clinical stage. Indeed patients at stages III and IV were less likely to use contraceptive methods. At these stages of the disease, patients were likely to suffer from opportunistic infections with certainly a reduction of sexual activities. This could impact on their demand for reproductive health services, including contraceptive use. Moreover, the odds of using contraceptives are even lower when women are followed up in a private centre. This could be explained by the fact that in our setting, HIV-infected patients in the advanced stage of the disease are most often referred for treatment in public hospitals. HIV-infected women who had children, regardless of number, were more likely to use a contraceptive method. This result is consistent with prior studies conducted in Uganda, 30 Malawi 25 and Ethiopia, 23 31 which reported that contraceptive use increased with the number of children. In fact, among women who gave birth at least once, it is very common for them to use contraceptive methods to space or limit births, especially for those who have already reached the number of children they want. limitations First, the sample may not be representative of the whole country because of cultural disparities across the country. Second, this study took place in health facilities with little chance of including patients who are less regular in the health centres and may have different characteristics. Sexually active women were defined as those engaging in sexual activity in the past 6 months; this definition used in the study could differ from other similar studies. In addition, we lacked information/analysis by contraceptive types, which can impact mother-to-child transmission. Finally, we relied on self-reported reproductive health behaviour through a face-to-face interview, which may introduce information bias. However, these biases would have had very negligible effects on the estimates in this study. --- ConClusIon This study demonstrated a high prevalence of contraceptive use among HIV-infected women followed up in health facilities in Centrale and Kara regions in Togo. Private health facilities seemed to favour the use of contraceptive methods, but not in women at clinical stages III and IV. Despite these results, there is a need for an intensified effort to increase the reproductive health service utilisation particularly in public health facilities in these regions. HIV caregivers must also integrate family planning services in HIV care during follow-up visits. --- Open Access the overall scientific management of the study, for analysis and interpretation, and the preparation of the final manuscript. All the authors read and approved the final manuscript to be submitted for publication.

Introduction Contraceptive use among HIV-infected women in Togo is poorly documented. We aim at assessing the prevalence of modern contraceptive use and associated factors among HIV-infected women in Togo. Design Cross-sectional study. setting The study was conducted in five HIV care centres in the Centrale and Kara regions in Togo. Participants We included 461 HIV-positive women aged between 15 and 49 years and who were sexually active. Main outcome measure The outcome variable was HIVinfected women who were using modern contraceptive methods. results A total of 461 HIV-infected women were interviewed, with an average age of 34.3 (±7.1). Among them, 332 (73.1%) women reported using contraceptive methods, mostly condom alone (74.7%) or in combination with hormonal contraceptive (16.9%). In multivariate analysis, education level (primary: adjusted OR (aOR)=1.99, 95% CI (1.05 to 3.76); secondary level and higher: aOR=3.95, 95% CI (2.03 to 7.67)), WHO clinical stage (stage II: aOR=0.7, 95% CI (0.37 to 1.33)), follow-up in private care facilities (aOR=2.54, 95% CI (1.22 to 5.29)) and having a child (aOR=2.51, 95% CI (1.41 to 4.5)) were associated with higher contraceptive use, while marital status (living in union: aOR=0.45, 95% CI (0.28 to 0.74)) and WHO stages III and IV (aOR=0.47, 95% CI (0.24 to 0.94)) were associated with lower contraceptive use. Conclusion About three-quarters of sexually active HIV-infected women in Togo were using contraceptive methods, and private health facilities favoured this contraceptive use. It is important to strengthen the implementation of interventions to increase the incentives for HIV-infected women to use contraception in Togo.

Introduction Initiatives targeting geographical areas of relative deprivation (area-based initiatives -ABIs) have been a consistent feature of public policy within Western countries in recent years (Randolph, 2004, Bradford, 2007, Larsen and Manderson, 2009, Berkeley and Springett, 2006). Since 1997 in particular, a central feature of ABIs in the United Kingdom (UK) has been an emphasis on co-ordinating service provision within and across sectors through various partnership arrangements (Sullivan et al., 2006). These partnerships have generated considerable research interest, which has revealed consistent difficulties in their development (Benzeval, 2003, O'Dwyer et al., 2007, Judge and Bauld, 2006, Beatty et al., 2010). Despite this attention, the experience of ABI partnerships remains under-theorised, limiting the extent to which the difficulties reported in the literature can be explained. Although interorganisational relations have been examined in detail in the fields of business and healthcare (Aveling andMartin, 2013, Dickinson andGlasby, 2010), ABI partnerships in public health present a set of specific issues given that they usually involve a large number of organisations working towards disparate aims within the confines of a time-limited intervention. This paper offers some empirically grounded theoretical insights into partnership processes among service providers within ABIs in order to shed light on why collaboration often remains elusive. It presents findings from a case study of an ABI delivered between 2007 and 2012 in a town in the north west of England. Target Wellbeing (TW) was a BIG Lottery funded initiative targeted at 10 geographical areas of health 'disadvantage' across the North West, defined as such by initiative co-This paper was published in Health 18 (6) 561-79 https://www.ncbi.nlm.nih.gov/pubmed/24695385 3 ordinators according to levels of self-rated physical and mental health, obesity rates, fruit and vegetable consumption, incidence of coronary heart disease and benefit claims (name withheld, personal communication, 2007 1 ). The multi-sector providers co-ordinating TW intended to improve the health and wellbeing of residents via a programme of activities in targeted areas. Each programme was made up of 8-10 projects designed to improve one or more of the following: healthy eating, physical activity and mental wellbeing. The projects were delivered by a range of voluntary and statutory organisations and each programme was managed by a lead organisation and a designated TW co-ordinator. Figure 1 shows the management structure for TW. --- Figure1 Target Wellbeing management structure This paper was published in Health 18 (6) 561-79 https://www.ncbi.nlm.nih.gov/pubmed/24695385 4 The specific question this paper examines is, how do relations between service providers develop over time when an ABI is introduced in an area? Ideas from figurational sociology were drawn on to conceptualise ABI collaborations as a figuration of interdependent people. The paper starts with a brief review of the literature on service provider relations within ABIs before introducing concepts from figurational sociology that were used to inform the research. The methodology is then described and the findings presented, describing and explaining the social processes that shaped co-ordinated working over time. The discussion examines how figurational sociology facilitated a more adequate understanding of the ways in which ABI partnerships develop and considers some of the policy and practice implications of our conclusions. --- The rhetoric and reality of provider collaboration Since the late 1960s successive UK governments have shown considerable interest in co-ordinating service provision through ABIs (for a brief overview see Burton, 1997, Stewart, 2001). The Labour Government of 1997-2010 argued that the health of people living in deprived areas could be improved through, among other things, better co-ordination of, and improved access to, services (Department of Health, 2004, National Institute for Health and Clinical Excellence, 2008, Department for Communities and Local Government, 2006, Department of Health, 2001). Coordination of local public services was a core feature of many Labour-funded initiatives including Health Action Zones (HAZs), Sure Start and New Deal for Communities. The Government claimed that the causes of worse health and social outcomes in deprived areas were interconnected and therefore required the co-ordinated expertise of a range of providers (Blair, 1997). HAZs, for example, were seen as This paper was published in Health 18 (6) 561-79 https://www.ncbi.nlm.nih.gov/pubmed/24695385 5 vehicles for innovating services by encouraging providers to work outside of their professional boundaries in the delivery of projects seeking to promote health (Bauld et al., 2005). Although there have been fewer government funded initiatives under the current Coalition Government, interest in local co-ordination of services has remained (Department of Health, 2010). Despite sustained policy interest, research indicates that ABI partnerships have not been implemented according to the expectations of policy makers (Berkeley and Springett, 2006) and that co-ordinated working between providers within ABIs has been limited (Hunter and Perkins, 2012). Some of the problems with partnership development reported in empirical studies are discussed below. --- Partnerships as social processes Much of the work on ABI partnerships has focussed on identifying "factors" that influence partnership development (Wildridge et al., 2004, 6) or, as Dowling et al. (2004) have argued, on the identification of barriers to implementing planned action. For example, differences in the governance structures of organisations across different sectors were identified as barriers to collaboration in the strategic development of HAZs (Unwin and Westland, 2000), but are the ways in which governance structures are established, maintained or challenged through human interaction within an ABI partnership are not understood. Similarly, several studies have shown that competition for funding between service providers within ABIs can undermine capacity for collaborative working (Milbourne, 2009, Carlisle, 2010) and that such competition can exacerbate "fear of outsiders" among service providers (Milbourne, 2009, 287). While such work is important in identifying problems in partnership development, we would argue, for the reasons set out below, that none of this work provides a model which offers an adequate understanding, on a more theoretical level, of the processes involved. We would further argue that such a general model is required for, without a continual interdependencewhat Elias referred to as "an uninterrupted two-way traffic" (Elias, 1987:20) between the development of detailed knowledge and synthesising models, the collection of detailed knowledge of particular situations will be of limited use, for it is only by the use of synthesising models that we can generalise from one situation to another. The limitations within the ABI literature can partly be explained by the focus on strategic partnerships between service co-ordinators at the regional or city level (Beatty et al., 2010, Carlisle, 2010, Henderson, 2011, Sullivan et al., 2006). This has directed the focus of research towards management and leadership issues within ABI partnerships and has deflected research attention from the social relations that develop between those involved in service delivery, which is central to our purposes in this paper. Checkland et al. (2009) suggest that more attention is needed on the social conditions that create barriers to implementing policy. Examining the ways in which providers have become interdependent with others over time, including colleagues and professional peers, might facilitate a better understanding of the ways in which they are constrained in their capacity for collaboration by emotional involvement in a particular set of relations. Pawson and Tilley (1997, 70) draw attention to the fact that social interventions are "introduced... into an existing set of social relationships." However, there has been a tendency within ABI evaluations based on a realistic methodology to view provider relations as static (Barnes et al., 2003). Furthermore, the cross-sectional study design of many ABI evaluations has meant that health partnership processes are often depicted as linear and predictable (see for example Boydell and Rugk<unk>sa, 2007, Lawless, 2002, Wholey et al., 2009). Asthana et al. (2002) identified a framework for evaluating HAZs, distinguishing between context, inputs, processes, outcomes and impacts. This framework, however, does not identify the connections between these elements, failing to recognise the complex ways in which these social processes are interrelated. Changing social relations (between those involved directly in the partnership and a range of others on the periphery) and their influence on the development of partnerships have not been adequately accounted for thus far. As Sullivan et al. (2006) note, researchers have struggled to explain how partnerships are influenced by unplanned events within, what is described as, the wider social context of the partnership. The influence of national policy changes on local-level partnerships (Beatty et al., 2010) suggests that the social networks in which ABI providers are embedded, beyond the immediate partnership, are likely to shape the way in which they work with other providers. Although the everyday microdynamics of partnerships have been researched and reportedas illustrated in the above examplesthey have not been adequately understood. This is partly because the everyday relations can only be understood when contextualised within broader, longer-term social processes. The literature indicates that there is no straightforward causal relationship between strategic aspirations for partnerships, planned activity and outcomes. Examining prospectively, as we do here, the planned and unplanned outcomes in an ABI, and the processes that connect them, provided an opportunity to develop understanding about processes of co-ordinated working within an ABI. Some key concepts within figurational sociology, which informed our theory of partnership development, are discussed below. --- Theorising service provider collaboration in public health ABIs Figurational sociology, a perspective which has been used to examine organisational change within the NHS (Dopson andWaddington, 1996, Mowles, 2011) Elias, 1991). Service providers within an ABI are interdependent with (at the least) other local and national providers, funders and policy makers, and their capacity for co-ordinated action is therefore both facilitated and simultaneously constrained by the actions of those people. Planned public health ABIs, therefore, are likely to produce consequences which no group or individual intended (Elias, 1994). Examining the figurations in which service providers are immersed presents an opportunity to explain why providers have been constrained in their capacity for collaborative working. A central dimension of figurations is power, conceptualized not as a substance possessed by particular individuals or groups but as 'a structural characteristic of human relationships' (Elias, 1978: 74). Power is never absolute but always a question of relative balances, for no-one is ever absolutely powerful or absolutely powerless. Power balances are also inherently unstable and continuously in flux. While most sociological perspectives draw attention to power relations, conceptualising an ABI partnership as a figuration of interdependent service providers draws attention to the ways in which their interdependencies are characterised by different balances of power. Where there are heavy imbalances of power, for example in relation to the professional status of providers within a crosssector group, some parties might be better able to exert more control over events than others. Based on empirical examination of resident relations within a small town, Elias (with Scotson) (1965) argued that power balances within a figuration could be influenced by one's status as either 'established' or 'outsider'. The introduction of a new set of projects into a small town as part of an ABI has much in common with the social changes examined by Elias and Scotson (1965). Observing caution that the 'established-outsider' concept might dichotomise the experiences of different groups (see Bloyce and Murphy, 2007), it might usefully be applied to examine power relations between providers in a small town. Finally, for the purposes of this paper, figurational sociology encourages analysis that incorporates the historical context of social relations within an area. Elias (1991) argued that because social phenomena emerge from interweaving human actions, it is impossible to locate their origins to any precise'moment' in time. The interweaving actions result in dynamic interdependencies and shifting balances of power between people over time. However, Elias also perceived that over time a person's place in a network of relations with others strongly influences her/his disposition, tastes, ambitions and expectations (Elias, 1991) or what he described elsewhere as habitus (Elias, 1996). Examining social processes prospectively, and thus developmentally, therefore has the potential to better explain unplanned events. Again, figurational sociology is not unique in pointing to the importance of historical context, but offers 'a set of sensitizing concepts... with the potential to draw many of the various threads of sociological thought together' (van Krieken, 2001, 353). --- Methodology A longitudinal, qualitative case study design was used to examine relations between service providers within a single town ('Seatown') targeted by TW. This provided an opportunity to trace the links between particular events (Maguire, 1988) in order to generate theoretical generalisations about the social conditions that shape partnership development (Yin, 2003, Dopson, 2003). The study had a commitment to a grounded theory approach (Strauss and Corbin, 1998) while also testing out a number of figurational ideas, using them as sensitising concepts as outlined above, to maintain a two-way relationship between inductive and deductive processes (Elias, 1978). Non-TW co-ordinators 5 Total 29 Interviews took place over 12 months. Discussion focussed on the history of provider relations in the town, perceived balances of power between providers, processes through which providers worked together, and ways in which co-ordination between providers was perceived to influence service provision locally. In order to explore changes over time, three interviewees were interviewed twice and a second interview was arranged with providers at four out of the six TW provider organisations in the town, albeit with a different person at the organisation in some instances. Participants for these follow-up interviews were purposively sampled according to their place in the network of providers; the aim was to explore changes to the network that had been identified through observations and documentary analysis. For example, analysis indicated that the person appointed as Seatown programme manager occupied a central position within the figuration of TW providers and so this person was interviewed twice. All fieldwork was carried out by the lead author. Interviews were audio-recorded and transcribed verbatim; the data were managed in NVivo. Coding was carried out by the primary author but on-going discussion with the secondary authors encouraged a greater degree of detachment from the data (Perry et al., 2004, Elias, 1987). Figurational ideas were used to sensitise the researcher to particular social processes taking place at every stage of the analysis. In this respect, the use of grounded theory facilitated a "constant interplay" between generating new ideas directly from collated data and testing existing explanations of human actions as Elias (1978, 34) encouraged. Observation and documentary data provided contextual information in which to situate interview accounts. Constant comparison of incidents across interview, observation and documentary data facilitated the synthesis of codes (Strauss and Corbin, 1998) which were used to capture the different ways in which particular ideas and issues emerged in the data (Bartlett and Payne, 1997). Following initial coding of the data, connections between codes were explored in order to develop explanations about what was taking place. The analytic concepts developed through this process formed the basis for theoretical development as Charmaz (2006) outlines. --- Findings TW projects within Seatown were commissioned by representatives from the Local Strategic Partnership (LSP) via a competitive bidding process. On the basis of 19 submitted bids, the LSP members selected eight projects to fund, delivered by six different organisations. Within these organisations, new or existing staff were appointed to deliver TW activities. In some instances, these staff had also been appointed to deliver other projects within the organisation, funded through other sources. The social dynamics between the providers that emerged as relevant within the analysis are outlined in Table 2. --- Organisational pull The analytic concept 'organisational pull' was developed from the data and captured the way in which TW providers were interdependent with others in their organisation, which constrained their capacity to work collaboratively with TW providers at other organisations. A shared commitment to the "mission" and values of their organisation was one of the ways in which these providers were interdependent (TW provider 14). Association with a particular area of expertise shaped how providers defined their professional identities. Consequently, staff identified more closely with the specialist organisations in which they worked than with TW more generally, as these were the organisations within which their professional identities were rooted. One TW coordinator commented that the TW providers "see themselves as... a member of [their organisation]...as opposed to, 'I'm a member of Target Wellbeing'" (TW coordinator 01). Where TW providers had experiences in common with their service users, this shaped their commitment to the client group with which the organisation worked. For example, one TW provider (TW provider 14) described how her work with young people at a mental health organisation was shaped by her own low self-esteem as a young woman. In these instances, as well as instances where providers had a long history of working with a particular client group, past experience had shaped providers' views about the priorities for services in 'deprived' areas. Providers at different organisations did not always share the same priorities. One TW provider perceived that others were "delivering their own agendas" (TW provider 12). The historically constituted social identities to which TW providers subscribed were therefore defined against other professional identities in the town. Analysis of observation and documentary data indicated that TW providers working on different projects at the same organisation worked together more frequently than did TW providers across different organisations. TW providers working on different TW projects within the same organisation reported working as "one big team" within their own organisations (TW provider 12) and, from the point of view of co-ordinators, "seem[ed] to merge together" (TW co-ordinator 01). Organisational pull also helped to explain the way in which TW providers were drawn together because of their dependence on one another for future work. Although many TW providers were employed through short-term TW funding, there was an expectation among them that should their organisation secure funding from other sources, this might enable them to secure more paid work. TW providers were keen to ensure that collaboration facilitated their own work in a particular field; one TW provider from an organisation based some distance from the town said, "We [as an organisation] want to develop links with [the town] and develop new projects from our contacts" (TW provider 02). Providers at other organisations were sometimes deemed by TW providers to pose a threat to the future success of their organisation. This was exacerbated by the introduction of individual project recruitment targets by TW co-ordinators, which heightened the sense of competition between TW providers. These targets (set by TW co-ordinators) were regarded by many TW providers as an important measure of success to the funders, partly because of monitoring arrangements (which were deemed to be more extensive than the delivery arrangements) and partly due to previous experiences of having funding withdrawn having failed to reach targets within past initiatives. Several TW providers expressed a fear that engaging in joint activities with other TW providers might threaten achievement of their own organisation's resident recruitment targets as the following quotation illustrates: I just don't think we're talking to each other as well as we could in terms of projects. And I think part of that is the fear of crossing over (Non-TW co-ordinator 02). Therefore, the deeply-rooted behaviours of providers interwoven with the actions of others in both the TW figuration and more complex figurations at regional, national and global levels, constrained the ways in which TW providers approached relations with providers outside of their own organisation. Although intra-organisational interdependencies had a strong influence on providers in the early stages of the initiative, relations between providers were dynamic and over time other processes emerged that mediated the influence of these interdependencies. Analysis of observation and documentary data indicated that TW providers across different organisations shared more information about their work with one another over time. This was particularly evident at quarterly programme meetings, where TW providers shared more details about their activities over time. During one of the last quarterly meetings, providers discussed how they had felt more inclined to work with TW providers at other organisations once they had begun --- 'Local' status The development and operationalisation of 'local' status were processes which influenced the balance of power between providers in the figuration. Being 'local' was expressed as having one or more of the following attributes: living locally, having an established history of working in the area, or working from a local office base. TW and non-TW providers and co-ordinators associated a number of positive characteristics with local status. Local status was associated with having a good understanding of the population targeted by the initiative. This understanding was considered to be valuable by TW providers because, it was explained, it was used to inform the development of activities and recruitment methods within projects. One TW provider explained that she had sought to appoint people who "live [d] in... local wards" to deliver TW activities because they were more likely to have local knowledge (TW provider 12). Others commented that living locally gave providers greater insight into resident needs. Local status was also associated with caring more about residents. The following quotation, from a non-TW provider at an organisation based in Seatown, demonstrates how local status was associated with an investment in its residents: I do what I do here and I care about it because I live here and my family lives here, I want there to be good services, you know... If I didn't live here, would I care in the same way or would it just be about the money? (Non-TW co-ordinator 02). Local status was explicitly defined against "outsider" status which was sometimes associated with poor understanding of residents' needs (TW provider 03). Analysis indicated that 'outsider' providers were perceived as posing a threat to the position of those based in Seatown. Some 'local' non-TW providers, for example, expressed agitation that providers based outside of Seatown had been chosen to deliver TW projects. One out-of-town TW provider described how providers based in Seatown had predicted that his organisation would be unable to deliver the TW programme successfully: The initial feedback... was [that] there was no chance we would be Local status was associated with legitimacy in terms of accessing Seatown resources. TW providers and co-ordinators articulated a view that being seen as local made other 'local' providers more inclined to refer their service users to TW. For example, TW providers whose organisation had connections with other providers in Seatown revealed this connection to people with whom they wanted to work. One TW provider described how the "reputation" of her organisation gave it a "real advantage" when working with other providers in Seatown; she said, "they don't just want anybody coming in and working with their [users]" (TW provider 12). Local status was cultivated by the ways in which providers worked with one another. TW providers with experience of working in the area revealed that, in some instances, they preferred to refer their service users to non-TW providers with whom they had established relationships than to other TW providers with whom they did not. Explaining why she preferred to refer her project users to one provider over another, one TW provider said: We've never really been able to engage with [one of the TW projects]... [another non-TW provider] will work longer with our [users] because of the working relationship we've got with them. So... there are probably other organisations that we already work with. (TW provider 14). Referring residents to 'local' providers with whom one already had an established relationship perpetuated a provider's status as local and served to prevent other providers from accessing resources. Such was the perception that being local was an advantage that one TW provider (TW provider 05) accounted for the difficulty she had in establishing relations in the town by wrongly assuming that her organisation was the only one that was not 'local'. This indicates that local and outsider status was used effectively to exclude some providers from the provider network in the town. --- Earning one's stripes Outsider status was not fixed but, rather, was part of a fluctuating balance of power between providers. Analysis indicated that local status could be earned over time, particularly through developing relations with others who were deemed to be local. There were several processes through which the balance of power between providers shifted over time. First, the development of "niche" activities by TW providers, which did not overlap with existing provision, facilitated the development of relations with non-TW providers (TW Provider 13). Over time, some TW providers made changes to the activities that they were delivering because they perceived that there was "overlap" with their provision and that of other TW and non-TW providers in the town (TW co-ordinator providers based outside Seatown were more likely to adopt this strategy. One TW provider, for example, described how "fitting in with their agenda" made it easier to access support from local non-TW providers (TW provider 10). This can be understood with reference to the sense of competition between providers in the area. The development of a specific niche for TW activities removed some element of competition between providers and increased the likelihood that TW providers could offer something to the clients or users of non-TW services. This illustrates the way in which 'local' providers were able to use their status to influence what was delivered within TW, how it was delivered, and by whom. Another way in which some TW providers became more accepted among providers in Seatown was through word-of-mouth endorsements from providers considered to be more 'local'. TW providers at an organisation with no history of working in the town asked a TW co-ordinator, based at the PCT, to arrange meetings for them with health practitioners in the town. One co-ordinator considered that these meetings provided legitimacy for TW providers and a "sort of reference... to actually get recognised as something that was kosher" (TW co-ordinator 01). These endorsements could start a process of discussion between providers, as one TW provider said, "It was certainly a door opener for us with the recommendation from the PCT" (TW provider 05). TW providers considered it important that other providers understood and valued their work and this could be achieved through word-of-mouth endorsements. The influence of word-of-mouth processes can partly be explained by the perception among local providers that the voluntary and community sector (VCS) in Seatown was underdeveloped. With few established networks between VCS providers prior to TW, word-of-mouth endorsements helped 'local' VCS and statutory providers to determine whether or not it would be helpful for them to work with the newlydeveloped TW projects. The sense of competition between providers also influenced their sense of wariness and word-of-mouth endorsements enabled providers to judge the extent to which providers with new projects might pose a threat to their own organisations. --- Discussion: Theorising Target Wellbeing In this paper we have drawn upon key aspects of figurational sociology in order to offer a more adequate understanding of processes of joint working, which have been a key feature of social policy within many Western countries. There has been an assumption that ABI partnerships have encountered problems due to implementation failure. By emphasising the complexity of the figurations within which ABI providers were immersed, this study has shown that the problems within this ABI partnership were not chance or accidental events, nor can they be understood in terms of poor leadership; rather, they can only be understood in terms of the unplannedand in this case unwantedoutcomes of the way in which networks of relationships between service providers developed over time. Organisational pull was a concept developed from the data and informed by the concept of figurations to explain processes that constrained partnership development. It reflects the way in which TW providers within an organisation were drawn to work together rather than with providers outside their organisation. Elias (1978, 15) argued that individuals "are directed to and linked to each other in diverse ways through their basic dispositions and inclinations," formed over many years through processes of socialisation or habitus formation. Working with the same people, or in a particular field of professional practice over many years, providers at the same organisation had similar priorities in terms of what they thought was needed in deprived areas such as those targeted by TW. These findings resonate with those made in the field of teacher education, where the term 'occupational socialisation' has been coined to explain the way in which learning processes in a particular field of occupation come to shape perceptions (Lawson, 1983). Shared dispositions could be seen to bind TW providers together in this study such that they developed a sense of allegiance to the work of their organisation. As Milbourne (2009, 291) has noted, "collaborative work often depends heavily on the commitment, dispositions and networks of individuals, and situated experiences." Through the longitudinal approach adopted in this study we have sought to show how these commitments, dispositions and experiences of service providers are shaped through the historically constituted figurations of which they are a part. Providers at the same organisation had a vested interest in the survival of their organisation, which became more apparent in the light of their fears about competition and funding. Competition for funding between organisations within ABIs has previously been shown to undermine capacity for collaborative working (Carlisle, 2010, Milbourne, 2009). The findings from this research extend this analysis to show that competitive processes between providers are on-going and do not necessarily recede once the commissioning process is over and that the sense of competition between providers at different organisations seemed to be exacerbated by TW monitoring and evaluation processes. One of the unintended consequences of setting resident recruitment targets at an organisational level was that TW providers were persuaded that such targets were vital to the funders, which limited TW providers' capacity to work towards other goals. Organisational pull therefore helps to explain how the interweaving actions of providers and co-ordinators led to consequences that co-ordinators, despite their apparent position of authority, could not control. The networks in which providers and co-ordinators were embedded (including those with funders and other providers) constrained the development of collaborative working. Competition was an aspect of a struggle for power between TW providers that was predominantly shaped by the status of providers as either 'local' or 'outsiders'. Milbourne (2009, 287) showed how competition for funding between providers can exacerbate "fear of outsider [providers]" in community-based initiatives. This research provides an explanation for Milbourne's findings by showing how the significance of outsider status in TW reflected power balances between providers. TW providers who were successfully able to claim 'local' status defined 'local' and 'outsider' status in dichotomous terms that served to reinforce their own privileged position, in much the same way identified by Elias and Scotson (1965, 81). As noted earlier, it is important to avoid conceiving of networks of relations as static: a project like TW is more adequately conceptualised as a social process with fluctuating balances of power. As such, the position of providers as 'outsiders' could, at least to some degree, be modified. TW providers who were able to earn the endorsements of some 'local' providers and adapt their activities to fit in with them were more likely to earn local status which facilitated collaboration. These findings support the claims made by Bloyce and Murphy (2007) that 'established and outsiders' might be most helpfully used to understand degrees of establishment in a community and suggest that a provider's status in a community is in a state of fluxshifting in response to new funding arrangements. A figurational view of power relations in constant flux seems, therefore, key to an understanding of ABI partnerships. --- Conclusions It is hoped that this study has provided a more adequate account of partnership development in ABIs than has hitherto been developed. Concepts from figurational sociology were used to inform the development of a framework that focused on the constraints on service provider and co-ordinator actions and helped to draw analytic attention to the ways in which shifting power dynamics over time shaped the way in which provider relations developed. A number of policy and practice implications can be drawn from this work. Although previous research has revealed the potentially negative impact of competition before, it remains the case that service co-ordinators are unable to control the unplanned outcomes that often emerge from competitive processes. Although ABI co-ordinators are relatively powerful, they are still heavily dependent on those who deliver projects. Complex interdependencies are likely to limit the ability of any one group to coordinate service delivery even in a relatively small geographical area. Joint working tended to be viewed as a managerial issue in this initiative, as shown by the complex monitoring arrangements that were set up. Less emphasis was placed on supporting social relations. Although monitoring processes are important, it was clear that the development of organisational targets were not conducive to collaboration between providers at different organisations. This raises questions about the ways in which joint working might be better nurtured. Co-ordinator definitions of success in this ABI were framed in terms of resident outcomes, rather than partnership development. Local status, which represented commitment to the area and legitimacy to some providers, facilitated the development of relations between particular providers. To some extent, the concept 'being local' reflects a power struggle between providers for resources, but, given the advantages associated with local status, it might be helpful to explore how this status could be nurtured to develop more supportive conditions for collaboration. Greater appreciation of the historically produced social networks within which providers are embedded provides a more adequate understanding of partnership working. However, these findings indicate that there is a need for more realistic expectations among policy makers about what can be achieved through short-term area-based partnerships.

Area-based initiatives (ABIs) have formed an important part of public policy towards more socio-economically deprived areas in many countries. Co-ordinating service provision within and across sectors has been a common feature of these initiatives. Despite sustained policy interest in ABIs, little empirical work has explored relations between ABI providers and partnership development within this context remains under-theorised. This paper addresses both of these gaps by exploring partnerships as a social and developmental process, drawing on concepts from figurational sociology to explain how provider relations develop within an ABI. Qualitative methods were used to explore, prospectively, the development of an ABI targeted at a town in the north west of England. A central finding was that, although effective delivery of ABIs is premised on a high level of coordination between service providers, the pattern of interdependencies between providers limits the frequency and effectiveness of cooperation. In particular, the interdependency of ABI providers with others in their organisation (what is termed here 'organisational pull') constrained the ways in which they worked with providers outside of their own organisations. 'Local' status, which could be earned over time, enabled some providers to exert greater control over the way in which provider relations developed during the course of the initiative. These findings demonstrate how historically constituted social networks, within which all providers are embedded, shape partnership development. The theoretical insight developed here suggests a need for more realistic expectations among policy makers about how and to what extent provider partnerships can be managed.

To combat the SARS-CoV-2 virus's transmission, global policies implemented restrictions that limited access to traditional funeral and mourning rituals. Families were unable to have contact with the deceased's body, coffins remained closed, and attendance at ceremonies was restricted (Aguiar et al., 2020). Therefore, several bereaved individuals were unable to participate in mourning rituals as desired, contributing to increased challenges in adjusting to their loss. Supporting this, a recent systematic review and meta-analysis indicated that throughout the COVID-19 pandemic, the overall prevalence of grief symptoms and disorders among the bereaved was 45.1% and 46.4%, respectively (Kustanti et al., 2023). Moreover, a study by Harrop et al. (2020) found that amidst the pandemic there were reports of higher levels of complicated grief symptoms compared to pre-pandemic levels. Mourning rituals have important therapeutic roles within the grief process. These include facilitating acceptance of the Susana Manquinho and Sara Albuquerque contributed equally and shared first authorship. 1 3 reality of loss, providing avenues for emotional expression and community support, maintaining a connection with the deceased, and deriving meaning from the loss (Albuquerque et al., 2021;Mitima-Verloop et al., 2021). Recent studies have shown that the inability to engage in these rituals due to pandemic restrictions has had a profound impact on bereavement experiences (Burrell & Selman, 2020;Chen, 2022). Pre-pandemic data suggest that dissatisfaction with funeral rituals and ceremonies is associated with a high risk for consequent depression, post-traumatic stress, and/or prolonged grief disorder (PGD; Mayland et al., 2020;Mitima-Verloop, et al., 2021). However, Burrell and Selman (2020) emphasize that the benefits of post-mortem rituals depend on the bereaved person's ability to adapt themselves to these rituals and express their goodbyes in a meaningful way, suggesting that the restrictions of these practices do not necessarily imply maladjustment to loss. In times of adversity such as the COVID-19 pandemic, flexibility in identifying alternative meaningful ways to honor the deceased becomes pivotal in the process of grief adjustment (Cardoso et al., 2020;Mergulh<unk>o, 2020). Recent research highlights the importance of coping and psychological flexibility, defined as the capacity to adjust emotional expression according to situational demands (Cohen & Katz, 2015), in adapting to the singular challenges posed by the COVID-19 pandemic in mental health (Hemi et al., 2023;Prudenzi et al., 2022). Also, psychological flexibility has been shown to contribute to reduced suffering, enhanced emotional regulation, and greater adaptability in dealing with the loss of a loved one (Bonanno & Burton, 2013;Cohen & Katz, 2015;Kashdan & Rottenberg, 2010). On the contrary, bereaved individuals with PGD display a less expressive and flexible repertoire than bereaved individuals experiencing adjusted grief (Bonanno & Burton, 2013). The psychological and behavioral flexibility to engage in alternative mourning rituals (AMR) and ceremonies has proven to be a protective factor against the severity of grief, anxiety, and loneliness and was associated with more post-traumatic growth (Cohen & Katz, 2015;Kashdan & Rottenberg, 2010;Knowles & O'Connor, 2015). Mourning rituals can be individual or collective (Mitima-Verloop et al., 2021). Individual rituals encompass creating something in order to express feelings (e.g., poems, drawings, paintings, books), visiting the grave of the deceased or the location of ashes, listening to music or watching a movie that evokes memories of the loved one, visiting a special place for the deceased, setting up an altar or designated spot, possibly with a photograph of the deceased, wearing something of the deceased, producing tattoos in tribute, and using objects of the deceased (Gibson, 2004;Mitima-Verloop, et al., 2021). The use of meaningful objects (e.g., photographs, clothing, jewelry, books) act as transitional objects with the deceased (Gibson, 2004;Wakenshaw, 2020). Transactional objects are symbolic material that provide security and connection to the bereaved. These practices can serve to facilitate the journey through the grieving process, allowing emotional connection and intimacy with the deceased (Goldstein et al., 2020;Wakenshaw, 2020). Examples of collective rituals are sharing stories with other people about the person who passed away, participating in a memorial service, paying tributes in public spaces, and using flowers, photos, and candles (Mergulh<unk>o, 2020). Often related to religious and cultural manifestations of mourning, the use of candles in the context of memorial ceremonies is a compensatory mechanism of caring for the loved one, facilitating reflection, serenity, expression of solidarity, grief, and compassion (Sedakova, 2015). Additionally, the internet and social media (e.g., Facebook, Instagram) provide new means of emotional expression and tribute to the deceased, thus constituting a collectively constructed mourning ritual (Christensen & Gotved, 2015). By creating specific online pages or platforms dedicated to the deceased with photos, videos, and other content, the grieving process is made public, favoring expressions of support and emotional connection, especially when physical contact is restricted (Brubaker et al., 2013;Meyer, 2016). Recent studies have emphasized the role of technology in facilitating collective mourning rituals during the pandemic, with virtual memorial services, online tributes, and social media platforms becoming important tools for grieving individuals to connect and share their grief experiences (Alexis-Martin, 2020; Costa et al., 2022;Sapalo, 2023). These activities enable the bereaved to overcome isolation and meet other bereaved people who have gone through a similar experience, with whom they can share memories and feelings, finding support and sometimes meaning in grief (Bailey et al., 2015). Research has demonstrated that AMR can alleviate the impact of physical contact restrictions by providing channels for grief expression, maintaining bonds with the deceased, and paying tribute to them (Cardoso et al., 2020). However, further investigation is required to fully understand the efficacy of these memorial practices in mitigating early symptoms of PGD, especially during the COVID-19 pandemic. Hence, the aim of the present study was to explore perceived limitations in funeral ceremonies and the moderating role of specific AMR (i.e., paying tribute on social media, lighting candles in memory of the deceased, using objects of the deceased) in the relationship between the perception of limitation in the funeral ceremonies (perceived limitations in performing the present body ceremony and in fulfilling the wishes expressed by the deceased concerning to funeral ceremonies) and the intensity of the early PGD symptoms. Based on the literature, the following hypotheses were established: a) higher levels of perceived limitation in the funeral ceremonies are associated with greater intensity of grief symptoms; b) the performance of AMR may attenuate the impact of the perceived limitation in the funeral ceremonies on the intensity of grief responses. --- Method --- Participants The sample was composed of 227 participants, aged between 18 and 77 years, of which females predominated. Regarding marital status, most participants were married or cohabiting. Regarding the characteristics associated with the death, the most prevalent degree of kinship for the deceased was that of the grandfather/grandmother. Most of the deceased had an oncological disease as a cause of death, and extreme proximity to the deceased is highlighted. The characteristics of the sample and the circumstances of bereavement are shown in Table 1. --- Procedure This study is an exploratory study with a cross-sectional quantitative design. Through the online link, participants completed the informed consent form, which presented the study objectives, who could participate, what would be asked for, the anonymity and confidentiality of data, the voluntary nature of participation, and fact that participating in the study may carry the risk of evoking painful memories and emotions. No financial rewards were provided to participants for their involvement in the study. After submitting the answers to the questionnaire, they were provided with the contact details of the psychological support line SNS 24 (Health Care Line) and of the grief unit in public hospitals across the country. The sampling method was non-probability snowball sampling, i.e., participants were asked to identify or invite other bereaved individuals, but as the study was publicized on social media, anyone who met the inclusion criteria was eligible. Sample collection for this study was initiated in October 2020 and completed in March 2022. The sample of the present study was composed of bereaved people who lost a close person throughout the pandemic period2. Those who agreed to participate and met the following inclusion criteria were included: 1) Being aged 18 years or older; 2) Loss of a person considered close during the pandemic period; 3) Death occurred during the pandemic period by COVID-19, after March 16, 2020. This study is part of a larger national project focused on the impact of bereavement on the SARS-CoV-2 pandemic. This project was approved by the Ethics Committee of the Administraç<unk>o Regional de Sa<unk>de Centro (no 55/2020) and Alentejo (5/CE/2021). --- Instruments --- Sociodemographic and situational questionnaire This questionnaire was designed within the scope of this research to collect information about the socio-demographic characteristics of the participants (e.g., sex of the bereaved, age, nationality, professional situation) and the characteristics/circumstances of the death (e.g., cause of death, proximity to the deceased, degree of kinship). --- Perception of limitation in funeral ceremonies Two items constructed within the scope of the study were used to assess the individual's perception of the limitations imposed by the restrictions on funeral ceremonies. One perception referred to performing a specific funeral ceremony and the other to the fulfillment of the deceased's wishes, i.e., the perception of limitation in performing the present body ceremony and the perception of limitation in fulfilling the wishes expressed by the deceased about the funeral ceremonies. The items were answered on a scale with the following response categories: none, little, some, a lot, extreme limitation, and not applicable. The 'does not apply' option of these variables was recoded to 'no limitation'. --- Alternative mourning rituals Participants were asked how satisfied they were with their use of any of the following examples of AMR: paying tribute on social media, lighting candles in memory of the deceased person, and using objects of the deceased person. The items were answered on a scale with the following response categories: none, low, moderate, high, extreme satisfaction, and not applicable. The "not applicable" option of these variables was recoded to "no satisfaction", and then two groups were created in each of these variables, i.e., 'did' or 'did not do' certain AMR. The content of the items pertaining to the perception of limitations in funeral ceremonies and alternative mourning rituals was developed based on a review of the relevant literature on mourning rituals and on the clinical expertise of the research team. The items were carefully crafted to encompass the diversity of these rituals, and feedback was retrieved from bereaved individuals regarding the clarity and relevance of the items. --- Prolonged grief disorder assessment instrumentreduced version (PG-4; Djelantik et al., 2017) To assess the intensity of the grief responses in the grief process, we selected the Reduced version of the Prolonged Grief Disorder Assessment Instrument (PG-4). This instrument focuses on the bereaved person's subjective perception of the frequency of symptoms associated with the mourning process (i.e., intense manifestations of longing and absence, persistent bitterness and resentment towards the loss and life in general, perception of life as empty or meaningless, numbness, or shock). Djelantik et al. (2017) statistically analysed the 13 items of the PG-13 (created by Prigerson et al., 2009) and highlighted 4 items that statistically identify risk indicators of prolonged grief (e.g., "I miss and miss the person I lost", "My life is empty or meaningless without the person I lost") answered on a scale of 1 (never) to 5 (always). The reduced version of this instrument (i.e., PG-4) was used due to the time post-mortem may include less than 6 months. The overall score (i.e., the sum of the items) was used. In the present sample, Cronbach's alpha was 0.86. --- Statistical analysis Firstly it was assessed the presence of missing values, followed by a descriptive analysis of the sample characteristics and of the responses using measures of central tendency (mean), dispersion (standard deviation) and distribution (skewness and kurtosis). The analyses were conducted in SPSS. First, correlations were performed to the main variables of the study to explore the relationship between the variables (Table 2). The magnitude of the correlations was analyzed using Pearson's coefficient, using Cohen (1988) criteria: weak (r <unk> 0.30), moderate (r between 0.30 and 0.49) and strong (r > 0.50). Second, to examine the moderating effect of the achievement of AMR (Fig. 2) six moderation models were tested using the macro for SPSS-PROCESS (Hayes, 2018). Perceptions of limiting restrictions (perceived limitation in performing present body ceremony; perceived limitation in fulfilling the wishes expressed by the deceased person regarding funeral ceremonies) were entered as independent variables, and the intensity of grief responses as dependent variable. The moderator variables corresponded to the AMR (paying tribute on social media, lighting candles in memory of the deceased person and using objects of the deceased person). Thus, for each independent variable, 3 moderations were tested. The effects were analyzed using the bootstrapping procedure. Confidence intervals (CI) are considered significant if they do not include zero. A significance level of 0.05 was considered. --- Results Table 2 shows the results of the correlations with the main variables of the study. The results indicated a positive and significantly strong correlation (Cohen, 1988) between the perceived limitation in performing the present body ceremony and the perceived limitation in implementing the wishes expressed by the deceased person in relation to the funeral ceremonies. The perception of limitations in performing the funeral ceremony was positively and significantly weakly correlated with the intensity of grief responses. The perception of limitation in implementing the wishes expressed by the deceased person regarding the funeral ceremonies proved to be positively and significantly weakly correlated with the intensity of grief responses. As regards the moderation analyses, three moderation models were tested for the variable perceived limitation in performing the ceremony (Table 3). Model 1 considered as moderator whether or not to paying a tribute in social media in the association between the perception of limitation in performing the present body ceremony (VI) and the intensity of grief responses. Results suggested that the model proved to be non-significant, explaining 4.27% of the variance of the bereavement response intensity results (F(3, 223) = 3.32, p = 0.02, R 2 = 0.04). There was no significant interaction effect, pointing to the lack of a moderating role of the variable performing tribute in social media in the relationship between VI and intensity of grief responses (VD). However, VI was found to have a statistically significant effect on the SD. Model 2 considered as moderator whether or not candles were lit in memory of the deceased person, in the association between the perceived limitation in performing a present body ceremony (VI) and the intensity of grief responses. The results suggested that the model proved to be non-significant, explaining 3.3% of the variance in the results of the intensity of grief responses (F(3, 223) = 2.54, p =. 06, R 2 = 0.03). There was no significant interaction effect, pointing to the lack of a moderating role of the variable lighting candles in memory of the deceased person in the relationship between VI and the intensity of grief responses. Model 3 considered as moderator the use or nonuse of objects of the deceased person in the association between the perceived limitation in performing the present body ceremony (VI) and the intensity of grief responses. The results suggested that the model proved to be significant and explained 8.83% of the variance of the results (F(3,223) = 7.20, p = 0.00, R 2 =. 09). The existence of a negative and significant interaction was observed between VI and using objects of the deceased person (Moderator variable). The existence of the moderating effect of using objects of the deceased person, suggests that the effect of VI on the intensity of grief responses is significantly different depending on whether or not the bereaved used objects of the deceased person. More specifically, for those who used the deceased's objects, the relation between the perceived limitation in performing the bodily ceremony was not significantly related with the intensity of the bereavement responses. On the other hand, for those who did not use this alternative form, higher levels of perceived limitation were associated with greater intensity of grief responses (Fig. 1). Model 4 considered as moderator whether or not paying a tribute in social media, in the association between the perception of limitation in fulfilling the wishes expressed by their deceased in relation to funeral ceremonies (VI) and the intensity of grief responses. The results suggested that the model proved to be significant and explains 9.61% of the variance of the results (F(3,223) = 7.90, p = 0.00, R 2 = 0.10). From the analysis of Table 4, it was observed the existence of a negative and significant interaction between (VI) and performing tribute on social media (moderator variable). The existence of the moderating effect of paying tribute in social media suggests that the effect of VI on the intensity of grief responses was significantly different depending on whether or not the bereaved performed tribute in social media, this relationship being graphically presented through Fig. 2. More specifically, for those who performed the alternative form of tribute on social media, the relationship between the perceived limitation in fulfilling the wishes expressed by their deceased regarding funeral ceremonies was not significantly related with the intensity of grief responses. In turn, for those who did not perform the funeral ceremonies, higher levels of perceived limitation were associated with more intensity in the grief response (Fig. 2). Model 5 considered as moderator whether or not lighting candles in memory of the deceased person in the association between the perceived limitation in fulfilling the wishes expressed by the deceased in relation to funeral ceremonies (VI) and the intensity of grief responses. The results suggested that the model proved to be significant and explained 8.82% of the variance of the results (F(3,223) = 7.19, p = 0.00, R 2 = 0.09). From the analysis of Table 4, the existence of a negative and significant interaction between VI and lighting candles in memory of the deceased person (moderator variable) was observed. The existence of the moderating effect of lighting candles in memory of the deceased person suggests that the effect of VI on the intensity of grief responses was significantly different depending on whether or not the bereaved lit candles in memory of the deceased person, and this relationship is presented graphically through Fig. 3. The results show that for those who lit candles in memory of the deceased, the relationship between the perceived limitation in fulfilling the wishes expressed by the deceased in relation to the funeral ceremonies was not significantly related to the intensity of grief responses. On the other hand, for those who did not perform the funeral ceremonies, higher levels of perceived limitation were associated with more intensity of grief responses (Fig. 3). Model 6 considered as moderator the use or non-use of the deceased person's objects in the association between the perceived limitation in implementing the wishes expressed by the deceased in relation to funeral ceremonies (VI) and the intensity of grief responses. The results suggested that the model proved to be significant, explaining 10.48% of the variance in the results of the intensity of grief responses (F(3, 223) = 8.71, p = 0.00, R 2 = 0.10). However, there was no significant interaction effect, pointing to the lack of a moderating role of the variable using objects of the deceased person in the relationship between VI and the intensity of grief responses. --- Discussion In this study, we aimed to study the impact of perceived limitations in funeral ceremonies during the pandemic and the moderating effect of AMR. Concerning the first objective, results confirm the hypothesis that higher levels of perceived limitations in funeral ceremonies were associated with a higher intensity of early PGD symptoms. This aligns with existing research that underscores the pivotal role of funeral ceremonies as cultural and psychological anchors, as organizers of the farewell process and important for grief elaboration (Albuquerque et al., 2021;Mitima-Verloop et al., 2021). Also, this finding highlights the need for a more comprehensive understanding of how cultural and social factors affect the grieving process. Moreover, our findings shed light on the connection between the inability to conduct present body ceremonies and the perception of limitations in fulfilling the deceased's wishes concerning funeral ceremonies. Present body ceremonies are often deeply emotionally charged events where mourners have the opportunity to physically see and be near the deceased's body. These ceremonies may provide a tangible connection between the living and the deceased, and when such ceremonies are not possible, individuals may feel that they are falling short of providing the level of respect and honor that they believe the deceased desired. Also, many cultural and religious traditions emphasize the importance of specific funeral rituals (e.g., Arslan & Bulduko<unk>lu, 2023;Burrell & Selman, 2020), including present body ceremonies, as a way to show respect for the deceased. When these rituals cannot be carried out, individuals may fear that they are failing to adhere to these cultural or religious expectations, thereby not fulfilling the deceased's wishes. The restriction of such rituals can prompt heightened distress, emphasizing the need to consider these emotional nuances in bereavement support. Regarding the second objective, data from the present study supported the hypothesis that the performance of Alternative Mourning Rituals (AMR) lessened the strength of the relationship between perceived limitations in funeral ceremonies and the intensity of early Prolonged Grief Disorder (PGD) symptoms. This finding can be explained by the therapeutic and coping mechanisms that such alternative rituals offer. These rituals allow mourners to express their feelings, share memories, and find validation, understanding, and companionship in a supportive environment with friends and family, fostering a sense of community and shared experience and reducing isolation (Burrell & Selman, 2020;Cardoso et al., 2020;Gabriel et al., 2021). Also, AMR may entail personalized rituals that reflect the unique relationship the bereaved had with the deceased. This may foster the creation of new meanings and narratives surrounding the loss and reduce the sense of unfulfilled expectations derived from the restrictions in funeral ceremonies (Mergulh<unk>o, 2020;Meyer, 2016). Engaging in AMR demonstrates adaptability and resilience (Worden, 2018), as it implies the ability to navigate challenging circumstances and find creative solutions to honor the deceased (Cardoso et al., 2020). In consonance with contemporary research (Cardoso et al., 2020;Mergulh<unk>o, 2020;Meyer, 2016;Worden, 2018), our study underscores the remarkable capacity of individuals to adapt and find alternative, meaningful ways to accomplish their needs of tributing the deceased and sharing their grief during the restrictions imposed by the pandemic. Alternative rituals highlight the flexibility and creativity of the human capacity to mourn and remember loved ones, even when Fig. 2 Effect of the tribute in the social media on the relationship between the perception of limitation in fulfilling the wishes expressed by the deceased concerning the funeral ceremonies and the intensity of grief responses Fig. 3 Effect of lighting candles in memory of the deceased person on the relationship between the perception of limitation in fulfilling the wish expressed by the deceased regarding funeral ceremonies and the intensity of grief responses traditional avenues are restricted or unavailable (Cohen & Katz, 2015;Prudenzi et al., 2022). However, results show that distinct mourning rituals (i.e., paying tribute on social media, lighting candles, and using the deceased's objects) have varying impacts on moderating the effect of restrictions on grief symptoms. This variation in impact can be explained by their unique characteristics and the emotional needs they fulfill. The symbolic and collective mourning rituals of using social media and lighting candles in memory of the deceased emerged as an influential moderator, buffering the relationship between perceived limitations in fulfilling the deceased's wishes and the intensity of grief responses. This means that these behaviors may play a restorative function in the face of the impossibility of fulfilling the deceased's wishes (Meyer, 2016). Social media and candle lighting rituals may serve as therapeutic outlets, allowing mourners to connect with others and reduce the emotional distress (e.g., Christensen & Gotved, 2015) possibly caused by unmet expectations and restrictions on traditional funeral ceremonies. Social media platforms offer an expansive and accessible network of friends, family members, and acquaintances who can provide emotional support and understanding during the grieving process (e.g., Costa et al., 2022;Sapalo, 2023). When individuals perceive limitations in fulfilling the deceased's wishes, the support received from social media connections can mitigate feelings of isolation and emotional distress (Alexis-Martin, 2020;Costa et al., 2022;Sapalo, 2023). Furthermore, social media provides a platform for individuals to express their grief openly and publicly and to honor the deceased by carrying forward their legacy, sharing memories, life stories, and their impact on others (Silva & Silva, 2021). It can also facilitate the creation of virtual rituals and memorialization practices (Gilbert & Horsley, 2021). Moreover, the act of sharing tributes on social media platforms can be used to maintain the connection in the digital realm and to continue to care for the loved one (Johnson & Wijdicks, 2018). In short, when circumstances restraint fulfilling the deceased's wishes, social media provides an alternative channel for such expression, involving emotional sharing, continuing bonds, community building, and preserving the deceased's legacy. Additionally, the act of lighting candles can contribute to lessen the distress linked to unfulfilled wishes in multiple ways. Lighting candles holds deep cultural and symbolic significance in many mourning traditions and is often accompanied by personal reflection, prayer, or meditation, which can provide comfort for the bereaved (Sedakova, 2015). Furthermore, the act of lighting a candle engages the senses, such as watching the flame, which can be soothing and grounding (Darbonne, 2023). It provides a tangible and sensory experience that allows individuals to connect with their emotions and memories of the deceased. It may also offer a dedicated time and space for mourners to connect and remember their loved one and share their thoughts and emotions. Additionally, candle lighting can be a communal activity, bringing together friends and family members to participate in the ritual (Sedakova, 2015), fostering a sense of togetherness and support. This shared experience can provide a supportive community that helps individuals cope with the emotional distress arising from unmet expectations and restrictions on traditional funeral ceremonies. On the other hand, the use of the deceased's objects surfaces as a distinct and important coping mechanism but with a different moderating effect. This individual mourning practice seems to be important in mitigating the limitation stemming from the absence of present body ceremonies. Deceased's objects hold personal significance and are tangible/concrete reminders of the person who has passed away (Wakenshaw, 2020). When individuals interact with these objects, they can physically touch and hold something that was part of their loved one's life. This concrete connection can help fill the void left by the absence present body ceremonies, establishing a tangible link to the loved one. Also, the use of objects that belonged to the deceased can instill a sense that the loved one continues to be an integral part of the daily life of the bereaved, offering a source of emotional and psychological comfort (Wakenshaw, 2020). Likewise, the deceased may have articulated specific preferences regarding the utilization or preservation of particular objects. Using these objects in accordance with the deceased's wishes, may provide the bereaved with a sense that they are honoring the person's choices, perpetuating their legacy, and paying tribute to their life. In summary, the use of the deceased's possessions can function as both a physical and emotional connection, aiding individuals in maintaining their bond with their loved one, fulfilling their wishes, and finding comfort in the absence of traditional ceremonies (e.g., Goldstein et al., 2020;Wakenshaw, 2020). --- Clinical implications The clinical implications of our findings are multifaceted, resonating with contemporary perspectives on grief and resilience (Kashdan & Rottenberg, 2010;Walsh, 2020). It underscores the need for clinicians to acknowledge the diverse ways in which individuals grieve and recognize the therapeutic potential inherent in mourning rituals (Kashdan & Rottenberg, 2010;Walsh, 2020). Specifically, findings from this study contribute to identifying the role of Alternative Mourning Rituals (AMR) in order to promote the bereaved's flexibility and mitigate the deleterious effect of ceremonies and rituals restrictions on bereavement outcomes. Considering that people were affected differently by restrictive measures amid the COVID-19 pandemic, clinicians should take into consideration the specific therapeutic functions of mourning rituals that better meet the bereaved's particular needs. For example, for those grappling with guilt tied to unfulfilled wishes of the deceased, interventions that encourage and facilitate public and symbolic forms of tribute, such as participation in social media memorials, offer solace and a sense of connection. These strategies provide a space for individuals to express their love and devotion to the departed while finding comfort in shared grief experiences. When confronted with the passing of a loved one, people may find it hard to accept the finality of the event. Promoting rituals that incorporate concrete representations of the deceased (e.g., such as present body ceremonies, visiting the grave) can provide a bridge between the abstract concept of death and the tangible world and help individuals understand the finality of death. --- Limitations and suggestions for future studies Despite the strengths of this study, it also has some limitations. One of them is the fact that the sample is not representative of the Portuguese sample (given that snowball sampling was used), namely, in terms of distribution between urban and rural areas. The sample was collected by convenience, and most of the participants were relatives of patients followed in palliative care at a central hospital, in an urban area. In addition, as is common in samples of bereaved people, the female gender is overrepresented in our sample, reflecting differences in coping styles with grief (Walsh, 2020). Thus, we questioned whether, in a rural and male sample, the function of mourning practices would be identical. Another limitation concerns the fact that we have not addressed the influence of religiosity on mourning practices. Taking into account the importance of this variable, it would be relevant to verify to what extent aspects such as faith professed by the bereaved or whether the person is practicing or not interfere with the moderating effect of AMR. Likewise, relational aspects such as previous relationships with the deceased and the quality of social support are also expected to interfere in this relationship between perceived restrictions and AMR. Also, it would be interesting to verify, through a qualitative study, the meanings attributed by the bereaved themselves to these tribute practices. Finally, future studies should focus on characterizing mourning rituals, namely in terms of frequencies and associated factors (e.g., time post-loss). --- Data availability The dataset generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. --- Declarations Competing interests The authors report there are no competing interests to declare. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Amid the COVID-19 pandemic, measures were enforced that constrained people's ability to engage in conventional funeral and mourning rituals, which in general serve as vital mechanisms for coping with the experience of loss. This study aimed to investigate how these limitations affected early grief symptoms and the influence of alternative mourning rituals (paying tribute on social media, lighting candles in memory of the deceased, and using objects of the deceased) on moderating the relationship between individuals' perception of these limitations (such as being unable to perform the present body ceremony or fulfill the funeral wishes of the deceased) and the intensity of the grief symptoms. The study involved 227 participants, aged between 18 and 77 years, who had experienced the loss of a loved one during the pandemic. Results demonstrated that higher levels of perceived limitation in funeral ceremonies are associated with greater intensity of grief responses. Paying tribute on social media and lighting candles in memory of the deceased person moderated the relationship between the perception of limitation in fulfilling the deceased's wishes in relation to the funeral ceremonies and the intensity of the grief responses. Using objects moderates the relationship between the perception of limitation in performing a present body ceremony and the intensity of the grief responses. Results underline the relevance of psychological flexibility, in particular concerning alternative morning rituals, that can be used as a way of buffering the impact of the perception of limitation in funeral ceremonies on the intensity of grief responses. This study offers a distinctive insight into bereavement during the pandemic, highlighting the role of flexibility in morning rituals in mitigating the deleterious effect of ritual restrictions on bereavement outcomes.

Introduction Social networks and digital media have become a major part of our daily lives, exerting an increasingly significant impact on individuals in general and, more specifically, on adolescents (Smith et al., 2016). Digitally driven communication fosters positive interactions and has multiple benefits, but it can also bring about new challenges (Englander & McCoy, 2017). The emergence of online communication has been linked to the global rise in messages with sexual content (Sweeny & Slack, 2017). Today, adolescents can explore their sexuality in new ways, redefining and normalizing more recent types of intimate relationships such as sexting-the sharing of self-produced sexual material through electronic means (Barrense-Dias et al., 2017;Schubert, 2014). Concerns about this phenomenon have grown, and it has attracted considerable attention from researchers, families, teachers, schools, and the media (Anastassiou, 2017;Gewirtz-Meydan et al., 2018;Van Ouytsel et al., 2015). This has occurred as a consequence of its potential negative impact and the effects it can have on adolescent well-being, where sexual content is disseminated without consent or where teens feel peer or partner pressure to engage in sexting behaviors (Klettke et al., 2014;Olivari & Confalonieri, 2017;Schubert, 2014). Even teachers agree that sexting could cause classroom disruption (O'Bannon & Thomas, 2014). The sexting phenomenon may play an influential role in the process of building new relationships as well as in the development of adolescents' sexual behaviors (Ringrose et al., 2012). Therefore, identifying the institutions that need to get involved and the types of actions that need to be taken is key to ensuring effective prevention and intervention in these areas. Thus, this systematic review analyzes the scientific evidence that helps identify the lines of action to address sexting. --- Definition and Prevalence of Sexting There is a lack of general consensus around the definition of sexting. Definitions vary depending on the type of behavior included, alluding to active sexting (such as sending or third-party forwarding) and passive sexting (receiving a message directly from the creator or via an intermediary). They also vary according to the content of the messages (sexual pictures, videos, or text) and the degree of sexual explicitness (suggestive or explicit) (Barrense-Dias et al., 2017). Therefore, the most restrictive definitions exclusively refer to sending sexually explicit pictures (Choi et al., 2016;Marume et al., 2018), whereas the most comprehensive definitions also cover other types of behavior, for example, content dissemination (Mitchell et al., 2012;Villacampa, 2017). In this context, sexting encompasses the sending, receiving, and forwarding of suggestive and explicit sexual pictures, videos, or text messages via cell phones, the Internet, or other electronic means (Mitchell et al., 2012). Because sexting is an evolving concept that has become increasingly complicated (Van Ouytsel et al., 2018), each study and the sexting behavior must be analyzed, as the action to be taken may vary depending on the behavior displayed. The estimated number of adolescents engaging in sexting is consistently definition dependent (Barrense-Dias et al., 2017). In a recent meta-analysis (Madigan et al., 2018), the average prevalence of sending sexual content was 14.8%; receiving sexts was 27.4%; forwarding a sext without consent was 12.0%; and receiving a forwarded sext was 8.4%. In recent years, sexting rates among youth have seen a rise with increasing age, and no significant gender differences in the rate of sending or receiving sexts have been observed (Madigan et al., 2018). --- The Need to Address Sexting The practice of sexting is characterized by its psychological, social, and behavioral consequences (Klettke et al., 2014), which can lead to ethical and socio-moral conflicts alongside other concerns about the privacy and protection of personal content (Schubert, 2014). Results of a recent meta-analysis suggest that the exchange of sexual messages, photographs, and videos through technological devices is associated with sexual behavior (sexual activity, multiple sexual partners, lack of contraception use, etc.) and mental health issues (delinquent behavior, anxiety/ depression, alcoholism, drug consumption, smoking, etc.), especially in younger adolescents (Mori et al., 2019). The consequences of sexting may affect the physical and psychological health of those involved, and adolescents may end up experiencing peer pressure and emotional difficulties (Olivari & Confalonieri, 2017;Van Ouytsel et al., 2015). However, most studies indicate that these relationships are cross-sectional, and the fact that sexting presents as a problematic behavior would seem to depend on the situation, such as the context of the relationship, and whether it is consensual or coercive (Temple et al., 2019). Although the bulk of the research focuses solely on consented parties (sending and/or receiving this type of content), the most detrimental action, and therefore the most important when trying to understand the consequences behind this phenomenon, would be the forwarding of sexual content by third parties (Livingstone & Görzig, 2014;Strassberg et al., 2017). A possible explanation is that sexual content can be spread quickly without consent, reaching undesired recipients, thus increasing its audience and affecting the victim's reputation (Van Ouytsel et al., 2014a, 2014b). Consequently, most efforts should be directed in this area to prevent and effectively intervene in sexting behaviors. As a result of this dissemination, sexting is also associated with other potential risks which can aggravate its possible consequences, such as blackmail, extortion, bullying, and cyberbullying (Döring, 2014;Kopeck<unk>, 2015;Medrano et al., 2018;Montiel et al., 2016;Strassberg et al., 2013;West et al., 2014;Woodward et al., 2017). The gender dynamics that arise from this phenomenon are also noteworthy. Differences in the roles of sexting are observed, and the different practices do not seem to affect boys and girls in the same way. Boys are perceived as those who ask for photographs, whereas girls are seen as those responsible for setting the boundaries (Symons et al., 2018). Moreover, girls usually experience a damaged reputation and tend to suffer the consequences of sexting more than their male peers, the latter even experiencing positive effects which can boost their popularity (Cooper et al., 2016;Dobson & Ringrose, 2016;Symons et al., 2018;Wood et al., 2015). In all cases, a robust response to any sexting-related behavior (sending, receiving, and third-party forwarding) is highly recommended. It is clear that the non-consensual forwarding of content to third parties is a type of behavior that must be avoided, and prevention strategies must be taught (Van Ouytsel, et al., 2014a, 2014b). However, young people also need to know how to act when this type of content is received (Mitchell et al., 2012); how to send sexual content safely, if this is indeed the intention; and how to fend off any attempts at peer pressure (Wurtele & Miller-Perrin, 2014). Thus, this phenomenon is seen as a challenge for educational institutions and teaching professionals (McEachern et al., 2012), given that incidents brought about by sexting can have a negative impact at school (Van Ouytsel et al., 2014a, 2014b, 2015). Although prevalence is higher outside of school walls, sexting also occurs during school hours (Schubert, 2014). What is more, sexting outside of this environment may also have an impact within the school context (McEachern et al., 2012). Because this phenomenon has the power to influence the interpersonal relationship building process, not to mention adolescent sexual development in general (Ringrose et al., 2012), the classroom is an ideal environment for educating individuals on how best to use Information and Communications Technology (ICT) and, more specifically, prevent any negative consequences of sexting (Mura et al., 2014;Theodore, 2011). Educating young people about sexting in schools has the potential to reach far more adolescents, as many young people do not learn about sexuality or this new reality at home. Furthermore, teaching professionals can train and inform parents, fostering positive relationships and building a strong school community that guarantees the continuity of education received by students in both key settings: home and school (Van Ouytsel et al., 2014b). Specifically, teaching professionals can play a crucial role when it comes to addressing this phenomenon proactively and when taking specific preventive actions to address the consequences of sexting (Bhat, 2018;Kopeck<unk>, 2012). --- Research Question Despite the importance of preventing the potential negative consequences of sexting (Van Ouytsel, et al., 2014a, 2014b), information on how to do this effectively remains scarce. As such, there is an obvious need to develop strategies based on scientific research findings (Livingstone & Smith, 2014), identifying areas and lines of action that can help researchers and educators create and evaluate programs to successfully address sexting. In an effort to bridge this gap, this study aims to systematically describe the available scientific evidence outlining the effective lines of action to tackle sexting. To this end, the present review is guided by the following question: What are the types and frequencies of proposed lines of action for sexting? --- Method --- Inclusion and Exclusion Criteria This systematic review includes all articles published up until 2018 that meet a predetermined set of inclusion and exclusion criteria. The specific criteria to determine suitable studies for inclusion in the review were: a Studies with a target or participant population of children or adolescents aged up to 19 years. b. Studies that consider sexting as their phenomenon of interest. c. Studies that include information about sexting education, prevention, and/or intervention. The specific criteria to determine studies that should be excluded from the review were: a. Studies with an adult target or participant population. b. Studies whereby the topic of interest is a sexual phenomenon other than sexting, such as grooming, sexual abuse, or pornography. c. Duplicate articles. d. Articles with insufficient information because the full text is not available. --- Search Strategy The search strategy used for this systematic literature review was based on the PRISMA statement (Urr<unk>tia & Bonfill, 2010)., child, minor, adolescent, teen, youth, student, prevention, education, and intervention. The following search query was entered for the title, abstract, and key words: "Sexting AND (child* OR minor OR adolesc* OR teen* OR youth* OR student*) AND (prevention OR education OR intervention)." Before selecting this strategy, other attempts were made, such as "sexting AND adolesc* AND (prevention OR intervention)." The results were examined to find the right balance between sensitivity and specificity. Database searches were conducted up to September 2018. --- Data Coding and Analysis Article coding was carried out in two phases using an analysis sheet. During the first phase-the abstract screening phasea check was run to ensure that the article abstracts met the inclusion criteria and not the exclusion criteria. To do so, the following data were collected: database, year of publication, authorship, journal/publication, article title, inclusion criteria (a) and (b), and exclusion criteria (a), (b), and (c). The second phase-full-text eligibility-was carried out on articles that only met the previous inclusion criteria. Checks were run to verify that they also met inclusion criterion (c) and that they did not meet exclusion criterion (d). In this case, the whole publication was analyzed: type of study, area of study (journal/editorial field of study), language, country (country of the participating sample or, if not, the first author's home institution), objective(s), methodology, recipients, sample, definition of sexting, instrument, instrument characteristics, evidence of action, type of action, evaluation of action, area of action, and results. Only those articles clearly stating their own definition of sexting were taken into account. Review articles describing the definitions adopted in other articles without taking a personal stance were not considered. In the case of review articles, only recommendations pertaining to the article were considered, and proposals collected from the analyzed articles were discarded. Coding was undertaken by the principal investigator. Subsequently, 57.2% of the articles were selected at random and codified equally and independently by a second assessor, easily exceeding the 20% recommended minimum (Garc<unk>a-Moya et al., 2018). During this phase, the sexting definitions found in the articles under study were also categorized, leading to the following category system: specific behaviors by definition (distribution/exchange; sending; sending and receiving; sending, receiving, and forwarding), the type of content (text message; pictures; text messages or pictures; pictures or videos; text messages, pictures or videos; not specified), and the degree of sexual explicitness (explicit; suggestive or explicit; not specified). The reliability level was high. The percentage of agreement in the abstract screening phase was 91.3%, reaching 92.1% in the full-text eligibility phase. Disagreements were discussed and resolved in a consensual manner. The articles selected were downloaded using the Mendeley 1.19.3 software program. Figure 1 shows a summary of the selection process. A total of 456 articles were identified for the systematic review. The abstract was taken into account across all reviews, and 308 articles were excluded as they met exclusion criteria (a), (b), or (c). The full text of the remaining articles was analyzed, and 57 articles were excluded as they did not meet inclusion criteria (c) or they met exclusion criteria (d). Therefore, 91 articles were included in this systematic review. An overview of the general characteristics corresponding to these studies is provided in Appendix. Once the included articles had been determined, the areas and lines of action found in each article were categorized. The category system resulting from categorizing the areas of action is shown in Table 1. Similarly, the category system resulting from categorizing the lines of action is shown in Table 2. The sum of the areas, lines of action, and relevant information found in the articles were also categorized by an external reviewer. The reliability level was high (percentage of agreement at 84.2%). Disagreements were discussed and resolved in a consensual manner. --- Results --- Overview of the General Characteristics of the Studies The general characteristics analyzed in the articles were: year of publication, geographical area of origin, subject area, definition of sexting, type of action recommended, and area of action where intervention is deemed necessary. Regarding year of publication, the reviewed articles were published between 2009 and 2018, as no articles pre-2009 were found on any of the 21 databases. Notably, an increase in publications was observed in 2014; however, wide frequency variability is found and a clear pattern cannot be determined. In terms of geographical area of origin, 52.2% of studies were conducted in North America, 27.8% in Europe, 10% in Oceania, 4.4% across several continents, 3.3% in South America, 1.1% in Africa, and 1.1% in Asia. Regarding subject area, 39.6% of studies were conducted in the field of health, 26.4% in psychology, 20.9% in other social science disciplines, 11% in education, 8.8% in sociology and political sciences, 4.4% in the field of communication, and 3.3% in the discipline of law. The definition of sexting varies depending on the specific behavior at play, the type of content, and the degree of sexual explicitness. In terms of the different behaviors, 39.2% of articles only refer to sending, 23% generally define sexting as a sharing or exchange process, 17.6% refer to sending and receiving, and 16.2% identify the three behavior types: sending, receiving, and forwarding. The type of content behind these messages also varies. Specifically, 29.7% of articles mention text messages, pictures, and videos; 29.7% refer only to pictures; 28.4% refer to text or picture messages; 6.8% refer to sexual content in general but do not specify the content; 2.2% refer to pictures or videos; and 2.2% refer only to text messages. Lastly, 48.7% include suggestive and explicit content; 43.2% include only explicit content; and 8.1% do not specify the type of sexual content behind the definition used (Table 3). In terms of the type of action recommended, 7.7% of articles recommend taking actions to address and prevent the potential consequences of sexting, but do not include any interventions or suggest any strategies to make this happen; 85.7% of articles do propose specific strategies that can be effective when addressing this phenomenon, but do not include any interventions; and 6.6% present specific interventions to tackle sexting. Regarding the area of action where intervention is required, 86.8% of articles indicate that intervention should be school-led. Next, 20.9% state that the family should intervene, whereas the same percentage of articles (20.9%) report that family intervention must also involve the school. Similarly, 19.8% state that action should be taken from a healthcare perspective; 13.2% from a political perspective; and 7.7% from across all areas in general. A total of 5.5% of articles state that family-led involvement should be health based, whereas the same percentage (5.5%) report how legal advice is required to lend a legal perspective on this phenomenon. Moreover, 4.4% of articles call for the participation of law enforcement agencies, and 2.2% note that collaboration from technology experts is also necessary (Table 1). --- What are the Types and Frequencies of Proposed Lines of Action for Sexting? The collected data have been divided into three different types of information: lines of action, initiatives developed, and main recommendations. Fifteen lines of action emerged after categorizing the strategies identified in the research papers as key aspects of tackling sexting ( In addition to these lines of action, only six articles set out specific initiatives to address sexting: the specific action under the 3rd Spanish Master Plan for the Coexistence and Improvement of School Safety; the Webrangers educational project; the action research project Image.me; a three-level strategic plan; the Sextorsion prevention course; and school assemblies about the risks of sexting. The specific action established under the 3rd Spanish Master Plan for the Coexistence and Improvement of School Safety by the Spanish Government's Ministry of Education, Culture, and Sport; the Ministry of the Interior; and the Ministry of Public Health and Social Affairs contains presentations about the safety issues and risks associated with the use of the Internet to be developed alongside high school students. A Spanish civil guard officer, who is a specialist in New Technology and Risks, delivered an hourlong presentation about the possible risks of Internet use, especially those derived from using social networking sites, such as cyberbullying, grooming, and sexting. After the presentation, students completed an ad hoc questionnaire about their social network involvement. Information was gathered about their usage time or external supervision; the presentation content they found most interesting; and their personal opinions about the role of the Spanish civil guard officer. The impact of the activity on sexting was not evaluated (Mart<unk>n et al., 2013). The Webrangers education project is delivered in partnership with Google Inc., Palack<unk> University Olomouc, and the NGO Google Education Group. It consists of a peer training program in which students interested in the topic are selected and given full-time training. This project covers risky Internet behaviors and the safe use of this tool to help prevent dangerous online conduct. Case studies are used to train students on the following core topics: cyberbullying, cyber-grooming, sexting, and skills for online interaction. After the training, students must create their own projects to raise awareness through Google Plus, Facebook, the project's Web site, and through workshops and activities for their colleagues and teachers. Information about their evaluation was not reported (Kopeck<unk> et al., 2015). Image.me is an action research project about sexting prevention. It uses social theater with young people as preventive medical care, focusing on peer education through media and digital literacy. This combination encourages critical thinking and promotes collaborative work between classmates. These activities are part of a wider research project about theater and scientific communication led by the Catholic University of Milan. Social theater was used as a form of social care and online risk prevention. Three art scenes tailored to the project's target audience were chosen in an attempt to envisage how to address sexting effectively. One of the results was the creation of a pet-puppet used to get to know young people at schools, clubs, and other informal contexts. Videos were also made to raise young people's awareness about the presence of sexting in their communities. Social theater was also used to communicate the research findings at the end of the project. An art scene was designed and used to discuss the information gathered, making it easier to understand and engage with. Information on its evaluation was not reported (Ferrari et al., 2016). The three-level strategic plan describes specific activities that schools can implement to tackle cyberbullying, sexting, and other risk behaviors on social media. The three-level model aims to meet the common needs of all students, the specific needs of some students, and other more specific and complex needs. Level 1 meets the general needs. It aims to provide a definition of the phenomenon and set out regulations planned for and by the education community, which also covers how to handle a conflictive sexting situation. In addition, as part of the school curriculum, the whole institution should be given training about this phenomenon and its effects in order to safely address any kind of online behavior. Assessment twice a year is also important. Level 2 offers strategies for students at risk of becoming a bully or a victim. Prevention groups are formed to work on specific skills according to the potential participants' needs. Finally, Level 3 is delivered to students who are already actively involved in sexting and have more complex needs. In this case, intervention should focus on their direct needs, such as individual advice geared toward specific abilities, meetings with family members, and disciplinary and/or legal action. In general, the activities need to stress the importance of educating those involved about the Internet and its dangers. Information about its evaluation was not reported (Davis & Schmidt, 2016). The Sextorsion prevention course is delivered in high schools to teach students how to practice safe sexting. The learning methodology combines traditional lectures which provide an overall description of the topic with active learning, and directly engages students in the learning process. The course content includes the conceptualization and characteristics of sexting and sextortion, its associated risks, empathy toward the victim, legal consequences, safe practices, and measures and protocols to deal with the phenomenon. In particular, active learning was carried out by simulating different sexting and sextortion scenarios, encouraging students to be more independent and building their ability to search for relevant information related to sexting and sextortion. Different resources such as videos, group discussions, cases analyses and simulations, and group reflections were used. Regarding assessment, students answered a pre-and postcourse questionnaire about their knowledge of the topic and how satisfied they were with the course. However, the impact of the activity was not reported (Palop et al., 2016). Finally, school assemblies were designed to educate high school pupils on the risks of sexting. The specific content of these assemblies was not provided. However, four years later, sexting patterns (except third-party forwarding) had not changed significantly. Most adolescents had exchanged sexually explicit pictures on their phones, and the common behavioral narratives remained very similar (Strassberg et al., 2017). The previously stated lines of action and initiatives were joined by other notable recommendations for tackling this phenomenon: Specifically, 19% of articles recommend staying away from scare tactics as a tool for intimidating young people; 11.9% recommend avoiding messages that promote the abstinence from and prohibition of sexting given their low level of effectiveness, adopting a more educational than authoritarian perspective; and 9.5% recommend not blaming and judging the victim or those who partake in this practice. Furthermore, 9.5% of articles highlight the need to start taking action early on in school and in preadolescence, and 7.1% recommend sexting assessment in schools to establish a baseline and to be able to promote strategies and actions based on the results obtained. Finally, 11.9% of articles recommend evaluating the impact of these strategies post-implementation to determine their effectiveness. Thus, the practices could be evidence based, promoting continuous improvement and adapting the strategies to the intervened context. --- Discussion Sexting has shaped itself into a new form of adolescent sexual exploration and expression (Schubert, 2014). However, it has also become a new challenge that professionals working with children need to understand in order to deal with it effectively (Bhat, 2018;Kopeck<unk>, 2012). This systematic review sought to gather information and describe the existing scientific evidence relative to the effective lines of action that address sexting, helping researchers and educators to design and evaluate sexting programs. There is scientific evidence to support the need for sexting intervention. Specifically, these efforts must focus on the different ways in which this phenomenon is experienced and expressed: sending, receiving, and forwarding. The most commonly used sexting definition in the analyzed articles had sending as the most studied behavior. However, this definition does not cover the phenomenon's complexity. Third-party forwarding of sexual content also plays a highly significant role in understanding the consequences of sexting (Livingstone & Görzig, 2014;Strassberg et al., 2017). Thus, it is important to include all three types of sexting behaviors 1 3 (sending, receiving, and third-party forwarding) in order to analyze each one on its own and to be able to identify the necessary actions for each behavior. Studies mainly focus on a particular geographical area (North America) and a specific subject area (Health), meaning that sexting research in other countries and in the educational field must also be encouraged. At a disciplinary level, the focus is primarily placed on the school setting (Livingstone & Smith, 2014;Van Ouytsel et al., 2014b). Furthermore, only six of the 91 articles feature a specific intervention and just one article evaluates the impact of this action. This may be due to the fact that literature and educational campaigns have mainly focused on analyzing sexting as a problem, promoting abstinence, and condemning the practice. Consensual sexting as an intimate means of communication in line with contemporary communication methods must be accepted (Döring, 2014;Strassberg et al., 2017). The development of specific programs that address sexting is the notable line of action. Undertaking both proactive and reactive activities is crucial (Albury et al., 2017). Some examples are: the use of case studies (e.g., Kopeck<unk>, 2015;Palop et al., 2016); discussions (e.g., Gregg et al., 2018;Siegle, 2010); educational campaigns, lectures, and workshops (e.g., Dobson & Ringrose, 2016;Hinduja & Patchin, 2012); the creation of information resources, a compilation of best practices (e.g., Döring, 2014;McEachern et al., 2012); real testimonies (Mart<unk>n et al., 2013;Van Ouytsel, et al., 2014a, 2014b); debates (e.g., Dobson & Ringrose, 2016;Van Ouytsel et al., 2015); and cross-curricular classroom projects (Laguado et al., 2018;Theodore, 2011). --- What are the Types and Frequencies of Proposed Lines of Action for Sexting? Fostering a safe and healthy use of ICT, the Internet, and social networks is also noteworthy. Because adolescents who use their cell phones as their main Internet connection and spend more time connected are most likely to receive sexting requests (Atwood et al., 2017), it is important to teach them how to use technology appropriately. This training should cover personal expectations about digital privacy (e.g., Albury et al., 2017;Soriano-Ayala & González-Jiménez, 2014); control over personal data on the Internet (e.g., Diliberto & Mattey, 2009;Patrick et al., 2015); safe online behaviors (e.g., Mura et al., 2014;O'Keeffe, 2016); and knowledge of rights and responsibilities when it comes to digital technology (e.g., Gámez-Guadix et al., 2017;Uhler & Smith, 2012). Efforts to incorporate sexting into sex education programs is also a fundamental part of handling this phenomenon, as sexting may be seen as a way to maintain intimate communication with a partner in a healthy relationship (Van Ouytsel, et al., 2014a, 2014b). On many occasions, sexting is used to show a romantic or sexual interest in another person; to build new emotional bonds; to delve deeper into the development of their sexual identity; and merely as another form of sexual activity in a long-distance relationship (Döring, 2014;Walker et al., 2011). Thus, addressing sexting as an integral component of sex education programs provides young people with information about the phenomenon and how to tackle it safely, instead of evading it or encouraging the negative views held by many adolescents about sexting. Given the correlation between sexting and traditional or digital risks, such as bullying and cyberbullying (e.g., Rodr<unk>guez-Castro et al., 2017;Woodward et al., 2017), introducing sexting into preventive programs that address other associated risksadopting an integrated approach-is also recommended (e.g., Dake et al., 2012;West et al., 2014). The promotion of sexual ethics is also linked to sex education, namely specific and key aspects that address sexting effectively. Its focus is on developing the necessary skills to build and maintain an intimate and ethical relationship (Walker et al., 2011), such as preventing coercion and pressure in a loving sexual relationship; fostering reflection on the importance of proper consent and real respect for a partner or intimate companion; and maintaining a critical attitude toward the exchange of non-consensual sexual content (e.g., Albury et al., 2017;Wurtele & Miller-Perrin, 2014). In terms of sexual ethics, gender roles and stereotypes must also be considered. Acknowledging the cultural norms and values that underpin social behavior is essential to successfully addressing the phenomena which play out in personal interactions. Thus, it is necessary to understand and question the heteronormative values associated with femininity and masculinity which form part of the digital culture and to determine the dynamics and roles played by individuals who engage in sexting (e.g., Karaian, 2014;Wood et al., 2015). It is particularly important to involve young people in analyzing the power imbalance between genders and the double sexual standard and to avoid the use of stereotypes and blaming women, in the strategies used to address sexting (e.g., Döring, 2014;Van Ouytsel et al., 2014a). Raising awareness about the impact and risks of sexting is also important. Sexting can lead to undesired consequences, which can turn into problematic scenarios alongside other risks such as bullying and cyberbullying (Frankel et al., 2018;Medrano et al., 2018;Ringrose et al., 2012). For this reason, young people need to be aware of the risks. However, we need to remember that, for some people, sexting is a romantic and enriching part of their relationship, although it can be dangerous (Ybarra & Mitchell, 2014). Furthermore, some studies suggest that many adolescents already have a clear picture of the consequences this phenomenon entails, and steps to ban and warn against sexting alone do not work to prevent the potential consequences (Lim et al., 2016). In this systematic review, only one evaluation of the proposed lines of action has been observed, which focuses on the lack of efficacy behind this strategy. It does, however, seem to reduce the extent to which this type of content is forwarded without consent, yet it does not effectively avert other possible negative consequences linked to sexting (Strassberg et al., 2017). Providing training to professionals who work with young people would also help them feel more capable of addressing sexting. It would equip them with the skills to react properly to difficult situations brought about by sexting. Thus, there is a need to stay up-to-date with the apps that adolescents use; discuss sexuality in a professional way (e.g., Van Ouytsel et al., 2014a, 2014b); have the resources to deal with this reality (e.g., Brown et al., 2009); know what sexting is and what drives adolescents to participate in it (e.g., Frankel et al., 2018); and be aware of the ensuing legal and moral obligations (e.g., Schubert & Wurf, 2014). From this perspective, sexting regulations and protocols should be developed in order to provide professionals with a common ground for dealing with this phenomenon. They must be given the necessary tools to act consistently and appropriately, making it easier for pupils to understand the differences between correct and incorrect sexting behavior (e.g., Krieger, 2017;Theodore, 2011). From this perspective and taking into account the other strategies, the adoption of disciplinary or legal measures is crucial in cases where this type of consent is shared non-consensually and where aggressive attitudes are exhibited (e.g., Davis & Schmidt, 2016;Russo & Arndt, 2010); for example, in cases where sexting is associated with bullying (e.g., Van Ouytsel et al., 2014b). Involvement by the entire education community and potential participants is a key aspect when it comes to fostering coherent and stable actions across the different settings that young people move within and between, thus joining efforts and strategies to tackle this phenomenon (e.g., Frankel et al., 2018). The peer group risk factors for sexting also play an important role among adolescents. During ad

Sexting has become a new form of intimate interaction in line with contemporary communication methods. This phenomenon often leads to positive outcomes, but it can also have negative repercussions depending on the situation, such as the context of the relationship, and whether it is consensual or coercive. Despite this, the main types of sexting behaviors (sending, receiving, and third-party forwarding) must be addressed in order to promote safe and healthy practices. However, the approach to tackling this phenomenon remains unclear. This systematic review sought to summarize the lines of action proposed or conducted in the scientific literature to address sexting, to help researchers and educators create and evaluate effective programs. A systematic search of 21 databases was conducted; only articles relating to sexting education, prevention, and intervention among child and adolescent populations were considered. In total, 456 articles were identified, 91 of which were included for the purposes of this research. The results highlighted a need to respond to the aforementioned sexting behaviors and to tackle the resulting conflict situations. Although interventions across different areas are recommended (e.g., health, family, policies, legal advice, law enforcement, technology experts, and even society as a whole), most studies agree that school is the most practical setting for intervention. Thus, the 15 lines of action identified in this systematic review must all be considered to effectively address sexting in childhood and adolescence.

order to provide professionals with a common ground for dealing with this phenomenon. They must be given the necessary tools to act consistently and appropriately, making it easier for pupils to understand the differences between correct and incorrect sexting behavior (e.g., Krieger, 2017;Theodore, 2011). From this perspective and taking into account the other strategies, the adoption of disciplinary or legal measures is crucial in cases where this type of consent is shared non-consensually and where aggressive attitudes are exhibited (e.g., Davis & Schmidt, 2016;Russo & Arndt, 2010); for example, in cases where sexting is associated with bullying (e.g., Van Ouytsel et al., 2014b). Involvement by the entire education community and potential participants is a key aspect when it comes to fostering coherent and stable actions across the different settings that young people move within and between, thus joining efforts and strategies to tackle this phenomenon (e.g., Frankel et al., 2018). The peer group risk factors for sexting also play an important role among adolescents. During adolescence, social status is particularly important (Chalfen, 2009;Ling, 2004), and sexting offers an opportunity to become more popular (Gewirtz-Meydan et al., 2018). It may be seen as a strategic move for adolescents to gain popularity among peers (Baumgartner et al., 2015). Thus, encouraging reflection on social pressure and the need for popularity, as well as being critical toward the content received via the Internet, is fundamental (e.g., Ahern & Mechling, 2013;Wolak et al., 2012). Similarly, the false beliefs that adolescents hold about sexting and the notion that all young people engage in this mainstream phenomenon should also be challenged. Adolescents believe that the messages circulating within their immediate surroundings and the media influence their predisposition to develop sexting attitudes, viewing this phenomenon as a normal practice (Smith et al., 2014). The perceptions and experiences of adolescents also make for a good starting point according to the literature, allowing us to understand the whys and hows of their behavior in order to respond to young people's actual needs (e.g., Livingstone & Görzig, 2014;Murray, 2014). Efforts to improve the school environment are also reported to have a positive effect on resolving difficult situations associated with sexting. Some examples of good coexistence practices include: maintaining high expectations for student performance; offering pupils the opportunity to participate and contribute in class, at school, and in the education community (e.g., West et al., 2014); implementing peer education/coaching (e.g., Ferrari et al., 2016;Siegle, 2010); and promoting student safety at school, for example, by adopting measures that encourage them to report worrying cases of sexting and other negative behaviors without fear of retaliation (e.g., Gregg et al., 2018;McEachern et al., 2012). Furthermore, when responding to the needs of young people, it is important to cater for diversity and to adopt specific measures that acknowledge, address, and integrate the particularities of vulnerable groups (e.g., Brown et al., 2009;Livingstone & Görzig, 2014). Finally, in addition to the discussed lines of action, recommendations are made to address this phenomenon early on in the school cycle, likely because the use of virtual networks increases gradually until the age of 13, when it comes into more general use (Garmendia et al., 2016). Furthermore, sexting is characterized by the developmental stage of adolescents' first romantic or sexual relationships (Fox & Warber, 2013;Van Ouytsel et al., 2016), which highlights the importance of educating young people and relying upon strategies suitable for minors at an early age. Fear tactics and abstinence should also be avoided, as they can make young people increasingly more interested in this practice, without giving them alternative approaches (Gómez & Ayala, 2014). In addition, they do not accurately represent the sexual reality of our contemporary society, preventing us from suggesting strategies to dissuade the negative consequences of sexting among young people (e.g., Döring, 2014). It is also advisable not to judge the victims and individuals who engage in this practice, but rather those who inflict harm and forward content without consent (e.g., Wood et al., 2015). Lastly, evaluation plays a significant role when addressing sexting in order to understand the reality within the corresponding context (e.g., Barrense-Dias et al., 2017;Davis & Schmidt, 2016) and to promote evidence-based practices. Drawing on measurable objectives and considering the definition used is also important. This would enable us to determine the impact and effectiveness of the strategies, so they could act as support mechanisms for professionals working with minors on a daily basis (e.g., Lim et al., 2016;Livingstone & Smith, 2014). This systematic review does present some limitations. Studies addressing this topic may not have been considered for the following reasons: Sexting was identified through another term; sexting was implicitly covered in intervention programs about other phenomena; or the full article could not be accessed. It is also possible that effective actions are still under development, currently at the "to be published" stage or on the lookout to be published. Future research could build on this review by including studies from other databases, and more comparative studies and further analyses into the nature and characteristics of sexting from an educational point of view would prove useful. There is also a need to evaluate the strategies and actions used to address sexting, with the aim to design and implement evidence-based initiatives that equip schools and teaching staff with effective tools to prevent and tackle the potential risks associated with this phenomenon. --- Appendix: General information on the articles included --- Declarations Conflict of interest The authors report no conflicts of interest. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Sexting has become a new form of intimate interaction in line with contemporary communication methods. This phenomenon often leads to positive outcomes, but it can also have negative repercussions depending on the situation, such as the context of the relationship, and whether it is consensual or coercive. Despite this, the main types of sexting behaviors (sending, receiving, and third-party forwarding) must be addressed in order to promote safe and healthy practices. However, the approach to tackling this phenomenon remains unclear. This systematic review sought to summarize the lines of action proposed or conducted in the scientific literature to address sexting, to help researchers and educators create and evaluate effective programs. A systematic search of 21 databases was conducted; only articles relating to sexting education, prevention, and intervention among child and adolescent populations were considered. In total, 456 articles were identified, 91 of which were included for the purposes of this research. The results highlighted a need to respond to the aforementioned sexting behaviors and to tackle the resulting conflict situations. Although interventions across different areas are recommended (e.g., health, family, policies, legal advice, law enforcement, technology experts, and even society as a whole), most studies agree that school is the most practical setting for intervention. Thus, the 15 lines of action identified in this systematic review must all be considered to effectively address sexting in childhood and adolescence.

Background The influence of social capital on health has been widely discussed in existing literature; however, these previous studies have produced conflicting findings regarding whether social capital (in the form of social networking and cohesion) can, through communication and social support, enhance health outcomes [1]. We analyzed the effects of social capital on women's health during pregnancy. In Japan, the social environment surrounding pregnant women has evolved in recent decades, with existing trends toward nuclear families, rural depopulation, and higher urban population densities being considered to have weakened intergenerational exchanges and community ties [2]. Importantly, it is unclear whether such weakened social ties negatively affect health during pregnancy. In particular, if there is a negative association between a lack of access to social capital and healthy pregnancy, this could mean that current societal trends are having a significant negative overall effect in this regard. Conversely, if social capital has a positive impact on health during pregnancy, it could, in conjunction with standard medical interventions, offer a means of alleviating physical and mental difficulties for pregnant women. Previous studies have investigated the effect of social capital on several indicators of health during pregnancy, including self-rated health [3]; 12-Item Short-Form Health Survey (SF-12) scores [4]; symptoms of depression and anxiety [4][5][6][7][8][9]; pregnancy complications [4,10,11]; oralhealth-related quality of life (OHRQoL) [12]; preterm birth [4,8,11,[13][14][15][16][17][18][19][20][21][22]; and low birth weight [4,13,20,21]. Of these investigations, those of self-rated health, SF-12 scores, symptoms of depression and anxiety, pregnancy complications, and OHRQoL have reported that social capital has a favorable positive influence in this regard. However, the analyses of the influence of social capital on preterm birth and low birth weight have produced inconsistent results, with some studies reporting positive relationships [4,8,11,[20][21][22] and others reporting no effect or a negative effect [13][14][15][16][17][18][19]. In the most relevant study to the present research, the SF-12 was administered to pregnant women in Berlin, and social support was consequently found to have a positive influence on their scores [4] (the SF-12 is a shorter version of the 36-Item Short-Form Health Survey [SF -36], which is widely used to measure functional health and well-being, and is based on self-reports from respondents). However, this previous study was conducted across a limited geographic area and examined fewer than 1000 respondents; thus, analysis of large-scale nationwide data is needed to assess the generalizability of the researchers' findings. Considering this, we sought to clarify the impact of social capital on health during pregnancy using a large nationwide sample. More specifically, our study analyzed data gathered by the Japan Environment and Children's Study (JECS), which is a nationwide birth cohort study sponsored by the Ministry of the Environment of Japan. The primary aim of the JECS is to analyze the effect of environmental risk factors on children's health, and the project is being conducted in 15 regional centers across Japan (Hokkaido, Miyagi, Fukushima, Chiba, Kanagawa, Koshin, Toyama, Aichi, Kyoto, Osaka, Hyogo, Tottori, Kochi, Fukuoka, and south Kyushu/Okinawa). As part of JECS, researchers recruited expectant mothers from these areas between 2011 and 2014, of which approximately 100,000 pregnancies registered. The project will continue to follow these parents and children until the children reach 13 years of age. The design of the JECS has been reported in detail elsewhere [23][24][25]. We used nationwide survey data from the JECS to investigate the impact individual and neighborhood social capital have on the physical and mental component summary (PCS and MCS, respectively) scores of the 8-Item Short-Form Health Survey (SF-8; another short version of the SF-36). This research is expected to make two main contributions: first, the results should clarify whether social capital has a positive impact on health; and second, the results obtained from this large-scale dataset should reveal general attributes of the Japanese population. Previous studies of social capital have discussed both its positive and negative effects. Examples of negative impacts would be the exclusion of outsiders and strong enforcement of local norms [1,26,27]. To determine its true impact, the effect of social capital should be verified based on validated measurement scores and data. If the scale of the positive impact is smaller than that of the negative impact, this may indicate that social capital has an overall negative impact. Conversely, if our investigation finds that the positive impact is large and statistically significant, this could contribute to the promotion of health-care policies that focus on the social capital of pregnant women. --- Methods --- Study design The JECS gathered medical records, questionnaire results, and biological specimens from pregnant women from pregnancy through to child-rearing, with the content of the data collected depending on the stage of gestation, parturition, and childcare. This method of investigation enabled researchers to determine participants' characteristics throughout the period in question. We used the data obtained from the questionnaires and medical records. Pregnant women completed the first questionnaire (M-T1) during their first trimester, and the second questionnaire (M-T2) during their second and third trimesters. These respondents answered the questionnaires and returned them in person at subsequent prenatal visits or by sending them via mail to JECS Regional Centers. Where possible, the centers addressed incomplete questionnaires by performing subsequent face-to-face or telephone interviews with the respondents [24]. The participating women also recruited their partners, and there are approximately half as many registered fathers in the dataset as there are registered mothers. We limited the data used in our analysis to the mothers' responses; this was to avoid the risk of sample selection bias that could be caused by including fathers' responses. M-T1 includes question items concerning family characteristics, disease, tobacco use, substance use, working status, working environment, and various other topics. Moreover, M-T2 contains question items pertaining to health status, dietary habits, tobacco use, sleep quality, home appliances, substance use, working status, education history, household income, and social capital. Finally, medical records following delivery (Dr-0 m) contain details regarding the newborn baby, obstetric and delivery complications, and other topics. From M-T1, we used the information regarding family characteristics, self-reported history of disease, and labor-force participation; moreover, from M-T2 we used the PCS and MCS scores, age, experience of stressful events, education history, household income, and level of social capital (stressful events included experiencing, over the course of the previous year, the death and/or illness of a loved one, the loss of the respondent's and/or spouse's job, the acquiring of a significant mortgage, divorce, moving home, and marital problems). The presence of obstetric complications was identified using information from M-T2 and Dr-0 m. In Dr-0 m, physicians reported the timing and diagnosis of obstetric complications; if the diagnosis of an obstetric complication was recorded prior to the respondent completing M-T2, we regarded the respondent as having experienced a pregnancy with an obstetric complication. --- Outcome measures In our statistical analysis, we considered the PCS and MCS scores as outcome variables. Specifically, the SF-8 PCS and MCS scores were calculated based on the respondents' answers to the question items in M-T2, which includes items assessing general health, physical functioning, role-physical, bodily pain, vitality, social functioning, mental health, and role-emotional. The PCS and MCS scores measure physical and mental functioning, respectively, with higher scores indicating better health status; the validity of the Japanese translation of these question items has been verified in previous research [28]. --- Exposure The main exposures are the variables measuring pregnant women's social capital. As previous studies conceptualized, we regarded the resources embodied by the individual's social network as individual social capital; and the resources formed by social cohesion, such as the stocks of trust or reciprocal relationships within the community, as neighborhood social capital [3,12,29]. The M-T2 questionnaire contained question items pertaining to individual communication and evaluation of trust in and support received from neighbors. Supplementary Table 1 (Additional File 1) shows the question items related to social capital. The contents of the questions concerning individual social capital (questions A to D) are similar to those of the six questions from the Social Support Questionnaire (SSQ) [30]. In response to these questions, the respondents provide information regarding how often and strongly they depend on others. The variables we extracted from the questions represented social capital in terms of social networking at the individual level. Moreover, the questions on neighborhood social capital (questions E and F) are similar to the "social cohesion and trust" components of a questionnaire used in the Project on Human Development in Chicago Neighborhoods (PHDCN) [31]. Questions E and F require respondents to evaluate their degree of trust in and the support they receive from their neighbors. The answers respondents provide to these questions imply group attributes, measured in terms of individual understanding. The variables we extracted from these questions were considered to reflect social cohesion --- Participants Research groups can access the JECS data through the JECS Program Office's regional centers. We used the "jecs-ag-20160424" dataset, which includes questionnaire responses from mothers and fathers and medical records from physicians from the time of registration to 1 month after parturition. Before beginning the statistical analysis, we excluded some portions of the dataset, in accordance with our research criteria. More specifically, the total number of pregnancies registered in the dataset was 103, 099. Women registered with multiple pregnancies within the survey period were included but, for each woman, we limited the data to that for the first pregnancy, which reduced the dataset to 97,454. Data from respondents who withdrew consent were eliminated; this resulted in a further reduction to 97,425 participants. Finally, we targeted data from only participating women who answered all question items related to the variables used in our analysis. Thus, we ultimately analyzed a dataset of responses obtained from 79,210 respondents. The study flow chart is shown in Fig. 1. --- Statistical analysis To estimate the causal influence of social capital on health, we adopted average treatment effect (ATE) estimation with an inverse probability weighting (IPW) estimator. In the analysis, it is required to consider the property that the richness of social capital would be endogenously determined by individual characteristics in general, and to measure the causal influence without bias caused by said property. Generally, ATE is based on the difference of average outcomes between treated and untreated groups in an intervention. In this research, we reckoned individuals' different levels of social capital as a kind of non-randomized treatment for respective individuals, and we applied the method of the ATE estimation. The IPW estimator is useful to control the bias caused by non-randomized treatment. In cases where the treatment is not randomized, a simple comparison of outcome averages between treated and untreated groups highlights both the effects of the treatment and the differences in characteristics between them. For the correction of the sample selection bias caused by the characteristic differences, the IPW estimator is valuable in ATE estimation. We regarded pregnant women with the lowest level of social capital as untreated samples and the women with other levels of social capitals as treated samples in the ATE estimation with the IPW estimator. When the level of social capital is determined by individual characteristics, the characteristics would differ between treated and untreated samples. Our analysis used the IPW estimator to avoid the bias caused by the characteristic differences. In IPW, the reciprocal of the probabilities of the assignment to the treated and untreated groups for respective samples are estimated and employed as the weighting variables in the calculation of averages within the groups. The variables control the effect of the nonrandomized assignment on the averages. The IPW estimator requires a regression equation for assigning samples between the treated and untreated groups. As a dependent variable, the equation has a dichotomous variable that is equal to "1" for treated samples and "0" for untreated samples. The individual characteristics considered to affect the sample assignment are used as the independent variables. The results of this estimation provide for each individual predicted probabilities regarding their level of belonging to the treated and untreated groups, respectively; this predicted probability is generally called a "propensity score." The inverse probability, namely, the reciprocal of the propensity score, is used in the IPW estimator. Thus, we respectively calculated the weighted averages for the The mathematical specification of the IPW estimator is as follows. For individual i, z i is a dichotomous variable that reflects his/her assignment to the treated and untreated groups, respectively. If a sample is assigned to the treated group, z i = 1; if the sample is assigned to the untreated group, z i = 0. x i represents the vector of the covariates for the assignment. The predicted probability that the sample is assigned to the treated group is described as e i = p(z i = 1| x i ), and the range of the probability is 0 to 1. This probability is individual i's propensity score. The predicted probability that the sample is assigned to the untreated group is 1 -e i. These predicted probabilities are generally obtained from the results of logistic regression analysis. When y i is an outcome of individual i, the weighted average of the outcome variable among the treated samples using the inverse probability is <unk>y 1 <unk> 1<unk>4 P N 1 i1<unk>41 z i y i e i = P N 1 i1<unk>41 z i e i. Moreover, the weighted average among the untreated samples is <unk>y 0 <unk> 1<unk>4 P N 0 i1<unk>41 <unk>1 -z i <unk>y i 1 -e i = P N 0 i1<unk>41 <unk>1 -z i <unk> 1 -e i. The ATE is calculated by <unk>y 1 <unk> -<unk>y 0 <unk> [32-37]. In the survey from which our data are sourced, the question items pertaining to social capital present three or more response choices, allowing for different levels of social capital. Based on their answers to these questions, the samples are divided into groups representing various levels of social capital. Each answer is transformed into a categorical variable. We calculated the predicted probability through multinomial logistic regression, with the categorical variable set as a dependent variable. The group with the lowest level of social capital was regarded as the untreated group; the other groups were regarded as the treated groups. The analysis considered the ATE between the groups with the lowest and some different medium levels of social capital, as well as the ATE between the groups with the lowest and highest levels of social capital. The following example illustrates the method we used to calculate the ATE, based on a question item for which an individual selects an answer from three options. Specifically, one of the items on the questionnaire regarding social capital is: "The number of friends or neighbors to whom you can talk casually about your concerns" (Supplementary Table 1, Additional File 1). The associated response options are: "none," "one or two," and "three or more." This question item creates three groups with different levels of social capital. In our analysis process, first, the multinomial logistic regression analysis is performed, which provides the probabilities that the respective options are selected by an individual. When the predicted probabilities are described as e i0 (for "none"), e i1 (for "one or two"), and e i2 (for "three or more"), respectively, for individual i, e i0 + e i1 + e i2 = 1. We calculate the weighted averages of the outcome variable for the respective groups using the reciprocal of the predicted probabilities. The weighted average of the outcome among untreated samples who select "none" is: <unk>y 0 <unk> 1<unk>4 P N 0 i1<unk>41 y i e i0 = P N 0 i1<unk>41 1 e i0. Similarly, the weighted averages among treated samples who select "one or two" and "three or more" are expressed as: To calculate ATE based on the IPW estimator, two assumptions must be satisfied, otherwise this ATE evaluation cannot be justified [38]. One assumption is that each sample has a positive possibility of receiving each treatment level. When there is at least some overlap between the estimated density of the propensity scores that treated samples are assigned to an untreated group and the estimated density of the propensity scores that untreated samples are assigned to an untreated group, the overlap assumption is not violated. If the estimated density for the treated samples has most of its mass near 0, while for the untreated samples the estimated density is near 1, these densities do not have an overlapping region, and the overlap assumption is violated [38,39]. <unk>y 1 <unk> 1<unk>4 P N 1 The second assumption is that the means of the covariates corrected by the IPW estimator are balanced between treated and untreated samples. When the means of the covariates of the treated samples are close to those of the untreated samples, the assumption can be considered as not being violated [38,40]. Before obtaining the ATEs, the validity of these assumptions must be checked. --- Results The summary statistics of the variables we used are shown in Table 1. These summary statistics indicate the individual characteristics of the participating pregnant women. The mean PCS and MCS scores were 45.64 and 49.13, respectively. Most participants were married (including common-law marriage). Approximately half did not have children prior to the current pregnancy. Of the pregnant women in this sample, over 80% had been diagnosed with at least one type of disease, approximately 20% had experienced at least one obstetric complication, and over 40% reported experiencing at least one stressful event in the past year. The summary statistics also revealed the sample profiles of individual and neighborhood social capital. For questions A to F, responding "none of the time," "none," or "disagree" indicated the lowest level of social capital. First, regarding individual social capital (questions A to D), for questions A to C 10-15% of the pregnant women responded "none of the time" or "a little of the time," thereby indicating that they had low levels of individual social capital. Moreover, the responses to question D indicated that approximately 40% of the respondents did not have three or more friends or neighbors with whom they could casually discuss their concerns. Regarding neighborhood social capital, responding "disagree" or "somewhat disagree" to questions E and F indicated low levels of neighborhood social capital; overall, approximately 45% of respondents answered "disagree" or "somewhat disagree" to these questions. A two-sample t-test that compared the respective mean health statuses of the two groups showed that both PCS and MCS scores were related to age, marital status, disease, obstetric complications, stressful events, labor-force participation, mothers' and fathers' academic histories, and household income (Table 2). For individual and neighborhood social capital, the difference in the two-sample mean was generally greater for the MCS score than for the PCS score. In addition, the group with a higher level of social capital had better physical and mental health statuses, with a few exceptions. The two-sample t-test of social capital revealed a correlation between social capital and certain individual characteristics (Table 3). More specifically, high individual social capital (measured in questions A to D) was found to be related to being married, having no previous children, experiencing no stressful events, having a higher education level, and having a higher household income (Table 3). Moreover, high neighborhood social capital (questions E and F) was found to be associated with older age, being married, having children, having no current disease, having no obstetric complications, experiencing no stressful events, not participating in the labor force, having a higher education level, and having a higher household income (Table 3). The ATEs for the PCS and MCS scores showed the degree that social capital affected physical and mental health. The weighted averages and ATEs were obtained through IPW estimation. The ATEs were obtained via subtracting the weighted average of the group with the lowest level of social capital from that of the group with other levels of social capital. The ATEs indicate how social capital influences the summary scores of the treated groups compared to the untreated group. If the ATE is positive and the magnitude is large, it can be concluded that social capital largely improves the summary score. Before assessing the ATE values, we needed to verify the validities of the assumptions. We first found that the estimated densities of the propensity scores for the treated and untreated groups overlapped in the respective question items; this finding indicated that the overlap assumption was not violated. Second, we calculated the standardized differences, shown in Supplementary Tables 2-1 (Additional file 2), 2-2 (Additional file 3), and 2-3 (Additional file 4), as performed in previous research [38,40]. These tables indicate the standardized difference between the means of the treated and untreated groups before and after correction by the IPW estimator. If the standardized difference calculated from the data weighted by the inverse probability is close to zero when compared with that from the raw data, the correction made to balance the treated and untreated groups can be regarded as appropriate. These tables show that the differences from the weighted data are approximately zero. Overall, the differences from the weighted data are smaller than those from the raw data; thus, we can conclude that the assumption of the balance between the treated and untreated groups was satisfied. The ATEs for the PCS scores indicated a significant negative impact of social capital in question A, and a positive impact of social capital in question D (Table 4). Moreover, a positive impact of social capital on the PCS score was observed for neighborhood social capital (Table 4). For example, the ATEs for question E were 0.50 and 0.90 for "somewhat agree vs. disagree" and "agree vs. disagree," respectively (a value of 0.90 indicates that the PCS score for respondents who most strongly feel mutual trust is 0.90 higher than that of respondents who feel no mutual trust). These tables show that the absolute values of the impacts of individual and neighborhood social capital on the PCS scores were between 0 and 1. These results indicate that social capital has a negligible effect on physical health during pregnancy. The ATEs of the MCS scores are reported in Tables 5. Except for "a little of the time vs. none of the time" for questions A and B, the statistically significant ATEs within pairs of different levels of individual social capital were all positive (Table 5). Further, all ATEs relating to neighborhood social capital were positive and statistically significant (Table 5). These ATEs show that higher levels of social capital have a larger positive impact on mental health. We can identify a proportional relationship between the level of social capital and the scale of the ATE. In questions A, B, E, and F, the largest differences in the MCS score associated with the highest levels of social capital were approximately 1.0-1.6. Moreover, for questions C and D, the largest differences caused by social capital were approximately 3.6 and 4.4, respectively. The maximum effects of individual and neighborhood social capital were approximately 4.4 (question D) and 1.6 (question E), respectively. The results for the PCS and MCS scores imply that both individual and neighborhood social capital have some degree of positive impact on mental health, with this being particularly true for neighborhood social capital. --- Discussion Our research contributes to existing literature by identifying the positive impact social capital has on the mental health of pregnant women. Furthermore, by using nationwide survey data collected across Japan, the generalizability of our findings relating to social capital and health during pregnancy is high; in comparison, a previous work that used scores obtained from the SF-12 [4] examined fewer than 1000 participants. Our results indicate that, for pregnant women, a lack of social ties is associated with worse health; this was especially notable in regard to mental health during pregnancy. This finding, showing that the mental health of pregnant women is improved by social capital, can have It is also important to explain the relevance of this study in terms of comparing its design with that of previous related research. Studies that have measured the effect of social capital on health can be categorized into four groups in terms of the data type examined: (1) individual social capital and health outcomes, (2) individual social capital and group-level health outcomes, (3) group-level social capital and individual health outcomes, and (4) group-level social capital and health outcomes [1]. Our study can be categorized into designs 1 and 3. PCS and MCS scores, individual-level social capital, and group-level social capital were measured through the participants' responses. The question items relating to social capital required participants to provide information regarding their communication network at the individual level and to evaluate their degree of trust in and support received from their neighbors. This evaluation indicates their neighbors' group attributes and is regarded as a collective factor. The variables of individual and neighborhood social capital indicate the degree of individual network resources available to participants and the participants' social cohesion, respectively. This study differs from previous studies that have analyzed the influence of social capital using JECS data [9,10]. One such study, which used the Kessler 6-Item Psychological Distress Scale (K6) as an outcome measure, did not consider the neighborhood social capital data obtained through questions E and F [9]. Another study considered gestational diabetes mellitus as an outcome; while this research used the question items we applied in our analysis, the researchers also examined responses to question items concerning the degree of regional public safety, mutual trust, and mutual assistance through principal component analysis [10]. The JECS question items for mutual trust and assistance are: "Would you say that most people can be trusted?" and "Would you say that most of the time people try to be helpful, or that they are mostly thinking of themselves?" This measure of generalized trust is debatable, because the question items do not specify a reference area for the respondent [41]. Further, questions that ask about generalized trust may cause respondents to report certain perceptions that are unrelated to their life within their communities [41]. Therefore, such question items are gradually being removed in favor of items that refer to familiar or personal trust [41]. The question items in the JECS survey regarding regional public safety and mutual assistance also present similar problems, as they do not clearly specify a reference area. Our study regarded both individual networking and social cohesion as social capital to be investigated, and our analysis purposely did not include question items concerning regional public safety and generalized feelings. A limitation to our analysis is that we did not obtain detailed information related to social capital, such as friends' and neighbors' characteristics. Network analysis of social capital, using the "position generator" and "resource generator" measurement instruments, could identify the effectiveness of individual network members [42]. However, identification of substantial functions among network members is difficult in analysis of nationwide survey data sourced from a large number of question items. --- Conclusions We used JECS data to analyze the impact of social capital on the health of pregnant women in Japan. We calculated ATEs of social capital on the PCS and MCS scores of the SF-8 using the IPW estimator. We consequently found that social capital has a degree of positive influence on MCS score. This result implies that --- Availability of data and materials The data used to derive our conclusions are unsuitable for public deposition owing to ethical restrictions and the specific legal framework in Japan. Specifically, it is prohibited by the Act on the Protection of Personal Information (Act No. 57 of 30 May 2003, amended 9 September 2015) to publicly deposit data containing personal information. The Ethical Guidelines for Epidemiological Research enforced by the Japan Ministry of Education, Culture, Sports, Science and Technology and the Ministry of Health, Labour and Welfare also restrict the open sharing of epidemiologic data. All inquiries regarding access to the data should be sent to [email protected]. The person responsible for handling inquiries sent to this e-mail address is Dr. Shoji F. Nakayama, JECS Program Office, National Institute for Environmental Studies. --- Supplementary information Supplementary information accompanies this paper at https://doi.org/10. 1186/s12884-020-03131-3. Additional file 1: Supplementary Table 1. Items that assess social capital. Additional file 2: Supplementary Table 2-1. Balance check using standardized differences for individual social capital. Authors' contributions HI designed the study. RM and KM analyzed and interpreted the data. RM wrote the manuscript. The JECS group collected the data and obtained the funding. KM, KH, ATs, ATa, HI, and the JECS group contributed critical revisions to the manuscript, and read and approved the final draft of the manuscript. --- Additional file 3: Supplementary --- Ethics approval and consent to participate The JECS comprehensive protocol was approved by the Review Board on Epidemiological Studies of the Ministry of the Environment of Japan (100910001) and by the ethics committees of all participating institutions. This specific study was also approved by the Ethics Committee of the University of Toyama (R2015094). JECS is conducted in accordance with the Helsinki Declaration and other national regulations, and written informed consent was provided by all participants. --- Consent for publication Not applicable. --- Competing interests The authors declare that they have no competing interests. --- Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Background: Previous studies revealed positive, negative, and no influence of social capital on the health outcomes of pregnant women. It was considered that such differences were caused by the disparities of outcome measures and sample sizes between studies. Our chief aim was to verify the positive influence of social capital on the health condition of pregnant women using established health outcome measures and large-scale nationwide survey data. Methods: We employed questionnaire survey data from 79,210 respondents to the Japan Environment and Children's Study, and physical and mental component summary scores from the 8-Item Short-Form Health Survey as outcome measures. We estimated the effect of individual and neighborhood social capitals on physical and mental component summary scores. To consider the property that the richness of social capital would be generally determined by individual characteristics, and to estimate the causal influence of social capital on health without bias caused by said property, we adopted average treatment effect estimation with inverse probability weighting. Generally, average treatment effects are based on the difference of average outcomes between treated and untreated groups in an intervention. In this research, we reckoned individuals' different levels of social capital as a kind of non-randomized treatment for respective individuals, and we applied average treatment effect estimation. The analysis regarded pregnant women with the lowest level of social capital as untreated samples and women with other levels of social capitals as treated samples. Results: For mental component summary score, the maximum average treatment effects in the comparison between the lowest and highest levels of social capital were approximately 4.4 and 1.6 for individual and neighborhood social capital, respectively. The average treatment effects for the physical component summary score were negligible for both social capital types.

INTRODUCTION ecently, the phenomenon of walled communities has increased and developed in many countries of the world in general and Iraq in particular, which expresses new characteristics of the urban product. Despite the multiplicity of urban patterns of walled communities from one country to another, and even from one city to another within the same country, they are similar in a number of characteristics, and investors have recently moved towards building walled communities in Iraqi cities, especially the city of Kut, which was discussed in the research. There are studies that indicated the importance of gated communities in achieving a number of dimensions of urban sustainability by ensuring privacy, security and a luxurious social life. Perhaps the new gated communities are good examples of the success of urban design in itself and its role in achieving social cohesion for residents. The idea or belief that social cohesion at the neighborhood level can be created or enhanced through urban planning and design is certainly not new. For example, the concept of "neighborhood unity" coined by Clarence Perry [1] was based on the assumption that the localization of facilities and services can help create neighborhoods that are characterized by greater face-to-face interaction between neighbors and a stronger sense of community than was prevalent in the modern urban environment. The same idea also influenced the work of neourban planners such as Clarence Stein and Michael Corbett, however, it is the writings and works of neo-urbanists that generated the current enthusiasm and wide interest in this very idea [2]. In order to create socially diverse neighborhoods, the new urbanists advocate careful integration of different dwelling sizes and different types of tenure [3]. Neighborhood design principles established to promote social cohesion among residents deal mainly with issues such as the integration of land uses, the location of services, the layout of streets, the design of streets and other public spaces, the width of residential lots, and the relationship between house and sidewalk [4]. Based on these design principles, new urbanists have developed typical zoning codes, which have been adopted by many municipalities around the world for their zoning ordinances. These design principles have also been incorporated into the urban and architectural codes that have guided the planning and regulation of construction in a rapidly growing number of new urban development's[5]. The research dealt with the walled residential community that was built by an investment company in the university residential neighborhood in Al-Kut as a case study, so the research problem is raised through several questions that lie in showing the extent of the impact of urban design on social cohesion and how can we design communities to increase social cohesion in the future, and whether Social cohesion can be enhanced through the personal characteristics of residents, and that the aim of the research is to demonstrate the impact of the diversity of residential use on the efficiency of the walled community, and that the hypothesis of the research lies in that the diversity of residential use is necessary to achieve social connections and contributes to achieving the requirements of balance in the urban design of walled communities. The research relied on the descriptive and quantitative analytical method to reach the results. --- GATED COMMUNITIES Gated Communities are societies that are characterized by calm and serenity and include the elite and the class. High-end social, which has prestigious lifestyles, and these methods are achieved through a set of traits Common elements of natural environment coordination, security, safety, distinctive architectural character and services [6]. It is also defined as those cities that are surrounded by walls and have several entrances that are controlled and controlled by security guards, or through electronic control points. These settlements are equipped with their needs of services such as large shopping centers, large commercial centers, recreational facilities, open areas and parks, in addition to artificial lakes and swimming pools [7]. There are some reasons that led to the emergence of this type of society, which are those related to the reality of the existing urbanization, represented in many urban problems, such as the general deterioration of urban spaces due to high densities, traffic congestion of roads, pollution, as well as the low performance and efficiency of services, represented by the decrease in the per capita share of green and open spaces. All of the above is considered as centrifugal forces that pushed some residents to leave their communities to escape to those communities in search of a better lifestyle [8]. --- Features of planning and managing gated communities Gated communities affect urban planning and management issues, but there are concerns about the monitoring and management of these gated communities, and there are questions about the powers of each of the residents, associations and private security agencies individually or with a joint responsibility with the residents of those communities as bodies responsible for managing such types of societies as these communities create a kind of large social division that in turn reflects a pattern of urban separation urban separation, which requires double services and facilities to serve residents within those communities and others to serve other classes outside them, and this also raises many concerns about the traditional role of government and the potential impact on urban management in the future [9]. --- The concept of urban management The residential community is a dynamic life system that has many components that vary according to time and place, but its components cannot be separated, otherwise a defect occurs, and it should improve the management of this system to achieve the needs of citizens, which is called urban management[10]. This system includes the public and urban services sectors in addition to the housing sector, and enters the process of providing services under the crucible of urban management in the residential community, including the elements of planning, implementation and management and the set of relations between the parties involved in it (the city apparatus, the urban plan, the state, the private sector, the population), as the time has passed when the state was carrying out all the burdens of development in planning, -up[11] The overlap of the parties involved in urban management without coordination or a general framework governing relations between them leads to a disjointed and uncoordinated result and does not achieve real sustainable development, as the management of urbanization has many elements for its success, such as participation, transparency and justice. All these elements must be integrated, and good management also requires adequate resources such as material funding and labor [12]. With good urban management, residential communities flourish and social and economic conditions in communities improve, and good urban management aims at the following[13]: Activate the participation of actors, stakeholders and residents in management and decisionmaking (decentralization of management). Accept the role of the government as an observer and regulator of the process, while acknowledging the participation of the population and all parties in decision-making. --- Urban Design and Social Cohesion Social cohesion is a concept that is difficult to define in its most general sense. It refers to a type of glue that binds a community together. Social cohesion in a neighborhood is often linked to building a neighborhood's sense of community. The literature most relevant to the study of the potential impact of urban design on social cohesion among the various components of society can be divided into two main directions. These two trends can be clearly identified in the international literature. The first approach focused on the relationship between urban design and social cohesion at the neighborhood level, or a sense of community [14].[This focus on the neighborhood has been based, in large part, on the assumption that tight-knit neighborhoods are viable units for implementing social and economic development programs and that, with the associated strong sense of community among their residents, they are important to the general well-being for the individual and the benefit of the family and society as a whole [15]. Within this first orientation, two distinct groups of literature can be identified. First, there is the literature associated with and defense of the principles of the New Urbanism movement. is the latest urban design movement that proposes specific design principles to create socially diverse neighborhoods with a strong sense of community among residents [16]. The second trend includes literature that has focused on the impact of design on the livability of urban public spaces outside the neighborhood scale. Following the early work of researchers such as Whyte and Appleyard [17], this literature has investigated the impact of design characteristics on the increased presence of people in public spaces such as plazas, squares, streets, and parks. This approach is relevant to the study of social cohesion between different residents. Our research will focus on the first approach. --- Case Study: The Gated Residential Community Project In The Univesity Neighborhood The residential investment university neighborhood project is one of the projects of the National Investment Commission in Wasit Governorate, Al-Kut District, according to the investment license 42/ S / 2018. The project is located in the city of Kut, on the right side of the Tigris River, in District 45 Um Halana. The use of the land according to the basic residential design (horizontal) With an area of 29.25 Hectare, about (292,500) square meters. The project includes 542 housing units and service facilities divided according to the sectorial design No. 887, Kut Municipality and its amendments. The lands in the project were re-divided for the purpose of reducing the number of lands with a large area into parts with smaller areas, and after dividing the lands allocated for building residential units, the total number became 776 housing units, note that the available areas before dividing the areas are 500 square meters, 240 square meters, and 200 square meters after the division, models have been added 170 square meters, 120 square As in --- ANALYSIS 3.1 The method of implementing the project The method of implementing the project is through the investing company, it is sold directly to the citizen through a sale and purchase contract between him and the company According to the investment law, the ownership of the residential land is transferred in the name of the company, provided that the reservation sign is placed in favor of the Kut Municipality Directorate until the completion of the construction of the housing unit and the transfer of its ownership to the beneficiary (to ensure that there is no speculation on the lands. And after collecting the amount of the housing unit according to the concluded contract, its ownership will be transferred to the beneficiary under a formal transfer transaction in the competent Real Estate Registration Directorate. The Central Bank's initiative for the years 2020 and 2021 to grant real estate loans to citizens (an interest-free loan for a period of 20 years) and to support investment projects contributed to the growth and success of the project, in order to increase the citizen's demand for a decent housing house. --- Land uses in the gated residential community The gated community was initially designed according to the activities and uses shown in Table No. ( 1), and with regard to residential use in the gated community, the housing units were divided according to the areas and numbers indicated against each of them, as indicated in Table No. (2). The investment license for the project was issued for the divided housing units, which numbered 542 housing units, in addition to the housing units belonging to the municipality, according to the percentage referred to above in the investment law, which numbered (39) units. Thus, the total number of housing units within the area designated for residential use is 581 housing units. Changes in land use in the gated residential community. A number of modifications took place in the land uses of the gated residential community, as shown in Table No. (3), and there was also a change in the residential use of the gated community, where the area of a number of housing units was divided as shown in --- Table( 1):- --- RESULTS --- Diversity Indices Diversity index is a quantitative measure to measure the number of different species present in a data set of population [18]. These indicators are statistical representations of the diversity of land uses in the gated residential community as well as the diversity in the area of housing units in different aspects (richness, parity, and control). A comparison was made to measure the diversity index of land uses in general and residential use in the form of the design stage and the implementation stage in which land uses and the area of housing units were modified within the residential use [19]. --- Shannon-Weiner Index A widely used indicator that takes into account species richness and evenness is the Shannon Wiener Diversity Index, originally proposed by Claude Shannon in 1948. It is also known as the Shannon Diversity Index [20]. H = -<unk> Pi * Ln * Pi <unk> n=1 (1) where Pi = proportion of individuals of types i, and Ln is the natural logarithm, and S = Types richness. The value of H ranges from (0) to H (max), whereas H (max) is different for each community and depends on species richness. Diversity analysis was conducted for the residential units of the gated community before and after changing the area of the housing units, as shown in Tables 5 and6. Through the results of the above indicator, the urban design of the gated residential community after changing the area of the housing units within the residential use in it is more diversified, and the modified housing units' design is richer in type than the previous design. --- CONCLUSIONS The research proved that there is a better understanding of the potential role of the diversity of housing units in gated communities in contributing to social cohesion among the various components of society in the city of Al-Kut through the case of good diversity that appeared in the residential use in the closed residential community, which reinforced the presentation of relevant literature. Through the current knowledge reached by the research, it became possible to clarify the directions of the future urban design required, where the literature was discussed and the results of the research indicated that it is possible to manipulate the design characteristics of the neighborhood environment to enhance the neighborhood cohesion or the neighborhood's sense of belonging to the community, despite the social homogeneity Culture among residents is an important precondition for neighborhood cohesion. The idea of trying to create diverse, culturally and socially cohesive neighborhoods through design and/or politics does not appear to be the appropriate or most appropriate way to achieve cohesion between the various components of the greater community, and to promote social cohesion of the greater community. In a community, the size of homogenous areas should be small enough to allow people to identify with others. According to this approach, there are two needs: The first: the need to determine the appropriate size for homogenous communities or residential neighborhoods, Second, the need to identify the types and nature of urban public spaces that may play the role of neutral settings that can attract people of different social groups and where communication and interaction between them can occur.

Perhaps the new gated communities are good examples of urban design in their own right, and the questions that the paper raises are how the diversity of housing unit space affects social cohesion and how we can design communities to increase social cohesion among its occupants in the future. The diversity of land uses facilitates better social cohesion s and bonds, as good street coordination and open spaces calming traffic contribute to creating a suitable built environment for residents to generate a sense of place. Gated residential compound models seek to create residential units that provide a safe, luxurious and private life for its residents. They achieved it simply by gated themselves. Studies have confirmed that the main reason that prompted investors to build gated communities is their desire to achieve quick wealth and that the diversity in the area of the residential unit in it contributed well to the speed of marketing the residential units built in it, so this research paper seeks to show the effect of the diversity of the area of the residential unit on the efficiency Residential use in gated communities. The research concluded that the gated community has achieved a state of balance in residential use through the diversity of housing units.

Introduction Families' daily lives are structured to a great extent by their social and economic resources. Geographical location of the home, work places and schools, as well as neighborhood structure and children's leisure activities are other important conditions. Exploring the ways children and parents perceive, experience and negotiate the organization of time and space in everyday life is the aim of this study. Specifically, in this article we examine how the children and parents in a middle-class, inner-city family negotiate the organization of time and space within the family. Within this case study we address this question with examples of how children and parents "do time" and "create space" in the domains of children's outdoor play and institutionalized leisure. Research on families' organization of daily life has often focused on the organization of time. In Sweden, Forsberg (2009) investigated parenthood and everyday organization in child-centered middle-class families, focusing on negotiations over the balance of work, leisure and family commitments between parents and children. Forsberg mostly focused on negotiations between parents (e.g., over who takes the children to day care and who goes to work early), but because in contemporary Sweden children are often expected to make their voices heard in family matters, he also addressed negotiations between parents and children (e.g., about when to do homework). Jurzcyk and Lange (2007) have also addressed the balance between work and leisure, stressing time but also space, and discussing the effects that flexible work hours and parents bringing work home have on children's well-being. U.S. sociologist Annette Lareau has done important work in the field of families' temporal organization by adding class to the analysis of the organization of daily life. In her famous book Unequal Childhoods (2011), Lareau compared how families with children from different social backgrounds in the United States organize their daily life and move between different social and geographical contexts. In addition to noting class differences in organized leisure activity participation, Lareau showed that the daily lives of middle-class families had a much faster pace than those of working-class families. Members of well-off families talked in terms of "deadlines" and the amount of "wiggle room" they had. Lareau (2011) also discussed differences between lower-and middle-class families in the balance between children's time for informal play and for institutionalized leisure activities. Working-class children could engage in informal play at their own pace and as they pleased. They exercised more agency over their playtime and were less tired and less stressed about tight schedules than middle-class children. In addition, their outdoor informal play led to children of different ages playing together. For middle-class children, time for informal play or hanging out in the neighborhood was more restricted (see also Karsten 2010;van der Burgt 2010). In her study, Lareau found that middle-class children sometimes had difficulty coping with unstructured time, feeling they had nothing to do. Time outside school was devoted to the development of different skills in sports and music useful in adult life, including what Lareau (2011, 39) called "white-collar work skills, including how to set priorities, manage an itinerary, shake hands with strangers, and work on a team." Institutionalized leisure activities are seen as desirable by middle-class parents and as an important part of their children's general education, a logic of childrearing Lareau (2011, 238) termed "concerted cultivation." Middle-class parents are thus willing to spend their leisure time on different institutionalized leisure activities for each child, engaging in "cultures of busyness" (Karsten 2005). Tillberg (2002) and Lareau (2011) have shown that children's leisure activities take up a lot of their parents' time, both in chauffeuring to practice and weekday competitions and in attending competitions on weekends. Tillberg argued that the amount of time and engagement parents put into their children's leisure activities and clubs results in these becoming a considerable part of the parents' identities and identity construction. In addition, Lareau stressed the great deal of time that children's leisure activities often take up for younger siblings, almost taking "hostage" of their time. The children in Lareau's middle-class families also did not spend much time with their parents and siblings because every family member had his or her own time-space schedule. Even while at home, middle-class family members seldom spent time in the same room. The location of the neighborhood in which a family lives also affects the ways members are able to organize their everyday lives in time and space. An inner-city neighborhood with cultural and commercial services near at hand often offers poor opportunities for children's outdoor play. Compact city policies are being adopted in Sweden and elsewhere in Europe that increase the percentage of apartments in the housing stock throughout the city. The increasing density leads in many neighborhoods to a decrease in green spaces and an increase in traffic (Björklid and Nordström 2007). Limited access to good outdoor play areas and prohibitions against unsupervised play in the streets affects crucial aspects of children's mental and physical health (Kyttä 2004;Hillman, Adams and Whitelegg 1991). Despite this, the New Urbanism trend has led to more families with children living in innercity apartments, a tendency seen in Sweden as well as cities such as Amsterdam, London and Paris (Boterman, Karsten and Musterd 2010;Brun and Fagnani 1994;Butler and Robinson 2003). According to Karsten (2007), urban parents have several reasons for remaining in inner-city locations. They may want to retain the urban lifestyle for themselves as well as for their children, to keep close to local social networks or to avoid time-consuming commutes from the suburbs. Another reason may be proximity to what they perceive as good schools (Butler and Robson 2003;Holme 2002;Boterman, Karsten and Musterd 2010). To allow their children to attend a good school, parents move to a home within that school's district or in a location that makes commuting an option. In Sweden, free school choice has led to local school markets and an increased commuting among urban children (Gustafson, Pérez Prieto and Löfdahl 2011;Bunar 2010;Kallstenius 2011). Information on which neighborhoods and schools are well regarded is usually gathered (and even constructed) through parents' social networks (Holme 2002). Research focusing on children's perspectives and experiences shows that not only parents but also children's social networks play an important role (Gustafson 2011;van der Burgt 2008;Reay 2007). Due to their age and dependent status, children are particularly restricted in their mobility, what they can do and where they can do it. Without assistance of a parent, children usually have a small activity zone within which they can move by themselves (Hägerstrand 1970). The densification of urban space, together with more emphasis on institutionalized leisure, is decreasing the time children spend unsupervised outdoors, both in Sweden (van der Burgt 2010) and in other European contexts (Karsten 2005;Valentine and McKendrick 1997). The ferrying of children from home to school to leisure activities results in what Zeiher (2003, 67) calls the "insularization of children's individual life spaces." "Doing Time" and "Creating Space" In daily life, the ways individuals are able to organize their lives in time and space are bounded by different kinds of interacting constraints (Hägerstrand 1970), such as the needs to eat and sleep (capacity constraints), the need to come together and link in time and space for specific activities, and the need to follow rules and regulations in certain domains (authority constraints). In a family, the different members' paths adjust to the family's common timetable and spaces, forming what Hägerstrand calls family "bundles." This implies interacting schedules and a conceptualization of the family as a relational system (Jurzcyk and Lange 2007; Karsten 2007). In addition, the time-space organization of families has to be understood in the context of the transition from the Fordist system, with its rigid external time structures, to a more flexible system, a process that "blurs boundaries" (Jurczyk and Lange 2007, 217) between different spheres. In early modernity the times and spaces of work and leisure were separated by clear boundaries, but such boundaries have in late modernity become more subtle and less clearly defined. A society with fewer fixed external time structures assigns more responsibility for time management to the individual (Jurczyk and Lange 2007). However, this blurring of boundaries does not mean that clear temporal and spatial boundaries have ceased to exist. In fact, Tillberg (2002) argues that the conviction that children should be engaged in institutionalized leisure activities has become such an important norm in today's society that such activities have become the dominating authority constraint in family daily life. An effect of the increasingly institutionalized daily life of children (Hengst 2007) amid the time bind of family routines is that children have little control over their free time and space. Children have been hidden inside family statistics, especially in research investigating society and families (Qvortrup 1997). In our research, however, we view childhood as socially situated and children as social agents (James and James 2004). Accordingly, agency is an important concept in our research and we stress both children's and parents' agency when it comes to the organization and negotiation of time and space within the family. Children's and parents' agency in regard to the organization and negotiation of time is emphasized by the concept of "doing time" (Jurczyk and Lange 2007). Jurczyk and Lange argue that an active "doing" of time-such as creating "family time," "quality time" or "own time"-is becoming more important for both children and parents. These authors also point to the relational aspect of agency in that children and parents demonstrate different capacities to "do time" depending on their social and economic resources. In the present article we aim to contribute to this discussion by emphasizing the importance not only of family members "doing time" but also of "making and creating space." Time and space are intimately connected dimensions (Massey 2005;Hägerstrand 1970) and omnipresent in that every action is taken in a certain place at a certain time. We argue that while researchers such as Jurczyk and Lange (2007) and Lareau (2011) do discuss space to some extent (the blurring/ articulation of spatial boundaries), they pay insufficient attention to the ramifications of space for agency. People do not only act passively in regard to spatial restrictions, they are also involved in the production of space; that is, the making and creating of space (Massey 2005). Parents' and children's agency and living conditions are not only intimately connected with time-space variations (Holloway and Valentine 2000), but their agency may also differ in different domains. According to Jurczyk and Lange (2007), for example, children often exercise relatively strong agency in commercial domains (e.g., toys, computer games, clothes), but it is much more difficult for them to influence their parents' work hours. --- Methodology The present article is based on a pilot study conducted to ascertain which aspects of families' time-space organization of daily life could be investigated with different research methods. While planning research on the time-space organization of families in neighborhoods with different class and geographical characteristics, we decided to do a pilot study of one family in particular-two parents, a son and a daughter-in a well-off, inner-city neighborhood in Uppsala, Sweden. This yielded not only valuable methodological insights but also valuable data on the time-space structures that frame this family's everyday life. In addition, the study suggested important concepts concerning parents' and children's agency in negotiating the family members' shared and individual time and space. The study had a qualitative mixed-method character, with a broad thematic approach to different family members' perspectives on time and space in daily life. We asked the parents and children to fill out time-space diaries across a seven-day week. We conducted separate interviews with the children and parents to encourage them to express their unique perspectives on and experiences of organizing and negotiating their time, spaces and activities. After the interviews, we took walks through the neighborhood with both the children and parents during which they showed us the places they frequented. Speaking with the parents and children separately yielded differing insights into how the family's decisions were negotiated (although we did not study the negotiation process itself). --- The Setting The centrally located upper/middle-class district of Kungsängen in Uppsala is a new residential neighborhood consisting of five-and six-story buildings with some sections still under construction. Apartments constitute 95 percent of the residences. The studied family lives on the fourth floor in one such building, in a five-room apartment with a large balcony facing a courtyard. Before moving to Kungsängen the family lived in a detached house in a fairly new development in the countryside, approximately a half-hour drive from Uppsala. The neighborhood consists of predominantly well-off Swedish inhabitants (only 14 percent of residents have a foreign background). A sushi bar and a posh clothing boutique reflect the neighborhood's economic level (Figure 1). Proximity to a wellstocked grocery, the Uppsala city center, the highway to Stockholm and public transportation including the train station provide easy access to commercial and cultural services and make Kungsängen a good location for commuters. There is a high rate of car ownership (52 percent), and the parking areas were built underground (Uppsala Municipality Statistics 2010)-an advantage for outdoor play since it reduces traffic dangers near the apartment buildings. A bus to schools in nearby areas serves the neighborhood. Children constitute 17 percent of the inhabitants and as a demographic group are slowly increasing (Uppsala Municipality Statistics 2010). However, Kungsängen was clearly not planned with children in mind. There is a lack of child services such as day-care centers, schools, and leisure activity sites. While there are a few playgrounds for small children, larger green spaces, playing fields, activity sites and playgrounds for older children and teenagers are lacking. Accordingly, the neighborhood does not meet the criteria for a child-friendly environment (Kyttä 2004). Indoors, the family's apartment has a large kitchen with a dining table and a spacious living room with a television and a Wii gaming console. The children have their own bedrooms, the parents share another and a spare room is used as an office with a computer. The shared laundry is located on the first floor. In the following section, we discuss the ways that the children-Anne, age 12 and Tom, age 9-and their parents, Hilda and Oskar, each 37-organize time and space in the domains of children's outdoor play and institutionalized leisure. We also relate this time-space organization to structural conditions such as neighborhood location and characteristics as well as social class. In addition, we provide examples of how the children and parents "do time" and "create space" for outdoor play and institutionalized leisure, which generally involves the balancing of different kinds of activities. --- Outdoor Play The opportunities and limitations for outdoor play are shaped by the family's apartment lifestyle and Kungsängen's structure and inner-city location. As noted above, playgrounds, larger green spaces, ball courts and activity sites for older children and teenagers are lacking. What spaces are available are small greens in the courtyards, a central plaza (Figure 2) with a circular green and paved area surrounded by the apartment buildings, and the streets that circle around the buildings. --- Figure 2. The Plaza In adjacent neighborhoods there are large green spaces and playgrounds for older children, but a river and major roadways surrounding Kungsängen effectively limit its children's access to these facilities. For outdoor play, then, the neighborhood is neither inviting nor very interesting for older children. The parents view this as the downside of the neighborhood. In addition, outdoor play is more complicated when living in an apartment on the fourth floor rather than the first or in a single-family house. The children and the parents spend less time outdoors now than they did when living in their country house with a garden. Discussing their choice of geographical location, we learned that the move to an apartment had not been their first choice. They had preferred a nearby area with detached houses, but due to Uppsala's limited stock of such housing a single-family home there was too expensive for the family. --- Choosing School over Outdoor Play and Mobility While the parents were not explicit as to why they decided to leave country living for an inner-city location, it is clear they actively searched for housing in the part of the city to which they eventually moved. They had two main reasons. One was proximity to the workplace; both Hilda and Oskar now travel short distances to work (Karsten 2007). The more important reason, however, was the proximity of what the parents considered to be two "good" schools, one in a neighborhood that could be reached on foot and another just a little further away. Even if the possibilities for children's outdoor play were taken into consideration by the parents, this factor was not given priority when choosing where to live. Instead, the parents placed greater value on access to a school with a reputation as a welldefined, workable environment (as reflected in effective procedures and sufficient order). In line with other research (Holme 2002), the parents' school preference was based on what they had heard from friends and colleagues at work. Specifically, the mother's best friend is a teacher at the school the children now attend, so it seems likely that her view of the school influenced the parents' choice. The children exercised little agency in the decision. Tom had preferred the nearer school because he would be able to walk each day, while now he has to take the school bus. This offers insight into the ways family social networks and the information that circulates within them affects children's time-space conditions and practices and, consequently, their well-being and competence. While gaining access to the "best" education and the experience of riding the school bus, Tom loses the experience and physical activity of walking to school. Thus, in line with other research (Karsten 2007; Jurzcyk and Lange 2007), we stress the importance of conceptualizing the family as a social network. --- Creating Space and Doing Time for Physical Activity Although Tom misses out on the physical activity of walking to school, his parents value the children's physical activity a great deal. Yet, the outdoor environment of inner-city Kungsängen struck their daughter Anne at first as dull and she preferred indoor activities. She does not have any friends in the neighborhood and there is not much for her to do there. Worried about Anne being physically inactive and gaining weight, her parents convinced her to join a soccer club. She now practices soccer on Tuesdays and Thursdays and then plays in weekend matches-a significant increase in physical activity. The soccer club practices in another neighborhood, so her father now drives her there and back twice a week. In this way, together they create both time and space in order to organize Anne's physical activity. The family's middle-class status (which means they have access to a family car and flexible work hours) allows it to conform to the time-space conditions required by this institutionalized activity (Lareau 2011;Karsten 2005;Tillberg 2002). Unlike Anne, Tom has no interest in getting involved in any institutionalized sports activity. According to his parents, this is not a problem with regard to his need for physical activity and fresh air. Oskar and Hilda view Tom as a typical and physically very active outdoor child, involved in informal play outside all the time with friends. This is also the way Tom describes himself and his leisure in the neighborhood. Although uninterested in organized soccer, he is keen on playing soccer with his friends in the small green spaces of the several courtyards. While outside, Tom is indeed very active, moving about all the time. He is also outside much more than Anne according to the data of their time-space diaries. Tom played outside in the neighborhood 280 minutes and Anne 60 minutes during the study's seven days. However, the time Tom actually engaged in outdoor play was quite limited compared to his other leisure activities. His passion for soccer showed itself more in the amount of time he spent watching matches on TV and playing soccer video games on the Wii console. In fact, Tom spent three times as much time with the TV and computers-860 minutes-than he did playing outside. Thus, even if he plays outside more than Anne and is physically very active while doing so, according to his diary he is more an indoor child than an outdoor child. The fact that Anne's soccer activity generally takes place outside serves to increase her time outdoors. So while Anne is viewed as an indoor child by her parents, the amount of time she spends outdoors is actually higher than her brother's: Anne is outside 17 percent of her time awake and not at school, compared to only 12 percent for Tom. In fact, it is the parents who spend the least time outdoors: only 6 percent for Oskar and 8 percent for Hilda of their time awake and not at work-biking, walking or standing and watching Anne's soccer matches. Apartment living seems to be negatively affecting the parents' physical activity levels more than the children's. --- Creating Play Space in a Place Not for Play When playing outdoors Tom exhibits an intense and creative use of space. Together with his friends-mostly boys of different ages but also Anne and sometimes a friend of hers-they play soccer and other games in the small green and open areas between the buildings. While playing soccer in a small space promotes the development of techniques for maneuvering the ball, Tom and his friends yearn for a large green space in the neighborhood where they could play "proper" soccer. In Kungsängen the lack of a large ball court means they have to be careful not to break any windows. They also have to avoid kicking the ball into a fenced-off construction site just behind the small green where a neighbor (a sports instructor) has erected a mini soccer goal (see Figure 3). This does not mean, however, that the children always stay within this small green space and obey safety rules. Tom showed us a hole in the fence where they sometimes enter the construction site to play. Also, in one of the courtyards there is a single swing and here both Anne and Tom express wishes for another so they could swing together or with someone else, compete for the fastest and highest swinger, etc. --- Figure 3. The mini goal As Christensen, James and Jenks (2000) show, children construct a sense of home and family by moving "in-and-out-and-around" the home. However, when living in a fourth-floor apartment, going in and out is not as easy as in a single-family house, and when Tom and Anne are outside their parents want them to stay there a while and not be going in and out all the time. Tom has found a way in his outdoor play to connect the inside of his home with the outdoors while remaining outside the apartment. When soccer games end and he wants to do something else outside, he must first return the ball to the apartment-and so puts it in the elevator to transport it up, while using the building's intercom to tell his parents upstairs that the ball is on its way for them to collect (Figure 4). This way he does not have to spend time climbing stairs or on the elevator to return the ball and then going back down and outside to continue play. In a way the elevator becomes part of Tom's play space; his creative use of it and the intercom, with his parents' help, gains him time to play. Thus he takes control over his own time and space and transforms places not meant for play into play spaces. --- Figure 4. Soccer ball in the elevator Another creative and effective way in which Tom recreates "places not for play" into "play spaces" is through the use of private courtyards in the neighborhood. As noted above, the lack of variety in play spaces (Kyttä 2004) leads Tom and his friends to bend some of the rules. Each apartment building has a courtyard intended for the residents of that building. The courtyards are accessible only by keying in a door code (Figure 5). In order to use shortcuts into the different courtyards and be able to play in more than just their own, Tom and his friends have discovered and memorized the door codes for each building. Thus, they create additional space for play. --- Figure 5. Coded-lock entrance to a building with a courtyard Institutionalized Leisure Anne's soccer activity consumes a lot of her leisure time both during the week and on weekends (17 percent of her time awake and not at school). Anne herself does not see this as a problem. On the contrary, she really enjoys the time spent and enjoys the sport so much that she plans to enroll in a secondary school with a soccer program. It is her hobby, she says, together with hanging out with her friends. A big part of why Anne enjoys organized soccer is that the soccer club is a place for meeting and socializing. Her teammates are also her friends and classmates and it is when she plays soccer that she meets them outside school. Thus, to Anne, playing at the soccer club is about doing and creating social time and space rather than time and a space for physical activity. The importance of the social dimension of institutionalized leisure is often ignored in the debate on children's leisure. It is also her parents' perception that Anne likes to go to the soccer club. Although at first they had to persuade her to start soccer for health reasons, they are happy to now see that her own driving force inspires her playing. Both parents are interested in Anne's soccer and like to go to her matches to cheer on the sidelines. Oskar describes himself as a leisure-club person who has always been engaged in some kind of team sport. He believes that playing a team sport is all about team spirit and being there for your team, so Anne's attendance at practice and matches is therefore very important to him. He is willing to help her attend practice by ferrying her to and from the soccer club twice a week. Hilda thinks that Anne's soccer is very time-consuming and wonders if Anne would rather skip practice or a match now and then, then concludes that this is probably her own perception rather than her daughter's. Anne herself does not speak of being stressed by or tired of her soccer activity, unlike children in Lareau's study (2011). Having one leisure activity and not several seems to allow Anne enough time and space for recuperation (Tillberg 2002). The ways she creates such time and space, between school, homework, meals and soccer, is by watching TV and chatting with friends on the Internet at home. --- Resisting Institutionalization Although Hilda and Oskar are confident that Tom does not need an institutionalized leisure activity to be physically active and get fresh air, they would still like him to engage in such an activity. They have tried to get him involved in a variety of sports and are currently trying to persuade him to start playing organized soccer. This can be understood as socialization into middle-class practices and in this respect the family is typically middle-class (Lareau 2011;Karsten 2005;Tillberg 2002). So far, Tom has refused to take part in any organized leisure activities or after-school clubs. Resisting institutionalization outside school, he manages to retain quite a bit of autonomous time and space. Not attending an after-school club allows him, for example, to be the first in the family to arrive home in the afternoon, make his own snack and eat it while watching TV or playing a computer game. This he greatly values, in line with Jurzcyk and Lange's (2007) discussion of the importance for children in mid-childhood of a degree of autonomy in determining, doing and creating their times and places as well as having time alone. --- Soccer as an Important Family Time-Space Constraint Looking at the family's time-space organization from a time-geographical perspective, it becomes clear that the members' paths on the weekend are bounded by Anne's soccer activity, which consumes a good deal of the family's common time as well as her parents' and her little brother's leisure while gathering the family at institutionalized leisure spaces. Anne's leisure activity is also an important authority constraint upon the family (Hägerstrand 1970;Tillberg 2002). Oskar drives her to practice twice during the work week and does not stay to watch but instead makes the 20-minute drive back and forth each time. Soccer practice also affects the way the family eats their meals. Oskar says that he often eats by himself, before, after or in-between his ferry runs. In order to have enough energy for practice, Anne has a large snack beforehand then eats dinner afterwards, either by herself or together with her father. On the weekends there are usually soccer matches somewhere in the area that both parents and often Tom attend. Although Tom did not talk about this himself, his parents say that he does not always like to come along to the matches. Due to their length (they often take the whole day) he cannot stay home alone and it is not always possible to stay at a friend's house for the whole day. Sometimes Hilda stays home for Tom's sake. Oskar, however, does not, attending Anne's matches whenever possible. This indicates that her soccer activity is primarily Oskar's parental responsibility. --- "Soccer Parents" Anne's soccer activity has grown to be an important part of her parents' social life as well. Hilda and Oskar like to go and watch the matches on weekends and to cheer Anne's team on from the sidelines. They do not often miss a match. Especially for Oskar this is an important part of parental identity. While Hilda was not very interested in watching soccer at the start, she is now very engaged in attending Anne's matches, both for the fun of seeing the team win and for socializing with other engaged parents. For both Hilda and Oskar the weekend matches have become social events as they have grown acquainted with a small group of parents who are as committed to their children's activity as themselves. Thus, Anne's soccer activity and the social network it has generated have become important parts of her parents' social life and a way of "doing parental time" and identity. However for Tom, who would rather stay home and play with his friends, Anne's soccer is a disadvantage. He misses out on social activities while those of his sister and his parents increase. Adjusting Work Hours for Children's Activities Tillberg (2002) has shown that parents adjust their work hours to children's leisure activities. Hilda and Oskar describe their family as one with "common" routines and members who like to be at home and in each other's company. They portray themselves as family-oriented parents and say they have adjusted their work hours and personal interests to family life. The way these parents "do family time" is to spend a lot of time at home during after-school hours. Hilda has adjusted her work hours and leisure time to family life. She works part-time in order to see to the children eating breakfast, packing bags and getting off to school, then arrives home again around four p.m. Oskar works full-time. Having flexible hours, however, he chooses to start work early in the morning, before everyone else in the family gets up, and comes home around 4:20 in the afternoon. He is then often the one who takes Anne to soccer practice, but this is not the explicit reason for the way he organizes his work hours. His main reason for starting and leaving work early is his preference for spending more time with other family members at home. He likes to start work early to be able to get home early, and then, being home, does not mind driving Anne back and forth to soccer practice. Oskar's time-space diary shows that he spends 71 percent of his time awake and not at work at home. Thus, while at first glance it appears that the parents' choices were based on logistical demands of family life, when we look more closely at the ways these parents organize their everyday time and spaces, we see that their choices were actually mostly made to obtain a home-oriented lifestyle and more time for the family to be together. According to Hilda, moving into town from the country has also contributed to more "family time" as well as more time with extended family and friends. Still, since Oskar often sits at the computer while home, sometimes for work and sometimes for pleasure, the spatial and temporal boundaries between work and family are not entirely clear. --- Conclusions Compact city policies and New Urbanism have in Sweden and other European countries resulted in the planning of urban neighborhoods without sufficient regard for the well-being of children and youth. In this article we have discussed results from a case study of how a middle-class family in a child-unfriendly inner-city neighborhood organizes time and space for outdoor play and institutionalized leisure. The ways children and parents within a family context "do" and "create" time and space for outdoor play and institutionalized leisure have not received enough attention in the literature. Our results show that parents and children have to adjust to certain conditions and time-spatial constraints in the domains of outdoor play, institutionalized leisure and family life, but that they also possess and can exercise agency regarding time and space within these domains. The neighborhood in which the study family lives does not meet the criteria for a child-friendly environment in terms of outdoor play space. While the parents value outdoor play as physical activity for their children, this factor did not steer their choice of neighborhood. Their middle-class status allows them to reach beyond the neighborhood for physical activity and institutionalized leisure. As a result, their daughter Anne's soccer activity has become an important authority constraint for family life in time and space. It has also become central to the identities and social life of her parents and the ways they "do parental time," while Anne herself enjoys the social dimension of organized soccer most of all. For her the soccer club is a social time and space. For her younger brother Tom, however, spending weekend days at his sister's soccer matches reduces social time with his own friends, reflecting the effect of institutionalized leisure on siblings' weekly paths. While Anne does not engage with the (for her) dull neighborhood, Tom recreates outdoor places into play spaces in a creative and physically intense manner. However, he also spends a lot of time indoors. The parents' construction of Tom as a physically very active outdoor child should be understood in relation to their daughter's limited independent play, her organized soccer activity and their own indoor and physically inactive everyday lives. In addition, although the parents are noticeably affected by the discourse of "concerted cultivation" (Lareau 2011) regarding institutionalized leisure (but not informal play) as an important part of children's general education, they justify Tom's disengagement by leaning on the discourse of physical health. So far Tom has been able to resist socialization into the middle-class practice of institutionalized leisure by being physically very active and creative while outdoors. By also refusing institutionalization in after-school clubs he simultaneously creates autonomous time for himself indoors. It is noteworthy that this family's interest in their daughter's leisure activity has not led them to engage in a "culture of busyness" (Karsten 2005). Although Anne's soccer activity is the hub of the family's time-space organization, they manage to maintain a home-oriented lifestyle at a rather relaxed pace and with plenty of time for recuperation and socializing with family and friends. They do not talk in terms of "deadlines" and "wiggle room" in the way Lareau's (2011) families did. This reflects the importance of understanding middle-class practices not as uniform but rather as --- diverse, depending on family and cultural contexts. On the other hand, if and when Tom's parents persuade him to join the soccer club, this family's "family time" at home will increasingly be converted into "family time" at the soccer club. Danielle van der Burgt, Ph.D. in Social and Economic Geography, is a postdoctoral researcher at the Department of Social and Economic Geography, Uppsala University, Sweden. Her main research interest is within the fields of social and urban geography, with a specific focus on children, youth and families and their relation to the built environment. She is particularly interested in children's and youths' everyday mobility and well-being in public

The New Urbanism trend has led to more Swedish middle-class families living in apartments in cities, with proximity to attractive schools and services but poor opportunities for outdoor play. Viewing families' and children's agency and living conditions as intimately connected with time-space variations, we investigate how children and parents in one urban family "do time" and "create space" in the domains of outdoor play and institutionalized leisure. Using a mixed-method approach, we find that within one family there are several ways of handling this. Initially a way of promoting physical health, the daughter's soccer activity has become the hub of the family's time-space organization and an important part of social life and identity for both the daughter and the parents. The son, refusing institutionalized activities, is considered a physically active outdoor child even though he spends more time indoors. When outside, he recreates child-unfriendly places in the neighborhood into spaces for play. We argue that using a mixedmethod approach gives insight into variations within families. The results of this case study show that parents and children have to adjust to certain conditions and constraints in the domains of outdoor play, institutionalized leisure and family life, but that they also possess and can exercise agency regarding time and space within these domains.

's soccer activity is the hub of the family's time-space organization, they manage to maintain a home-oriented lifestyle at a rather relaxed pace and with plenty of time for recuperation and socializing with family and friends. They do not talk in terms of "deadlines" and "wiggle room" in the way Lareau's (2011) families did. This reflects the importance of understanding middle-class practices not as uniform but rather as --- diverse, depending on family and cultural contexts. On the other hand, if and when Tom's parents persuade him to join the soccer club, this family's "family time" at home will increasingly be converted into "family time" at the soccer club. Danielle van der Burgt, Ph.D. in Social and Economic Geography, is a postdoctoral researcher at the Department of Social and Economic Geography, Uppsala University, Sweden. Her main research interest is within the fields of social and urban geography, with a specific focus on children, youth and families and their relation to the built environment. She is particularly interested in children's and youths' everyday mobility and well-being in public space. Other interests are emotional geographies of fear and risk, experiences of place, gender, parenting, segregation, and physical planning, as well as families' organization in time and space. In her research she uses mainly qualitative methods such as in-depth interviews and maps but also quantitative methods such as surveys and time-space diaries. Her current research projects focus on urban families' time-space organization of daily life and on the social construction of fear of urban violence in families with teenage children. She is coordinator of Children, Young People and the Built Environment, a Swedish national network for researchers and professionals. --- Katarina Gustafson, Ph.D. in Education, is a senior lecturer in Child and Youth Studies, Department of Education, Uppsala University, Sweden. Her research field is children's everyday lives and identity within families, schools and neighborhoods. She takes a childhood geographical approach, focusing on the importance of place in children's daily identity work. Along with children's agency and living conditions she is also interested in segregation, urban childhoods, special education, school markets, school transportation, families' organizations in time and space, and children's mobility and identity work. In her research, together with traditional ethnographical research methods, she also works with children's own documentations such as narratives, photos, maps, etc. She is part of the research group Studies in Childhood, Learning and Identities as Interactional Practices (CLIP).

The New Urbanism trend has led to more Swedish middle-class families living in apartments in cities, with proximity to attractive schools and services but poor opportunities for outdoor play. Viewing families' and children's agency and living conditions as intimately connected with time-space variations, we investigate how children and parents in one urban family "do time" and "create space" in the domains of outdoor play and institutionalized leisure. Using a mixed-method approach, we find that within one family there are several ways of handling this. Initially a way of promoting physical health, the daughter's soccer activity has become the hub of the family's time-space organization and an important part of social life and identity for both the daughter and the parents. The son, refusing institutionalized activities, is considered a physically active outdoor child even though he spends more time indoors. When outside, he recreates child-unfriendly places in the neighborhood into spaces for play. We argue that using a mixedmethod approach gives insight into variations within families. The results of this case study show that parents and children have to adjust to certain conditions and constraints in the domains of outdoor play, institutionalized leisure and family life, but that they also possess and can exercise agency regarding time and space within these domains.

INTRODUCTION The use of skill-mix in dentistry, where different dental team members employ different skills, is now well established in the UK and elsewhere. 1 Although the extent to which it is used varies internationally, the rationale for using skill-mix focuses on the potential for increasing access and efficiency of services. [2][3][4][5][6][7] In the UK it has been estimated that 70% of all visits and 60% of all clinical time in primary care could be provided by dental therapists. 8 This model of care has been given increasing prominence over the last two decades in the UK following positive published reports. 9,10 Subsequent legislative changes permitted dual-trained dental hygienists and therapists (referred to as Objectives To investigate public awareness and the social acceptability of dental treatment provided by dental therapists in the UK. Method A telephone survey of a representative quota sample of 1,000 UK adults. Results 10.4% of participants were aware of dental therapists as a professional group, of whom none correctly identifi ed their permitted duties. 61.3% were willing to receive simple restorative treatment from a therapist, with acceptability predicted by being male [OR 1.44 (95% CI 1.09-1.90)], being younger [OR 1.024 (1.016-1.032)] and having a perceived need for treatment [OR 1.49 (1.09-2.5)]. Fewer were willing to allow a therapist to restore a child's tooth (54.7%, p <unk>0.001, McNemar's test) with acceptability predicted by being younger [OR 1.026 (1.018-1.034)]. Those receiving some private treatment were less likely to report acceptability of simple restorative treatment for themselves [OR 0.61 (0.46-0.81)] or for children [OR 0.54 (0.41-0.72)]. 48.2% of participants expected to pay less for treatment provided by therapists, with acceptability of equal costs predicted by the participant being male [OR 1.81 (1. 50-2.40)]. Conclusion These fi ndings identify a need for education and reassurance of the public on the training and permitted duties of dental therapists and the rationale for skill-mix in dentistry. and their roles. This is particularly important now given recent recommendations to increase the use of skill-mix. 21 The term 'acceptability' is often used synonymously with'satisfaction'. However, the acceptability of a service or professional group should be conceptualised more broadly. People can only express satisfaction with a service or professional when they have experienced care. However, the views of people who have not experienced care become crucial if the use of a professional group is to be expanded. For example, healthcare quality assurance requires a service's social acceptability and legitimacy be considered as a key part of service quality assessment. 22,23 Similarly, the psychology literature refers to the importance of assessing the social validity of healthcare interventions where the social acceptability of the goals, procedures and outcomes of treatment is assessed. [24][25][26] Conceptually this can be applied to services or professional groups such as DCPs. In both, the views of those who have and have not experienced care should be sought. A preliminary study of the social acceptability of skill-mix in South Yorkshire reported low levels of awareness of dental RESEARCH therapists as a professional group and of their permitted duties. Although the acceptability of some procedures was relatively high, more invasive procedures and those provided for children were regarded as less acceptable. 27 In addition, qualitative data suggested a possible association between dental anxiety and lower levels of acceptability of skill-mix. 28 However, nothing is known of the prevalence of these views at a national level. Therefore the aim of this study was to investigate public awareness and social acceptability of the use of dental therapists in the UK. --- METHOD A telephone survey was undertaken by a market research company (GfkNOP) using structured interviews on a representative quota sample (n = 1,000) of UK adults (18 years and over). Precision estimates were undertaken to calculate the desired sample size. These were based on data from the earlier study 27 which identifi ed that 15% of participants were aware of dental therapists as a professional group. A sample of 1,000 was selected to provide a 95% confi dence interval that a population proportion of 15% would provide estimates <unk> 2.2%. This level of precision was regarded as suffi cient for the study. The content of the questionnaire was informed by a review of the literature, anecdotal reports and data from qualitative interviews. 27,28 Areas of inquiry included the following explanatory variables: participants' age, sex, socio-economic 29 and educational status; dental attendance patterns and access to care; perceived treatment need; and levels of dental anxiety (assessed by the Modifi ed Dental Anxiety Scale). 30 The outcome variables included awareness and knowledge of dental therapists and their permitted duties, acceptability of treatment provided by therapists for adults and children, and expectations of cost of treatment. Only closed questions were used. The questionnaire was piloted with 20 volunteers in face-to-face interviews and the market research company undertook further telephone piloting before commencement. Minimal modifi cations were requested. Potential participants were then telephoned out of normal working hours (5.30pm-9.00pm) by random dialling in postcodes to obtain a quota sample that was representative of the Offi ce for National Statistics mid-2005 population estimates for the UK. 31 The sample was weighted to ensure appropriate proportional representation of England, Scotland, Wales and Northern Ireland. To achieve a quota sample of 1,000 UK adults, 6,937 unique telephone numbers were called. Of the numbers called, 1,704 were called back as they were either busy or engaged (n = 466) or the participant requested the interview take place at a different time (n = 1,238). Only 91 calls resulted in no contact being made (due to wrong numbers, no answer after a number of call attempts, or the number being out of service). Of those contacted, 5,828 declined to participate and a further 18 stopped the interview. The subject matter, purpose and likely duration of the survey were explained. Potential participants were informed that they could decline involvement in the survey at any stage during or after the interview. Having been asked the questions --- RESEARCH therapists were able to extract deciduous teeth and 69.2% that they could administer local anaesthetic (Table 1). No participant correctly predicted all the permitted duties of dental therapists. Having received information on the role of therapists and their training, 61.3% of participants would be happy to receive simple restorative treatment but 20.4% would not be happy to receive any treatment from them (Table 2). Overall, participants regarded dental therapists providing care for children as less acceptable (Table 3). Fewer participants regarded dental therapists providing fi llings ( for children as acceptable than they did for themselves. More than half of participants found dental therapists extracting deciduous teeth to be unacceptable (Table 3). Thirty-nine percent of participants expected to pay the same for treatment provided by dental therapists but 48.2% would expect to pay less. Signifi cant predictors of the acceptability of care provided by dental therapists and its cost are reported in Table 4. Male participants, younger participants and those with perceived treatment need were more likely to fi nd having their tooth restored by a therapist acceptable. Younger participants were more likely to report a therapist restoring a child's tooth acceptable. Those receiving some private treatment were more likely to fi nd treatment provided by dental therapists unacceptable for themselves and for children. Men were more likely to accept paying the same (rather than less) for care provided by a therapist. Although levels of acceptability varied in different socioeconomic groups, chi-square test for trends did not identify a clear association. --- DISCUSSION The aim of this study was to investigate public awareness and social acceptability of the use of dental therapists in the UK. It found low levels of awareness of therapists as a professional group and of their permitted duties. Although more than half of participants accepted the prospect of being treated by therapists, signifi cantly fewer were willing for them to treat children. The low level of awareness of dental therapists reported here is similar to the fi ndings of the South Yorkshire study. 27,28 Before 2002, therapists were employed primarily in the salaried dental services. Consequently, despite the recent step change in the number of training places, it is unlikely that participants would have encountered them and this may explain the limited awareness. Indeed, no participant correctly identifi ed all of their permitted duties. The fi ndings of this and the South Yorkshire study suggest that participants were confusing dental therapists with dental hygienists. Our earlier qualitative study also suggested that procedures perceived as more invasive were regarded as less acceptable when performed by DCPs, where more emphasis was placed on the importance of qualifi cations, familiarity and trust in the clinician performing the task. 28 The proportions of participants in this study willing to accept local anaesthesia and simple restorative treatment were consistent with the South Yorkshire data 27 (61.3% v 57.0% and 68.5% v 64.0%, respectively). However, the proportion of participants not willing to have any treatment provided by a therapist differed (20.4% v 7.0% respectively). Younger participants were happier to receive treatment from dental therapists and for them to treat a child. The odds ratios (Table 4) indicate that approximately 2.6% more people found treatment provided by dental therapists for adults or children unacceptable for every year of life. Although high levels of satisfaction have been reported with skill-mix in general healthcare, [32][33][34][35][36] some older patients have expressed a preference for seeing the doctor in particular circumstances. 34 Lower acceptance of skill-mix in older age groups must be a consideration when delegating care. The earlier qualitative study had suggested that lower levels of acceptability of skill-mix might be observed in the more dentally anxious. 28 As previous studies have identifi ed that those who do not access dental services are often dentally anxious, [37][38][39] such an association would limit the use of skill-mix in increasing access to care. No association between dental anxiety and acceptability was found in this dataset. Logistic regression identifi ed that perceived need for treatment predicted acceptability of receiving dental therapy treatment about demographic data and the duties of a dental therapist, all participants were provided basic information about therapists indicating that they were professionally trained to undertake certain tasks once a dentist had examined the patient and prescribed particular treatments. All interviews took less than ten minutes to complete. The analytical strategy aimed to identify putative associations between explanatory variables and the fi ve outcome variables. Data were analysed in two phases. Initially descriptive and appropriate bivariate analyses (chi-square and McNemar tests) were undertaken to describe and compare key descriptor variables. The second phase identifi ed explanatory variables that predicted the main outcome variables. As the key difference between dental hygienists' and dental therapists' competencies is the provision of simple restorative care, willingness to have simple fi llings performed by a therapist was used as an outcome measure of the acceptability of care provided by them. Bivariate analyses (chi-square tests) of the possible predictors of the acceptability of providing this treatment were used to pre-select variables for forward stepwise logistic regression models. All variables with a relationship of p <unk>0.2 were entered into the models. All analyses were performed using SPSS version 15 and alpha level was set to 0.05. Ethical approval for the study was granted by the University of Sheffi eld, UK. --- RESULTS Of the 1,000 participants, 486 (48.6%) were male. Their mean age was 47.0 years and 28.7% had a child under 16 years of age. Seventy-one percent attended regularly for check-ups and 28.3% perceived they had treatment need. Of those reporting diffi culty with access to a dentist (10.1%), 70.4% cited a lack of local availability of a National Health Service (NHS) dentist as the cause. Overall, 10.4% of participants said they were aware of dental therapists as a professional group. Unsurprisingly, participants with relatives or friends who had worked in a dental team were more aware of dental therapists (42% v 13%, p <unk>0.001, McNemar's test). No other variable predicted awareness. Of those that had heard of a therapist (n = 104), only 38 (36.5%) thought that RESEARCH (Table 4). Bivariate analyses had suggested that irregular attendance and perceived need were associated with acceptability, but lack of access to care was not, suggesting some confounding with irregular attendance. Taken together, these fi ndings are encouraging if skill-mix is to be used in areas where access to dental services is poor and perceived needs are high. The survey participants perceived that dental therapists providing treatment for children was less acceptable than for adults. This fi nding is consistent with the preliminary study, 27,28 although levels of acceptability were slightly higher nationally. Almost half of participants would not want dental therapists to provide restorative care, administer local anaesthetic or extract a child's deciduous tooth. Given that dental therapists are often employed to treat children and adolescents in the UK and elsewhere, 40,41 this fi nding is important. Our qualitative data suggest that more negative views about dental therapists treating children were related to assumptions of inexperience and questionable technical competence. However, positive experiences of being treated by therapists could modify views. 28 Those within the dental profession opposed to the use of dental therapists have expressed concerns about DCPs treating children. 42,43 If skill-mix is to be used to increase access to care for children, it is important that these views are considered. Reassurances about dental therapists' training, qualifi cations, regulation by the General Dental Council, competence and supervision requirements need to be carefully communicated, emphasising that the quality of care received should be at least as good as that provided by a dentist. Similarly, care in the communication of the rationale for the use of skill-mix is required. Signifi cantly, both this study and the preliminary study found that younger participants were more likely to fi nd dental therapists treating children as acceptable, which is encouraging if skill-mix is to be used more in the future, especially as older adults are less likely to have children. Interestingly, those who received some private dental treatment were less likely to accept dental therapists. These fi ndings converge with our qualitative data, 28 which suggested a spectrum of views being held on the nature of dental services, ranging from a private service view of dentistry to a public service view. Those with a public service view tended to be more positive about skill-mix, seeing the potential for the effi cient use of resources in an NHS service. Conversely, those with a private service view often took the standpoint of a consumer, preferring to be treated by a dentist if the cost for treatment would be unaltered. Such fi ndings have profound implications for the employment of therapists in the treatment of adults and children if the proportion of care delivered on private contract continues to increase, but also in the NHS where charges relate to the treatment required rather than the dental team member providing care. Such consumerist concerns have been expressed elsewhere in the belief that the use of dental therapists would lead to a two-tier dental service, where treatment by dentists is reserved for those that can afford it. 43 Our earlier qualitative study 28 suggested that dental and medical services are perceived differently -those with a public service view of medical services often held more consumerist views of NHS dentistry owing to the patient charges levied for treatment and some welcomed the introduction of a scale of fees relating to the clinician providing care. This study has identifi ed that more participants would expect to pay less for treatment from therapists than those who would expect to pay the same. A similar fi nding was reported in the South Yorkshire study. Evaluation of health services increasingly emphasises the importance of user views, 22,23,[44][45][46] the most contemporary of which go beyond evaluating patient satisfaction with services experienced. The concepts of social acceptability 22 and social validity 24 require a broader assessment of the social desirability and appropriateness of a service and its outcomes. The fi ndings of this study, consistent with the earlier study, raise questions about the current social acceptability of the use of dental therapists in the UK. They have identifi ed that awareness of dental therapists as a professional group and their permitted duties is low. Therefore, it is doubtful whether many would be able to give informed consent for treatment provided by therapists should perceptions remain unchanged. There is a role for both the Government and the profession to communicate the rationale for using skill-mix in terms of increasing effi ciency and access while maintaining its effectiveness and thus improving the service's quality overall. 23 Our qualitative data suggested that once the rationale for skill-mix is explained, views on the use of dental therapists were modifi ed. 28 Similar arguments could also be used by dentists solely working in the private sector, if effi ciency improvements are refl ected in patient charges to patients. 28 Research into the potential for the use of skill-mix in private care would also be benefi cial. Telephone surveys have been used in national dental surveys 47 and are frequently used in health and social surveys, particularly in North America. [48][49][50] Although every effort was made by the market research company to achieve a representative sample, like all surveys, a risk of sampling bias exists which tends to result in an overrepresentation of the views of white participants of higher income and educational attainment. 51 This telephone survey reported 26.6% of participants as having perceived need for treatment, which is similar to the proportion (25.4%) reporting perceived need in a recent largescale postal survey of adults (n = 10,864), 37 suggesting that the impact of any sampling and response bias was similar in both. This study and the earlier preliminary study have identifi ed a number of areas for future inquiry. Further research is needed into the broader aspects of the quality of services provided by dental therapists in terms of their effi ciency, effectiveness, their impact on equity and accessibility and their acceptability. 22,23 Research into their acceptability should include an investigation into the views of patients, parents/ guardians and children on the experiential acceptability of care provided by dental therapists. The theoretical and methodological diffi culties of such research are well documented 52,53 and mixed-method approaches have been recommended. 54,55 Given the increasing proportion of provision of dental services in the private sector, models of how skill-mix can be best employed in both NHS and private practice should be explored and how the delegation of care is best undertaken and communicated in both should be investigated. --- CONCLUSION This study confi rmed that the awareness of dental therapists and their permitted duties was low across the UK. More than --- half of participants accepted the prospect of treatment provided by dental therapists; signifi cantly fewer were willing for them to treat children. A fi fth of participants would not be willing to receive any treatment from a therapist. Common predictors of acceptability were being younger and if the participant had a perceived need for treatment. More participants expected to pay less for treatment provided by dental therapists than those who expected equal costs. Such factors need to be considered, as skill-mix is to be increasingly used in the delivery of dental services. The authors would like to thank Irene Quoico of GfkNOP and Dr Sarah Baker for her support with the statistical analysis.

dental therapists in this paper) to work in all sectors of dentistry 11 and the number of training places for dental therapists has increased signifi cantly. Recently the General Dental Council (GDC) has defi ned the permitted duties of all members of the dental team. As well as describing the core treatments and procedures each professional group may conduct under the prescription of a dentist, a wide range of additional treatments that appropriately trained dental care professionals (DCPs) may perform was identifi ed. 12 Such skills potentially extend the list of permitted duties for each type of DCP. Sequential UK studies indicate increasing acceptance by dentists of the use of skill-mix, [13][14][15] although there is still ignorance of dental team members' roles. 14,15 However, the views of patients and public are missing from any consideration of potential changes in service provision. US data suggest patient satisfaction with care provided by DCPs, [16][17][18][19] however these cannot necessarily be generalised to the UK. Indeed, the acceptability of care provided by DCPs has been identifi ed as a priority for research. 20 Furthermore, little is known of the public awareness of dental therapists

health promotion interventions requires collaboration with communities. 1,2,3 Thus, collaboration was formed between health behaviour scientists and CALD community leaders, advocates and bicultural workers (hereafter referred to as 'community leaders') to discuss the challenges faced by CALD communities in communicating COVID-19 health messages, and to develop evidence-based strategies to inform and empower CALD communities to adapt their behaviours in line with COVID-19 recommendations. --- Participatory research methods CALD community leaders, advocates and bicultural health workers have a wealth of knowledge and experience working across diverse communities in Australia. Developing research collaborations with those who have lived experience of the issue being studied is now a well-established method in health research, under the banner of 'participatory research'. 3 Using methods of participatory research, this paper has been co-produced by academics and CALD community leaders from end to end: from data collection through to the co-production of an academic manuscript. Purposive and snowball sampling were used to invite CALD community leaders to collaborate. The academic authors initially invited leaders known to them and those contributing to the public conversation about the issue in the media to take part. They then asked the core sample of seven who had agreed to participate to invite other CALD community leaders with relevant experience to take part. Five semi-structured dialogues were held online via Zoom teleconference in July 2020 with various combinations of CALD community leaders (12 in total), focusing on the following three questions: Australia documented its first confirmed case of coronavirus disease 2019 (COVID-19) on 25 January 2020, and nearly 500 cases were being reported daily by late March. The Federal Government promptly adopted a controlled adaptation strategy and, as of July 2020, every state and territory in Australia appeared to have contained the virus, except for Victoria. After Victoria relaxed restrictions in late May, metropolitan Melbourne and the Mitchell Shire located north of Melbourne experienced a second wave of COVID-19 in July 2020 and restrictions were reintroduced. The experiences during the second wave highlighted the need to engage culturally and linguistically diverse (CALD) communities. Widespread outbreaks of COVID-19 in several public housing towers, populated largely by people from CALD backgrounds, suggested that attempts to communicate the recommendations of the Chief Medical Officer may not have reached and/or been understood by all CALD community members. It is often assumed that providing people with information and explaining why people should perform a behaviour will result in compliance. 1 However, behavioural science has demonstrated that achieving desirable behaviours in a target population is accomplished by understanding the needs of the population and the barriers and enablers to behaviour change. Effective services and interventions draw upon evidence-based practices while taking local context into account. However it is not just what goes into the program, but how it is delivered that matters: implementation science offers techniques to embed evidence-informed strategies into frontline delivery. Designing and implementing successful groups, and worked with all participants to develop insights that were applicable to the many different CALD communities represented in the research collaboration. Results: Three key findings emerged: 1) partnerships between CALD leaders, communities and government are critical for effective health communication; 2) shifting behaviour requires moving beyond disseminating information to designing tailored solutions; and 3) the diverse needs and circumstances of people and communities must be at the centre of health communication and behaviour change strategies. --- Lessons learnt: The collaborative process we undertook in this study enabled us to identify key challenges experienced and solutions offered by CALD leaders in communicating health information throughout the COVID-19 pandemic. Partnering with communities that are subject to health messaging can reduce inequalities in healthcare communication by enabling the development of strategies that help align human behaviour with the recommendations of health experts. This -along with sustained partnership and collaboration with CALD communities, understanding the cultural context, and the appropriate tailoring and delivery of communications -will ensure health-related messages are not lost in translation. The lessons provided in this paper are applicable not only to the current pandemic but also to postpandemic social and economic recovery. These challenges demonstrate the complexity of developing and implementing mass public health communication strategies during public health crises. 5,6 Messages need to be tailored when developed, so they are presented in a meaningful, relevant and applicable way to all population groups. This will be most effectively achieved by codesigning solutions with consumer and community involvement. 7 What are the elements of effective messaging for CALD communities? --- Key points Below we outline solutions offered from CALD community leaders, advocates and bicultural health workers that have been synthesised from the participatory process. In order to see what insights applied across the varied experiences of CALD communities, community leaders' responses were recorded for three topics that were the focus of the dialogues: 1) disseminating messages; 2) designing messages; 3) building trust between communities and government. Solutions were synthesised through an iterative mapping exercise using Miro, a collaborative online whiteboard tool (San Francisco, CA: Miro). Community leaders generated a range of recommendations, for example, about how COVID-19 messages could be disseminated more effectively, and these recommendations were sorted by AW in an initial map (Figure 1a, available from: doi.org/10.26180/14122295. v1). The second phase of mapping added explanations community leaders provided about why each of these recommendations would be effective, and drew connections between specific examples and the principles underlying their effectiveness (Figure 1b, available from: doi.org/10.26180/14122295.v1). This mapping process was repeated with community leaders' responses regarding how messages could be designed more effectively (Figure 2, available from: doi. org/10.26180/14122295.v1) and how trust could be built between communities and government (Figure 3, available from: doi.org/10.26180/14122295.v1). To arrive at practical recommendations, authors then prioritised principles of effectiveness according to three criteria: 1) how many specific examples were linked to an explanatory principle (as indicated by arrows in Figure 1b); 2) the generalisability of a principle across CALD communities; 3) consistency with relevant research evidence. Five recommendations for health officials were distilled from this prioritisation process: 1) partner with communities to tailor messages; 2) use trusted messengers; 3) communicate via appropriate channels; 4) avoid blame and stigma and 5) establish structures for meaningful partnership, such as an advisory body. These recommendations are described below. • What has been done by the government to communicate with or engage your communities about COVID-19? • What needs to happen to make sure that messages effectively reach diverse communities, given that the COVID-19 situation is rapidly changing and messages need to be delivered quickly? • How can we use technology to effectively communicate with people from diverse backgrounds? After completion of the audio-recorded dialogues (mean length 60 minutes), transcripts of the dialogues were analysed by AW and HS using inductive, interpretivist qualitative methods to identify key themes. 4 This grounded-theory approach was used to identify key themes which were circulated to all in the research collaboration for further refinement before finalisation. A 'live' online document was created to allow for multiple authors to contribute to the paper in conjunction with email consultations at four points: first, a draft prepared by AW, AK, DG, and HS was shared with all co-authors for further input/comment; next, initial feedback was incorporated into the paper by AW, and a revised draft was circulated for three more rounds of comment and approval. This iterative process allowed for collaboration throughout analysis and interpretation. The following describes the insights gleaned from this process. --- Communicating COVID-19 health messages: challenges faced by CALD communities While CALD community leaders commended the efforts of the federal and state governments to combat COVID-19, they suggested that challenges remain when communicating public health messages about the pandemic. Challenges may occur due to a lack of the following: 1. Translation of information into all of the various languages spoken by community members 2. Testing of translated materials by CALD groups to ensure that messages are understood by people who might also have limited health literacy 3. Tailoring translation and messages to specific communities, thereby contextualising messages 4. Using trusted messengers to deliver information (e.g. community leaders, advocates) --- Overcoming barriers Drawing upon the principles recommended above, the authors considered how the barriers to effective communication identified at the beginning of this article might be overcome. A number of solutions were generated during the dialogues, which were refined during collaborative writing, to arrive at the following possible solutions: 1) Providing CALD community leaders with effective channels for contacting health authorities, to ask questions or request additional messaging support when needed (e.g. how to adjust behaviour during religious ceremonies or community events) 2) Preceding any major announcement by convening a diverse group of CALD representatives that could flag any major logistical and cultural considerations in advance and suggest solutions (e.g. importance of sending female health workers for door-to-door swab testing; culturally appropriate food support for migrant families in lockdown) 3) Deputising and resourcing community leaders and health advocates to take the lead on translation and health communication within their communities 4) Collaboration with ethnic language media to convey important health, safety and community messages 5) Reporting measures of ethnicity to the National Notifiable Diseases Surveillance System 6) Having proactive solution-focused consultations, rather than reactive consultations in response to a crisis, to create opportunities for 'collective --- Recommendations for health officials 1) Beyond translation of health messages: why partnership and tailoring messages to community values are important Translating COVID-19 health information and prevention strategies into different languages is necessary, however information alone does not lead to behaviour change. 1 The process of language translation must also consider the audience's frame of reference, and the context in which the translated material will be used, which can be done by partnering with community leaders to understand the barriers to understanding and adopting the information. 8 For example, working with CALD representatives will ensure that information is translated into the most appropriate dialects and that it addresses common misconceptions (e.g. the cost of testing) and that difficulties in adhering to disseminated health information can be identified. --- 2) Translated messages must be delivered by trusted messengers Translated information should be delivered using trusted messengers, as the effectiveness of any message depends upon how an audience perceives the credibility of its source. 9,10 For many of the CALD communities in Australia, religious practice is important, and faith leaders can have great impact when promoting and modelling desired behaviours. 11 3) Messages need to be delivered using appropriate and accessible channels Recognising that health communicators are competing for the attention of their audience, information should be delivered in ways that are accessible and appropriate to the target community. 12,13 For example, the WeChat platform plays an important role in disseminating COVID-19 information to the Chinese community in Australia, providing up-to-date information in Chinese languages. Similarly, other CALD communities rely on community message boards, ethnic language print, radio and TV media. --- 4) Blame and stigmatisation must be avoided Infectious disease outbreaks create feelings of fear that can exacerbate racist and xenophobic behaviour, as has been documented during the COVID-19 pandemic. 14 Care should be taken to ensure that extra efforts to reach and support communities that are disproportionately affected by COVID-19 are not interpreted as singling out particular groups as more likely to spread disease. Not only could this increase instances of racism, it would be --- Conclusion Bringing together CALD community leaders and health behaviour researchers enabled us to synthesise recommendations that draw on insights from behavioural and implementation science, as well as the lived experience and professional expertise of community members. A participatory, collaborative research approach enabled us to identify barriers to communicating COVID-19 directives to CALD communities in Australia, and to codesign possible solutions to overcome those barriers through an iterative process of dialogue, analysis and synthesis. The research found that sustained partnerships and collaboration with CALD communities, both directly and through respected leaders and organisations, will enable more effective communication strategies. Working in this way provides the opportunity for timely responses to urgent public health needs, such as those experienced during the COVID-19 pandemic. --- Peer review and provenance Externally peer reviewed, not commissioned. --- Competing interests None declared. --- Author contributions AW, BK, DG and HS drafted the article, assisted by LZ, MK, WS, FM, MM, MA, MK, MG, DC, EC and EM who are community leaders, working with culturally and linguistically diverse communities during the COVID-19 pandemic. All authors read and checked different versions of the paper as it was developed, contributing to the interpretation of data, the intellectual content of the paper, and providing final approval.

Objective: To consider the challenges of communicating COVID-19 directives to culturally and linguistically diverse (CALD) communities in Australia, and present evidence-based solutions to influence policy and practice on promoting relevant health behaviours; to advance participatory research methodologies for health behaviour change.We present a case study of a participatory research collaboration between CALD community leaders and health behaviour change scientists during the COVID-19 crisis. The goal was to better understand the role of community leaders in shaping health behaviours in their communities and how that role might be leveraged for better health outcomes. Methods: This article is the culmination of a series of dialogues between CALD community and advocacy leaders, and health behaviour change scientists in July 2020. The academic authors recruited 12 prominent CALD community leaders, conducted five semi-structured dialogues with small

INTRODUCTION Violence against women permeates the history of humanity from its earliest days. However, three decades ago it started to be seen as an inherent problem of the health sector, and it is currently understood as a public health problem and a violation of women's human rights (1), affecting them as regards quality of life and health. It is conceptualized by the Convention of Belém do Pará (1994) as "any action or conduct based on gender that causes death, harm or physical, sexual or psychological suffering to women, both in the public and private spheres" (2). Situating the rural context, situations of violence against rural women assume different facets, which can be visualized through daily social relations, exemplified by the subordination of women to their husbands and/or partners, male domination, and sexual division of unequal work (3). In addition, a study carried out in cities of Rio Grande do Sul, which aimed at analyzing the conceptions of domestic violence against rural women in the expression of Community Health Agents, pointed out that women living in rural areas are permanently exposed to different forms of violence. In addition to physical violence, psychological, moral and sexual aggressions, as well as the overload of work and deprivation of freedom on the part of the husband and/ or partner, are mentioned (4). In the case of the elderly women, this population is more vulnerable to experiencing situations such as those mentioned above, due to the low educational level, physical limitations due to diseases and the aging process, besides the subjective aspects of the elderly rural women (5)(6). Another aspect that contributes to the non-confrontation of violence against women is the absence of collective resources for the social care and protection in the rural settings, together with the geographical distance from the urban centers (7). Most of the time, rural women do not have access to police stations, health services and social assistance, for example; and they end up not reporting violence situations, which contributes to the invisibility of the problem. In view of the above, discussing the issue of violence against elderly women in the rural context is relevant, since there are difficulties in identifying the occurrence of these events, both by women and other segments of society, especially in the health sector. Giving a voice to this population with generational and cultural particularities can contribute as a guideline for the development of strategies for coping with and preventing violence in rural areas, as well as for the (re) construction of public policies in this area. In this perspective, the theory of Social Representations is inserted as a theoretical and methodological set allowing a better understanding of the object under study, understanding it as a social phenomenon. The Theory of Social Representations is a typical product of present times and of paradigmatic transition, as well as the feminist and other theories that, in their own way, emerge as new conceptual tools to analyze angles of reality highlighted by new looks (8). In view of the exposed problem, this study was guided by the following research question: What are the social representations of violence against rural women, from the perspective of the elderly in the rural context? In order to answer to this question, this study aimed at understanding the Social Representations of violence against rural women, from the perspective of the elderly, considering generation and gender as an influence in this aggravation. --- METHOD This is an exploratory-descriptive study, with a qualitative approach. The qualitative method applies to the study of history, relationships, representations, perceptions, and opinions resulting from humans' view of themselves, their ways of living and building their artifacts, as well as their feelings and thoughts (9). The Theory of Social Representations has been chosen as a theoretical-methodological reference. This reference makes it possible to make something unfamiliar, or even the non-familiarity, and to this end it uses two mechanisms called projective techniques: anchorage and objectification, which act on the basis of memory (8). The study was developed with elderly women living in rural areas, linked to the rural Family Health Strategy (FHS) of a city located in the northwest region of Rio Grande do Sul, in the months of July and August of 2013. Previous contact was made with the nurse responsible for the rural FHS, who, through a Community Health Agent (CHA), has organized a workshop with elderly women living in rural areas. The participants of the study were 12 elderly women residing in the rural area of the study scenario city. The following inclusion criteria have been used: to be aged 60 years or over, and to present cognitive conditions to respond to the interview. For the generation of data, projective techniques have been chosen (8), and semi-structured interviews have been used. With the projective techniques, a collage workshop was organized and held in a community hall suggested by the CHA as a place more geographically accessible to the elderly participants. In this, magazines and newspapers were made available to women, so that they could express, through the use of images (cut and pasted), thoughts and Rev Ga<unk>cha Enferm. 2017;38(4):e68209 meanings about violence against rural women, thus, allowing the objectification. For the anchorage, the elders were asked to verbally present their knowledge about the phenomenon. The workshop was recorded in audio, with an average duration of two hours. The observations were recorded in field diary by the researcher and a research assistant, academic of the undergraduate course in Nursing. Also, in order to complement the data obtained through the workshop, semi-structured interviews have been conducted with the 12 elderly women who participated in the workshop. The sample size has been defined by the data exhaustion criterion, that is, the closure of data collection through the interviews occurred when all the eligible participants participated in the study (10). For the interviews, an operative script has been divided into two parts: one related to the sociodemographic data of the elderly, containing closed questions; and the other part consisting of open questions regarding the object of the study, that is, the social representations of violence against rural women. Thus, during the semi-structured interviews, so that women could verbalize about situations of violence against rural women, it was necessary to instigate them from questions such as: What do you understand by violence against rural women? How do you conceptualize violence against rural women? Are you aware of the forms (types) of violence against existing rural women? Do you know or have you witnessed any type of violence against women in rural areas? It should be emphasized that the semi-structured interviews had previous scheduling, being carried out individually in the house of the participants, with an average duration of one hour. The home environment has been chosen as the locus for the interviews, considering that the elderly women could feel more at ease and comfortable to answer the questions at their own environment. During the data collection period, other people, especially neighbors, spouses and/or companions were sometimes present at the women's home for the interview. In these situations, the interviews were not carried out, since the presence of another person violates the ethical principles of the study development, as well as could interfere in the answers, besides exposing the elderly women, considering that violence against women is an issue that is surrounded by censorship. In these situations, the researcher rescheduled a new date for the interview, in order to perform it when the elderly woman was alone at home. It is important to observe that the interviews have been recorded in digital media after the consent of the elderly, guaranteeing the authenticity of the data for the analysis. It should be highlighted that prior to the data collection procedure, the participants have been clarified about the objectives, methodology, risks and benefits of the study, and the researcher read the Term of Free and Informed Consent (TFIC). After this procedure, all the elderly women who agreed to participate in the study signed the TFIC in two copies, remaining one with the participant, and the other with the researcher responsible. In order to ensure the participants' anonymity, they were identified by the letter I followed by the ordinal number according to the order of the interview (E.g.: I1, I2, I3... I12). The data obtained through the workshop has been analyzed in the light of the theoretical reference of the Social Representations, based on the methodological steps (11) below. The 1 st moment understood as objectification: through the process of making something thought into something real, through the collage, it was possible to form the figurative nuclei, which correspond to the themes that have the property of being represented and that express the veracity of ideas; 2 nd moment -anchorage: it corresponds to the interpretation of the elderly about the material generated in the collages, assigning a meaning to the images produced by them. Such a moment allows the collection of symbolizations and perceptions from the verbal descriptions about the knowledge regarding the phenomenon in question; later, the themes that represent this knowledge have been identified, constituting the symbolic nuclei. The 3 rd momentvalidation by the participants of the configurations given by the researcher to the social representation of violence against elderly rural women, confirming or not the identification of nuclei of this representation (11). The 4 th moment provides for the systematization of information, grouping the main themes into figurative and symbolic nuclei (8). Then, in the light of the theoretical reference, the anchoring about the social representation of violence against rural women for the elderly has been obtained. The semi-structured interviews have been submitted to the content analysis proposed by Minayo, which is divided into three stages: pre-analysis, exploration of the material and treatment of the results obtained and its interpretation (9). Regarding the pre-analysis, it began with the organization of the material collected for further analysis. Initially, the complete transcription of the data obtained from the audio recordings of the workshop and semi-structured interviews has been carried out in the Word text editor, producing the corpus of the study. Afterwards, this material was printed, to facilitate the manipulation by the researcher. Then, the listening of the recordings was carried out, as well as the floating reading, Rev Ga<unk>cha Enferm. 2017;38(4):e68209 which allowed generating the initial impressions about the material to be analyzed, coming from the two techniques of data collection. Continuing the analysis, a sequence of detailed readings has been made, from which it was possible to highlight, with a text-mark brush, the excerpts in which the speeches of the participants were similar. The exploration of the material was carried out from the segmentation of similar information contained in the transcribed material, which provided the constitution of the themes, from which the registration units emerged. These units are associated with the meaning of the phrase, which can be obtained through words, phrases and expressions (9). Then, after the registration units have been defined, it was possible to categorize the three thematic categories through the condensation of the recurrent speeches and the isolated speeches, both relevant to the study. During the treatment phase of the results obtained and interpretation, the researcher proposed inferences and performed interpretations about the results, based on the research objective. The research originates from a course completion work (12), and it has obeyed the ethical recommendations in force for conducting research involving human beings. The study has received a favorable opinion from the Research Ethics Committee of the Universidade Federal of Santa Maria (UFSM) with the protocol No. 17148813.0.0000.5346. Therefore, all the recommendations of the Resolution No. 466/12 of the National Health Council, which regulates the development of research with human beings, have been respected. --- RESULTS E DISCUSSION Regarding the sociodemographic characterization of the elderly women, the age ranged from 62 to 83 years old, with an average of 68.83 years old. As for marital status, eight were married or had a partner, and four were widows. With regard to schooling, nine elderly women had incomplete primary education and three had not attended school at any point in their lives. These data converge with a nationwide survey, which pointed to the fact that the majority of the rural women over the age of 60 are not literate or have only incomplete elementary education (5). Also, in the direction of these findings, an international study revealed that the level of education is closely linked to situations of violence, so the lower the educational level, the greater the chances of experiencing situations of violence (13). The discussions obtained through the analysis of the empirical material collected during the workshop and the semi-structured interviews have been grouped into three categories: violence anchored in extreme acts of corporal ag-gression and distant from the rural context, violence anchored in male domination and gender culture, and violence anchored in the context of life and work. --- Violence anchored in extreme acts of corporal aggression and distant from the rural context The first meaningful nucleus identified in this study presents the issues related to violence in the rural context, exposed by the elderly as events distant from their reality and visualized in most of the speeches by extreme acts of corporal aggression, signaling the difficulty of recognizing the problematic in this context. The speeches express their representations: A lot happens, but it happens more in the big city [...] we live in the countryside, there are no cases here. (I4) There is a lot of violence against women, but here, thank God, there is nothing to tell about it, because there are no cases. Violence against women exists a lot, but we are not aware. (I9) What we see is the boyfriend killing the woman, [...] we see it on television, there are fights at home and they also kill the woman. (I12) --- [...]There are many who also beat the woman [...] we watch on television. (I1) The testimonies reveal the representation of violence against women anchored in the social dimension, in which it occurs daily. However, as it is not understood as a violence, it becomes banalized and naturalized by the elderly, since this event is a common phenomenon, introjected to their daily life and that is often overlooked. In this sense, a study developed with rural Pakistani women has revealed the different forms of violence suffered by them. They are understood by psychological, sexual and physical violence. These, in turn, contribute to the findings of the study in question, since violence in the rural context is understood as intrinsic to their daily routine, that is, they do not perceive that they are being abused (13). In this context, identifying situations of violence becomes more difficult, since the women who experience it, most of the time, do not identify that they are being abused. Thus, they perceive this violence as something natural, which is sometimes present in their lives since childhood, from relationships in the family environment, for example (14). In this sense, a study has identified that women living in Rev Ga<unk>cha Enferm. 2017;38(4):e68209 the rural context were treated with brutality by their companions, however, they did not understand such brutality as violence and, rather, as a common behavior that was part of their daily lives (5). And the images they presented in the collage workshop represented the violence anchored in extreme acts, related to death and the great centers. If we lived in the big city, I think it would be a lot, kids kill parents, women get raped, but not here, there are no such things here. (I4) --- Violence is rape and death, this is the worst thing, because either you surrender or give your body or you are dead. (I2) It shows on the television those women full of bruises because the husband had hit them. [...]. (I11) The elements reveal the representation of violence as associated with violent attitudes, such as death, rape, and its occurrence as a phenomenon of the urban world, not being visualized in the daily context of the elderly women living in rural areas. A study, developed with managers in the cities of the southern half of Rio Grande do Sul, has pointed out that they also related violence to large centers and poor regions, confirming that for some managers there was no violence in rural areas (3). These data (re)think about the non-recognition of violence in the daily lives of people living in rural areas, and its invisibility as a public health problem. --- Violence anchored in male domination and gender culture In this nucleus of meaning, contradictory aspects have been identified, because although in the first the elderly rural women represent violence against women as a phenomenon distant from the rural context, in this nucleus they mention marks of symbolic violence and gender asymmetries in their experiences, mainly related to power and male domination. The statements of the elderly revealed that the male--female relationship in the rural context is anchored in chauvinism, in which man is given the role of sovereign and dominator, holder of physical strength and other characteristics representing power. But the woman is "graced" with attributes related to fragility and submission, in which she can hardly manifest and challenge her companion. On the other hand, the elderly women recognized this behavior as belonging to the past, mentioning achievements such as employment and retirement, which were important for the economic development of women. The-se elements can be identified in the following statements: The woman is humiliated, the woman suffers all kinds of humiliation... it seems that the man is becoming more manly. (I8) In the past women had nothing, there was no retirement, they had no job [...] if the husband said: "I'm the one who orders!", the woman would lower her head, she would never said anything, because the man was the owner, the one that ordered. (I4) Over the years the woman began to take over.. ---. she no longer admits being lowered by men. (I6) The speeches reveal the representation of violence, which are anchored in the psychological and moral perspective, considering that for women many deprivations are imposed when compared to men. Such representation is anchored in the ideology that man is the one who has power in the context of the conjugal relationship, and the woman, as a passive subject, must accept all that is imposed, without questioning. A study, which sought to outline the current generation of older women in Pernambuco, indicates that women in the rural context are associated with responsibility to the family, children and the home, a context in which, according to the women studied, "to suffer a lot" is part of their destiny, because they are women (5). It reveals the gender violence, which is a problem that transposes the barriers of the domestic context, involving the community and the work. The relationships in society are permeated by the structural asymmetries of gender, in which the reproduction of the relations of subordination of the woman to the detriment of the man is visualized. The older women living in rural areas are seen from the perspective of work, subordination and obedience. The speeches demonstrate their role, as housewives, mothers, care takers and workers: --- She had to work, and would be hit if she would not [...]. (I9) She just has a few cows, and she defends herself with it. Only with that. Because she depends on the husband. A little is this [...] but a woman who works, who takes care of the house, the children, why be so abused? (I8) The dimensions expressed by the participants regarding situations of violence between rural men and women reveal the power and authority of man as a provider and Rev Ga<unk>cha Enferm. 2017;38(4):e68209 head of the house. But the rural woman is seen from the perspective of subordination, obedience, and service (15). Regarding the division of labor in the rural context, this occurs in an asymmetrical way, in which the woman, by assisting her husband in the field, is seen as the "wife of the producer", and her work is represented as aid, favoring; thus, favoring the invisibility of the work and the lack of feminine professional identity (16). This aspect, elucidated by the speeches, points out that most rural women are financially dependent on their partners, and they are not encouraged to seek other means to provide for their livelihoods. In this way, they feel obliged to continue to bear the different manifestations of violence, in addition to the violation of their rights (4). It is observed in the expression of these women discontentment by the absence of recognition of the domestic work on the part of their companions. Violence is seen as an old problem, which presents itself in different forms, based on patriarchal representations of women, thereby compromising relationships, highlighting the inequalities of power, as well as the threat to the physical, moral and psychological integrity of women (17). In this direction, the home environment is the space in which older women suffer more abuse when compared to men. And considering that the phenomenon of gender violence against the elderly occurs primarily at the heart of the family, in Brazil there are still no records about this kind of violence, given that issues such as these are not identified as a social phenomenon (18). Also, it is possible to visualize the aspects related to the influence of culture and generation on the reproduction of violence against the elderly women residing in rural areas. The speeches represent the above: I remember that since my family, my parents, that the father was the macho, he was the boss, so the mother was quiet [...], because she had no time for anything [...](I4) [...] the neighbors were generally all the same... they were in charge of women and they obeyed [...] --- this is also violence. (I5) It is observed in the narrative that the elderly women represent violence in family relations, in which the role of women in this space is based on submission and coercion, and there is no possibility of challenging the facts. This is anchored in the reproduction of violence taking into account culturally imposed ideologies. Corroborating with these findings, a study with 354 rural Nicaraguan women reveals that 20% of these women have observed situa-tions of physical, psychological or sexual violence against their mothers (19). In many instances, the representation of violence against women in the rural context anchored in gender heterogeneities is seen, in which the social role of women is restricted to the home environment, based on the relationships of subordination and obedience, whereas man is seen as the reference in the family unit. These facts reaffirm the gender stereotypes already instituted, derived from social and cultural ideologies. --- Violence anchored in the context of life and work The third nucleus of meaning reveals the representational elements directed to the life and work context of women living in rural areas, which shows the deprivation of the right to study, depending on the work. The speeches express these representations: Those years we did not study. They did not want us to study, we had to work. (I3) I studied until the middle of the year of the third grade, then,my father took me to work in the field, because at that time the parents did not offer education. (I10) --- Oh, I never went to class [...] my father never wanted me to. (I7) The speeches point out that the aging generations of women had difficulty accessing school when they were young. Traditionally, these women did not study because they continued to live in the countryside, married to men of the same environment, continuing their lives and production established in their families. Nowadays, there is a tendency among rural women to leave home to seek new alternatives, both for study and for paid work. This empowerment of rural women is associated with their participation in social movements, which tends to grant them greater autonomy (20). The insertion of rural women in the context of work is anchored in the fragility of being a woman in a scenario in which gender inequalities stand out in the male-female relationship. The role of the woman is restricted to the home environment, to the care of the children, the house, and in the aid of the crop. This is reinforced by the form of social organization of work in rural areas, where farming, considered a productive site, is a symbolic masculine domain; while the space of the home, considered as reproductive and in some ways unproductive, is the feminine symbolic domain (7). Rev Ga<unk>cha Enferm. 2017;38(4):e68209 The representation of these elderly women is that this form of social organization and work generates suffering, given the overload of work attributed to the rural women and the lack of recognition of their own work, as rural workers, which is observed in the following statements: The woman suffers enough [...] to have her children, to take care of the house, to help in the fields. (I4) The woman should not suffer anything, because she is everything, she raises the family, she raises her children, she has to take care of her husband, she has to do the chores. (I9) In this direction, by not recognizing the double/triple working hours of women, the established social representation is based on a discourse that makes the majority of women internalize and incorporate, through a subtle process, the "consent", thus, establishing the complicity with the masculine power, generating consequences by the excess of work (3). The dynamics of life with double or triple working hours of the rural women occurs daily through the development of activities in the fields, in the home and in the care of the children and husband and/or companion. In this way, their work day is continuous, starting at dawn, extending until dusk. The activities developed involve the productive and domestic environments, the latter considered unproductive, that is, without economic value. Women, however, do not receive recognition for their working hours, given the gender culture (20), in which they assume the position of "helping" the husband. --- FINAL CONSIDERATIONS The situations of violence against women have always been present in the historical construction of society and are still current; however, they require strategies for their recognition and confrontation, especially when it comes to the rural context. In this sense, it was possible with this study to understand the social representations of violence against rural women from the perspective of the elderly women who lived in rural areas. On one hand, the elderly women represented violence as distant from their life context, relating it to extreme acts of corporal aggression and to the great centers. On the other hand, they represented violence as present in their relationships, showing themselves vulnerable to domination and subordination towards men, being still seen as the "fragile sex", not having the right to question and to impose their desires. It is worth considering that the elderly women evoked these aspects when they were indu-ced to think about violence against women, which leads them to infer that they perceived themselves involved in this phenomenon. It was also possible to visualize the elderly women in a context in which the attributes related to the house and children care, help in farming activities, are evidently introjected, evidencing the overlapping of roles and the overload of work, characterizing their lives as suffering. Thus, this study has provided an approximation with the different faces of violence perpetrated against women living in rural areas, and the determining and triggering factors of this phenomenon. Knowing the social representations of violence against women in the perspective of the elderly has as theoretical contributions that it can help in the (re) formulation of actions to confront violence in the health sector, as one of the sectors that can make up the network of attention to rural women in situations of violence, with a view to promoting their right to life and health. The nurse stands out as a member of this sector, which, in the rural setting, is in a privileged position to identify and intervene in situations of violence against women, since they are inserted in the territory through the Family Health Strategy. This research presents characteristic limitations of qualitative and descriptive studies. The contradictory elements that have appeared in the testimonies of the elderly women may be due to the collection scenario, or to the fact that one of the techniques of data collection approached the phenomenon of violence against rural women in a workshop with a group of women that knew each other, which may have inhibited the participants' speeches. Future research is recommended in order to give greater visibility to the violence perpetrated against women in the rural context, given the complexity of this problem and consequences for the health/illness process of the women who experience it. Moreover, the data presented here offer possibilities for the development of interventional studies, aiming at suggesting measures to promote the quality of life and health of rural women. The theme of violence against women is an object of contemporary study for the fields of health and nursing, increasingly present in scientific productions. However, researchers should be encouraged to produce research on this theme in the rural setting, given the specificities of this scenario and the population that lives in it.

Objective: Understand the social representations of violence against rural women, from the perspective of the elderly, considering how the generation and gender influence this aggravation. Methods: Qualitative study, based on the Theory of Social Representations.It was carried out with 12 elderly rural women from a city of Rio Grande do Sul, in July and August of 2013. The data obtained through a projective technique and semi-structured interviews has been analyzed through the reference of the Social Representations and content analysis. Results: On one hand, the elderly represent violence as distant distress, anchored in extreme acts of aggression. On the other hand, they represent it as present in their relationships, showing themselves vulnerable to the domination of men in the context of life and work.It was revealed the invisibility of violence against elderly women in the rural context, where the male figure overlaps with the female.

Introduction Print/TV media have often been blamed for the widespread perception in Anglo-Saxon countries that benefit claimants are not 'deserving'. In the UK, longstanding concerns over newspapers (Page, 1984:40) were recently revived amidst a surge in both negative newspaper coverage and'reality' television (e.g. Baumberg et al., 2012;Paterson et al., 2016;Reeves and de Vries, 2016). In the US, claimants have long been represented as disproportionately Black and undeserving (Gilens, 1996;Gilens, 1999;Misra et al., 2003), which has been causally related to harsh attitudes (Gilens, 1996;Hannah and Cafferty, 2006;Iyengar, 1990;Iyengar, 1991;Nelson and Oxley, 1999;Rose and Baumgartner, 2013;Shen and Edwards, 2006;Sotirovic, 2000Sotirovic,, 2001)). Yet there is a problem with these accounts: when people are asked to justify perceptions of benefit fraud, they rarely mention the media, and instead cite people they know -"we've got a neighbour who does it" (Briant et al., 2011:64). Indeed, in Briant et al.'s (2011) study, every respondent claimed to have 'first-hand knowledge' of fraudulent claims, while in Golding and Middleton (1982: 172-3), twice as many justified fraud perceptions via personal observation than media stories. Moreover, when scholars study public conversations about social issues in-depth, they find a "clear hierarchy, with personal experiences and anecdotes at the top, and media stories and statistical information at the bottom" (Rolfe et al., 2018:59; see also Gamson, 1992:123). If most people know undeserving claimants, and these perceptions are trusted more than the media, then the media's role may have been over-stated. In this paper, I present and test a new theory that resolves this contradiction, by suggesting that the deservingness we'see' in everyday interactions is influenced by the media-influenced frames we use to make sense of the world. Surprisingly, given well-developed literatures on media effects, this argument is noveland if it is correct, it would rebut this potential challenge to media effects, explaining how the media can be powerful even in the midst of widespread everyday tales of undeservingness. I focus on disability benefits, partly because disability is the major category of benefits for working-age people internationally, 1 and partly because it is an ideal case study: it is often unobservable and fluctuating, with ambiguous signs of deservingness that are susceptible to framing effects. I use novel data to experimentally test if media-style frames can cause people to respond differently to hypothetical contact with a vignette claimant (Study 1), to test if perceptions of claimants in everyday life match my hypotheses (Study 2), and to test if perceptions vary within the different discursive contexts of the UK vs. Norway (Study 3). To begin, however, I situate my hypothesis within the wider media effects literature. --- Existing theories of the media and benefits attitudes To date, researchers have explored two main theories to explain how media coverage influences benefits attitudes: 2 1. Cultivation theory suggests that the more time people spend consuming media, the more their worldview resembles it (Morgan and Shanahan, 2010). 3 Suggestive evidence supports this: attitudes to claimants are more hostile among those who consume more negative media sources (Baumberg et al., 2012;Gilens, 1996;Sotirovic, 2000). However, establishing causality through these cross-sectional studies is difficult, given that people choose their media sources based on pre-existing needs/beliefs (Baumberg et al., 2012;Gilens, 1996). 2. Framing theory suggests that media 'frames'that is, the way that media representations organise the many pieces of information relating to an issue (Scheufele and Tewksbury, 2006)affect how people think about claimant deservingness. There are many framing studies on welfare attitudes, typically using survey experiments that present people with different narratives (Avery and Peffley, 2003;Hannah and Cafferty, 2006;Iyengar, 1990;Iyengar, 1991;Nelson and Oxley, 1999;Shen and Edwards, 2006;Slothuus, 2007). These permit stronger causal inference, and mostly show that frames impact somebut not alloutcomes (Avery and Peffley, 2003;Hannah and Cafferty, 2006;Nelson and Oxley, 1999;Shen and Edwards, 2006;Slothuus, 2007). Other methodologies also show mixed results, e.g. convincing natural experiments show both significant (Reeves and de Vries, 2016) and null effects (Hedegaard, 2014a). Nevertheless, the overall balance suggests that negative frames are likely to negatively impact attitudes. Despite this considerable evidence on benefits attitudes-media links, there is almost no evidence that has looked at the connection between media coverage and people's direct experiences (indeed, few studies have examined this even in the wider media effects literature). Those that do have mostly assumed that the media matters most where people have little direct experience of the phenomenon in the story, often citing Lippmann 1922's argument that the media have power because "our opinions cover a bigger space, a longer reach of time, a greater number of things, than we can directly observe" (Gilens, 1996:515). This assumption is explicit within some studies on benefits (Sotirovic, 2000;2001:754); for example, Gilens (1996:530-1) argues that if personal experience does not explain the perceived racial make-up of claimants, media representations must be responsible. A smaller number of researchers have studied other interactions between media representations and experiences. Some have argued that the media matters most when it resonates with people's experiences (within cultivation theory and also e.g. Gamson, 1992:125-134). Similarly, Hopkins has convincingly argued that the combination of high local levels of immigration and a 'politicizing agent' (such as the media) affects people's attitudes (Hopkins, 2011). Yet while qualitative studies show how people counter/support media stories about welfare with personal experiences (Briant et al., 2011;Paterson et al., 2016), these ideas have otherwise not been applied to studies of benefits attitudes. In my theory, in contrast, it is not that the media supplants our direct experiences, or is only powerful if it resonates with thembut rather that the media fundamentally shape our interpretation of these experiences. This is a novel argument; I now develop this into a testable theory. --- Developing a new theory Our theory starts from two observations: we know that stereotypes influence judgements about the people we interact with (Shrum, 2009:54), and that media representations influence such stereotypes (Roskos-Ewoldsen et al., 2009). It is therefore likely that media representations change the way that we judge known others via their influence on stereotypes. Yet surprisingly, this hypothesis has never been testedneither for welfare attitudes nor more broadly. Gamson (1992:125) comes closest when noting in passing that "even our personal experience is filtered through a culturally created lens : : : We walk around with hyperreal images from movies and television and use them to code our own experiences"but this is not developed further. I hypothesise: H1: Media frames influence how known welfare claimants are judged. If media frames suggest that known claimants are mainly fraudulent, then people will'see' fraud in their everyday contact and judge claimants harshly. There is almost no previous evidence on H1. While there is considerable evidence of the effect of frames on general welfare attitudes (see above), it is not obvious that general effects will carry over to judgements of specific claimants. I know of only one framing experiment that looked at perceptions of an unrelated vignette claimant, whichdespite explicitly describing the vignette to respondents as a 'perfect example' of the storyfinds only inconsistent effects (Avery and Peffley, 2003). Secondly, my theory makes specific predictions about the circumstances in which these media effects will be found. Prima facie, stereotypes seem more likely to influence judgements based on ambiguous information; otherwise the direct information will outweigh the stereotype (Aar<unk>e and Petersen, 2014). In the case of known welfare claimants, ambiguity will tend to be greater for less well-known acquaintances vs. close friends (our 'peripheral' vs. 'core' network; Morgan et al., 1997). I hypothesise: H2: Media frames primarily influence judgements of peripheral network members (e.g. acquaintances) where ambiguity is greater, with little/no effect on judgements of core network members (e.g. close family). While this echoes various strands of previous work, the hypothesis itself is novel. For example, intergroup contact positively influences attitudes (the 'contact hypothesis'; Hewstone and Swart, 2011), and more specifically on welfare, those who know service usersgreater policy 'proximity'will be more positive towards services (Hedegaard, 2014b). However, neither theory contrasts core vs. peripheral relationships, 4 nor do they examine how media shapes the interpretation of contact. Still, H2 parallels wider accounts that knowing claimants cancontingentlyinfluence welfare attitudes. Third, if my hypothesis holds at the individual level, we would also expect it to hold at the societal level. That is, in countries where stereotypes of welfare claimants are more negative, I expect that claimantsparticularly in peripheral networkswill be judged more harshly. To test this, I compare the UK and Norway: newspaper reporting of benefits fraud is common in the UK but rare in Norway (Larsen and Dejgaard, 2013). I hypothesise that: H3: Judgements of known welfare claimants will be more positive in Norway (vs. the UK), particularly for peripheral (vs. core) network members. We must also bear in mind wider cross-national differences in attitudes (general perceptions of undeservingness are lower in Norway 5 ) and policy (disability benefit claims in Norway are 2-3 times as prevalent; see Appendix D). But there is no reason to think that these wider differences will lead to cross-national differences in how peripheral vs. core network members are judgedhence H3 provides a useful additional test of my theory. --- Applying this to disability benefit claimants We focus on disability benefits, partly because they are the major category of working-age benefits (they are noticeably more prevalent than unemployment claims in high-income countries 1 ), and partly because they are an ideal test of my hypotheses. This may seem surprising, as the deservingness literature conventionally argues that disabled claimants in general are perceived to be deserving (Jensen and Petersen, 2017;van Oorschot, 2000van Oorschot,, 2006)), usually explained by lower average levels of control and higher levels of need (within the 'CARIN' model of van Oorschot and Roosma, 2017). However, this evidence is often misinterpreted: it shows that some disabled claimantsparadigmatic "genuine" cases evoked by terms like'sick and disabled people' (van Oorschot, 2006) are perceived to be deserving, but does not necessarily mean that all disabled people are seen in this way. In fact, wider evidence shows that only some disabled claimants are judged to be deserving. Most directly, vignette studies show substantial variations in perceptions of different disability benefit claimants (Geiger, 2021). We can also see disabled people being differentially judged in e.g. political debates (Morris, 2016;Pennings, 2011;Soldatic and Pini, 2009) and everyday life (Hay, 2010;Holloway et al., 2007). These distinctions are usually based on whether a claimant is perceived to be 'genuinely' disabled, which is often unclear to external observersthere are few demonstrable external signs of impairments related to e.g. pain or mental ill-health (Baumberg et al., 2012:26-7;Holloway et al., 2007). Moreover, disabilities are often fluctuating, which makes even those outward signs of disability on one day an unreliable guide to capacities on another (Boyd, 2012). Disability claimants are therefore an ideal test of H1people make sharp judgements about the disabled people they know, but the outward signs of disability from interpersonal contact are ambiguous, and this ambiguity allows space for media-influenced stereotypes to influence judgements. Disability benefits are also a good test for H2: qualitative studies show that disabled people's core network tends to support their accounts, with scepticism usually reserved for more peripheral relationships (Hay, 2010:265) notwithstanding that people are occasionally judgemental about close friends/family (Holloway et al., 2007(Holloway et al., :1460)). In the conclusion I consider how far my findings are likely to extend to other types of claimant. In the rest of the paper, I test these hypotheses across three studies, all using a purpose-collected survey of nearly 4,000 people in the UK/Norway (described below). Study 1 tests if an experimentally-assigned newspaper-style story influences judgements of a specific vignette claimant (testing H1). Study 2 examines perceptions of 'non-genuine' disability claims in everyday life, testing how these vary according to media use and relationship type (testing H2). Study 3 also examines perceptions of 'non-genuine' disability claims in everyday life, but now testing differences across countries (testing H3). The studies complement one another's strengths/weaknesses: Study 1 allows confident causal inferences but weak ecological validity; whereas Studies 2 and 3 provide more tentative causal inferences but with strong ecological validity. Ethical approval was given by the lead author's institution, and replication code/data are publicly available (see Appendix I). Study 1: Survey experiment H1 hypothesises that media frames will influence how a specific disabled benefit claimant is judged. To test this, Study 1 uses a survey framing experiment, investigating if a newspaper-style article on disability benefits fraud affects judgements of a vignette disability benefit claimant. As a test of specificity, I further check that the fraud framing has no impact on judgements of non-disabled claimants (see below). Framing experiments allow strong causal inferences, and have been widely used to study the impact of the media on attitudes (see above). --- Data The experiment was embedded within a survey of 3,836 people in the UK and Norway using YouGov's opt-in panels. To achieve an approximately representative sample, panel members were invited to participate according to sociodemographic criteria, and the resulting sample was then weighted to known population totals (see Appendix E). The diversity of this sample is a strength compared to the student samples used in most previous welfare framing experiments (Hannah and Cafferty, 2006;Nelson and Oxley, 1999;Shen and Edwards, 2006;Slothuus, 2007), with rare exceptions (Avery and Peffley, 2003;Iyengar, 1991). I conducted the framing experiment in both countries because they vary considerably in welfare discourses (see H3 above); if we find similar effects in both then this supports the generalisability of my findings to other high-income countries. --- Procedures and measures Our experiment uses a sample newspaper story that had previously been developed by Ford and collaborators (embedded within the survey used for e.g. Kootstra and Roosma, 2018) based on stories in UK newspapers. The story is shown in Figure 1 below: it was seen by a random half of respondents (the 'benefits cheat frame' group), but not others (the control group). We then asked respondents to judge the deservingness of vignette claimants, each representing 'hypothetical contact' with a claimant. Rather than choosing a single vignette to represent all claimants, I used a multiple factorial design to investigate the effects of the benefits cheat frame on different claimant types. Most vignettes described disabled claimants, but some described non-disabled unemployed people, who I hypothesised would not be affected by the disability benefits cheat framing (see below). Each disabled claimant vignette contained eight experimentally-varied dimensions, while each unemployment vignette contained six dimensions (see Table 1). I obtained 8,605 disability vignette judgements and 2,468 unemployment vignette judgements from 3,836 people in Feb-May 2017. Respondents were then asked whether each vignette 'deserves to receive support from the Government while [he/she] is out of work?', giving answers on a 0-10 scale. To test H1, I examined whether those in the fraud frame group judged the disability vignettes (but not the unemployment vignettes) more harshly than those in the control group. More formally, I regress deservingness on the fraud framing vs. the control group using a simple OLS model (using cluster-robust OLS to account for the clustering of vignettes within respondents). I do not weight the data as this can increase bias in survey experiments (Mutz, 2011:114-123), nor do I include control variables as this is unnecessary and may decrease power (Mutz, 2011:124-6). In sensitivity analyses, however, I test if the results are robust to the inclusion of weights and sociodemographic controls (age, gender, children, marital status, qualifications, working status, and own benefit claims; see Appendix A). --- Results The average effect of the 'benefits cheat' story on perceived deservingness is shown in Table 2 below. This supports H1: reading the 'benefits cheat' story makes respondents judge a vignette claimant more harshly (-0.31, 95% CI -0.46 to -0.15). Table 2 also shows that this effect is found in both the UK and Norway, despite differing welfare/wider contexts, suggesting the effect is generalisable across highincome countries. To further test my hypothesised mechanism, I test whether these results are specific to disabled benefit claimants and do not carry over to non-disabled unemployed claimants (given that the story focused on the genuineness of disability, not unemployment). The results are shown in the final column of Table 2, which shows that there is no effect of the benefits cheat story on judgements of nondisabled unemployed vignettes (the difference between the impacts on disabled and unemployed vignettes is strongly significant, p<unk>0.001). This confirms that the impact of the disability benefits cheat story is specific to hypothetical contact with disabled benefit claimants. In further sensitivity analyses (Appendix A), the findings Please read the following news story that was published in a British newspaper: --- Benefit cheat received thousands John Smith -Deputy Editor ANDY Green, a benefits cheat, received £158,000 in benefits because he claimed to be unable to leave the house, although he spent several months per year living abroad. Mr. Green told the Department of Work and Pensions that due to his condition ofagoraphobia he was unable to leave the house by himself, could not travel to strange places and was often afraid to be amongst other people. Anti-fraud examiners were alerted by a member of the public that Mr. Green had actually bought an apartment in Thailand. While he would spend several months per year in his new house, back in the UK his benefits would continue. Upon investigation, officers found out that besides the apartment, Mr. Green also frequently attended gambling events, where he would gamble with taxpayers' money. are unchanged if I include controls/weights; or exclude those failing attention checks. Overall, I find clear support for H1, although the effect is moderately small (0.3 on a 0-10 scale, <unk>10% of the standard deviation of 3.06). However, it would be surprising to see a larger effect. The size of effect is not dissimilarand is more consistent across measures/sensitivity analysesto previous studies that have been interpreted as showing evidence of framing effects (e.g. Slothuus, 2007). Moreover, the newspaper-style story is muted in tone and makes no attempt to generalise 'Andy Green' to claimants in general (unlike real newspaper stories; Baumberg et al., 2012); respondents therefore have to connect a story about one person to a completely different person. Finally, while my vignettes were somewhat ambiguous, they were perhaps less so than everyday contact with peripheral networks. These suggest that my estimates may be lower bounds on real-world effects (though these must be weighed against the limitations of survey experiments), and I therefore turn to examining real-world effects directly. Study 2: Real-life deservingness judgements Study 1 provides convincing evidence that media frames can influence judgements of specific hypothetical claimants. However, this comes at the expense of ecological validity: like most social experiments, it invokes one frame without any competing ones, and removes temporal/other barriers between frame and judgement (Barabas and Jerit, 2010). To complement this, Study 2 examines how people's real-world deservingness judgements relate to media consumption in the UK. It also goes further in specifically testing H2, which hypothesised that media effects would be found for peripheral network members (e.g. acquaintances) but not core network members (e.g. close family), because of the greater amount of ambiguity when judging peripheral network members. --- Data and measures of real-world deservingness judgements We use the same YouGov survey used in Study 1, but here focus only on the UK (as my measure of welfare media coverage is only available in the UK; see below). For each of four types of social relationship (close family, close friends, distant family, and neighbours), I asked respondents (full question wording is given in Appendix G): • If they personally knew anyone with that relationship (e.g. close family) that they think has probably claimed disability benefits in the past year; • If they knew any claimants with that relationship (e.g. close family) that they thought 'are not genuinely sick or disabled' (hereafter 'non-genuine' claimants). Table 3 shows that nearly half (45.4%) of UK/Norway respondents reported knowing a disability benefit claimant per semost commonly close family or close friends (17.0-22.6%), and to a lesser extent distant family and neighbours (7.0-7.1%). Yet when looking at non-genuineness, people were more likely to know a non-genuine neighbour than a non-genuine close family member (6.0% vs. 4.1%). This is similar to recent findings in the US (Fang and Huber, 2020). To test H2, we need a fair comparison of the prevalence of non-genuineness for different relationship types, which takes account of the potentially different size of each group. Our outcome variable is therefore reports of knowing non-genuine claimants only among respondents that had already said they knew a claimant of that type, as shown in Table 4 below. 6 Using this outcome variable, we can see a clear pattern in the final column of Table 3, where claimants in core networks are less likely to be seen as non-genuine than those in peripheral relationships (30.2% of neighbours are seen as non-genuine vs. 13.6% among close family). --- Measures of media use To measure media framing of benefit claimants, I use a hand-coded content analysis of a 20% sample of all UK newspaper articles on benefits 1995-2011 (1,291 articles in total), kindly shared by Declan Gaffney (see Baumberg et al., 2012). While television is another pivotal media source and there has been an increasing use of online news sources, I focus on newspapers as in the UK they have long been particularly polarised around welfare (Page, 1984:40), and have therefore been the • Fraud, • 'Shouldn't be claiming' (not fraud), • Never worked/hasn't worked for very long time, • Large families, • Bad parenting/anti-social behaviour, • Claimants better off on benefits, • Claimants better off than workers, and • Compulsion of claimants. The resulting 'newspaper negativity' score is merged into the survey data on the basis of the newspaper that respondents report reading most regularly (see Appendix B). Note that this is a measure of general newspaper negativity to welfare, rather than the disability-specific negativity of the benefits cheat newspaper story in Study 1. The Gaffney measure reflects a decade-long average several years prior to my survey (with Reeves and de Vries, 2016 showing that short-run changes are possible). I therefore validate Gaffney's measure in the present-day using my survey: after giving half of respondents the mock article on disability benefit fraud in Study 1, I asked them how often they had read similar newspaper articles, and compared this to the newspapers they regularly read. Across 8 major UK newspapers, newspaper negativity 1995-2011 correlates reassuringly strongly with 2017 reports of disability fraud articles (r = 0.86; see Appendix B). --- Analytical approach We regressed'self-reported contact with a non-genuine claimant' on newspaper negativity (treated as a continuous variable) alongside sociodemographic control variables (gender, age, marital status, presence of children in the household, education, working status, own benefits claims, and disability; see Appendix G) using standard logit models, allowing covariates to have different effects for each type of interpersonal contact. (In sensitivity analyses I further control for ideology and social trust). Note that I have 0-4 observations per person, depending on whether they reported knowing claimants within each relationship type, and I use cluster-robust standard errors to account for this clustering within individuals. To avoid possible misinterpretations, all results are presented as average marginal effects (Mood, 2010). --- Results H2 predicted that those who read negative newspapers would perceive greater nongenuineness among neighbours but not among close family/friends. The results are shown in Figure 2, and strongly bear out my hypothesis. There is no systematic relationship between newspaper negativity and perceptions of close friends or family, but a strong relationship for distant family and particularly neighbours (e.g. a 1% rise in newspaper negativity raises perceived non-genuineness among close friends by only 0.1% [95% CI -0.3 to 0.6%], but raises it among neighbours by 2.1% [95% CI 1.0 to 3.1%]; see Appendix H). In further analyses I also control for social trust and political ideology (which may confound the relationship between newspaper negativity and claimant judgements). These still show strong support for H2, only slightly attenuated from my main results (e.g. a 1% rise in newspaper negativity raises perceives perceived non-genuineness among close friends by <unk>0.05% [95% CI -0.5 to 0.5%], but raises it among neighbours by 1.6% [95% CI 0.75 to 2.5%]; see Appendix A). These control variables enable us to compare the judgements of readers of different newspapers who are otherwise identical (in sociodemographics, social trust and ideology), but I should stress that this is cross-sectional observational research, and it is possible that there is both unobserved confounding or reverse causality. This reflects the different trade-offs of Study 1 and Study 2: the latter enhances ecological validity at the expense of confident causal inference. Study 3: Comparing real-life judgements in the UK vs. Norway Our final hypothesis was that Britons would be more judgemental of acquaintances vs. close family members than Norwegians, given differences in welfare discourses (H3). This uses the same YouGov data and real-life judgements as Study 2 (but unlike Study 2, now looking across both the UK and Norway). Again, I focus on perceived non-genuineness only among respondents that report knowing a claimant of that type per se (see discussion under Study 2 and below). Our analytical approach here is simple: I regress perceptions of non-genuine claimants on respondents' country using logit models, with weights but no controls. (In sensitivity analyses I further include the same controls as in Study 1). As before, I use cluster-robust standard errors to account for the clustering of responses within individuals, and present all results as average marginal effects. --- Results Contrary to expectations, Figure 3 shows no clear difference in perceptions of neighbours between the two countries (about 30% of neighbours who claim disability benefits are perceived to be non-genuine in both). Also against expectations, Norwegians perceived more non-genuineness among close family and friends, again shown in Figure 3. There is some evidence that the difference between perceptions of close family/friends vs. neighbours was slightly stronger in the UK than Norway (the specific prediction of H3)but this was small and imprecisely estimated. Sensitivity analyses that include sociodemographic controls produce effectively identical results. This is not because my choice of case studies is incorrectin my survey, Britons do indeed report seeing more disability fraud articles than Norwegians, albeit by less than expected (42.3% in the UK vs. 28.2% in Norway report seeing articles like the benefits cheat frame in Study 1'very often' or 'quite often', a difference of 14.2% [95% CI 9.7-18.7%]). Yet this does not necessarily mean that the more positive public discourse in Norway has no impact on perceptions of known disability benefit claimants (contra H3), because of three complexities. Firstly, many respondents initially said they did not know a claimant of a particular relationship type per se, but then reported knowing a non-genuine claimant of that type (see Study 2/ Table 4/note 6). This was more common in the UK, and as these cases of perceived non-genuineness are excluded from my main analysis, this makes cross-national comparisons more difficult. It is possible when testing H3 to use a simpler outcome measure than Study 2, because we can assume that the size of different networks is similar in the two countries (i.e. that people know similar numbers of neighbours in the UK and Norway). If I therefore adopt a simpler outcome measurereported non-genuineness per sethen I find that Norwegians were only slightly more likely than Britons to know a non-genuine neighbour (17.7% more likely [95% CI 3.7-31.8%]), but much more likely than Britons to know a genuine neighbour (80.7% more likely [95% CI 55.5-105.8%]). This provides some support for H3, but I must stress this is a post-hoc exploratory analysis, rather than my initial approach. Secondly, at least 2-3 times as many people claim disability benefits in Norway compared to the UK (see Appendix D); in my survey 29.1% of Britons vs. 61.5% of Norwegians report knowing a disability benefit claimant. Assuming that people in Norway do not have greater morbidity, this means that people with less severe disabilities are more likely to be claiming disability benefits in Norwayhence the average severity of disability of claimants in Norway vs. the UK will be lower. Given that severity is linked to perceived deservingness (Geiger, 2021), this will ceteris paribus inflate the rate of perceived non-genuineness in Norway vs. the UK. This may explain why I see greater levels of perceived non-genuineness in Norway for all relationship types. Third, differences in severity are unlikely to be the full explanation, because the meaning of 'non-genuine' appears to differ between the two countries. To explore this, I asked respondents, "Think of the non-genuine claimant that you know best. How could you tell that they were not a genuine claimant?" Norwegian responses were translated into English, and then all responses were coded to an inductivelyderived coding frame (in the absence of 'don't know/refused' options, <unk>20% of responses in each country were un-code-able or blank). Results for code-able responses are shown in Table 5, showing three summary codes (see Appendix C for further details): • 'Able to do more than they claim': the claimant was believed to be less sick/ disabled than they claim. This referred to either general functioning (being healthy, having an active life); specific tasks they have been observed doing (e.g. walking, odd jobs/gardening, sport/partying/socialising, holidays, shopping); lying (e.g. not using a wheelchair/stick when they think they cannot be seen, or vaguer accusations of'malingering'); or the claimant admitting fraud/ being well. • 'Lazy/bad attitude': the claimant was believed to be lazy, or not trying hard enough to find work (or more rarely, not trying hard enough to get better). In Norway this occasionally included references to people being young, with the implication that efforts should be greater at their age. • 'Could work with support/adaptations': the claimant was believed to be capable of doing some (but not all) work, or comments that they could be working without attaching any blame (sometimes explicitly saying that it is society's fault that they are given inadequate support). This shows a strikingly different pattern of justifications in the two countries. In the UK, accusations of non-genuineness were primarily because people were thought to be 'faking it' or exaggeratingover three-quarters (77.9%) of justifications were on the grounds that the person was not as sick/disabled as they claimed to be. While this was still relatively common in Norway, it was much less dominant than in the UK (41.1% of justifications). Instead, Norwegians were much more likely to justify their perceptions of non-genuineness on the grounds that the person was lazy (or at least, not trying hard enough to work), or that despite their genuine sickness/disability, they still had the ability to work (22.6% of justifications in Norway, but virtually absent in the UK). I revise the theory in the light of these findings in the Concluding section. In the meantime, the evidence does not provide clear support for/against H3. --- Conclusions To the extent that welfare benefit claimants are seen as undeserving in the US/UK, it is often argued that the media are partly to blame. Yet this is seemingly inconsistent with the way US/UK publics primarily justify perceptions of undeservingness through their own contact with claimants, rather than through media reports. In this paper, I have tested the hypothesis that the media may still play a role by shaping how we interpret such contactparticularly contact with peripheral network ties, where outward deservingness cues are few and ambiguous, and where media-driven stereotypes may therefore play a role. I focus on disability benefits, which are not only the

The media are often blamed for widespread perceptions that welfare benefit claimants are undeserving in Anglo-Saxon countriesyet people rarely justify their views through media stories, instead saying that they themselves know undeserving claimants. In this paper, I explain this contradiction by hypothesising that the media shapes how we interpret ambiguous interpersonal contact. I focus on disability benefit claimants, which is an ideal case given that disability is often externally unobservable, and test three hypotheses over three studies (all using a purpose-collected survey in the UK and Norway, n = 3,836). In Study 1, I find strong evidence that a randomly-assigned 'benefits cheat' story leads respondents to interpret a hypothetical disability claimant as less deserving. Study 2 examines people's judgements in everyday life, finding that readers of more negative newspapers in the UK are much more likely to judge neighbours as non-genuinebut with effectively no impact on judgements of close family claimants, where ambiguity is lower. However, contra my expectations, in Study 3 I find that Britons are no more likely than Norwegians to perceive known claimants as non-genuine (despite more negative welfare discourses), partly because of different conceptions of what 'non-genuineness' means in the two countries.

justify their perceptions of non-genuineness on the grounds that the person was lazy (or at least, not trying hard enough to work), or that despite their genuine sickness/disability, they still had the ability to work (22.6% of justifications in Norway, but virtually absent in the UK). I revise the theory in the light of these findings in the Concluding section. In the meantime, the evidence does not provide clear support for/against H3. --- Conclusions To the extent that welfare benefit claimants are seen as undeserving in the US/UK, it is often argued that the media are partly to blame. Yet this is seemingly inconsistent with the way US/UK publics primarily justify perceptions of undeservingness through their own contact with claimants, rather than through media reports. In this paper, I have tested the hypothesis that the media may still play a role by shaping how we interpret such contactparticularly contact with peripheral network ties, where outward deservingness cues are few and ambiguous, and where media-driven stereotypes may therefore play a role. I focus on disability benefits, which are not only the most commonly-claimed working-age benefits in highincome countries, but are also an ideal case study given the often ambiguous external signs of disability. We tested this using three complementary approaches: • Study 1 used a survey experiment to test if negative newspaper frames influence respondents' deservingness judgements of a hypothetical person, using a UK-Norway survey of nearly 4,000 people and a wide range of vignettes. Confirming H1, I found that a disability benefits fraud frame did indeed make people judge disability vignettes more negatively. • Study 2 used perceptions of non-genuine disability claims in real life and how this related to newspaper readership in the UK. This also supported H1: newspaper negativity was associated with a greater probability of viewing known claimants as non-genuine. This also supported H2: these associations were much stronger when judging neighbours, and there was effectively zero association of newspaper negativity with judgements of close family, where perceivers have high information and low ambiguity. • Study 3 also used perceptions of non-genuine disability claims in real life, but focused on a comparison between the UK vs. Norway (media discourses being more hostile to claimants in the former). Contra H3, I did not find that Britons were more likely to judge neighbours to be non-genuine claimants than Norwegians. In further exploratory analyses, it seemed that this may be partly because Britons were more likely to make reporting errors; partly because the average severity of disability benefit claimants in Norway is noticeably lower; and partly because in Norway there was a different conception of what 'nongenuine' meant (see below). These studies are complementary: survey experiments allow confident causal inferences but are weakly generalisable; whereas real-life associations are strongly generalisable but require more cautious causal inferences. Nevertheless, we must bear in mind the limitations that apply to each studythat framing experiments are some distance from everyday deservingness judgements; and that there is the possibility of unobserved confounding and reserve causation when looking at everyday deservingness judgements. Moreover, while YouGov opt-in panels have been widely used in academic research (e.g. Aar<unk>e and Petersen, 2014;Jensen and Petersen, 2017;Kootstra and Roosma, 2018), and are preferable to the student samples in most previous framing experiments, there is always a risk that the samples are unrepresentative. --- Theoretical revisions Overall, the evidence mostly supports my theory that the media shapes how we interpret our interactions with disability benefit claimants. Nevertheless, H3 was not unequivocally supported, and the post-hoc analyses above suggest that the theory needs to be revised. Welfare state institutions not only shape media framings (Larsen and Dejgaard, 2013), but also the make-up of benefit claimants (extending Larsen, 2006 to consider disability). The lower threshold for disability benefits in Norway means that claimants are fundamentally different groups of people in Norway vs. the UK (indeed, disability itself is defined differently; Kapteyn et al., 2007;O'Brien, 2015). There are two plausible explanations here. More simplistically, media discourses in Norway are more generous (and lead to more generous judgements under uncertainty), but are counterbalanced by the lower average severity of disability (which leads to harsher judgements). But a more complex interpretation is that the Norwegian system is underpinned by (and underpins) a fundamentally different framing: rather than focusing on whether disability is genuine vs. fraudulent, it focuses on partial work capacity that people have an obligation to make use of, and which the state has an obligation to facilitate (McKowen, 2020). Judgements under ambiguity are therefore not just about genuineness, but also whether worklessness among genuinely disabled people is due to individual effort vs structural constraints. Going forward, I believe that it is essential to study how benefit claimants are judged under ambiguity, and how the media shapes this. However, these judgements are more complex than my original account allows. In comparative perspective, further research should therefore: • Extend the research here using multiple media framings, multiple dimensions of judgements and multiple claimant types. I hypothesise that media frames will influence judgements of other claimant groups (e.g. control over unemployment is often unobserved), but multiple dimensions of deservingness will be involvedas they are even for disabled claimants; • Test and develop the theory by studying mechanisms in more detail. This includes varying ambiguity more directly (e.g. following Petersen and Aar<unk>e, 2013), but also qualitatively studying day-to-day social interactions, examining what external cues are given by claimants to which audiences, and how various cues are interpreted by audiences with different worldviews; • Explore comparative differences further by studying the role of welfare state institutions in constructing media narratives and claimant groups (and their interactions)while being sensitive to the many factors that influence media reporting (Baumberg et al., 2012:Ch4;McKendrick et al., 2008). --- Implications Our findings deepen the existing challenge for those seeking public support for more generous benefits policies. Such campaigners in Anglo-Saxon countries are already aware of the challenges of a hostile media context; but one implication of my results is that welfare attitudes will be sticky even in the face of new media frames, which may appear to conflict with 'common sense' grounded in everyday experience. Progressive reformers can choose three different responses to this challenge (Baumberg, 2012). One is to pursue radical reform in moments of crisis, where worldviews rupture. A second is to use'sequencing', where institutional reforms nudge discourses in a more progressive directionfor example, to gradually increase welfare universalism, which will later create wider constituencies of interest for increased generosity. However, more research is necessary to understand how people judge one another within different welfare regimes (see previous section). Finally, we should note that past radical reforms did not require a 'golden age' of pro-welfare attitudes (Hudson et al., 2016). Even if hostile welfare attitudes are buttressed by interpretations of ambiguous everyday experience, they are neither immutable nor unchallengeable. Most of all, though, my theory rebuts a potential challenge to the impact of the media in framing welfare deservingness. When people say that their neighbours cheat benefits, this is not decisive evidence of widespread fraudinstead, these judgements are themselves part of the tangled web of media-fuelled stereotypes, political discourses and institutional design. 2 Other mechanisms have been extensively explored in the wider media effects literature, including agendasetting and priming (Scheufele and Tewksbury, 2006). However, these are rarely discussed in the benefits literature (beyond brief mentions in Gilens, 1996;Reeves and de Vries, 2016). 3 While the original formulation refers to television in general, many cultivation researchersincluding those studying benefits (Sotirovic, 2000) have focused on particular sources. 4 Interestingly, while Hedegaard finds a positive impact of core network contacts on welfare attitudes, Gelman and Margalit (2021) find null effects found for core+peripheral contacts combined. 5 In 2016, 64% of Norwegians vs. 43% of Britons agreed many people manage to obtains benefits or services they are not entitled to -Authors' analysis of weighted ESS 2016 data via http://nesstar.ess.nsd.uib.no/ webview/ 6 Some people reported knowing a non-genuine claimant when they had not initially reported knowing a claimant of that type (see final row of Table 4). This was least common among close family and most common for neighbours (where 24.4% and 64.5% of reported non-genuineness for close friends/neighbours respectively was among those who did not initially report knowing people of that type that claimed). This reflects two types of reporting errors: a response error (where people reported the closest claimant they knew, missing out more peripheral relationships) and memory errors (where people did not think of a person until prompted to think of a non-genuine claimant).

The media are often blamed for widespread perceptions that welfare benefit claimants are undeserving in Anglo-Saxon countriesyet people rarely justify their views through media stories, instead saying that they themselves know undeserving claimants. In this paper, I explain this contradiction by hypothesising that the media shapes how we interpret ambiguous interpersonal contact. I focus on disability benefit claimants, which is an ideal case given that disability is often externally unobservable, and test three hypotheses over three studies (all using a purpose-collected survey in the UK and Norway, n = 3,836). In Study 1, I find strong evidence that a randomly-assigned 'benefits cheat' story leads respondents to interpret a hypothetical disability claimant as less deserving. Study 2 examines people's judgements in everyday life, finding that readers of more negative newspapers in the UK are much more likely to judge neighbours as non-genuinebut with effectively no impact on judgements of close family claimants, where ambiguity is lower. However, contra my expectations, in Study 3 I find that Britons are no more likely than Norwegians to perceive known claimants as non-genuine (despite more negative welfare discourses), partly because of different conceptions of what 'non-genuineness' means in the two countries.

INTRODUCTION Entrepreneurial dynamics is often associated with identifying opportunities and discovering and developing new business ventures, as reported by several previous studies (Audretsch & Keilbach, 2005;Bailey, 2012;Kilkenny, 2006;Nordberg, Mariussen, & Virkkala, 2020;Nwankwo & Okeke, 2017;Ruef & Lounsbury, 2015) regarding entrepreneurial dynamics by pursuing various commercial goals that may conflict with moral behavior. The emergence of more moral forms of entrepreneurship typically exhibits a clear agenda for better outcomes that focuses on "innovation and efficient use of resources to explore and utilize the opportunities, thereby fulfilling social needs sustainably" (Asuquo & Enya, 2020;Nordberg et al., 2020;Sahoo & Mondal, 2020;Squazzoni, 2009). Meanwhile, the term entrepreneurship refers to a method that provides solutions to various social issues to create economic, social, and environmental values (Al-Dajani, 2013;Escandón-Barbosa et al., 2019;Gandhi & Raina, 2018;Maksum et al., 2020;Prodanov, 2018;Snyder, 2008). In general, Entrepreneurial Dynamics and Typology..... (Handono, Hidayat, Purnomo, and Toiba) research on entrepreneurship emphasizes the discussion concerning rural development, consisting of individual, institutional, marketing, and organizational management aspects. However, explicit attention to the role of service-providing entrepreneurs in this discussion has been limited (Adro & Franco, 2020;Cucari, Wankowicz, & Falco, 2019;Mottiar, Boluk, & Kline, 2018), although several authors have emphasized the importance of entrepreneurship and rural development comprising the role of the individual, product diversity, and contribution to the overall product goals (Galv<unk>o, Mascarenhas, Marques, Braga, & Ferreira, 2020;Komppula, 2014;Liang & Paul Dunn, 2014). In addition, the role of entrepreneurs as a contributor to rural development is a less-discussed area of research. Prior research by Korsgaard et al. (2015) identified the role of entrepreneurs and village development only at the stage of involvement and development of rural development without further discussion at a later stage highlighting the contribution of the entrepreneurs. As highlighted in the previous study by Dhewanto et al. (2020), the role of the entrepreneurs as a stakeholder has contributed to goals and competitiveness, having only two subtitles: competitive microenvironment and supporting factors and resources. Entrepreneurship refers to developing new ventures that contribute to the development of goals by creating competition, cooperation, specialization, innovation, investment, growth, risk-taking, productivity, and others. However, among many contributions, problems and challenges are inseparable for an entrepreneur. For example, it was reported that owners and managers did not have the skills, expertise, or resources in most small businesses in rural areas (Futemma et al., 2020;Mahon & Hyyryläinen, 2019;Nwankwo & Okeke, 2017) and that entrepreneurship in rural areas could not recognize and neglect about the wider old business, resulting in the fact that failure or challenges were only a consequence of their actions. However, the role of entrepreneurs is deemed necessary in ensuring that stakeholders' expectations are fully met. Particularly in the agricultural sector, rice farming is significantly influenced by individual performance characteristics and entrepreneurship of farmers; farmers could enhance their quality related to farming motivation and entrepreneurial abilities, as well as be capable of maximizing the performance of rice farming (Bao, Dong, Jia, Peng, & Li, 2020;Imelda, Hidayat, & Aritonang, 2022;Matarrita-Cascante & Suess, 2020). The relationship between the role of entrepreneurs (as part of economic institutions) and village development in previous research was only discussed in a partial aspect. Entrepreneurs (as business actors) serve as a mechanism for creating development in rural areas characterized by organization (Gaddefors, Korsgaard, & Ingstrup, 2020). Elaborating further, Moore and Westley (2011) described entrepreneurship as individual activities, whereas Korber & McNaughton, (2018) defined entrepreneurs as having tough characters. In addition, other studies unveiled that entrepreneurs played complementary roles in strategies for building coalitions, designing decision-making forums, and management (Meijerink & Huitema, 2010), as well as having special abilities or entrepreneurial competencies (Ataei, Karimi, Ghadermarzi, & Norouzi, 2020). Moreover, entrepreneurship relates to the capability of mobilizing ideas, resources, and support from external sources for the benefit of the rural community (Richter, 2019) and having a profound impact (Carayannis, 2020), driving and Rural Development Research transformation and contributing to the sustainable development of rural areas (Castro-Arce & Vanclay, 2020), considering the direction of development of rural destinations (Mottiar et al., 2018), and engaging with the social system (Korsgaard & Anderson, 2011). Furthermore, both entrepreneurs and entrepreneurship affect social change (Galappaththi, Galappaththi, & Kodithuwakku, 2017), social innovation (de Fátima Ferreiro, Sousa, Sheikh, & Novikova, 2021), and dynamics of change (Lamberte, Llanto, Meyer, & Graham, 1994). Sustainable change contains hope for the availability of economic, social (Yang, Cai, & Sliuzas, 2010), political (Sievert et al., 2022), and environmental resources for future generations (Liu et al., 2022). Rural entrepreneurs must consider how agriculture can preserve ecological conditions, drive economic growth, and support political networks in the local community. Some of these aspects (economic, social, political, and environmental) must be interlinked, thereby strengthening rural entrepreneurs in their respective interests. Following the theory cited by Cochran (1965), entrepreneurship is closely related to sociological aspects. Moreover, the actual economic development emphasizes the role of cultural values in determining the supply of entrepreneurs. Several points under Chochran's theory are entrepreneurs as a personality of community capital, personality of capital as a derivative of social conditioning, expectations of the role of entrepreneurs, types of parenting and schooling, and the influence of the intrinsic character of the executive. The typology of entrepreneurship is based on the type of business, business sector, characteristics of entrepreneurs, and business linkages with various sectors. However, there is no description or typology of rural entrepreneurs (RE) based explicitly on economic assets, politics, social relations, and the dynamics of rural entrepreneurs in rural development. Therefore, this research aims to analyze the dynamics of rural entrepreneurship and describe the typology of entrepreneurship. --- RESEARCH METHOD --- Research Type This research employed a qualitative method, aiming to understand field phenomena representing relevant concepts following studies or theories of entrepreneurship and social development in rural areas of Malang Regency. A qualitative method refers to the understanding that cannot be quantified (counted). Qualitative research is characterized by several points: (a) immature concept due to the lack of theory and previous research; (b) the idea based on the existing theory can be inaccurate, imprecise, incorrect, or biased; (c) there is a need to understand, interpret, predict and describe phenomena and develop theory; and (d) the nature of the phenomenon may not correspond to a quantitative measure (Creswell, 2014;Denzin & Lincoln, 2018;Morse, 1991). This research was conducted in three villages in Malang Regency, possessing a variety of potentials for tourism, agriculture, animal husbandry, and other businesses; highly potential natural resources (topography, soil types, and others); and a community with a unique diversity of activities in managing local natural resources. The selected locations comprised (1) agricultural industry (Kucur Village), (2) agritourism (Gubugklakah Village), and (3) Entrepreneurial Dynamics and Typology..... (Handono, Hidayat, Purnomo, and Toiba) traditional area (Bambang Village) in Malang Regency, East Java, Indonesia. This research was conducted for three years, from 2019 to 2022. --- Data Sources Data sources in this research included business actors (farmers, breeders, entrepreneurs, private sector, and local government concerned) and all parties involved in social, political, economic, and environmental activities. Other data sources from 90 figures who mastered the research topic at each location were determined by purposive sampling (intentionally), as illustrated in Table 1. In its mechanism, extracting informant data required a lengthy time as it was related to implicit information. Upon completing the extraction, an abstract was constructed, and various statements conveyed by the informants were revealed in more depth. In the final step, and Rural Development Research all information regarding the role of entrepreneurs in rural development in Malang Regency was explored. Table 2 lists the characteristics of age, gender, and employment status of the 90 informants. --- Data Collection Techniques Data were collected through observation, in-depth interviews, and Forum Group Discussions (FGDs). 1. Various phenomena, problems, and needs of business actors and any changes in each research location were observed. 2. In-depth interviews were focused on critical informants, including people who knew the history of the development of land-cultivated objects, initially unnoticed by the community and naturally cultivated, later becoming the central point of commercial fields for village or rural residents. The key informants also included the surrounding community, who fully understood the strategies of farmers, entrepreneurs, the private sector, and the government in accessing various natural resources, leading to alterations in each studied village. 3. FGDs were conducted with business actors and agricultural entrepreneurs in each village. These FGDs aimed to obtain an overview of the entrepreneurs' role in rural development. --- Data Analysis Techniques The interactive model proposed by Miles, Huberman, and Saldana was employed to analyze the data, encompassing three steps: (1) condensing data, (2) presenting data, and (3) drawing conclusions and verification. Data condensation refers to selecting, focusing, simplifying, abstracting, and transforming data (Miles, Huberman, & Salda<unk>a, 2014). Figure 1 illustrates the interactive model. --- FIGURE 1. INTERACTIVE MODEL BY MILES, HUBERMAN, AND SALDANA (MILES ET AL., 2014) Figure 1 demonstrates three research processes: data analysis through data condensation, data display in matrices, and conclusion drawing and verification. --- Propositions Following the research goals, several propositions were compiled and tested for validity. 1. The dynamics of RE depends on ownership of and access to social, economic, and political assets. --- Data Collection Data Display Data Condensation Conclusion Drawing/Verifying Entrepreneurial Dynamics and Typology..... (Handono, Hidayat, Purnomo, and Toiba) 2. The typology of entrepreneurs is considered capable of utilizing all or at least three assets: (1) typology of entrepreneurs with social, economic, and political assets and an entrepreneurial spirit; (2) typology of entrepreneurs with social, economic, and political assets; (3) typology of entrepreneurs with social and economic assets; and (4) typology of entrepreneurs with only economic assets. --- RESULTS AND DISCUSSION --- Dynamics of Development of Rural Entrepreneurship in Indonesia The dynamics of entrepreneurial development in Indonesia has been inseparable from various aspects and periods, starting from the Old Order, New Order, Reformation, and the present (2022), as described in Table 3. --- TABLE 3. VARIOUS ASPECTS OF RURAL ENTREPRENEURSHIP IN VARIOUS PERIODS (REGIME) --- Aspect Old Order New Order --- Based on the central policy of the local government and local creativity Table 3 presents the development, indicating trends from time to time in both aspects of the driving force of entrepreneurship and the nature or actors and typology of entrepreneurial development starting from state domination (centralistic) to the village-level (decentralistic or regional autonomy) implementation per individual entrepreneur. Development overcomes poverty, powerlessness, and inequality (Dhewanto et al., 2020). During the centralistic era, development was actualized through industrialization and commercialization between developed and developing countries, in which all forms of centralized policy were limited by the government elite (Old Order), including food policy that prioritized production quantity without regard to quality and the needs of the people. The centralized policy focuses on maximizing production at the central level (Achjar & Panennungi, 2010;Antlöv, 2003). Complexity, handling scale, and higher dependence on communication allow the development of rural local potential to be hampered (Dante, 2017;Pratama et al., 2021). Furthermore, the centralized policy eventually shifted to decentralization, enabling each region to make policy arrangements through the local government. Decentralization does not and Rural Development Research mean weakening the role of the central government but rather strengthening it by responding to the regional potential, predominantly rural areas. In line with the results of prior research (Dante, 2017;Pratama et al., 2021), the effectiveness of local government is supported by central government mechanisms, highlighting the setting of entrepreneurial potential according to the character, resources, and potential of the village (Yin, Chen, & Li, 2022); for example the phenomenon of tourist villages, arts-culture villages, and villages with other nonagricultural businesses. Since decentralization has been enacted, there has been a shift in the definition of development to a redistribution of growth with an approach to the basic needs of the community, requiring further analysis of economic development (Yin et al., 2022), human resources, infrastructure development, and projected environmental conditions for village and inter-regional development programs (Feisali & Niknami, 2021;Fuller-love, Midmore, Thomas, & Henley, 2006;Rădulescu, Marian, & Moica, 2014). The assessment of regional rankings is expected to describe the development in a particular area by identifying positive and negative transformation trends, comparing the level of stability with other regions, and assessing the effectiveness of regional management. In addition, proper management plays a vital role in ranking regions based on established criteria to analyze economic development, human resources, infrastructure development, and environmental conditions. Hence, development effectiveness is adjusted to the ability of various concerned parties or multi-disciplines to identify, control, and transform opportunities into sustainable entrepreneurship (Nugroho et al., 2021;Purnomo, 2020). The ranking of regional development is inseparable from the roles and contributions of many regional actors to build innovation systems through diverse knowledge and skills (Haugh & Talwar, 2016;Lang & Fink, 2019;Tellman, Eakin, Janssen, de Alba, & Turner, 2021). In this case, the development of each region also provides opportunities for community development, especially for each actor or individual practicing entrepreneurship. Regional and community development serve as a process based on initiative, creativity, and independence, along with government activities to enhance the concerned community's social, cultural, and economic conditions to become the integrity of the nation's progress. The process is characterized by essential elements, such as (1) community participation to improve their lives based on their strengths and abilities and (2) services and technical assistance from the government to generate initiative, along with determination to help oneself and willingness to help others. Thus far, in every sustainable development process, local stakeholders provide the participating community with a foundation of sustainable principles (social equality, economic prosperity, and ecological sustainability) (Kitchen & Marsden, 2009;Ley, 2017). This process was actualized in plans and programs, and their implementation was based on "empowerment" through guidance, coaching, and technical assistance to foster selfreliance and identity as human resources with the strength and ability to fight for their quality of life. Empowerment could positively affect behavior change, development, and contribution in the community (Alsop, Bertelsen, & Holland, 2005;Wilmsen et al., 2012) through development communication between elite governments as policymakers and the rural Entrepreneurial Dynamics and Typology..... (Handono, Hidayat, Purnomo, and Toiba) community as the object of development (Haji, 2021). Therefore, the results of case studies from the three villages unveiled variations in development communication between entrepreneurial drivers and each actor's role, forming a typology for each village based on the performed business, as displayed in Table 4. Table 4 demonstrates that the three villages (Kucur, Gubugklakah, and Bambang) were considered agricultural villages, but typology variations existed in each village. Initially, the entire community in the three villages was considered agricultural cultivators, with their respective superiority in agricultural products. In particular, Kucur Village had superior cultivation of orange trees. Gubugklakah Village possessed a superior apple crop. Meanwhile, Bambang Village owned two superior products: coffee and corn. These three villages also had vegetable commodity farming businesses, such as carrots, cabbage, and onions. However, people's livelihoods have developed to be more diverse. The potential of each village depicted variations in their typology, such as a traditional village, agricultural industry, and agritourism, respectively, with the following description. --- Agricultural Industry Kucur Village had potential in the agricultural and non-agricultural sectors, evidenced through the existence of various occupations, such as tour guides for Bukit Jabal and Lembang Gunung Sari (LGS) tours, managers of coffee farming products, breeders, wood artisans, and coolies as well as construction workers. Referring to the rising typology of Bambang Village, the people of Kucur Village had adopted many innovations, which, in this case, were inseparable from several entrepreneurial drivers (government, non-government, and individual actors). Moreover, one of the business actors (ED) in Kucur Village had developed digital marketing-based marketing of agricultural products, which could reduce the high margins of intermediaries to farmers. Likewise, the Mandiri Tani Republic (RTM) successfully conducted experiments on processing coffee into coffee powder and marketed it offline and online. Agriculture 4.0, or the agricultural industry, was actualized through a shift from traditional agriculture to smart agriculture practices, implementing the extensive use of the Internet of Things (IoT), providing future opportunities for farmers (Futemma et al., 2020;Morris & Bowen, 2020). In these agricultural industry-related efforts, the people of Kucur Village understood the strategies of digital-based entrepreneurial sustainability. --- Agritourism Gubugklakah Village, with its beautiful landscape and closeness to the Bromo Tengger-Semeru National Park (TNBTS), has great potential for various tourism activities. These include agritourism, homestays, transportation services such as jeeps, and opportunities for locals as tour guides. This potential open up other business opportunities in the trade sector, such as through shops and small stalls along the village's roads. In addition, a multi-function forest area could serve as a tourism forest, providing occupations for the local community (as tourism guides and culinary business managers). Entrepreneurship developed in this village was considered tourism-based, assisted by the Ministry of Tourism and Creative Economy policy in activating villages through developing tourist villages as a form of commercial urban agriculture. This business model combines agricultural production and tourism with developing agritourism functions from integrated and sustainable economic to environmental and social benefits (Kumalasari, Gutama, & Pratiwi, 2018;Yang et al., 2010). Entrepreneurial Dynamics and Typology..... (Handono, Hidayat, Purnomo, and Toiba) This study disclosed that the people of the three villages became the embodiment of rural entrepreneurs due to their skills in pursuing the agricultural and non-agricultural sectors. Entrepreneurship in the three villages stemmed from the community's ability to navigate opportunities and adapt them to each village's potential. Thus, village potential development was performed with a commercial orientation to boost the local community's living standard and contribute to village development. The variety of entrepreneurship in the three villages was inseparable from the "actor's role" concept to develop village potential as a livelihood diversification. Developing entrepreneurship required a network of several actors from farmers, government, universities, and stakeholders to support change activities and entrepreneurship development. --- Traditional Farming Village Branding The heterogeneous livelihoods of Bambang Village's people were conducted by utilizing abundant natural resources of sandy soil, dairy farms, and timber trees. Through this typology, the people of Bambang Village were considered business actors residing in a comfort zone due to the availability of resources in the forest. However, they tended to be less reactive or spontaneous or categorized as conservative, meaning maintaining the existing traditions and customs rather than prevailing circumstances. For example, sand mining business actors who ignored ecological sustainability focused on making profits. Hence, when a disaster such as a landslide occurred, they made efforts to deal with it. The conservative model assumes a reluctance to innovate, especially in preventing severe challenges, and tends to depend on external funding (Huttunen, 2019;Zhu et al., 2021). --- Typology of Rural Entrepreneurs (RE) The three observed villages had implemented several development programs, including community livelihood activities. The role of actors in entrepreneurship was marked by the involvement of government elites, economic institutions, business actors, and socio-cultural actors, as illustrated in Table 5. Table 5 exhibits the role of RE based on economics, politics, and social relations, contributing to village development in political, economic, social, and strength aspects to gather resources. Business continuity by entrepreneurs in rural areas depends on the involvement and mutual collaboration between 'development relations' and 'empowerment' (Fuller-love et al., 2006;Rădulescu et al., 2014;Suastika, 2017) through the identification of values and social problems adapted to local resources (social, economic, and cultural). The details of the actors from the three villages are as follows. 1. Social politics: Village government political elite (village heads, village secretaries, village treasurers, hamlet chiefs, and other village officials) 2. Economy (institution): Head of BUMDES, chair of Village Unit Cooperative (KUD), chair of LMDH, chair of Ladesta, chairs of farmer groups and farmer group associations 3. Social relations: Business actors (orange, apple, corn farmers, and others) and farmers. The observation results listed in Table 5 disclosed that the local government held the power to contribute to village development from political, economic, and social perspectives. Following the theory cited by Cochran (1965) entrepreneurship is closely related to sociological aspects, in which actual economic development becomes the emphasis on cultural values roles, determining the supply of entrepreneurs. Some points underlying Chochran's theory encompass (1) entrepreneurs as a personality of community capital, implying that an entrepreneur is part of society that reflects one's skills, style, and motives; (2) capital personality as a derivative of social conditioning, in this case, the power of an entrepreneur in his business can play a social role to create an innovation and conditioning in society; and (3) expectations of the role and role of entrepreneurs, and types of parenting and schools and their influence on the intrinsic character of the executive. However, the reality in society depicted that one's internal factors in growing up were highly influential in parenting and determining life goals in the future. The current decentralized policy gave the village government autonomy in managing its potential to improve its people's quality of life. Thus, as a local leader, the village government could influence individual entrepreneurs to develop village potential into business opportunities with a socio-cultural approach to the local community. When perceived from the personality of the government elite, the Javanese people tend to follow the decisions of their leaders with the assumption that the leaders, who are informative, creative, fair, motivating, responsible, and compromising on a common consensus, serve as the "fathers" for their community (Muzayyanah, Syahlani, Suranindyah, & Haryadi, 2014;Sutiyo & Maharjan, 2017). K* G B K G B K G B K G B Political In addition, Table 5 also illustrates how economic institutions contributed to political, social, and economic development in the three villages. Economic institutions in the three villages included (i) BUMDES with an average low contribution. There was a strong level of power to collect resources by BUMDES of Kucur, assisting the farming community in the Entrepreneurial Dynamics and Typology..... (Handono, Hidayat, Purnomo, and Toiba) village to access the leased land of BUMDES. Meanwhile, in the other two villages, the average contribution of BUMDES was considerably medium and low because it was limited to distributing subsidized fertilizers, with no participation in other innovations. Furthermore, the second economic institution (ii), forest farmer groups and farmer groups, had an average medium level of contribution in political, social, and economic aspects concerning the daily activities of the farming community to work on taxed land (owned or leased), such as in Komplangan or Tetelan land, owned by Perhutani. There were several strengths in the role of economic institutions, including political, social, and resource gathering in the villages of Gubugklakah and Bambang, because these two villages were projected to sustain their tourism forest development, especially in Gubugklakah. In Kucur Village, access to Perhutani's arable land was limited because the community considered Komplangan or Tetelan land an addition; hence, the cultivated commodities were easily maintained. The third economic institution (iii), Ladesta or tourism village institutions, possessed an average high level of contribution in Gubugklakah Village, considering the entrepreneurial typology of the village, "agritourism," thereby significantly contributing to the continuity of village development. In addition to the aforementioned economic institutions (i, ii, and iii), the role and contribution of individual business actors (from the political elite, agricultural entrepreneurs, skippers, and farm workers) were deemed pivotal. Business actors from the political elite could contribute to village development, depending on one's power to influence the actions and behavior of other entrepreneurs. Meanwhile, business actors, from farmers and skippers, contributed moderately to rural development, as they created job opportunities for other people in their village. For example, farming actors and skippers with more land area and capital ownership required labor from the farm workers to work on the land. However, farm workers did not contribute much to village development due to their tendency to meet the necessities of life for their households compared to operational village programs. The relationship among the business actors led to various typologies and connections, as portrayed in Figure 2. Figure 2 demonstrates that each typology of RE based on politics, economics, social relations, and an entrepreneurial spirit presented rural development dynamics. Economicand political-based typologies of RE strongly influenced the development of the village community. Furthermore, the typology of RE based on social relations and an entrepreneurial spirit had less influence on community development due to various roles and contributions to village development. The contribution of this diversity was precisely the impact not prominently visible and was considered to have a significant impact on rural development. Apart from that, in the blue indications, a reciprocal relationship existed between one indicator and another; for example, economic asset-based RE boosted entrepreneurship because, in reality, in rural businesses, anyone with access to significant economic assets had power in all social, economic sectors as well as politics within rural society. Elaborating further, Wiesmann and Hurni (2011) Livelihood or entrepreneurial activities in the three villages were oriented toward village development, emphasizing the role of RE by involving strategies for action, change, and innovation. This actor approach model comprises four core components: (i) action as a dynamic interaction between activities, (ii) action strategy as a combination of actions, (iii) exposure to activities and means for dynamic (socio-economic) action conditions, and (iv) enforcement of the action meaning in institutions that determine standard values, norms, evaluation of specific actions, and results of actions by actors. Through this approach, as proven in research, every actor or business actor is capable of maximizing actions in dealing with dynamic conditions by implementing the strategies through institutions and further achieving village development (Hidayat, 2017;Hidayat et al., 2019;Prastyo & Hidayat, 2016). Furthermore, the analysis results revealed that RE positively impacted rural development in various aspects, indicating that a better role of sustainable entrepreneurs could lead to better human resources, integration, and collaboration in village development. Entrepreneurial Dynamics and Typology..... (Handono, Hidayat, Purnomo, and Toiba) These results are in line with several other studies (Adro & Franco, 2020b;Dhewanto et al., 2020;Escandón-Barbosa et al., 2019;Nwankwo & Okeke, 2017;Richter, 2019;Steyaert & Hjorth, 2006;Yin et al., 2019), emphasizing that the role of entrepreneurs in rural areas could develop the self-quality or competence, finance, networking, and other aspects. Additionally, better village development is apparent to affect the social structure (village government), leading to the creative spirit, good leadership, innovation, and sound managerial practice in any management sphere (Barraket, Eversole, Luke, & Barth, 2019). The spirit of entrepreneurship in leaders (village heads) or individuals certainly provides a better outcome in developing sustainably and environmentally friendly villages (Kilkenny, 2006;Kitchen & Marsden, 2009;Sahoo & Mondal, 2020). The findings of this study follow Imelda et al. (2022), revealing the performance of task-oriented farmers. Furthermore, the entrepreneurial character of farmers who are work-oriented and optimistic about results would boost the performance of lowland rice farming (Handono & Puspita, 2019;Nugroho et al., 2021;Purnomo, 2018;Purnomo et al., 2021). This notion means that the enthusiasm and role of entrepreneurs in rural development refers to the the result or target orientation. Hence, entrepreneurs' ability and responsiveness to market challenges and local wisdom have been expected to increase. In addition, this study's findings align with the entrepreneurs' attachment to the village as an essential aspect of development commitments. The village community's feeling emphasized by rural entrepreneurs is essential to the expected outcomes and cooperation (Cucari et al., 2019;Hazarika, 2016;Komppula, 2014;Lang & Fink, 2019). --- CONCLUSION Conclusion The three observed villages had three key actors who possessed the ability to accumulate these resources: the local administration, successful entrepreneurs, and tourist activists. Briefly, this study discovered that the dynamics of rural entrepreneurship from the New Order era to the Jokowi era exhibited various rural developments, particularly in the direction of RE in the social, economic, political, and social network aspects of entrepreneurship. There were three typologies of RE: (1) political asset-based RE, (2) economic asset-based RE, and (3) social network-based RE. The RE with the most potential to drive social development in rural areas of the three typologies exhibited significant political influence, abundant economic assets, and broad social networks. Therefore, providing access to economic assets and social networks and increasing entrepreneurial spirit become the primary keys to encouraging social development in rural areas. --- Recommendations This study's findings are expected to apply to the government, NGOs, or other development agents in formulating more effective development strategies for RE. Furthermore, the recommendations for further research include integrating and collaborating internally and externally (entrepreneurs) in rural development in an integrated and and Rural Development Research sustainable manner based on local wisdom and creativity. Recommendations for RE contributing to the development must be carried out on various aspects and potential of the village, apart from impacting village development that will indirectly affect sustainable business development. In addition, future researchers are encouraged to explore the phenomenon of business actors in rural areas by implementing the existing entrepreneurial theories (by Weber, Schumpeter, and others). --- Authors' Contributions: SYH: Writing the original draft; preparation, creation, and presentation of the published work, explicitly writing the initial draft (including substantive translation). KHY: Project administration; management and coordination responsibility for the research activity planning and execution. MPR: Conceptualization; ideas, formulation, or evolution of overarching research goals. HTB: Review and editing; preparation, creation, and presentation of the published work by those from the original research group, specifically critical review, commentary, or revision-including pre-or post-publication stages. --- Conflict of interest: The authors declared no conflict of interest.

The term "rural entrepreneurs (RE)" has become a social and rural development factor in developing countries, particularly Indonesia. This research aimed to analyze the dynamics of rural entrepreneurship and describe the typology of entrepreneurship. This research was conducted in three typologies of villages, such as (1) agricultural industry, (2) agritourism, and (3) traditional area, in Malang Regency, East Java, Indonesia. Data were collected through village observation, in-depth interviews with 90 informants, and Focus Group Discussions (FGDs). The dynamics of rural entrepreneurship were affected by various factors: (a) politics, (b) social networks, and (c) economics. Furthermore, three typologies of RE were identified: (1) political-based RE, (2) economic assetbased RE, and (3) social network-based RE. Meanwhile, entrepreneurs could accumulate rural resources, comprising local government and social-economic networks. In summary, the primary keys to encouraging social development in rural areas were providing access to economic assets and social networks and boosting entrepreneurial spirit.

Introduction Religious groups draw people together, sustaining a sense of community during a crisis. A mature body of sociological research demonstrates that ongoing meaningful interactions bond members to their community (Collins, 2014;Kanter, 1972;Stark & Finke, 2000). Government policies and medical guidelines that interfere with in-person interaction, such as the directives to maintain "social distance" and isolation, affect these community bonds. The World Health Organization and the US Center for Disease Control and Prevention encouraged citizens to limit contact with non-household members. Where that was impossible or impractical, they recommended that everyone maintain at least six feet of space between non-household members. Experts proclaimed that these strategies were the best tactics for limiting the transmission of COVID-19 at the beginning of the pandemic (CDC, 2020b). However, these strategies also prevented people from gathering in collective worship, which is for many religious people an essential element of their spiritual life. As Ecklund (2020) notes, "COVID-19 complicates the usual response of religious organizations in times of crisis because the prescription for protecting against the disease requires physical isolation from organizations whose main purpose is to gather." While some religious groups complied with the mandates, others resisted the government COVID-19 decrees. In particular, many US religiously conservative groups (e.g., evangelical Protestants and Christian nationalists) objected to government interference with their faith (Ali et al., 2020;Baker et al., 2020;Duran et al., 2020;Perry et al., 2020aPerry et al.,, 2020b;;Pew Research Center, 2020;Stein, 2021). Pointing to puzzling inconsistencies (such as bars and restaurants facing fewer restrictions than churches), these resisters argued that COVID policies infringed on their right to worship (Ali et al., 2020;Baker et al., 2020;Duran et al., 2020;Perry et al., 2020aPerry et al.,, 2020b;;Pew Research Center, 2020;Stein, 2021). Researchers found that religious conservatives and devoutly religious people and places with higher numbers of them were less likely to wear masks, social distance, or comply with stay-at-home orders (Adler Jr et al., 2021;Corcoran et al., 2022;Gonzalez et al., 2021;Hill et al., 2020Hill et al.,, 2021;;Perry et al., 2020a;Pew Research Center, 2020;Scheitle & Corcoran, 2021;Schnabel & Schieman, 2022). Surveys fielded in March and June 2020 indicated that white evangelical Protestants in the USA were among the least likely to comply with these policies and the most likely to support the restrictions' termination (Pew Research Center, 2020). Christian nationalists (i.e., those who ascribe to a politically infused conservative religious ideology) were particularly resistant to COVID-19 mandates viewing them as contrary to economic prosperity and liberty (Perry et al., 2020a(Perry et al.,, 2020b)). This resistance was likely also rooted in cultural distrust of medicine, the government, and science (Armer & Radina, 2006;Baker et al., 2020;Galanter, 1999;Gasta<unk>aduy et al., 2016;Glassman, 2018;Gonzalez et al., 2021;Hill et al., 2020Hill et al.,, 2021;;Miller & Karkazis, 2013;Offit, 2015;Stein et al., 2022;Thompson & Kisjes, 2016;Whitehead & Perry, 2020). In response to these restrictions, some congregations simply ignored the guidelines (Baker et al., 2020;Perry & Grubbs, 2022). Others canceled services. Many adapted, some through hybrid services (e.g., online and in-person) (Almog, 2020;Baker et al., 2020;Church of England, 2020;Gjelten, 2020;Hartford Institute for Religion Research, 2021;Lange, 2020). Most of the research on religion, behavior, and pandemic policy, focuses on US evangelical Protestants and Christian nationalists. Fewer studies examine how pandemic policies affect closed religious communities (CRCs) 1 -religious communities that limit interactions between members and 2020 from all Ohio and Pennsylvania communities, home to the two largest Amish populations in the world. The newspaper publishes regular dispatches from scribes in Amish and Mennonite communities that provide updates on the authors' communities. While some communities did alter their church practices, we find that church services went on as usual for many. --- The Case: The Amish and Mennonites Anabaptist groups, including the Amish and Mennonites, separated from mainstream Protestants in sixteenth-century Europe due to differences in core beliefs. Anabaptists believed the church and the state should operate as separate entities; the church free from state power. Additionally, the name of Anabaptists, or re-baptizers, reflects the belief in adult baptism. These groups believe individuals should consciously choose to join the church when they are of an age and capacity to do so. After suffering much persecution in Europe, Amish groups migrated to the USA. Mennonite groups dispersed and settled around the world. Some groups emigrated to the USA (Kraybill, 2001;Loewen & Nolt, 2010;Nolt, 2016;Scott, 1996). Amish and Mennonite groups are distinct, with many variations within each group.2 Even so, Anabaptist groups share specific core values. These core values not only serve to unite people within the Amish and Mennonite communities but also distinguish boundaries between the CRCs and broader society. Sectarian characteristics emphasize differences that separate Anabaptist groups from the dominant culture, including behavior, dress, and social rituals. Anabaptists prioritize following God's law while emphasizing the separation of church and state. Sunday church services are recognized as a sacrament and communal worship as a sacred duty within Anabaptist communities, centralizing the role of religious beliefs (Kraybill et al., 2013). Amish and Mennonite people generally obey state and national laws but vary in the extent to which they will refuse directives that conflict with their understanding of God's higher law (Hostetler, 1993;Yoder, 2003). For example, some conservative Amish groups refuse to use reflective tape on their buggies based on theological grounds, despite state laws mandating such things for safety (Anderson, 2014;Yoder, 2003). Amish and Mennonite groups follow Biblical teachings while emphasizing the importance of living their faith through daily actions. Members of Amish and Mennonite groups build and maintain bonds within the church, organized as a spiritual and social community. The bonds translate into a strong community where members rely on one another for support (Hostetler, 1993;Kraybill, 2001;Nolt, 2016). While they share core values, Amish and Mennonite groups differ in significant ways. Many Amish congregations are constrained by geographical boundaries, as the primary means of transportation in Amish communities is horses and buggies. Approximately 20-40 families constitute an Amish congregation, and all congregations in a broader area form a settlement. The geographical closeness ensures Amish families can travel to members' homes within the same community by horses and buggies in a reasonable time frame. Mennonites are not constrained by geographical boundaries, as most Mennonite groups accept the automobile as the standard mode of transportation (Hurst & McConnell, 2010;Kraybill, 2010). In addition to different modes of transportation, technology usage varies across and within Amish and Mennonite groups. For example, many Amish people do not have access to a home telephone. Some Amish groups might use a shared community telephone outside the home for communication purposes, while the most conservative Amish groups generally rely on non-Amish neighbors if they need to make a telephone call (Kraybill, 2001;Scott & Pellman, 1999). In contrast, many Mennonite groups have no restrictions on home telephones. The prevalence of cell phones provides greater access to communication technology; however, many Amish churches limit cell phone usage to work-related issues (Ems, 2014). These guidelines encourage Amish people to keep cell phones out of the home. In-person collective rituals are a cornerstone of the Anabaptist faith. Amish services are full-day events held in members' homes. A typical Sunday includes worship, which lasts approximately three hours, followed by a meal and youth gathering. Church members spend the remainder of the afternoon visiting one another. People will often attend the services of other congregations on "off" Sundays when the home congregation does not hold church (Hostetler, 1993;Kraybill, 2001;Nolt, 2016). Most Mennonite church services are shorter than Amish services, including Sunday School meetings, and take place in a communal building. Mennonite churches often emphasize social time before or after the service. Though welcome, visitors are not as common in the Mennonite service, as there is no "off" Sunday. Mennonite congregations vary in size, ranging from 75 to 200 members, depending on affiliation (Christian Light Publications, 2020;Miller, 2019). The size of Amish church gatherings, with 20 to 40 families crowding into a home for worship and the post-service meal, creates opportunities for viral outbreaks once one member becomes infected (Thompson & Kisjes, 2016). For example, during the summer of 2014, a measles outbreak ravaged one of the largest Amish communities in the USA, spreading primarily through church contacts (Gasta<unk>aduy et al., 2016). Amish and Mennonite groups prioritize religious rituals, as these meetings represent a core value within the community. They are reluctant to disrupt their spiritual practices, even when confronted with viral contagion. Their priority on community fellowship has pronounced implications during the COVID pandemic. Likewise, patterns of (non)compliance to CDC guidelines among the Amish and Mennonites provide insight into the relationship between religious values and secular public health mandates. --- 3 Journal of Religion and Health (2022) 61:4260-4281 --- Data and Method We collected data from The Budget, an international Amish/Mennonite correspondence newspaper that focuses on US communities and currently has a circulation of around 18,500 (Stein et al., 2021). The Budget is "an important institution, serving as the major means of communication among Amish settlements" (Hostetler, 1993, p. 377). The Budget, published weekly, includes dispatches from scribes-writers who live in Amish/Mennonite communities3 across the US and world. Scribe letters typically follow a general pattern, including a report on the weather, church news, community news items-births, deaths, member health information, visiting practices, and a narrative section including any other topic of interest (Adkins, 2009;Galindo, 1994). While some scribes submit an entry every week, many write every other week or once a month (Stein et al., 2019). Most writers report on one congregation, but some cover two or more from their area. Amish intentionally keep their communities small to facilitate Sunday worship in homes. Consequently, when congregations grow too large, they split to maintain a manageable number of families (Stein et al., 2020). The scribes who report on more than one congregation are likely reporting on congregations that have split. Alternatively, some writers report on neighboring congregations that do not have a scribe to represent their congregation. We use "community" to refer to any congregation covered by the writer. We read and content-analyzed all Ohio and Pennsylvania scribe entries published between March 18th and May 6th, 2020. We coded eight editions of The Budget-March 18, and 25, April 1,8,15,22,and 29, and May 6th. Scribe letters typically report the previous week's events, which means the March 18 th -May 6th editions reflect the period from mid-March through the end of April. Ohio and Pennsylvania house the two largest Amish populations in the world and roughly 44% of all Amish congregations in the USA (Young Center for Anabaptist & Pietist Studies, 2020). It is important to note that The Budget started publishing CDC recommendations on COVID-19 in the March 11th edition, exposing readers to the guidance. We identified religious affiliation based on the location of church services mentioned in the entries. Most Amish congregations hold Sunday church in members' homes, and scribes report the site in their dispatch (e.g., "church was at Elmer Yoder's last week"). In contrast, many Mennonites use dedicated church buildings, and Mennonite scribes usually provide their church's name in their entries' header. Groups classified as Mennonite in our study include groups that meet in a communal building for church services. This classification contains groups not traditionally defined as Mennonite, such as Beachy Amish-Mennonite groups. The classification also includes horse and buggy Mennonites who use a meetinghouse for church services (Kraybill & Hurd, 2006). 4We counted the number of entries reporting in-person Sunday church to measure non-compliance with social distancing and isolation guidelines. In-person Sunday church violates the CDC's guidelines to avoid or limit contact with non-household members and avoid large gatherings. We coded whether Sunday church occurred, was canceled, or altered due to COVID-19. Specifically, we classified these alterations into broader themes. It is important to note that some communities tried different strategies as the pandemic progressed, so themes and subcategories are not mutually exclusive. Some scribes accounted for multiple weeks, reporting that their church met in person some weeks and canceled church other weeks. On "off" Sundays, those without services, Amish writers often list community members who attended church elsewhere as visitors. We classified these reports as indicating inperson Sunday church activity as any in-person rituals can spread the virus within the community. We also collected data on the number of times visitors (i.e., not community members) were mentioned as having attended in-person services. Of the 1,503 entries we coded, 78% (1178) mentioned Sunday church, whether in-person, canceled, or remote. Since Sunday church is our analytic focus, we excluded entries not mentioning Sunday church. Of these 1178 entries, we estimate the percentage that reported each category and subcategory. We provide excerpts from entries to illustrate these (sub)categories and have replaced names of people and congregations with initials. Because the number of entries varies by the community (i.e., communities that submit more entries have a greater influence on the results), we also report the percentage of communities falling into each of these (sub)categories to capture the extensiveness of these (sub)categories (Krueger, 1997). If we classified any entries tied to a particular community in a (sub)category, we counted that community as reporting that (sub)category. This allows for all communities to have an equal influence on the results. It is important to note that while most communities comprise one congregation, some have two or more affiliated congregations, and the same scribe reports on them all. There are 334 communities in our sample. Our university's Institutional Review Board determined that this study does not meet the definition of human subjects research. --- Findings We present results for two units of analysis-entries and communities. Using entries allows us to capture the number of times in-person Sunday services were held or canceled, reflecting the prevalence of (non)compliance with social distancing guidelines. However, some communities practiced more (non)compliance than others. Since there is variation in the number of entries submitted, some communities have 1 3 Journal of Religion and Health (2022) 61:4260-4281 a greater influence on the results when entries are used as the unit of analysis. Thus, we also report results with communities as the unit of analysis, which represents the percentage of communities that reported a given (sub)category at least once, that is, the prevalence of the (sub)categories across communities (Krueger, 1997). --- Business as Usual Tables 1 and2 present the prevalence of Budget entries and communities in the study period that mention Sunday church (either in person or in a modified form). Sixty-five percent of entries reported in-person Sunday church, and 94 percent of all communities had at least one entry that reported in-person Sunday church. Thus, nearly all religious communities in our sample held Sunday church at least once in-person during mid-March and April. Of entries mentioning in-person Sunday church, approximately 97 percent exclusively offered it in-person, with only 3 percent offering it in-person with a remote option. The prevalence of in-person Sunday church varies by religious affiliation, with 73 percent of Amish entries reporting inperson Sunday church compared to roughly 40 percent of Mennonite entries. Church visitors-people who visit a community to which they do not belongare an important part of Sunday church among the Amish and Mennonites. Of the entries that mentioned in-person Sunday church, roughly 61 percent mentioned visitors, corresponding to approximately 82 percent of communities identifying visitors in at least one of their entries. More Amish entries reported visitors (65%) than Mennonite entries (38%). Scribes will often note out-of-state visitors as well, "We had a large crowd at church yesterday. Visitors were J. and D. G. from Minerva, L. and R. M. and family from Tazewell, W. and R. A. G. from Guys Mills, PA, J. and C. S. from Holmes County [...]" (OH, Mennonite, 3/18). Not only did members of different households within the same communities gather for in-person Sunday church, but people from other communities also attended church (some crossing state lines). The importance of Sunday church is evident amongst the Amish and Mennonites, as nearly all communities reporting into The Budget continued church after the shutdown. One Amish scribe announced that sickness in the host household led the community to relocate, but not cancel, Sunday church: "Church was planned to be at M.'s, but was changed due to sickness. (Amish, OH, 3/25). In some instances, the cancelation of services led to more visitors in communities still meeting regularly: "Visitors today were P. M.s of Somerset, who came since their church was canceled because of Pennsylvania's quarantine concerning the COVID-19" (PA, Mennonite, 3/25). An Amish scribe in PA wrote, "In the present situation we were thankful for the privilege and blessing of weekend meetings. N. G. and C. S., both from Ohio, were our speakers" (3/25). Visitors were present, and the speakers traveled from out of state. While some people continued attending Sunday church outside their communities, others were cautious. In OH, a Mennonite scribe notes, "Our church attendance was smaller, with folks taking precautions to avoid getting sick" (3/18). A Mennonite writer in PA clarified that their communities exercised precaution, "Yes, we had church yesterday. The news message was put on the church line not to attend if you had any cough or cold" (3/25). The church line, also referred to as a conference line, is a telephone service where people have a pass code or pin number to call in and listen to a presentation (live or prerecorded). In this case, the church line allowed people to call in and listen to the live church service. Some scribes noticed illness affecting their church attendance: "Church on Sun. was at S. M.s. [...] Lots of people were home sick and others were coughing. Some had fever in church, not realizing how sick they were before they went. Lots of little children sick [...]" (PA, Amish, 3/25). While the scribes recognized the impact of sickness on attendance, there was no mention of canceling church or restricting travel. --- Alterations to In-Person Church Thirteen percent of entries mentioning in-person Sunday church indicated altering church in some manner. Of communities describing in-person Sunday church, 16 percent reported adjusting church in at least one of their entries. Of the entries mentioning alterations, roughly 47 percent said they canceled the traditional post-church meal. This alteration was more common in Amish communities. Churches also deployed a variety of social distancing strategies. Seventeen percent of entries indicating alterations to Sunday church reported only allowing members of the same household to sit together. An Amish scribe in OH writes, "Last Sun. West district did gather in a big shop for services, sitting families together and using the 6' method" (4/22). County health departments distributed social distancing guidelines to Amish communities, as noted by an Amish scribe in OH, "E. H., bishop in South district, got a letter from the health department saying that it's okay to have church services with regulations such as families seated together and no noon meals" (4/22). Dividing the congregation into smaller meeting groups was one of the more prevalent strategies for social distancing. A Mennonite scribe in PA wrote, "Our congregation divided into 4 groups for church again. We met at R. R.'s, B. S.', C. E.'s and the Church. [...] A few of our families do not want to have church via of the conference line, so the Ministry felt they could meet our needs best this way (4/29). Twenty-three percent of the alteration entries referenced prohibitions against social contact, specifically no handshaking or Christian greeting (i.e., holy kiss). Social contact limitations were more common among Amish communities in our sample. Even though these churches restricted social contact, they did not ban visitors, as documented by an OH Amish scribe, "L. M. church was at B. M.s. Visitors were J. M.s and lots of young folks. [...] There was no food served and no handshakes. Quite a few were missing" (4/8). Fifteen entries mentioned other forms of "social distancing." Finally, another prominent form of social distancing was only allowing a small number of people in the congregation, whereas everyone else participated remotely. Eighteen percent of the alteration entries mentioned this: "Our Church continues to share services via Zoom and call in. Eight to ten gather at the church while the rest of us listen in from home" (OH, Mennonite, 4/22). At least one of the scribes reported having their elderly members attend Church in person, yet, they are among 1 3 Journal of Religion and Health (2022) 61:4260-4281 the most at risk of severe COVID-19 complications (CDC, 2020c). Overall, more commonly reported among Mennonite communities than Amish, this form of social distancing allowed limited numbers to enjoy in-person church. --- Cancelations/Postponements While many communities continued having in-person church, others canceled for some time. Table 3 presents the prevalence of entries and communities that canceled in-person services and those using adaptive technology. Forty-one percent of entries reported canceled in-person church, corresponding to 66 percent of communities with at least one entry mentioning canceled in-person church. Of those, only 37.5 percent of entries and 31 percent of communities mentioned remote church. Cancelation varied by religious affiliation, with 33 percent of Amish entries reporting canceling in-person church compared to 68 percent of Mennonite entries. They also differ in whether they offered technology to replace the canceled in-person service, with Mennonites being much more likely to do so (79% versus 14% of entries). An Amish scribe in OH notes, "The coronavirus flu is probably the main talk worldwide. A decision was made by the Ohio Steering Committee, the Health Department, and bishops to cancel our church services the next few Sundays" (3/25). Many scribes noted the impact of cancelations on their community. A Mennonite scribe in OH writes, "I do believe we were designed for community and to have face to face conversations. I miss my church family and my extended family and look forward to a time when we can interact freely" (4/8). Several scribes indicate the importance of fellowship. An Amish writer in OH notes, "No church news, since it's been 4 weeks since we've last had church, since the coronavirus has been around. We definitely do miss not having our church services, and the fellowship with our other church people" (4/22). Another Amish scribe writes, "We skipped a few weeks of writing, since we did not have any church news to report. We just hope and pray that we can have church service again soon. We really miss the worship and fellowship together" (OH, 4/29). --- Church Using Technology While the above alterations to in-person Sunday church are generally evident across Amish and Mennonite communities, Mennonite communities were more willing to use adaptive technology. Approximately 17 percent of entries mentioned using technology, and 21 percent of communities had at least one entry describing technological adaptations for church, whether in addition to in-person services or in place of them. Of the entries mentioning technology, 87 percent reported that members connected to a conference line/telephone. One OH Amish scribe noted, "The past few Sundays we had the privilege of listening to some very inspiring sermons via conference calls. In these trying times we adapt to circumstances the best we can. Let's continue to work together in this and also be respectful of the guidelines set in place, even if it means canceling church services, etc." (4/22). Roughly 13 percent of the entries mentioning technology described virtual church hosted on live streaming services such as Zoom or Facebook. Additionally, three entries mentioned drive-in Church. A Mennonite scribe described how, "The pulpit was brought outside, and you could just sit in your car. They had some connection via the car radio that you could hear the message." (PA, 3/25). --- Discussion During the beginning of the pandemic, when social distancing and isolation were vital for saving lives, how did CRCs, particularly those that limit communication technologies, respond to the mandates? The current study examined this question in the context of the Amish and Mennonites. We found that Sunday church continued as usual for many communities, although some canceled church or adapted services. More Mennonite communities altered church than the Amish, which is consistent with the fact that they tend to allow more technology. Religious rituals can amplify the spread of COVID-19 and become superspreader events (Majra et al., 2021). The CDC recommended cessation of in-person religious gatherings or substantial alteration following social distancing guidelines (CDC, 2020d). Some religiously conservative groups met the moratoria on worship with resistance. Numerous studies demonstrate a negative association between religious conservatism (or intense religiosity) and healthy pandemic behaviors (e.g., mask wearing, social distancing, and staying at home) (Adler Jr et al., 2021;Baker et al., 2020;Corcoran et al., 2022;Duran et al., 2020;Gonzalez et al., 2021;Hill et al., 2020Hill et al.,, 2021;;Perry et al., 2020a;Pew Research Center, 2020;Scheitle & Corcoran, 2021;Schnabel & Schieman, 2022;Stein, 2021). Of course, religious resistance to public health mandates during pandemics is nothing new. Similar forms of resistance occurred in some congregations during the 1918 influenza pandemic (Marisam, 2007;Schoch-Spana, 2000). In both the 1918 influenza and COVID-19 pandemics, some congregations made headlines for violating mandates to close, alter, or restrict the size of their worship services (Galishoff, 1969;Green, 2020;Luscombe, 2020). What is new is the ability of some congregations to move their religious services online when faced with such mandates and the encouragement of governmental agencies to do so (CDC, 2020d). This solution makes such restrictions on gathering particularly burdensome for the Amish, and similar groups, who cannot shift to online services due to their sacramental limits on technology use. The Amish and Mennonites believe in abiding by the government's law when it does not conflict with God's law (Hostetler, 1993;Yoder, 2003). In cases of conflict, they comply with what they believe is God's will; however, they also believe it is God's will to do no harm. We found that several churches canceled church services early in the pandemic based on advice from local health departments but returned to in-person church a few weeks later when the mandates had lifted. Thus, many churches complied when they believed the government mandated them to do so, but 1 3 Journal of Religion and Health (2022) 61:4260-4281 not when the measures were mere recommendations. While research suggests that Christian nationalists were more likely to claim that COVID-19 social distancing guidelines violated their freedom to worship (Haynes, 2021;Perry et al., 2020b), a violation of rights was not prominent in the scribe letters. Instead, the Amish and Mennonite scribes emphasized the importance of personal interaction to maintain the spiritual and social bonds across community members. The sacred duty of communal worship in some Anabaptist communities superseded the government recommendations to limit social interaction. In May 2020, the CDC reported an outbreak of COVID-19 in an Amish community in Ohio (Ali et al., 2020). The CDC's contact tracing identified six in-person religious and social events occurring in the two preceding weeks. Our data suggest that some Amish traveled out-of-town for church because state authorities prohibited church in their location. Such travel undoubtedly amplified the virus's spread by connecting geographically distant and otherwise separate communities. Studies have found that USA states that are more religious had, on average, higher mobility during the beginning of the pandemic and were influenced less by stay-at-home orders (Hill et al., 2020(Hill et al.,, 2021)). Our findings advance this research by suggesting that, at least among the Amish, some people traveled out-of-town specifically to attend church services in locations where they weren't prohibited. More research is needed on how restrictions on religious gatherings in locales may facilitate travel to less restrictive locations and impact COVID-19 infection rates. Religious conservatism and intense religiosity have generally been a barrier to healthy pandemic behaviors in the USA (Adler Jr et al., 2021;Corcoran et al., 2022;Gonzalez et al., 2021;Hill et al., 2020Hill et al.,, 2021;;Perry et al., 2020a;Perry & Grubbs, 2022;Pew Research Center, 2020;Schnabel & Schieman, 2022). Yet some Amish and Mennonite congregations found ways to adapt to the COVID-19 health recommendations while maintaining their religious commitments. Adaptations that still allowed in-person services were better received and implemented among the Amish and Mennonites. Some ceased the communal meal following services, had only family sit together, prohibited physical contact, or had rotating small groups attend in person. These concessions were easier to accommodate as they still allowed them to hold to their sacrament of in-person fellowship. Yet, many communities did not alter church services in these ways. Examining what factors affect whether communities within the same religious affiliation voluntarily decide to enforce social distancing guidelines and implement them are fruitful avenues for future research. For many religious communities, concerns over COVID-19 combined with government restrictions led to canceling in-person services and transitioning to remote services through technology (Oxholm et al., 2020;VanderWeele, 2020). Indeed, the CDC recommended this (CDC, 2020d), and many congregations complied. A 2021 nationally representative sample of US congregations found that 80% of congregations offered a hybrid option for religious services (i.e., both online and in-person), 5% of congregations had services solely online, and 15% had services exclusively meet in person (Hartford Institute for Religion Research, 2021). Ninety percent of evangelical Protestant congregations offered a hybrid option (Hartford Institute for Religion Research, 2021). Unfortunately, recommendations to shift to virtual services fail to consider religious communities without the financial resources to do so and communities, like the Amish and Mennonites, whose religious beliefs prohibit or restrict it (Hostetler, 1993;Nolt, 2016). It is thus not surprising that few Amish communities used technology to hold church. On the other hand, Mennonite communities, which are more open to technology, reported using it more often for services than the Amish. While the Amish and Mennonites share many beliefs and practices, including the centrality of face-to-face interaction to bond community members together, the difference in their beliefs regarding communication technologies profoundly affected how they altered church based on social distancing and isolation guidelines. The Amish groups without technology resumed face-to-face church when government restrictions were lifted, while Mennonites and Amish groups with access to technology had greater flexibility to stay remote. As such, traditional Amish groups had a greater risk of exposure to COVID-19. More research is needed on how they and other religious groups that restrict communication technologies have responded to the challenges of COVID-19 restrictions. --- Study Limitations There are several limitations to this study. We only have data on Amish and Mennonite communities from Ohio and Pennsylvania submitted to The Budget. Even so, nearly half of all US Amish congregations reside in those states, as do the two largest Amish settlements (Young Center for Anabaptist & Pietist Studies, 2020). Additionally, as The Budget is a prominent means of communication across settlements (Hostetler, 1993;Nolt, 2008), we expect that most communities submit entries, and those submitting represent typical Amish communities. Since entries typically provide information on Sunday church and members who interacted at church, our data may undercount the number of times church was canceled as scribes may submit entries less frequently due to not having church. While there may be fewer entries per community, it is unlikely that a scribe would fail to submit an entry over two months. Additionally, some writers reported the number of weeks church was canceled in their community. Thus, the statistics we report on the prevalence of themes by communities (e.g., the percent of communities that reported canceling church) aid in correcting for undercounting canceled church when using the number of entries as the unit of analysis. Although collecting data from The Budget entries come with limitations, it also has several advantages over other data collection methods. The Amish and conservative Mennonites are CRCs that limit communication technology. Data collection through surveys or interviews is difficult during normal times and nearly impossible during a pandemic

At the onset of the COVID-19 pandemic, government and medical guidelines emphasized social distancing to limit exposure. These guidelines significantly impacted closed religious communities, particularly those opposed to modern technologies, such as Amish and Mennonite communities. How did these religious communities respond to COVID-19 policies in the USA? We draw data from Ohio and Pennsylvania scribe entries published in an Amish/Mennonite correspondence newspaper. While some of these communities altered church rituals to comply with government directives, others maintained communal worship without disruption. Mennonite communities were more likely to conform to guidelines.

& Pietist Studies, 2020). Additionally, as The Budget is a prominent means of communication across settlements (Hostetler, 1993;Nolt, 2008), we expect that most communities submit entries, and those submitting represent typical Amish communities. Since entries typically provide information on Sunday church and members who interacted at church, our data may undercount the number of times church was canceled as scribes may submit entries less frequently due to not having church. While there may be fewer entries per community, it is unlikely that a scribe would fail to submit an entry over two months. Additionally, some writers reported the number of weeks church was canceled in their community. Thus, the statistics we report on the prevalence of themes by communities (e.g., the percent of communities that reported canceling church) aid in correcting for undercounting canceled church when using the number of entries as the unit of analysis. Although collecting data from The Budget entries come with limitations, it also has several advantages over other data collection methods. The Amish and conservative Mennonites are CRCs that limit communication technology. Data collection through surveys or interviews is difficult during normal times and nearly impossible during a pandemic when in-person data collection was dangerous. The Budget represents a unique source of information that allowed us to collect data safely and ethically on 334 Amish and Mennonite communities in OH and PA. --- 3 Journal of Religion and Health (2022) 61:4260-4281 --- Conclusion While research on US conservative religious groups' reactions to the pandemic indicates a general resistance to government mandates during the pandemic (Adler Jr et al., 2021;Corcoran et al., 2022;Perry et al., 2020aPerry et al.,, 2020b;;Pew Research Center, 2020), research on CRCs generally, and those with restrictive technology specifically, are less prevalent. The CDC identified that it is vital to understand CRCs to reduce the spread of COVID-19 (Ali et al., 2020). This study is one of the first to examine how the COVID-19 pandemic and resulting social distancing mandates have affected CRCs that restrict technology due to their religious beliefs and how they respond. The data used in this study provide an unparalleled lens into the lives of the Amish and Mennonites during the COVID-19 pandemic. More research is needed on how members of these communities cope with the pandemic, given limited opportunities for interaction (DiGregorio et al., 2021) and how they adapt to the availability of a COVID-19 vaccine (Scott et al., 2021;Stein et al., 2022). This study highlights how the broader religion and health literature must consider how religious restrictions on technology use affect the ability of certain religious groups to transfer their rituals to virtual formats, which may, in turn, affect the transmission of COVID-19 and other infectious diseases. It also impacts what intervention strategies are likely to be the most effective. Our findings suggest that interventions focusing on modifying in-person religious rituals to reduce transmission will be more likely to be implemented than interventions emphasizing virtual rituals. --- Data Availability Newspaper data is publicly available. --- Declarations Conflict of interest The authors have not disclosed any competing interests. --- Ethical approval Not Applicable. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

At the onset of the COVID-19 pandemic, government and medical guidelines emphasized social distancing to limit exposure. These guidelines significantly impacted closed religious communities, particularly those opposed to modern technologies, such as Amish and Mennonite communities. How did these religious communities respond to COVID-19 policies in the USA? We draw data from Ohio and Pennsylvania scribe entries published in an Amish/Mennonite correspondence newspaper. While some of these communities altered church rituals to comply with government directives, others maintained communal worship without disruption. Mennonite communities were more likely to conform to guidelines.

Introduction The current health crisis has heightened the importance and value of healthy living in preventing and treating health conditions 1. There have been many approaches to promote healthy lifestyles 2 ; some have focused on cardiorespiratory and muscular fitness to prevent and treat cardiovascular disease 3, others conceptualize it as the integration of habits that prolong the healthspan of individuals 4. Activities such as physical activity, consuming nutritious and healthy food, not smoking or consuming recreational drugs, and maintaining an appropriate body weight are essential to this health. The current health crisis has also pushed mental health as one of the priorities since lockdown considerably increased psychological disorders in teenagers and young adults 5. The change in habits amid the pandemic had direct repercussions on these key pillars of wellbeing. Habits are critical for healthy living behaviors; these consist of automated responses that develop through repetition and are often performed without little or no forethought. Figure 1 depicts the four stages necessary for habit formation, starting from a decision that is sustained through time until it becomes an automatic behavior 6. The strongness and automaticity of an established habit may pose difficulty in changing it because short-term behavioral gains may be lost in the long-term context 7. This has to be taken into consideration when designing strategies for habit promotion or formation. Habits are forged closely in the familiar context, but socio-cultural norms and perceptions may shape them throughout time. While habits in early childhood are influenced directly by parents, self-care gradually shifts towards self-direc-tion 8. Elements related to the individual, such as self-consciousness in teenagers and young adults, and interpersonal elements such as belonging to a group or peer pressure may impact the perception and formation of new habits. Habits related to health are no exception 9. --- Raising challenges in Latin America Although all individuals are entitled to equal rights and opportunities, including the right to a healthy life, the integration of habits to prolong the health span seems to challenge underserved communities 10. Latin America, a culturally diverse region with a vast territorial extension, has many challenges in promoting healthy habits. The social context is characterized by deep inequities, where factors such as ethnicity, rural residence, or low educational level translate to inadequate access to health services 11. The eradication of poverty and the reduction of inequality in all its dimensions was a pressing issue in the political agenda; however, the pandemic aggravated these conditions. Factors such as informal employment, limited access to education, crime, urbanization patterns, lack of economic growth, and weak health or social protection system increase the vulnerability 12. Socioeconomic status may be one of the main contributing factors to this phenomenon. Insecurity in neighborhoods discourages children and parents from staying outdoors in parks or streets 13. Social deprivation from the community and inequality in access to health are very real problems. The monthly mean income per family in Latin America varies between 773 USD in Panama, which is one of the highest and can be as low as 175 USD, reported in Cuba 14. Other countries such as Brazil average 335 USD, Colombia 244 USD, and Mexico 518 USD 14. This forces families where two parents are present; both have to work to provide, which sometimes increases the frequency of the consumption of prepackaged or fast food. The consumption of healthy foods may also be limited due to their higher cost. According to the World Health Organization, the region's pressing regional challenges are obesity, depression, and anxiety 15. A prominent factor of the complexities in the area is the alarming rate of undernutrition and the highest prevalence of childhood obesity 16. The Latin American Federation of Obesity Societies (FLASO, for its initials in spanish), issued a consensus statement compiling data of a prevalences above 30% in obesity, which is increasing at a faster rate than in the rest of the world 17. According to Ng et al., the highest prevalence of obesity is found in El Salvador and Paraguay for women with a prevalence of 33% and 30% respectively 18 ; and Uruguay and Chile for men with a prevalence of 23.3% and 22%, respectively. Alarming statistics from Argentina show that in children, this problem impacts 9.9% of the population. Mexico holds the second place in the world with 40% of obesity in adults 18. Diets that are high in fat and sugar but low in micronutrient, and an increase in sedentary lifestyles are in part responsible for this phenomenon. Another level in which being overweight has its toll is on emotional and psychological dimensions 19. In the emotional domain, depression is a major challenge. It refers to persistent feelings of sadness and loss of interest, affecting a person's behavior and often impacting daily activities and causing physical or other emotional problems 20. Recent data shows an increasing mental health problem with a prevalence in teenagers and young adults around 10% in Latin America 21. Studies have found that its prevalence is as high as 4.7% in Argentina and Bolivia, 5.8% in Brazil, 4.7% in Colombia, Chile and Uruguay are in 5%, Peru 4.8%, and Mexico and Venezuela follow closely with 4.2% 22. Another increasing mental health issue is anxiety, which refers to an unpleasant emotional state characterized by feelings of tension, apprehension, and worries and by activation or arousal of the autonomic nervous system 23. Studies have found that its prevalence is around, 6.3% in Argentina, 9.3% in Brazil, 6.5% in Chile, 5.8% in Colombia, 3.6% in Mexico, 7.6% in Paraguay, 5.7% in Peru 5.7%, and 4.4% in Venezuela. The pandemic has also taken a toll on mental health and wellbeing in the young population. A rapid survey conveyed by the United Nations Inter-national Children's Emergency Fund (UNICEF) showed that in Latin America, 27% of the population between 13 and 19 have experienced symptoms of anxiety and 15% of depression 24. A change in emotional wellbeing may also promote other unhealthy activities such as alcohol, tobacco, and other recreational drug abuse which is a rising concern among the young population 25. According to the World Health Organization, worldwide, more than a quarter of all people aged 15-19 years which represent 155 million adolescents are current alcohol drinkers. Alcohol and drug use in children and adolescents is associated with neurocognitive alterations that can lead to behavioral, emotional, social, and academic problems in later life. Regarding tobacco use, most studies trace its start to adolescence. Globally, at least 1 in 10 adolescents uses tobacco, although there are areas where this figure may be higher 26. There have been many legislations that criminalize itsproduction and distribution; however, drugs such as cannabis are still heavily used in Latin America with prevalence rates between 4% and 13% 27. It is also widely used in young people, estimates from 2018 are that at least 4.7% of people aged 15-16 years used it at least once during the year 26. Overall, the socioeconomic inequities, as well as some aspects of the cultural and educational settings make the Latin America region a very particular ecological niche with specific and rising challenges that will need to be addressed from different approaches. --- Strategies to promote healthy living in child and adolescents According to the Healthy Living for Pandemic Event Protection (HL-PIVOT) network, adhering to healthy living behaviors (HLB) has a direct impact on a person's quality of life 8. This provides insight not only in physical, dietary aspects, sleep habits, etc. These key elements, as well as the emotional sphere, have yielded to a holistic view of health (Figure 2). --- Physical sphere Many interventions focus on this sphere. By understanding the obesity-promoting factors stakeholders can develop targeted interventions to mitigate the problem. The major contributors to the increased rate of obesity and overweight may be understood in two levels, food, and physical activity. The food environment changes regarding global, national, neighborhood, or in the school environment. Regarding physical activity, factors that impact negatively are the reduction of walking as a transportation mode, the increase in access to motorized vehicles, and the increase in sedentary leisure activities 28. Habits that predispose to overweight and obesity rely on the establishment throughout development and growth which makes it difficult to overcome them with cross-sectional interventions. Several programs have succesfully implement programs on prevention and reducing obesity prevalence in the region. The Plan of Action for the Prevention of Child and Adolescent Obesity (PAHO) was signed in Latin America to promote an active lifestyle, encouraging the use of bike paths or physical activity programs at school 29,30. This initiative pushed the implementation of fiscal policies on foods with a high energy content and low in nutrients. In Mexico, it resulted in taxation of sweetened soft drinks and high-calorie food 31. In 2016, Chile promoted labeling the front of the packages with a black-and-white warning sign if it exceeds defined limits of calories, saturated fat, sugar, and sodium 32. This measure was adopted by Mexico in 2020 and is also being promoted by the Colombian government to promote healthier food choices 33. In 2013, the Mexican health ministry designed and implemented the campaign called: Chécate, M<unk>dete, Muévete to inform people about food and nutrition through public awareness. Positive and motivating messages and informa-tion were transmitted on television or radio to raise awareness about the importance of prevention through commercial spots. These messages also motivated people to attend preventive medical check-ups, as well as for receiving guidance in health promotion 34. However, when the effectiveness of this campaign was analyzed the program was far from effectively decreasing the prevalence of overweight, obesity, and diabetes mellitus. Some of the limitations included poor outreach to rural communities and the transmission of an average of 13,000 advertisements of junk food products and 42 different brands transmitted in parallel 35. These factors might explain why the program did not reach the originally planned impact and its cancelation afterward. In 2011, Brazil launched the Academia da Sa<unk>de Program 36 as a healthcare promotion strategy that works with the implementation of public spaces known as hubs where physical activity practices are offered to the population. These centers are part of the Primary Health Care network and have specialized exercise equipment and qualified professionals to advise the users. As a point of attention in the territory, they complement comprehensive care and strengthen health promotion actions in conjunction with other health programs and actions such as the Family Health Strategy, the Extended Family Health Centers, and Health Surveillance. --- Lifestyle sphere Lifestyle habit changes and promotion may be the most difficult areas to address in Latin America. On one hand, the family context does not foster healthy habits, which could be related to the economic struggles of the population, on the other hand, insecurity and crime throughout the regions add another level of complexity. Interventions focused on these spheres may probably be the most effective in developing healthy lifestyle habits in the long term, hence promoting a healthier status of our population. Mexico has second place in child obesity, many interventions have been made to prevent childhood and adolescent obesity. Unfortunately, there have been no results that indicate the effectiveness of these strategies. One focuses on adding an 8% tax on food with an energy content exceeding 275 Kcal per 100 grams and 0.05 USD per liter on sugar-sweetened beverages 37. Another strategy focuses on food labeling. These labels indicate the food's caloric input, the excess of saturated fats, sugars, and sodium to raise awareness and provide guidance to the consumer. The Mexican strategy involved banning radio or television advertisements during hours in which children may be a significant part of the audience. Although different approaches were implemented, habits have not significantly modified the prevalence of overweight and obesity. Mexico still has one of the highest rates of child obesity worldwide 37. In order to mitigate it, a threemonth-long educational intervention was designed called Healthy Recess aimed to promote healthy eating and physical activity in children. This strategy also allowed the identification of the perceptions of the target population and the implementation of health-promoting interventions. Among the activities was a traffic light system where students had to associate cards with images of food or physical activities. Green, yellow and red colors allowed students to classify food or activities according to their healthiness. Regardless, results showstudents choose their food based on flavor regardless of their caloric intake or nutritional value. This may be due to the lack of parental guidance and school environments not offering enough healthy alternatives, or the lack of healthy alternative in lunch boxes 38. San Pedro de Pinta is a program created in 2011 in San Pedro municipality in Mexico with the purpose of recovering the public space to promote sustainable mobility, social and family interaction, and overall a healthier lifestyle 39. This program takes place every Sunday starting at 7:00 a.m. until 1:00 p.m; families can walk with their pets and ride bikes in this avenue, where vehicular circulation remains closed and becomes 5 km-long-park. The initiative has been well received with approximately 12,000 people attending every Sunday. It teaches new generations a different city model; but more importantly, it has also created spaces that enable the establishment of healthy habits that lead to a healthier lifestyle in the long term. Unorthodox interventions have explored technology as the basis for reaching young adults 40. This aspect could be beneficial because of the increasing number of internet users in the young sector of Latin America. The Exergames in Brazil implemented a strategy with two groups, one of them promoted physical activity-promoting interventions plus virtual games that promote healthy habits, and the second group did solely physical interventions. No difference between the two groups 41. --- Emotional sphere Perceptions of health habits tend to be associated with the nutritional state or BMI parameters, but the multidimensional concept of health is often overlooked. Particularly in adolescents, the focus has been mainly on physical appearance and average weight, but participation in school activities to exercise a sense of belonging with friends and the school community is important as well 42. The current health crisis has brought significant changes in this sphere as well. Social distancing and the lack of socialization in school contributed to this. A study by Genta et al. evaluated the sleep habits and quality of Brazilian adolescents and concluded the pandemic negatively affected their sleep quality 43. A web-based survey conducted during the first peak of the pandemic in Brazilshowed the negative impact on mental health, with a high prevalence of psychiatric symptoms in young adults in Brazil. This study also showed that the young population was more susceptible to depression, anxiety, and stress during the health crisis 44. With almost all our efforts focused on mitigating contagions and infection, emotional wellness has been overlooked. But a direct impact in the emotional sphere of wellness has been demonstrated. Among the consequences, an increase in depression (50%), anxiety (44.6%), and insomnia (34%) in heavily affected countries like China and Italy. Frontline health workers are also at risk of developing post-traumatic stress disorder (PTSD) as a consequence of the coronavirus out-break 45. Schools have dealt with closure mainly focusing their efforts in continuing the curricular delivery, but children and young adults have also been victims of the psychological consequences of the pandemic. Some institutions have implemented programs to mitigate the consequences of social distancing in their students. Therefore, the objective of this study was to assess the perception that adolescents and young adults in Latinamerica have of the importance of healthy living behaviors (HLB) in the physical, emotional, and lifestyle spheres. --- Methodology The approach considered for this study was quantitative and descriptive with a cross-sectional design. This process allows us to identify relevant variables and data to understand a complex phenomenon. The sampling strategy was a convenience sample that consisted of 192 young adults in three countries of Latin America: Brasil, Colombia, and Mexico. As an instrument, an eight closed-ended question survey was designed following the HL-PIVOT framework of healthy living. An additional open-ended question was included to represent the psychological wellbeing of the students, which has become a priority amid the pandemic. The first question considered that participants ordered, according to their perception of importance, the eight elements of the spheres in the Healthy Living Behaviors model: physical activity, nutritional state, access to healthy food, body habitus, sleep habits, avoidance of noxious habits, emotional wellness, and spiritual wellness. The rest of the questions focus on participants' self-assessment in each of these behaviors. These items consider a 1-100 scale for participants to intuitively assess their current state. To analyze the trends in the items and the different spheres, descriptive statistics were considered. The survey also included an open-ended question that asked participants to describe how their perception of wellness evolved as a result of the pandemic. This question added some depth to understanding the perception of the students and context. As part of the analysis, we considered a content analysis, identifying themes, categories and codes to describe similarities and differences in the participants' perceptions. As part of the ethical considerations of the study, it is important to highlight that the participants were informed about the purpose and granted written informed consent. The study followed all the applicable regulations and recommendations of the Comité de ética en investigación de la Escuela de Medicina del Instituto Tecnológico y de Estudios Superiores de Monterrey ethical committee that grants approval of the research protocols. The study was conducted following the Declaration of Helsinki. --- Results Participants' ages ranged from 15 to 25 years old; however, 67% of them were in the 18-21 years range. Regarding healthy living behaviors, in general, participants prioritized the eight elements of the spheres in the Healthy Living Behaviors model. The elements that participants prioritized higher were: emotional wellness (5.9), nutritional state (5.6), and physical activity (5.2). These results are presented in Table 1. If the data is contrasted by country, 60% of Brazilian participants identified physical activity as the first choice, which was prioritized as the first choice only by 8.86% of Colombians, and 8.25% of Mexican participants. The results of Colombian and Mexican participants prioritized emotional wellness as the first choice, with 36.25% and 40.21% respectively. According to the results, participants self-assess higher on their nutritional status with a weighted mean of 77.94. The lowest self-assessment was found on avoidance of noxious habits with a mean of 29.09. These results are presented in Table 2. The results of the open-ended question included in the survey were interesting as well. Participants reflected on how their perception or priorities in healthy behaviors had changed due to the pandemic. Five categories emerged from thematic analysis: 1) awareness of needing emotional wellness, 2) transitioning to a self-care approach, 3) increase in the importance of physical activity, 4) gaining a holistic health and wellness perspective, and 5) perception that their wellbeing has being worsening during the pandemic. The highest number of participants, 28.9% stated that their awareness of needing emotional wellness increased amid the pandemic. Some extracts of their reflections are the following: My priorities changed and I focused more on my mental health (participant 4, Mexico). I value my psychological wellness because it is important for remaining focused in school (participant 12, Colombia). Emotional wellness has an impact on all aspects of your life. When you are not mentally healthy, you can't perform in your life as usual (participant 48, Mexico). According to participants, communities are now prioritizing self-care higher. I make myself a priority in my life, I consider this important to achieve my goals in life (participant 65, Colombia). I prioritize the things that make me feel happy and make me feel good (participant 1, Brasil). Close to 14.91% of them stated that it was the current health crisis which brought the opportunity to focus on the importance of this element: I now have me-time and value my self-care (participant 22, Colombia). I have spent a lot of time with myself and have learned the importance of taking care of myself (participant 75, Mexico). The importance of physical activity increased as participants spent more time in their homes. Close to 14.42%, the majority emerging from participants in Brazil, commented on this theme: I have made my physical wellness a priority in my life (participant 3, Brasil). Being sedentary kills... you have to exercise yourself (participant 1, Brasil). I have made physical activity and healthy eating a priority in my daily routine (participant 27, Mexico). Several participants took into consideration more than one of the spheres of healthy behaviors, addressing the importance of gaining holistic health wellness (11.89%). For example, some overspossed the physical and mental spheres: The pandemic made me prioritize controlling my emotions, stress, anxiety and opened my eyes to the importance of having daily physical activity, socializing, and having recreational activities to be emotionally well (participant 20, Mexico). I knew before the pandemic, but now I have seen a stronger correlation between every aspect of my health. You can't have good mental health without physical activity, a healthy diet, or relaxing activities (you can't have one without the others). I also learned to prioritize activities that make me feel good and step aside from the things that emotionally drain me (participant 10, Brasil). Others described the importance of balancing them and revisiting to analyze the needs in that specific moment of time: I give more importance to socializing, my mental health, and psychological wellness (participant 18, Mexico). It came to my attention that wellness components are not static, they change every day and the needs of each one also change on a daily basis. One component does not compensate for another (participant 6, Brasil). On the other hand, several students reported a perception of their wellbeing had being worsening during the pandemic (6.79%) My personal wellness has worsened amidst the pandemic (participant 21, Colombia). It is tough to be ok [in a wellness state] since the pandemic started (participant 4, Brasil). Some describe the causes that have impacted, for example the motivation: I don't feel motivated to exercise (participant 2, Brasil). The pandemic has affected me physically and psychologically (participant 1, Colombia). --- Discussion In this study, mental health was the highest-ranked item in participants' priorities, followed by an excellent nutritional state and physical activity. Concerns and assumptions of the psychological consequences of this pandemic have been highlighted in editorials in different countries. Now an emergence of the first reports of the toll that quarantines and social distancing measures had are being published. These range from anxiety, stress, an increase in substance abuse and can be as severe as depression, some others report an increase in suicidal idiation and suicide 46. Undoubtedly, the current health crisis has affected the change of this perception, which was also supported by the high frequency of answers related to emotional health in the themes of the comments received. Physical activity has been one of the most traditional concepts associated with healthy living, but it is interesting to observe that emotional wellbeing was higher in the scale of importance. The results on the thematic analysis reflect the awareness of the importance of this sphere. However, the challenges in the region such as the increase of crime and the unsafety in open public spaces, might have impacted the opportunity of participants to practice exercise 13. There is also the fact that quarantine led to social distancing and the consequent decrease of spaces available for exercising. Participants self-assessed low in some of the HLB, for example in the consumption of drugs and the integration of noxious substances as part of their habits, the study depicted a consequence of the pandemic. Substance abuse poses another interesting and raising challenge among youngsters worldwide. The complex economical and social context of Latin America and the increasing rate of substance abuse bring into focus the need for assessing and educating the population. With increasing rates of alcohol consumption 27, and an increase in the stressors amid lockdown and the pandemic it remains to be seen if this has an impact on this parameter when pandemic is over. Probably interventions in the near future will be needed regarding these behaviors. A surprisingly high assessment of the participants in their nutritional state might be the result of the taxation policies that the different countries have implemented. This is quite important as energy-dense and nutrient-poor diets are factors that studies have linked with overweight and obesity. Still, the promotion of healthy food has still to overcome the economic and financial challenges of the Latin American context. The social interaction restrictions and quarantine measures, resulted in indoor physical activities during the pandemic in young people, even of those who were active before isolation, diminished considerably. The results of this study are similar to the results of Brito et al. 47, as the authors demonstrated that the majority of adolescents were unable to maintain daily physical activities, despite claiming to be motivated. Some of the root causes might be lack of physical space or a safe setting to exercise. There was a clear shift in the participant's perception regarding the importance of some aspects of wellbeing amid the pandemic. The awareness of participants of the need to prioritize aspects of health such as emotional wellness. These shifts may be due to quarantine but also influenced by all the interventions that promote a healthy lifestyle that have been implemented throughout the last years. Physical aspects, which have been reported as priority, had the second importance for our participants. As an encouraging first step, life-longlearning habits depend on self-directed commitment 42. The socioeconomic background in Latin America may contribute to a sedentary lifestyle and the growing rates of obesity. Still, the habit developmental process needs to be promoted, and the economic or social limits are yet to be addressed. Interestingly, a more holistic conception of health arose among our participants. Many of them stressed the importance of health being composed of more than one aspect of HLB, such as emotional wellness, socialization, physical activity, and nutrition. New strategies should focus on providing wellness programs with these intersections. School space might be an ideal environment to foster healthy habits that can permeate the family and the community, especially in younger individuals 6. Strategies become more effective when they include awareness that comes from the parents so that they serve as models of healthy lifestyles. Such strategies should include in the design, the specific needs of the context and target population but also the promotion of the habits in the long term and the familiar context. Traditional healthcare has relied on the reductionist approach, nevertheless there has been a shift in this tendency and now healthcare revolves around numerous extrinsic or intrinsic aspects that influence the network of health also known as an holistic approach 2. This was clearly perceived by some of the participants that established a correlation between each aspect of the HLB and acknowledged the interdependence between them in the reflections on the comment section. Perhaps this might be due to all the interventions of HLB promotion this population has been immersed in throughout their formative years. Main limitations of our study may reside in the cross-sectional approach that we used. The student's perception was assessed during one specific period of time and this might not be a reflection or their overall wellness. A longitudinal approach might be needed to better understand it. Also, this study took place during the pan-demic which per se could constitute a bias in the awareness on the importance of mental health. Perhaps during a less stressful time, the students' perception of the priority of mental wellness aspect could change or be scored lower in the importance scale. This is not a qualitative study but offers a glimpse into the students point of view. --- Conclusion Even though geographically different, the Latin American context shares a common pattern. The adoption of healthier behaviors face innumerable barriers, including the lack of social support, cost of adopting healthy behaviors, difficulty in daily routines, especially in time management, cultural preferences, and environmental restrictions. With social distancing amid the pandemic, a major shift in the perception of holistic wellness has emerged. Taking into focus the emotional sphere of a person's life and promoting the inclusion of different aspects of the healthy living behaviors to have an integrated perspective of health and its components. Many factors may hinder the implementation and promotion of the HLB. These need to be addressed from different approaches. First, considering the background and the ecological niche in which the population resides, it may be the main obstacle that needs to be overcomed. Second, taking into focus the problem around which strategies will be implemented. Third, taking into account the importance of promoting health as a whole, and lastly the long-term requirement to habit forging, otherwise, they will not be effective. --- Collaborations AR Garcia, M Lopez and A Mendez contributed to the design and implementation of the research. MM Diaz-Lopez, AR Garcia, A Mendez and DB Dibai contributed to the colection of the data. AR Garcia and A Mendez analyzed the data. MM Diaz-Lopez, AR Garcia, A Mendez, M Lopez and DB Dibai did the writting of the manuscript. M Lopez and A Mendez proofread and edited the final version of the manuscript. All authors provided critical feedback and helped shape the research.

Resumo O completo bem-estar físico, mental e social se refere a um novo conceito de saúde, muito além da ausência de doenças. Se os indivíduos são confrontados com profundas desigualdades sociais, em que fatores como etnia, moradia em zona rural ou baixo nível educacional se traduzem em um acesso menos adequado aos serviços de saúde, podendo ser um desafio adicional para aderir a comportamentos de vida saudável. O objetivo deste estudo foi avaliar a percepção de adolescentes e adultos jovens na América Latina sobre a importância dos comportamentos de vida saudável (CVS) nas esferas física, emocional e de estilo de vida. A abordagem é considerada quantitativa e descritiva, com desenho transversal. A estratégia de amostragem foi por conveniência. A amostra foi composta por 192 adultos jovens em três países da América Latina: Brasil, Colômbia e México. Os resultados mostram que os participantes priorizam um maior bem-estar emocional. Eles afirmam que a crise de saúde mudou a percepção, pois agora valorizam mais a saúde emocional e compreendem a importância da combinação dos diferentes aspectos do bem-estar. O contexto latino-americano desafia o desenho de estratégias com uma perspectiva holística da saúde, com complexidades nos domínios econômico e sociocultural. Palavras-chave Comportamento saudável,

of Medicine (2009) report underscored the importance of considering developmental competencies, such as educational and occupational outcomes, given that they enable the individual to be successful in subsequent developmental tasks and maintain resilience when faced with adversity. The few evaluations of programs for children from divorced families that have measured program effects on educational outcomes found that program-induced improvements in parenting led to children's enhanced academic functioning and achievement (Forgatch & DeGarmo, 1999;Wolchik, Sandler, Weiss, & Winslow, 2007). Further, only one study has examined possible mechanisms through which these programs may affect educational outcomes. Zhou et al. (2008) found that improvements in effective discipline mediated the effects of their preventive intervention, the New Beginnings Program (NBP), on grade point average at the six-year follow-up when the youth were adolescents. To date, researchers have not examined whether programs for youth from divorced families affect the formation of occupational and educational goals in adolescence, a critical developmental task (Barber & Eccles, 1992;Beal & Crockett, 2010;Card, Steel, & Abeles, 1980;Erikson, 1968) that provides the foundation for educational and job attainment in later developmental stages (Cheeseman Day & Newburger, 2002;Harackiewicz, Barron, Tauer, & Elliot, 2002). For instance, Harackiewicz, Barron, Tauer, Carter, and Elliot (2000) found that young adults' academic goals predicted their later educational achievement. Also, Judge, Cable, Boudreau, and Bretz (1995) found that individuals who reported ambitious goals for their occupational futures experienced greater objective job success, earned more and received more promotions than those who were less goal-driven. This study used data from the NBP, a randomized experimental trial of a preventive intervention for divorced families, to examine whether this program affected adolescents' educational and occupational goals. A secondary goal was to examine whether mediators of the program effects on educational and occupational goals could be identified. Two aspects of positive parenting, mother-child relationship quality and effective discipline, and four aspects of youth behaviors, externalizing problems, internalizing problems, self-esteem, and academic competence and performance, were tested based on research linking them to parental divorce and educational or occupational goals and data showing that the NBP had a positive effect on these variables. Below, we biefly describe the effects of the NBP. Then, we present research that links divorce to the potential mediators and research linking the potential mediators to academic and occupational goals. --- NBP The NBP was developed to mitigate several negative outcomes associated with parental divorce, including children's mental health problems, substance use, and social problems, by modifying risk and protective factors that have been linked with the negative post-divorce outcomes (Wolchik et al., 1993). A randomized experimental trial of the NBP, which included a mother program condition, a mother program plus child program condition, and a literature control condition, assessed both short-term and long-term effects. The trial found that the effects of the two conditions on mediators and outcomes at posttest and short-term follow-up did not differ (Wolchik et al., 2007). Thus, these two conditions have been combined in subsequent analyses of the NBP. Program effects were found at posttest on mother-child relationship quality, effective discipline, and mother/child report of children's externalizing problems and internalizing problems (Wolchik et al., 2000). The six-year follow-up showed that adolescents in the NBP condition had fewer sexual partners, lower rates of mental disorder, lower levels of internalizing and externalizing problems and substance use, and higher grade point averages (GPA) and self-esteem than participants in the literature control condition (Wolchik et al., 2007). Many of the program effects at posttest and follow-up were stronger for those with higher levels of baseline risk (Dawson-McClure, Sandler, Wolchik, & Millsap, 2004;Wolchik et al., 2000;2002;2007). --- Links between divorce, putative mediators, and educational and occupational outcomes Divorce is associated with diminished parenting, including decreased levels of warmth and responsiveness, less effective communication, and the use of harsh or coercive discipline (e.g., Astone & McLanahan, 1991;Hetherington, Cox, & Cox, 1985;Simons & Johnson, 1996). Theory and research also suggests that quality of parenting is related to adolescents' educational and occupational goals, aspirations, and engagement (Bryant, Zvonkovic, & Reynolds, 2006;Jodl, Michael, Malanchuk, Eccles, & Sameroff, 2001). For example, attachment theory proposes that, following the establishment of a secure caregiver base, children will feel safe and comfortable to explore their environments and individuate without risk to the parent-child bond (Ainsworth, 1989;Paquette, 2004). Eccles et al. (1993) further conceptualizes parents as providers of behavioral reinforcement, resources, and educational opportunities as children embark on their path to career success. Numerous studies have found support for an association between quality of parenting and youths' educational and occupational goals, aspirations, and engagement (e.g., Astone & McLanahan, 1991;Barnard, 2004;Jodl et al., 2001;Rodgers & Rose, 2001). Illustratively, Schmitt-Rodermund and Vondracek (1999) found that parental involvement in children's activities was prospectively related to more career exploration and planning in adolescence. Glasgow, Dornbusch, Troyer, Steinberg, and Ritter (1997) also showed that neglectful parenting predicted adolescents' lowered educational expectations one year later. It is well documented that children from divorced families are at an increased risk for externalizing behavior problems (e.g., Amato, 2000;2001;Amato & Keith, 1991;Hetherington, 1993) and that these problems are linked with later negative academic and occupational outcomes, both in adolescence (e.g., Andrews & Duncan, 1997) and young adulthood (Fergusson & Horwood, 1998). Fergusson and Horwood (1998) proposed that early-onset externalizing problems may lead to substance abuse and association with deviant peers, which may contribute to a lack of life opportunities in the domains of education and work. Masten et al. (2005) proposed that conduct problems in childhood could impede learning and alienate teachers and peers, which may produce deficits in educational and occupational functioning later in life. Notably, Masten et al. (2005) found that childhood externalizing problems were linked with low academic achievement and competence seven years later, when participants were adolescents. Risi, Gerhardstein and Kistner (2003) also found that children's aggression toward peers was related to a decreased probability of graduating high school 10 years later. Parental divorce has also been shown to be related to children's internalizing problems (e.g., Amato, 2000;2001;Amato & Keith, 1991;Hetherington, 1993), but the support for the link between children's internalizing problems and their educational and occupational outcomes is limited. Rapport, Denney, Chung, and Hustace (2001) found that anxiety and depression in childhood were related to later academic achievement, and that these relations were mediated through intellectual functioning and performance in the classroom. McLeod and Kaiser (2004) also showed that internalizing problems in school-aged children were related to a decreased likelihood of graduating from high school. Conversely, a number of studies have shown that externalizing problems in childhood were more predictive of later educational and occupational outcomes than were internalizing problems (Capaldi & Stoolmiller, 1999;Masten et al., 2005;McLeod & Kaiser, 2004). Masten et al. (2005) proposed that mental health problems in childhood, whether internalizing or externalizing, can inhibit success with developmental tasks, such as educational and occupational goals, through their influence on disruptive behavior and lack of engagement in the classroom. Studies have also shown that parental divorce is associated with lower self-esteem (Amato & Keith, 1991;Amato, 2001;Storksen, Roysamb, Moum, & Tambs, 2005) and decreased academic achievement during childhood and adolescence (e.g., Amato, 2001;Amato & Keith, 1991;Teachman, Paasch, & Carver, 1996). Researchers have demonstrated that both academic self-esteem and general self-esteem are linked to academic achievement and occupational goals (Ahmavaara & Houston, 2007;Emmanuelle, 2009;Pullmann & Allik. 2008). Theoretically, Baumeister, Campbell, Krueger, and Vohs (2003) suggested that individuals with higher self-esteem will be more likely to persevere when faced with failure, and Wigfield and Eccles (1994) proposed that one's belief in his or her abilities should determine expectations for success. Flouri (2006) showed that children's self-esteem at age 10 was related to their educational attainment 16 years later. Trzesniewski et al. (2006) also found that adolescents' low self-esteem was linked with a decreased likelihood of attending college and more work-related problems in adulthood. Bandura, Barbaranelli, Caprara, and Pastorelli (2001) found adolescents' academic self-efficacy was linked contemporaneously with academic performance and that academic self-efficacy and academic performance were related to the adolescents' choice to pursue challenging careers one year later. --- Contribution of the current study This study extends previous research by examining whether a preventive intervention for youth from divorced families has positive effects on educational and occupational goals in adolescence. In addition, it examines whether program effects on educational and occupational goals are accounted for by program-induced changes in mother-child relationship quality, effective discipline, and youth's externalizing and internalizing problems, self-esteem, and academic achievement. Examining whether prevention programs have positive effects on educational and occupational goals and identifying the program components that mediate these changes have theoretical and applied implications (Ginexi & Hilton, 2006;Sandler, Wolchik, Winslow, & Schenck, 2006). Currently, 10 million children live in divorced or separated households (National Center for Health Statistics, 2005). Thus, identifying programs that affect educational and occupational goals of these youth could have important implications for reducing the public health burden of parental divorce. Further, identification of the components of the program that accounted for change in these outcomes can provide guidance for program refinement and dissemination (Kazdin & Nock, 2003). This study advances existing knowledge in two important ways. First, it focuses on the developmental antecedents of occupational achievement and educational attainment in adulthood, which have significant implications for economic status and mental health throughout the lifespan (Cheeseman, Day & Newburger, 2002;Harackiewicz et al., 2002). Given the lack of previous research linking prevention programs and educational and occupational goals rather than educational attainment or occupational achievement, this study addresses a gap in the literature. Second, its use of data from a randomized trial allows a test of whether experimentally-induced changes in parenting and youth variables account for experimentally-induced effects on educational and occupational goals, thus strengthening the causal inference between these variables over those that can be drawn from previous work which has been correlational (Cole & Maxwell, 2003;Cowan & Cowan, 2002;Rutter, 2005). --- Method Participants Families were primarily recruited through divorce decrees obtained through public court records; about 20% of the sample responded to media advertisements. Participation was solicited by letters and follow-up phone calls to assess eligibility. Families that met eligibility criteria were asked to participate in an in-home recruitment visit. Eligibility criteria for participation in the trial included the child was living with the mother at least 50% of the time; the custody arrangement was expected to remain the same for the duration of the study; the divorce occurred within the last two years; the mother was not remarried, did not plan to remarry, and did not have a live-in partner; both mother and child were fluent in English; there was at least one child between the ages of 9 and 12 living in the home; and neither the mother nor child was currently receiving mental health services. In families that included more than one child between the ages of 9 and 12, one child was randomly selected for the interviews. Children who scored within the clinical range on measures of depression or externalizing problems or who endorsed current suicidal ideation were excluded and referred for treatment. The sample consisted of 240 families that were randomly assigned to one of three conditions: mother-only program (MP) (n = 81 families), dual-component program (MPCP) (n = 83 families), or literature control condition (LC) (n = 76 families). Of the families contacted by phone, 48% (n = 671) met the initial eligibility criteria. Of these families, 68% (n = 453) completed the recruitment visit; 75% (n = 341) of the recruitment visit completers agreed to participate in the intervention study; 92% (n = 315) of these families completed the pretest. We found 16% (n = 49) to be ineligible at the pretest interview; an additional 8% (n = 26) withdrew before assignment. Thus, 36% (n =240) of the eligible families were randomly assigned to condition. Analyses revealed that participating families reported significantly higher incomes (p =.03) and maternal educational level (p =.01), and had fewer children (p =.01) than refusers (n = 59) (Wolchik et al., 2000;2002). At the six-year follow-up, 218 (91%) families were interviewed. Attrition analyses comparing those who attrited between pretest and the six-year follow-up (N = 22) to those who remained in the study on baseline demographic variables and children's internalizing and externalizing problems revealed no significant attrition or condition x attrition interaction effects (Wolchik et al., 2002), indicating that attrition did not pose a threat to internal or external validity. At pretest, children were, on average, 10.34 years of age (SD = 1.1); 50% were female. Mothers' ethnicity was 90% Caucasian, 6% Hispanic, and 4% other. Average annual household income was $20,001 -$25,000; 47% of the mothers had completed some college. Legal custody arrangements were 63%, 35%, and 3% sole maternal, joint, and split, respectively; families had been separated an average of 26.7 months and divorced an average of 12.3 months. Baseline equivalence between the experimental and control conditions in regard to children's gender and age, mothers' ethnicity, household income, length of time since separation and divorce, custody arrangements, and children's internalizing and externalizing problems was examined, using <unk> 2 tests for the categorical variables and t-statistics for the continuous variables. No significant differences were found. In the families who participated in the six-year follow-up, youth were between the ages of 15 and 19 (M = 16.9, SD = 1.1); 49.5% were female. Mothers' ethnicity was 89% Caucasian, 6% Hispanic, and 5% other. Average annual household income was $50,001 -$55,000. Legal custody arrangements were 53%, 46% and 1% sole maternal, joint, and paternal, respectively. Families had been separated an average of 8.4 years (SD = 1.4) and divorced an average of 7.2 (SD =.55) years. --- Procedure Families were interviewed on five occasions: pretest (T1), posttest (T2), and 3-month (T3), 6-month (T4), and 6-year (T5) follow-ups. The pretest occurred prior to randomization to condition. In the present study, data collected at T1, T2, and T5 were used. At each assessment, confidentiality was explained, parents (and at the six-year follow-up, adolescents 18 or older) signed consent forms, and children signed assent forms. Mothers and youth were interviewed separately. Families received $45 at pretest and posttest; at the six-year follow-up, parents and adolescents each received $100. --- Experimental Conditions The MP targeted positive parenting (i.e., mother-child relationship quality and effective discipline), interparental conflict, and mothers' attitudes toward the father-child relationship. There were 11 group sessions (1.75 hour each); five focused on the quality of the motherchild relationship and three focused on effective discipline. Two individual sessions (1 hour each) focused on the mother's use of the program skills with her children. Sessions were led by two Master's-level clinicians and used didactic and experiential learning techniques that were based on social learning and cognitive behavioral research. The groups consisted of 8 to 10 mothers. The MPCP consisted of concurrent but separate groups for children (CP) and mothers (MP). The 11 sessions in the CP targeted adaptive coping skills, negative cognitions, and motherchild relationship quality. Social learning and cognitive behavioral research provided a foundation for program exercises; didactic material was presented and modeled by group leaders or videotapes. Youth practiced the skills in the context of games, role-plays, and, for the communication skills, in a conjoint exercise with their mothers. Groups, which consisted of 8 to 10 children, were led by two Master's-level clinicians. The MP in both conditions was identical, with the exception of the conjoint exercise on communication skills. Children and mothers in the LC each received three books about children's post-divorce adjustment and a syllabus to guide their reading. See Wolchik et al. (2000Wolchik et al. (, 2007) ) for more information about the conditions. --- Measures Demographics-Mothers responded to demographic questions such as their children's age and living arrangement, and their own ethnicity, income, and level of education. Data taken from T1 were used in the analyses. Mother-child relationship quality-Measures of mother-child relationship quality assessed at T1, T2, and T5 were used. Mothers and youth completed a revised version of the Acceptance (10 items) and Rejection (10 items) subscales of Schaefer's (1965) Child Report of Parenting Behavior Inventory (CRPBI; Teleki, Powell, & Dodder, 1982). Parallel motherand child-report versions were used. A sample item is "My mom isn't very patient with me." Coefficient alphas were acceptable (child report <unk>'s for Acceptance =.82,.84, and.90; <unk>'s for Rejection =.82, 81, and.86 at T1, T2, and T5; mother report <unk>'s for Acceptance =.73,. 77, and.82; <unk>'s for Rejection =.74,.72, and =.73 at T1, T2 and T5, respectively). The rejection items were recoded and then the rejection and acceptance items were summed. CRPBI scores have been shown to distinguish between delinquent and normal children (Schaefer, 1965). Mothers and children completed the 10-item Open Communication subscale of the Parent-Adolescent Communication Scale (Barnes & Olson, 1985). A sample item is "Mom is always a good listener." Coefficient alpha's were acceptable (child report <unk>'s =.85 at T1,.87 at T2,.91 at T5; mother report <unk>'s =.71 at T1,.72 at T2,.83 at T5). Scores on this measure have been positively linked with psychological adjustment in adolescents (e.g., Young & Childs, 1994). In addition, mothers and children completed an adaptation of the 7-item Dyadic Routine subscale of the Family Routines Inventory (Jensen, James, Boyce, & Hartnett, 1983). A sample item is "You had time each day just to talk with your kids." Coefficient alpha's were acceptable (mother report <unk>'s =.66 at T1,.63 at T2,.84 at T5; child report <unk>'s =.71 at T1,.76 at T2,.76 at T5). Scores on this measure and children's adjustment problems have been shown to be negatively related (Cohen, Taborga, Dawson, & Wolchik, 2000). All measures used the time frame of the past month. The six measures were standardized and averaged to obtain a composite of mother-child relationship quality. Effective discipline-Effective discipline scores at T1 and T2 were used; T5 scores were not used because the program did not affect discipline at T5. Mothers reported on inappropriate discipline (5 items; <unk>'s =.75 at T1,.77 at T2), appropriate discipline (9 items; <unk>'s =.59 at T1,.59 at T2), and discipline follow-through on the Oregon Discipline Scale (11 items; <unk>'s =.78 at T1,.76 at T2; Oregon Social Learning Center, 1991). Sample items include "When your child misbehaved, how often did you yell?" (inappropriate discipline), "When your child misbehaved, how often did you restrict privileges?" (appropriate discipline), and "How often did you feel that it was more trouble than it was worth to punish your child?" (follow-through). Responses on the appropriate and inappropriate items were used to compute a ratio of appropriate-to-inappropriate discipline. Similar discipline measures have been shown to correlate with adolescents' mental health problems (e.g., Patterson & Forgatch, 1995). Also, mothers and children completed the 8-item Inconsistency of Discipline subscale of Teleki et al.'s (1982) adaptation of the CRPBI (Schaefer, 1965), which used the time frame of the past month. A sample item is "It depended on your mother's mood whether a rule was enforced or not." Coefficient alphas were adequate (mother report <unk>'s =.82 at T1,.80 at T2; child report <unk>'s =.74 at T1,.73 at T2). These four scales were standardized and averaged to create a composite score. Externalizing problems-Externalizing problems at T1, T2, and T5 were assessed using a composite of 33 mother-reported items (<unk>'s =.88 at T1,.86 at T2,.89 at T5) from the Child Behavior Checklist (CBCL; Achenbach, 1991a;Achenbach & Edelbrock, 1981) and 30 child-reported items (<unk>'s =.87 at T1,.83 at T2,.84 at T5) from the Youth Self-Report (YSR; Achenbach, 1991b). The time frame used was the last month. A sample item is "Argues a lot." Scores on the CBCL and the YSR have been shown to distinguish children receiving psychological services from normal controls (Achenbach, 1991a). Mother-and child-reports were standardized and averaged to obtain a composite score. In addition, teachers completed the six-item acting out subscale (<unk>'s =.90 at T1,.90 at T2,.89 at T5) of the Teacher-Child Rating Scale using the time frame of the last month; a sample item is "Disruptive in class." Internalizing problems-Internalizing problems at T1, T2, and T5 were measured using a composite of 31 mother-reported items (<unk>'s =.88 at T1,.85 at T2,.86 at T5) from the Child Behavior Checklist (CBCL; Achenbach, 1991a;Achenbach & Edelbrock, 1981), 28 child-reported items (<unk>'s =.88 at T1,.90 at T2,.89 at T5) from the Revised Children's Manifest Anxiety Scale (RCMAS; Reynolds & Richmond, 1978), and 27 child-reported items (<unk>'s =.81 at T1,.83 at T2,.85 at T5) from the Children's Depression Inventory (CDI; Kovacs, 1981). Sample items for these measures are "Likes to be alone" and "You worried a lot of the time," and "I am sad once in a while", for the CBCL, RCMAS and CDI, respectively. Scores on the RCMAS are correlated with other measures of anxiety in children, including the Trait Anxiety score from the State-Trait Anxiety Scale for Children (e.g., Carey et al., 1994). Scores on the CDI have been shown to differentiate children who are clinically depressed from non-depressed psychiatric child patients (e.g., Kovacs, 1985), and scores on the CBCL have been shown to differentiate children referred for psychiatric services from non-referred children (Achenbach, 1991a). A composite score of internalizing problems was created by standardizing and then averaging the three measures. Self-esteem-Self-esteem scores at T1 and T5 were used; T2 scores were not used because the program did not affect self-esteem at T2. Youth completed the 6-item global self-esteem subscale of the Self-Perception Profile for Children (Harter, 1982; <unk>'s =.78 at T1,.86 at T5). A sample item is "Some kids like the kind of person they are." Scores on this measure have been negatively related to children's depressive symptoms (Renouf & Harter, 1990). Academic competence-Scores at T1 and T5 were used; T2 scores were not used because the program did not affect academic competence at T2. Mothers and children completed the 6-item academic competence subscale of the Coatsworth Competence Scale (Coatsworth & Sandler, 1993). Coefficient alphas were adequate (mother report <unk>'s =.91 at T1,.90 at T5; child report <unk>'s =.78 at T1,.81 at T5). Scores on this measure have been linked with other measures of competence and with mental health outcomes (Coatsworth & Sandler, 1993;Spaccarelli, Coatsworth, & Bowden, 1995). A sample item is "Your child had problems learning new subjects at school." Mother-and child-report scores were standardized and then composited by taking the mean. In addition, adolescents' cumulative unweighted grade point average (GPA) for all classes taken in high school was collected from school transcripts at T5. At T1, academic competence was assessed using only the Coatsworth measure. At T5, a composite was created by standardizing the scores for GPA and academic competence and averaging them. Educational and occupational outcomes-At T5, educational expectations were assessed with the question "When you think about your future, what is the highest level of education you expect to attain?" from the Future Expectations Scale (Linver, Barber, & Eccles, 1997). The five response options ranged from completing high school to attending post-college graduate or professional school. To assess job aspirations, youth were presented with a list of 28 occupations and asked "If you could have any job you wanted, what job would you like to have when you are 30 years old?" (Possible Jobs Scale; Tucker, Barber, & Eccles, 1997). Occupations were subsequently scored according to level of prestige, with higher scores reflecting more (highest prestigious jobs). --- Results --- Analytical Procedure Structural equation modeling (SEM) with Mplus software (Version 5.1 Muthén & Muthén, 1998-2007) was used to evaluate the program effects on educational expectations and job aspirations and to test the mediation models. Missing data were handled using full information maximum likelihood (FIML) estimation. FIML yields estimates that are less biased than the conventional listwise deletion or mean substitution in handling missing data (Collins, Schafer, & Kam, 2001;Schafer & Graham, 2002). Program effects on educational expectations and job aspirations were examined separately. Following the establishment of program effects on the two outcomes, mediational analyses were conducted to identify potential mediators of these effects. Because previous analyses demonstrated that youth who were at greater baseline risk for developing future adjustment problems benefited from the program more than those at lower risk (e.g., Dawson et al., 2004;Wolchik et al., 2000;2007), we first examined if the program x risk interaction effects on educational expectations and job aspirations were significant in accordance with Aiken and West's (1991) multiple regression procedure. A moderated effect was considered to occur if the interaction was significant, and the Johnson-Neyman procedure (Potthoff, 1964) was employed to probe the region in which the intervention and control groups differed significantly on the outcome variables (see Aiken & West, 1991). We then conducted mediated moderation analyses (Muller, Judd, & Yzerbyt, 2005;Preacher, Rucker, & Hayes, 2007), assessing whether the mediation process accounted for this moderation. If the mediated moderation effect was significant, we probed the simple mediation effect following the procedure outlined in Tein et al. (2004). Specifically, we examined whether the mediation effect was significant at different levels of the moderator (e.g., one standard deviation below [-1SD] and one standard deviation above [+1SD] the mean). This procedure does not artificially dichotomize the sample into high and low risk groups and thus provides greater power for examining the simple effect. The baseline risk index consists of the baseline variables that were the strongest predictors of adolescent adjustment outcomes in the LC group: externalizing problems and a composite of environmental stress measures that assessed the following divorce-related stressors, child-experienced negative events, interparental conflict, decreased contact with father, per capita income, and maternal distress (see Dawson-McClure et al., 2004 for a more detailed description of this index). Both three-wave longitudinal and two-wave half-longitudinal models (Cole & Maxwell, 2003) were employed to test mediation. In both approaches, the predictor was the program condition and the outcomes were educational expectations and job aspirations at 6-year follow-up. In the three-wave longitudinal models, the mediators were those for which positive program effects occurred at posttest: mother-child relationship quality, effective discipline, internalizing problems, and externalizing problems. In the half-longitudinal models, the mediators, which were measured concurrently with the outcomes, were those variables for which positive program effects occurred at the 6-year follow-up: mother-child relationship quality, self-esteem, internalizing problems, externalizing problems, and academic competence. Separate mediation models were first tested for each mediator variable and each outcome variable. When significant effects were found for more than one mediator, multi-mediator models, which included each mediator that was significant in the single mediator models, were tested to determine whether there was unique prediction of the mediator above and beyond the other mediators. In all models, the baseline measures of the mediators were controlled. Because GPA was not measured at pretest or posttest, T1 academic competence was used as the baseline proxy of the T5 academic competence/GPA composite. Variables measured at the same assessment point were permitted to correlate with one another. MacKinnon's (2008) guidelines for mediation were used in which mediation is established if the path from the independent variable to the mediator (a path) and the path from the mediator to the outcome controlling for the independent variable (b path) are significant. According to the simulation study by Fritz and MacKinnon (2008), this method provides a more robust test of mediation than Baron and Kenny's (1986) method that requires the path from the independent variable to the outcome without controlling for the mediator is also significant. In cases where the a and b paths were significant, the statistical significance of the mediation effect (a*b) was tested against the confidence interval, [CI: ab <unk> (significant critical value)*(SE ab )], where SE ab is the standard error using the PRODCLIN procedure (see Fritz and MacKinnon, 2008;MacKinnon, 2008). If zero is not contained within the 95% CI, it can be concluded that the mediated effect is significant. --- Preliminary Analyses A Box's M analysis, including all of the study variables, was conducted to determine whether the MP and MPCP conditions could be combined as they were in previous studies (Wolchik et al., 2007;Velez, Wolchik, Tein & Sandler, in press;Zhou et al., 2008). The Box's M analysis is considered a conservative omnibus test that assesses whether the variance and covariance matrices of two groups differ significantly (Winer, 1971). If Box's M is nonsignificant, it can be concluded the relations among the variables do not differ significantly across groups. The results showed that the variance/covariance matrices did not differ significantly (Box's M = 5.63; F(3) = 1.85, p =.14; <unk> 2 (3) = 5.54, p =.14). Thus, the MP and MPCP conditions were combined for the analyses. Dummy codes were created for the LC (0) and MP + MPCP (1) conditions. The diagnostic indices of leverage (Mahalanobis' distance), distance, and influence (DFFITS and Cook's Distance) were calculated to identify potential outliers or influential data points (Cook, 1977;Neter, Kutner, & Wasserman, 1989;Stevens 1984). These analyses revealed no outliers or influential data points; thus all cases were retained in the analyses. Descriptive statistics for all study variables are provided in Table 1, and the correlations among the study variables and potential covariates are presented in Table 2. The following variables were selected as potential covariates based on previous research indicating that they were significantly related to the mediator or outcome variables: child's age and gender, mother's and father's highest level of education, months since separation, and months since divorce. A path from the covariate to the outcome or the mediator was included in the SEMs that examined program effects or mediation effects if the zero-order correlation of the covariate with the outcome or the mediator was significant. If the path from the covariate to the mediator or the outcome was nonsignificant in the model, the covariate was dropped. As shown in Table 2, age was significantly correlated with T2 teacher-reported externalizing problems, such that younger children exhibited higher levels of externalizing problems. Gender was significantly correlated with T2 and T5 teacher-reported externalizing problems, such that boys scored higher than girls. Gender was also significantly correlated with T5 mother-child relationship quality, T5 educational expectations and T5 job aspirations, with males having lower scores than females on these measures. Mothers' level of education was significantly positively correlated with T2 effective discipline, T5 academic competence, and T5 educational expectations. Fathers' level of education was significantly positively correlated with T2 mother/child-reported internalizing problems, T5 academic competence, T5 teacher-reported externalizing problems, T5 educational expectations and T5 job aspirations. Fathers' level of education

This study examined whether the New Beginnings Program for divorced families led to improvements in youth's educational goals and job aspirations six years following participation and tested whether several parenting and youth variables mediated the program effects. Participants were 240 youth aged 9-12 years at the initial assessment, and data were part of a randomized, experimental trial of a parenting skills preventive intervention targeting children's post-divorce adjustment. The results revealed positive effects of the program on youth's educational goals and job aspirations six years after participation for those who were at high risk for developing later problems at program entry. Further, intervention-induced changes in motherchild relationship quality and youth externalizing problems, internalizing problems, self-esteem, and academic competence at the six-year follow-up mediated the effects of the program on the educational expectations of high-risk youth. Intervention-induced changes in youth externalizing problems and academic competence at the six-year follow-up mediated the effects of the program on the job aspirations of high-risk youth. Implications of the present findings for research with youth from divorced families and for the public health burden of divorce are discussed. It is well documented that parental divorce is associated with multiple problems for youth that extend into adulthood, including internalizing and externalizing problems, interpersonal difficulties, poor physical health, and substance use (e.g., Amato, 2001;Chase-Lansdale, Cherlin, & Kiernan, 1995). Several studies have found that parental divorce in childhood is also linked with negative educational and occupational outcomes across the life span, such as a decreased probability of graduating from high school, after controlling for income, parental educational attainment, ethnicity, and other demographic variables (e.g., Sandefur,

the outcome or the mediator was included in the SEMs that examined program effects or mediation effects if the zero-order correlation of the covariate with the outcome or the mediator was significant. If the path from the covariate to the mediator or the outcome was nonsignificant in the model, the covariate was dropped. As shown in Table 2, age was significantly correlated with T2 teacher-reported externalizing problems, such that younger children exhibited higher levels of externalizing problems. Gender was significantly correlated with T2 and T5 teacher-reported externalizing problems, such that boys scored higher than girls. Gender was also significantly correlated with T5 mother-child relationship quality, T5 educational expectations and T5 job aspirations, with males having lower scores than females on these measures. Mothers' level of education was significantly positively correlated with T2 effective discipline, T5 academic competence, and T5 educational expectations. Fathers' level of education was significantly positively correlated with T2 mother/child-reported internalizing problems, T5 academic competence, T5 teacher-reported externalizing problems, T5 educational expectations and T5 job aspirations. Fathers' level of education was significantly negatively correlated with T2 mother/child-reported externalizing problems. Time since divorce was significantly positively correlated with T5 self-esteem and T5 job aspirations; adolescents whose parents had been divorced longer had higher self-esteem and higher job aspirations. --- Analyses of Program Effects In the model in which educational expectations was the outcome, program, risk, and the program x risk interaction were included as predictors, and child gender and mother's and father's highest level of education were included as covariates based on the results of the correlational analyses. Mothers' level of education became nonsignificant and was thus dropped from the model. Although the program main effect was nonsignificant (<unk> =.05, p =.41), the program x risk interaction effect was significant (<unk> =.39, p <unk>.01). The Johnson-Neyman procedure (Aiken & West, 1991) revealed that for youth who had risk scores beyond 1.14 SD above the mean, the intervention and control conditions differed significantly, such that youth in the intervention had higher expectations than those in the control condition. Approximately 12% of the sample was in this region. In the model in which job aspirations was the outcome, program, risk, and the program x risk interaction were included as predictors, and child gender, time since divorce, and father's level of education were included as covariates based on the results of the correlational analyses. Similar to the findings of educational expectations, the program main effect was nonsignificant (<unk> =.12, p =.07) and the path from program x risk to job aspirations was significant (<unk> =.28, p <unk>.05). The Johnson-Neyman procedure revealed that for youth who had risk scores beyond.56 SD above the mean, the intervention and control conditions differed significantly, such that the youth in the intervention had higher aspirations than those in the control condition. Approximately 26% of youth in the sample were in this region. --- Mediation Models Because the program effects on the two outcomes were moderated by the baseline risk, we conduced mediated moderation analyses, which included program x risk interactions to the mediator (a' path) and the outcome (c' path) in the SEM. Figure 1 illustrates a theoretical mediation model. 1 illustrates a theoretical mediation model. Table 3 shows the results of the mediation models in which the program effects on T5 measures of educational expectations and job aspirations were mediated by the prospective effect of the five potential mediators measured at T2: mother-child relationship quality, effective discipline, mother/child-reported internalizing problems, mother/child-reported externalizing problems, and teacher-reported externalizing problems. As shown, all the mediation models fit the data adequately. The program had significant effects (a path) on all of the mediators except teacher-reported externalizing problems. Risk did not moderate any of the program effects on the mediators (a' path). After controlling for the program effect, none of the mediators had significant effects on educational expectations or job aspirations (b path). The direct effects from the program x risk interaction to the two outcomes remained significant (c' path). Because none of the models had significant a (or a') and b paths, the mediated effects were not assessed for significance. 4 shows the results of the mediation models in which the program effects on educational expectations and job aspirations were mediated by the concurrent effects of the six T5 potential mediators: mother-child relationship quality, mother/child-reported internalizing problems, mother/child-reported externalizing problems, teacher-reported externalizing problems, academic competence, and self-esteem. All of the mediation models fit the data adequately. There were significant program x risk interaction effects (a' path) on all of the mediators beyond the significant main effects (a path) on mother/child-reported and teacher reported externalizing problems. With the exception of teacher-reported externalizing problems, all the mediators had a significant effect (b path) on educational expectations. Only mother/child-reported externalizing problems and academic competence had a significant effect (b path) on job aspirations. --- Three-wave longitudinal models-Figure --- Two-wave half-longitudinal models-Table For those models that had both significant a' (indicating that the program effects on the mediators were moderated by risk) and b paths, simple mediation effects were tested. The findings of the simple mediation effects indicated that for youth with high, but not low risk, significant mediation effects were found for mother-child relationship quality (95% CI:. 0068,.1022), self-esteem (95% CI:.0361,.1740), mother/child-reported externalizing problems (95% CI: -.2666, -.0633), mother/child-reported internalizing problems (95% CI:.0073,.1174), and academic competence (95% CI:.0522,.2324) on educational expectations. Significant mediation effects also occurred for academic competence (95% CI:.0353,.1832) and mother/child-reported externalizing problems (95% CI:.0107,.1927) on job aspirations for high risk youth. To assess the unique mediation effect of each mediator, mother-child relationship quality, self-esteem, mother/child-reported externalizing problems, mother/child-reported internalizing problems, and academic competence were entered simultaneously into the SEM predicting educational expectations. The fit of the model was satisfactory: <unk> 2 (35) = 44.63, p =.13, RMSEA =.03, SRMR =.03, CFI =.98. All the a' paths remained significant: program x risk to mother-child relationship quality, <unk> =.26, p =.03; to self-esteem, <unk> =.53, p =.00; to mother/child-reported externalizing problems, <unk> = -.54, p =.00; to mother/childreported internalizing problems, <unk> = -.50, p =.00; to academic competence, <unk> =.34, p =.01. The b paths from academic competence (<unk> =.41, p =.00) and from self-esteem (<unk> =.19, p =.02) to educational expectations were significant. The paths from the other three mediators were nonsignificant. Probing of the simple mediation effects indicated that academic competence (95% CI:.0421,.2088) and self-esteem (95% CI:.0129,.1450) independently accounted for the effects of the program on educational expectations for the high-risk youth. Academic competence and mother/child-reported externalizing problems were entered simultaneously into the SEM predicting job aspirations to assess for unique mediation effects. The fit of the model was satisfactory: <unk> 2 (9) = 10.29, p =.33, RMSEA =.02, SRMR =.02, CFI =.99. Both a' paths remained significant: program x risk to academic competence (<unk> =.33, p =.01); to externalizing problems (<unk> = -.53, p =.00). The b path from academic competence (<unk> =.27, p =.00) to job aspirations was significant; the path from externalizing problems was nonsignificant. Probing of the simple mediation effects indicated that academic competence independently accounted for the program effects on job aspirations (95% CI:.0223,.1537) for the high-risk youth. --- Discussion This study examined whether a parenting-focused intervention for divorced families affected youth's educational expectations and occupational aspirations six years following participation and tested whether several parenting and youth variables mediated the program effects. The results indicated that, for adolescents who were at high initial risk for developing later problems, those in the program had both higher expectations for their educational attainment and higher job aspirations compared to their counterparts in the control condition. None of the posttest variables examined mediated the effects of the program on educational expectations and occupational aspirations. However, mother-child relationship quality as well as youth externalizing and internalizing problems, self-esteem and academic competence at the six-year follow-up mediated the effects of the program on high-risk adolescents' educational expectations. Also, measures of academic competence and externalizing problems at the six-year follow-up mediated the effects of the program on job aspirations for high-risk adolescents. When the significant mediators were entered simultaneously into the models predicting educational expectations and job aspirations to assess for unique mediated effects, only academic competence remained a significant mediator of program effects on job aspirations. Both self-esteem and academic competence uniquely mediated the effects of the program on educational expectations. This is the first study to examine the effects of a preventive intervention on the educational goals and occupational aspirations of youth in divorced families. The findings extend the results of previous studies, which have shown that prevention programs improved risky behaviors, substance use and mental health outcomes, and academic performance of youth in divorced families (DeGarmo et al., 2004;Pedro-Carroll & Alpert-Gillis, 1997;Pedro-Carroll, Sutton, & Wyman, 1999;Wolchik et al., 2000;2002;2007), to include a domain of functioning that has significant consequences for adult educational and occupational success (Cheeseman Day & Newburger, 2002;Harackiewicz et al., 2002). In the context of the consistent finding that youth from divorced families exhibit lower achievement in the domains of work and education (e.g., Biblarz & Gottainer, 2000;Hetherington, 1999) and experience more economic difficulties (Caspi et al., 1998), these findings have important implications for reducing the public health burden of divorce. The finding that program effects occurred for youth who were at high risk, but not low risk, of developing mental health and other problems is consistent with a growing body of research on the effects of prevention programs (e.g., Stice, Shaw, Bohon, Marti, & Rohde, 2009;Stoolmiller, Eddy, & Reid, 2000;Wolchik et al., 2000;2007). Screening for level of risk may be an effective way to increase the likelihood of benefits of interventions for divorced families (Dawson-McClure et al., 2004). It is important to note that support for meditational relations only occurred in the models in which the mediators and outcomes were measured concurrently. Thus, the significant pathways must be viewed as providing preliminary support for meditational relations (Kraemer, Yesavage, Taylor & Kupfer, 2000). Prospective mediational effects were not found for any of the posttest measures. It is possible that mediational pathways would have been detected if the time lag between the posttest and follow-up assessments had been shorter than six years. It is also possible that variables that were not assessed at post-test, such as monitoring and supervising of school-related activities, school performance, and completion of homework, may be related to occupational and educational goals in mid-tolate adolescence. Several dyadic and youth variables assessed at the six-year follow-up mediated the effect of the NBP on the educational expectations of high-risk youth. The mediational effect for mother-child relationship quality is consistent with previous work with this data set that has shown mediational effects of this variable for internalizing and externalizing problems and mental disorder symptom count (Zhou et al., 2008). Supportive parenting may provide adolescents with the confidence and self-worth necessary to develop ambitious long-term educational goals. The mediational relations between academic competence and educational expectations are consistent with previous studies linking academic achievement in middle and high school with later educational outcomes (Huurre, Aro, Rahkonen, & Komulainen, 2006;Strenze, 2007). These findings suggest that successful academic experiences may lead youth to aspire to achieve ambitious goals later in life. The mediational pathway for selfesteem and adolescents' internalizing and externalizing problems are consistent with research showing associations between self-esteem (Ahmavaara & Houston, 2000;Emmanuelle, 2009), internalizing problems (McLeod & Kaiser, 2004;Rapport et al., 2001), externalizing problems (Asendorpf, Denissen, and van Aken, 2008;Masten et al., 2005) and academic outcomes. It is possible that high self-esteem affects persistence in mastering academic tasks, which then affect educational goals. Similarly, aggressive or withdrawn behavior may prevent adolescents from learning effectively in school and inhibit positive relationships with teachers (Masten et al., 2005). These processes may affect academic performance, which then influences educational goals. It is notable that the effect of externalizing problems was obtained for mother/child-reported but not teacher-reported externalizing problems. One explanation for the difference in findings across reporters is that the high school teachers observed the adolescents for only one class period per day, which may have restricted the range of behaviors they could observe. Academic competence and mother/child-reported externalizing problems assessed at the sixyear follow-up mediated the effects of the NBP on high-risk adolescents' job aspirations. These findings are consistent with research linking academic success with later career success and socioeconomic attainment (Strenze, 2007) and suggest that career goals may be one mechanism through which school grades in adolescence contribute to later occupational attainment. In the multiple mediator models, only academic competence uniquely mediated program effects on educational expectations and job aspirations for high-risk youth. The lack of the contribution of the other variables may be due in part to the small sample size. A larger sample may be required to detect smaller mediation effects when multiple mediators are tested. The current study has limitations that can inform future research. First, given that all significant findings were found for the models where the mediators and outcomes were measured concurrently, it is not possible to draw causal inferences (Cole & Maxwell, 2003). It is possible that educational and occupational goals affect the hypothesized mediating variables or that there are reciprocal relations between educational and occupational goals and the proposed mediators. Future research that includes assessments in which the potential mediators have temporal precedence but are more proximal to the outcomes than those in the current study are needed to identify causal relations. Second, the sample was almost exclusively Caucasian and middle-class. Studying the effects of this program and others for youth from divorced families on educational and occupational goals using samples that are diverse in terms of ethnicity and socioeconomic background is an important future direction. --- Implications for Theory and Intervention The current study demonstrated that a prevention program for divorced families had longitudinal effects on the educational expectations and occupational aspirations of high-risk adolescents, and that these program effects were partially mediated through programinduced effects on academic achievement, self-esteem, externalizing problems, internalizing problems, and the quality of the mother-adolescent relationship. The finding of long-term effects of this program on educational and occupational goals highlights the importance of including follow-up assessments to identify both enduring effects and outcomes that are specific to developmental stages that occur after program participation. There is considerable evidence suggesting that the benefits of preventive interventions continue to unfold over time (DeGarmo et al., 2004;Gillham, Reivich, Jaycox, & Seligman, 1995;Wolchik et al., 2007). Further, the findings of this study underscore the need to study the effects of prevention programs on educational and occupational goals, as these outcomes have been previously linked to performance and attainment outcomes in these domains (Harackiewicz et al., 2000;Judge et al., 1995). To our knowledge, the current study represents the first test of the effects of a preventive intervention on educational and occupational goals and aspirations, and it is also one of the few studies linking these outcomes with youth self-esteem, internalizing and externalizing problems, and the motherchild relationship. --- Summary Children from divorced families are at an increased risk for decreased academic and occupational achievement, relative to their peers from non-divorced families (Biblarz & Gottainer, 2000;Caspi et al., 1998;Hetherington, 1999). The findings from the current study indicated that the educational and occupational goals of high-risk adolescents from divorced families were enhanced through a preventive intervention that focused on improving parenting skills. In addition, these findings suggest that intervention-induced effects on several intrapersonal and interpersonal risk and protective factors, such as self-esteem, academic competence, externalizing problems, internalizing problems, and mother-child relationship quality, were associated with improvements in high-risk adolescents' educational expectations and occupational aspirations. These results suggest that the public health burden due to divorce may be reduced through the widespread implementation of parenting-focused preventive interventions for this at-risk population

This study examined whether the New Beginnings Program for divorced families led to improvements in youth's educational goals and job aspirations six years following participation and tested whether several parenting and youth variables mediated the program effects. Participants were 240 youth aged 9-12 years at the initial assessment, and data were part of a randomized, experimental trial of a parenting skills preventive intervention targeting children's post-divorce adjustment. The results revealed positive effects of the program on youth's educational goals and job aspirations six years after participation for those who were at high risk for developing later problems at program entry. Further, intervention-induced changes in motherchild relationship quality and youth externalizing problems, internalizing problems, self-esteem, and academic competence at the six-year follow-up mediated the effects of the program on the educational expectations of high-risk youth. Intervention-induced changes in youth externalizing problems and academic competence at the six-year follow-up mediated the effects of the program on the job aspirations of high-risk youth. Implications of the present findings for research with youth from divorced families and for the public health burden of divorce are discussed. It is well documented that parental divorce is associated with multiple problems for youth that extend into adulthood, including internalizing and externalizing problems, interpersonal difficulties, poor physical health, and substance use (e.g., Amato, 2001;Chase-Lansdale, Cherlin, & Kiernan, 1995). Several studies have found that parental divorce in childhood is also linked with negative educational and occupational outcomes across the life span, such as a decreased probability of graduating from high school, after controlling for income, parental educational attainment, ethnicity, and other demographic variables (e.g., Sandefur,

Ethnography-based Anthropology and science-fiction are commonly regarded as separate genres of intellectual and literary production. The former is seen as scientific and the latter as artistic. Anthropology deals with actually existing forms of social life and social worlds whereas science fiction deals with "fantasy". Both a positivistic stance on science (especially social science and the humanities) and a romantic stance on art (especially literature) have been profusely questioned and challenged since the post-modern turn in both social sciences and the arts. This article aims at illustrating the possible porosity between Ethnography-based Anthropology and the literary genre of science fiction with a focus on sex, gender, and sexuality and resorting to a feminist and queer perspective. Anthropology aims at making the exotic familiar to make the familiar exotic, as the well-known saying goes. Science-fiction aims at imagining worlds that do not exist but does so based on the historical, social, and cultural experience of both author and readers. That experience includes the diversity of human social constructs as recorded in the ethnographic archive and in anthropological analysis. In this sense, can one say that they both work in a utopian mode? By this I mean the utopias and dystopias proper -those of science fiction -and the altertopias of Ethnography. By altertopias I do not mean that the ethnographic record relates to a non-existing reality (it does not, of course, since it reports on the actual life of existing human groups), but rather that the world thus reported is alien to the reader, albeit not so to both anthropologist and the human group with whom the ethnography was produced. In both genres and in both forms of intellectual production, we are faced with alternative worlds. The portrayal of those worlds itself opens a space for a critique of the reader's common-sense reality. Feminist and queer imaginations, specifically, work for changes in social relations and identities that necessarily engage in altertopian endeavors, since they are in opposition to actually existing heteronormativity, and both Anthropology and sciencefiction are of enormous potential as fields of inquiry and imagination with a politically transformative potential. Anthropologists can immediately think of Haraway's (2016) speculative and science-fiction inspired work on the future, or the ethnographically based theorization of Marilyn Strathern (2001) and her proposal to overcome the Western template of understanding gender as based on the sexualized concepts of "men" and "women". Their political effect can not be underestimated: one imagines the future as an overcoming of present challenges, the other presents an alternative cultural construct that chellenges our certainties. In this sense, politics requires imagination: "we could have been, or we could be like them" (the ethnographic other, or alter); "we could have been or become like them" (the other, or alter, in science-fiction). --- * The literary work of Ursula K. Le Guin (1929Guin ( -2018) ) has been the subject of extensive analytical and critical work, the review of which is outside the scope of this article. I would rather focus -based on my anthropological experience and on my predilection for Le Guin's oeuvre -on aspects of her approach to gender and sexuality, as well as political-economic utopias and dystopias. Specifically, I have been paying attention in her work to three aspects. Firstly, how it shows that our gender and political-economic orders can be seen as dystopian, including those utopian political projects that were implemented or experimented with; Secondly, how her work calls for utopian models that are small-scale, ecological, and egalitarian -and very much inspired in the ethnographic archives but without conceding to pre-modern nostalgia; Thirdly, how productive the porosity between Anthropology and science-fiction can be for our discipline's contribution to the imagination of a feminist-and queer-inspired world, as well as a critique of our present situation. Ursula K. Le Guin's science-fiction oeuvre deals, among others, with two fundamental topics of concern here: gender and sexuality, on the one hand; and political-economic utopia and dystopia, on the other. Can her feminist approach to the subject and subjects of her stories be seen as a form of queer imagination avant la lettre? Also, does her approach to anarchist utopias, counterpointed with capitalist and anthropocenic dystopias, avoid the simplistic dichotomies of the political debates of the pre-Fall-of-the-Berlin-wall era? From a biographical point of view, Le Guin's upbringing in her father's (Alfred Kroeber) ethnographic field sites and museum institution has -as she has admittedinspired her work. For this purpose, and to highlight the aspects mentioned in the previous paragraph, I will focus on her novels The left hand of darkness (1969), The word for world is forest (1972), The dispossessed (1974), Always coming home (1985), and, generally, in what critics have called the Hainish Cycle in her work. My reading of Le Guin has been centered, throughout the years, on the abovementioned cycle of novels and stories. The cycle was not intended as such by Le Guin, it was not "a series" that she had planned. It is rather an expost connection, rather than collection, of novels and short stories set in several different imaginary planets. These planets' populations were the result of colonization by the original humans from planet Hain. Both Hainish genetic experimentation and local evolutionary processes led to different sub-species of humans and different cultural complexes in different planets. Furthermore, in many novels we get to know a planet through the narrative of a Mobile, an ethnographer of sorts, who is supposed to report on the population, without interfering in their lives and history, in order to help the Ecumene (the cooperative, anarchy-like federation of planets colonized by Hain) decide when and if to establish contact and provide technology (mainly communication technology that allows for cooperation and exchange of knowledge). * Baker-Cristales (2012) uses the wonderful expression "social science fiction" when referring to Le Guin's ethnographic sensibility. She notes how "ethnographic writing (is) neither fiction nor science" (Baker-Cristales, 2012, p. 15) and resorts to Jameson's (2005) notion of how "the belief that storytelling and imagination are entirely distinct projects from science and truth-telling is a product of a particular, very circumscribed, social-historical system -modern capitalism" (Baker-Cristales, 2012, p. 17). As Jameson (2005, p. 286) has conveyed, rather than predicting, science fiction's purpose is to "defamiliarize and restructure our experience of our own present, and to do so in specific ways distinct from all other forms of defamiliarization". Baker-Cristales notes how often the protagonists of Le Guin's novels are anthropologists by another name. "By using ethnography as a textual model for fiction, Le Guin aims to transcend some of the limitations of fiction -the end of the story" (Baker-Cristales, 2012, p. 19). In the words of Fredric Jameson (1975, p. 4), "one of the most significant potentialities of Science Fiction as a form is precisely this capacity to provide something like an experimental variation on our own empirical universe". This experimentation is usually codified as analogy and extrapolation (Jameson, 1975). It is complemented by world reduction, a taking away of certain characteristics, whether technological, institutional, or cultural, of our habitual cultural world (Modern, Western, Capitalist...), thus setting the stage for speculating the outcome of that removal, a sort of "take away this variable and see what happens". In The left hand of darkness, Le Guin (1969) describes a world without gender. People in planet Gethen -who are human, as everyone in the Hainish Cycle planets, albeit modified by local evolution -are hermaphrodytes1 and do not experience permanent libido or sex drive. Like in many species we know, they have estrus, or cycles of sexual availability, called kemmer. When on kemmer, if two persons are attracted to each other one of them will develop secondary female characteristics and the other male characteristics. There is no way to tell which will be which and a person can "become" both male and female in different kemmer cycles in their lifetime. Jameson (1975, p. 7-8) also commented on the Gethen: Rather than stand in favor of a wider tolerance for all kinds of sexual behavior, it seems more appropriate to insist [...] on the feminist dimension of her novel, and on its demystification of the sex roles themselves. The basic point about Gethenian sexuality is that the sex role does not color everything else in life, as is the case with us, but is rather contained and defused, reduced to that brief period of the monthly cycle when, as with our [non-human] animal species, the Gethenian are in "heat" or "kemmer". So, instead of eliminating sex, Gethenian biology eliminates sexual repression, according to Jameson (1975, p. 8): For if Le Guin's Gethen does not do away with sex, it may be suggested that it does away with everything that is problematic about it [...] The dream of some scarcely imaginable freedom from sex, indeed, is a very ancient human fantasy, almost as powerful in its own way as the outright sexual wish-fulfillments themselves. Le Guin connects Gethenian absence of a permanent sex drive and the absence of gender to the wider theme of societal organization and conflict: there being no gender oppression or violence, Gethen is a society uninterested in growth and power, and does not therefore see the point in war between either factions or nations. The theme of gender and sexuality, and its connection to wider societal predispositions, is prevalent in Le Guin's works -from societies where women are in power, to those where there is no distinction between sexual orientations. In The dispossessed (Le Guin, 1974), the anarchist organization of the planet Anarres is not depicted only on the grounds of communal possession, central organization of the distribution of labor and resources, or on the ecological care and avoidance of disproportionate growth or accumulation. It is also characterized by what could be called as "free love" and the absence of structured marriage, cohabitation, or marriage alliances -as opposed to its nemesis planet, Urras (and also its origin, Anarres being a colony settled by radicals that have fled from Urras), where capitalism is full-fledged as well as gender and sexual structures very similar to our contemporary patriarchal ones. Is Anarres a utopia like so many others in mainstream, run of the mill science-fiction? Not quite. Le Guin seems to want to say that it is an experiment, an on-going one, where hesitation and doubt have a place. That is what the main character, Shevek, embodies, when he decides to play the role of the scientific emissary to Urras, where he is both attracted to and disgusted by the niceties of an affluent society, while back home in Anarres he is also ambivalent about the political choices and, most of all, the unavoidable fact that even in progressive utopias a new hegemony tends to be established, social pressure to conformity occurring as well. The above-mentioned reference, by Jameson, to the plot strategy of "world reduction" (let's remove capitalism, the nation state, industrial production, cars, and airplanes, and see what life could be like) is applicable to Le Guin. She does something like this in Always coming home, a novel that is not part of the Hainish Cycle. It is an invented ethnography. It follows the style of early ethnographic efforts and is reminiscent of Malinowski's writing template. However, it is also an archaeology of the future, not a rendition of the bast or of a contemporary culturally different society. It depicts California in a future when people have gone back (but, as certainly Le Guin would say, forward) to indigenous ways of living, especially in what regards keeping small-scale, ecologically balanced societies, with close social ties based on communal modes of existence. Indigenous but of today -and tomorrow. This neo-tribal society exists while an Internet-like technology is available, kept by automated entities that do not interfere with human life but simply collect, organize, and provide information. As Baker-Cristales (2012, p. 24) The connection between Le Guin's concern with gender and sex, on the one hand, and her concern with apparently utopian modes of social, economic, and political organization, on the other, was also observed by Jameson (1975, p. 12-13): The existence of modern technology in the midst of an essentially feudal order [in The left hand of darkness] [...] It becomes difficult to escape the conclusion that this attempt to rethink Western history without capitalism is of a piece, structurally and in its general spirit, with the attempt to imagine human biology without desire... Donna Haraway (2014) also refers to the connection between Always coming home and the problematization of the Anthropocene (actually, the Chthulucene, in her formulation), when she acknowledges her inspiration on Le Guin -together with anthropologists Marilyn Strathern and Anna Tsingand how she helped her realize that utopias shouldn't be big systems and signifiers, but actual small-scale human attempts at relation and connection with other species and the world. Haraway (2016) is the leading contemporary 2 See Trouillot (2003, p. 19). Trouillot's work can be a source of critique of homologies between Western utopian thought and narratives of the cultural difference of colonized peoples. Le Guin's social science fiction avoids the trap of both the noble savage and the savage primitive. author acknowledging the influence of (and dialogue with) Le Guin, when she writes that "Le Guin's carrier bag theory of narrative comes to the rescue, along with biologist Deborah Gordon's theories about ant interactions and colony behavior, to elaborate the possibilities of ecological evolutionary developmental biology and nonhierarchical systems theories for shaping the best stories. Science fiction and science fact cohabit happily in this tale" (Haraway, 2016, p. 7, my italics). Haraway is referring also to stories that deal with non-human characters, but her assessment would certainly aply to stories starring humans in the Hainish cycle. What is at stake is the radical change of perspective. Le Guin (1989, p. 97-98) herself wrote inspiringly on this: Copernicus told us that the Earth was not the center. Darwin told us that man is not the center. If we listened to the anthropologists, we might hear them telling us, with appropriate indirectness, that the White West is not the center. The center of the world is a bluff on the Klamath River, a rock in Mecca, a hole in the ground in Greece, nowhere, its circumference everywhere. Perhaps the utopist should heed this unsettling news at last. Perhaps the utopist would do well to lose the plan, throw away the map, get off the motorcycle, put on a very strangelooking hat, bark sharply three times, and trot off looking thin, yellow, and dingy across the desert and up into the digger pines. In Returns: becoming indigenous in the twenty-first century, James Clifford (2013) revisits the well-known story (among anthropologists) of Ishi, and he does so from a contemporary perspective, one that does not see Native American realities, especially in California, as something of the past, as exterminated, but rather as growing and expanding albeit in new ways. He retells the story of Ishi, popularized as "the last wild man", who lived part of his life in Alfred Kroeber's (Le Guin's father) museum at Berkeley. He also tells the story of Theodora Kroeber's (Le Guin's mother) popularization of Ishi's story and symbolism in her books, especially her children's book 3. He goes on to narrate the story of Orin Starn's investigation of Ishi's brain at the Smithsonian Institution and the Indian efforts, in California, for repatriation of the remains, as well as the polemic at Berkeley over the need or not for apologizing for Kroeber's and the university's actions in the past (a campaign championed by Nancy Scheper-Hughes, and countered by one of Le Guin's brothers, literary scholar Karl Kroeber). And, of course, he cannot avoid mentioning Ursula. Why and how? Because in a way he establishes an opposition and contrast between Le Guin's work, especially the abovementioned Always coming home (Le Guin, 1985), and "Kroeber's purified pre-contact reconstructions of California" (Clifford, 2013, p. 106). In a sub-chapter titled "Utopia", he says how Ursula's work "draws on folklore, popular culture, Taoism, post-sixties feminism and environmentalism" (Clifford, 2013, p. 111) and how much these are themes that are central to Ishi's world: --- * For many years I used Le Guin's (1972) novel The word for world is forest in Introductory Anthropology classes. It helped, through the guise of science fiction, to convey the sense of doing ethnography as a process of cooperative knowledge building between anthropologist and the social groups (through specific persons) -and how much ethnography-based knowledge is the outcome of processes of porosity and juxtaposition between scientific, literary, interactional, and interpersonal processes that result in a new type of knowledge that is quintessentially human(istic). The altertopia of the context of The word for world is forest mimics the process of ethnographic discovery, as well as depicts the conflicts of power resulting from colonial and capitalist encroachment in many communities. It does so through the guise of fiction, something that is more effective on young students than descriptions of our "real" History (always prone to trigger emotional or ideological reactions that hinder learning). Clifford (2013, p. 119) focuses precisely on this book: In Le Guin's parable [...] anthropological humanism emerges as both essential and impotent in situations of colonial/anti-colonial antagonism. Lyubov [the main character, an anthropologist of sorts] is unable to reconcile inter-personal loyalty, political commitment, and scientific comprehension: he will not emerge unscathed with his intercultural understanding. The word for world is forest depicts capitalist and colonialist exploitation of native resources (in a planet of the Ecumene, the loose federation centered in Hain), together with extreme racism and extermination of the "savages", whose rich culture -and, indeed, alternative (altertopian) social and ecological organization -is dismissed 4. It is at the very other end of the line connecting Kroeber's anthropology and the Native American experience of Kroeber's time, on the one hand, and today's Native American situation as well as contemporary utopian visions on the verge of the Anthropocene, on the other. Then, in Kroeber's time: the "last wild man", the unavoidable result of the triumph of civilization, something to feel sorry about but unstoppable for supposedly good reasons ("evolution", "civilization", the colonial/capitalist utopia...), and anthropologists whose mission was to salvage and register that which was about to disappear. To salvage the savage, so to speak. Now: the renaissance of Native American identities, with casinos and local and state politics, performative arts, comedy, self-representations, newly invented or consolidated tribal entities, mixed and urban identifications, etc., and anthropologists who consider the difficulties of cross-cultural dialogue and believe not only in constant, ongoing mixture, cultural reinvention and in decolonizing knowledge, but also in the unavoidable doubts and conflicts that utopias and civilizational projects entail -as does Le Guin. As Clifford (2013, p. 121) says, poetically: "Ishi in a loincloth, Ishi in work clothes, Ishi with feathers, Ishi in a suit and tie". Becoming indigenous after colonization (Clifford's expression) is, after all, what we can find in Always coming home. In her theoretical essay inspired by that novel, "A non-Euclidian view of California as a cold place to be", Le Guin (1989, p. 85, my italics) says: I am not proposing a return to the Stone Age. My intent is not reactionary, nor even conservative, but simply subversive. It seems that the utopian imagination is trapped, like capitalism and industrialism and the human population, in a one-way future consisting only of growth. All I am trying to do is figure out how to put a pig on the tracks. --- And in the same essay: Utopia has been Euclidian, it has been European, and it has been masculine. I am trying to suggest, in an evasive, distrustful, untrustworthy fashion, and as obscurely as I can, that our final loss of faith in that radiant sandcastle may enable our eyes to adjust to a dimmer light and in it perceive another kind of utopia [...] It may look very like some kind of place Coyote5 made after having a conversation with his own dung. [...] [A yin utopia] would be dark, wet, obscure, weak, yielding, passive, participatory, circular, cyclical, peaceful, nurturant, retreating, contracting and cold. (Le Guin, 1989, p. 88-90). --- * Maybe the trigger for this article was my difficulty in imagining the bodies of Gethen people. That is how "imaginative" became ivaginative, for in my efforts to picture Gethenians I always ended up picturing a vagina, certainly because it is an altertopian organ for my male body. It is important to distinguish my use of the differently spelled "ivagination", which is mostly playful, from the philosophical concept of "invagination". For Derrida (1980), invagination describes narratives that fold upon themselves. It is an aspect of "différance", opening the "inside" to the "other" and denying a stable identity to both. But I should be more accurate: not just the vagina. A clitoris will engorge exponentially in the "male" person during kemmer, and the vagina will widen significantly in the "female". And both gonads -testicles and ovaries -reside inside both persons. As does the uterus. One thing becomes obvious when the picture turns clearer: the gestational potential that resides in the uterus is in both per-sons. They are, in "our" sense, in our intrinsically biologically gendered world view, both women, in the sense that they are both potential gestating humans. Would I dare say that Le Guin was wrong, that Gethenians are all women? And that the presence of gender is tantamount to the presence of the masculine? Its absence would be general, human femininity, humanity as feminine. What, in my reading, is common to all Gethenians is the uterus. The Gethenian theme can thus be seen not just as a matter of gender balance, or gender oscillation, or gender interchangeability, or any variation thereof, but as referring to gender itself as a temporary phenomenon beyond being human. That could be seen as Queer avant la lettre -before queer became, in common parlance, a rather flat synonym for individual choice and performativity. Social Science Fiction and the Ethnographic mode seem to go hand in hand, instead of being at odds. Le Guin's novels, like all good science fiction, succeed in being internally consistent and logical. Science-fiction is not mere fantasy, for things in it must make sense according to its own established premises. In Social Science Fiction one must be even more careful because the material upon which the fictional work is done is human experience, relationality and sociality, and it needs to make sense to us, readers of the here and now. And it always does, at least with Le Guin, because the worlds she imagines are the worlds that we are constantly imagining in the "what ifs?" that we pose resorting to the materials at hand -bodies, social constraints, institutions, political projects, "nature", possible ontologies.... The ethnographic mode is one of creating intelligibility, through defamiliarization and refamiliarization, only achievable because the material used is common to interpreted and interpreter. In a world where the effects of colonialism-cum-capitalism have created new indigeneities; in a world where modern utopias seem to have become more disillusions than illusions; in a world now marked by all that the neologism "Anthropocene" encapsulates; in this world, Le Guin's invented knowledge of other worlds, and the anthropologists' knowledge of worlds created by others, provide materials and imaginative designs for altertopias. If Le Guin's father, Alfred Kroeber, was "salvaging the savage", Kroeber's daughter, Le Guin, provides us a glimpse of what it can be to "savage the salvage". The poetic notation of that change would be the replacement of the masculine m of imagination by the feminine v in ivagination -not an "image" but a life-giver, a genital. 6 Le Guin (2017, p. 1035) herself wrote thus about her Gethenian experiment: I was not reccomending the Gethenian sexual setup: I was using it. It was a heuristic device, a though-experiment. Physicists often do thought-experiments. Einstein shoots a light ray through a moving elevator; Schrodinger puts a cat in a box. There is no elevator, no cat, no box. The experiment is performed, the question is asked, in the mind. Einstein's elevator, Schrodinger's cat, my Gethenians, are simply a way of thinking. Following her inspiration, and in order to take seriously my approach to the porosity between science-fiction and Anthropology, what follows is my own work of fiction. It is clearly inspired in the Gethenian theme -in fact, it can be understood as a "take" on The left hand of darkness. It was meant to be passed on to students of gender and sexuality as part of a pedagogical sequence consisting of: a) the presentation of culturally diverse gender orders and regimes (Connell, 1987); b) their exposure to Le Guin's fiction (not just The left hand of darkness, but also A fisherman of the inland sea 7 ); c) my own fiction, as a way of putting together the possibility to imagine and the conventional and systematic nature of social organization that can be described by anthropologists with internal coherence. Fiction meets Anthropology in Social Science Fiction in the process of ivagining worlds. * 6 As mentioned before I have been reading mostly Le Guin's novels from the so-called Hainish Cycle. I have not dealt with her vast work on the fantasy genre, the Earthsea series. But as I was writing this essay, I was reading her early work, before both series. It is based on Orsinia, an imaginary Mitteleuropa country and the action takes place in the period of liberal and nationalist revolutions, namely against the Austrian Empire. It is very curious to see Le Guin's fascination with a specific historical past and geographical location, especially one where and when "utopias" thrived -nationalism, socialism, zionism, communism, utopian socialism -that today have clearly gone awry. --- Evolution From the notes of Eva g Sirdi, written approximately 150 years ago in the margins of a copy of the "Holy Ala'm" and discovered in the region of the Great Lakes of Northern Barbaria, selected and transcribed by Bei ag Arik, Doctor in Human Evolution by the University of Golmak. "It was the morning of my return from my second expedition and once more I was arriving empty handed. Seven years before, at the end of my first expedition, Ad had been waiting for me with an understanding smile. Ad had received my letter, sent by commercial ship, in which I told my frustration for not having found what I was looking for. Ad's smile had then been also one of encouragement: 'Do not worry, there will be a next time and, who knows, you'll then succeed'. Seven years and one more letter full of disappointment later, how would it go? One can take one, or two, or maybe even three consecutive disappointments. You take a deep breath and get back up on your feet again. But can the person you share your life with see the other fail time and time again? Would the smile be understanding and encouraging once more?" "Like seven years before, you could hardly see the harbor through the mist. The journey to the Northern Hemisphere must be done in the right time of the year, to take advantage of the favorable ocean currents and avoid the harsh climate. Consequently, the return trip always took place in the Southern Fall months, a season of thick fogs, penetrating humidity, and a feeble, pale light, as if filtered by layers of gauze. The contours of harbor and town took a while to focus. First you would see the fickle light of the whale-oil lanterns held by relatives waiting for you. Some higher, some lower, probably held respectively by adults and children. Slowly, the dark shapes of the Araken's palace up in the Hill became perceptible, together with the sails of other ships. But even before the bodies of the cherished ones became visible, you could hear a band playing a welcoming saruetz." "Expeditions to the Northern hemisphere were always State events. An expedition was an expensive endeavor and the satisfaction of curiosity, or the search for knowledge (what, so the saying goes, the ancients called Science) were the State's last priority. The Araken, our State council, wanted us to find precious metals more than anything. In that regard the expedition had gone quite well: two months of slow and painful sailing to the North; the good fortune of having found a potentate of Northern Barbarians enjoying a period of peace with their neighbors and therefore willing to negotiate with us; the childish naiveté with which those barbarians accepted our gifts of worthless monkey skins -apparently so important for their ritual bartering in prestige competitions between their chiefs -in Exchange for metals whose value they were far from even imagining; the sumptuous reception that was offered us; and, at last, a one and a half month return trip with exceptionally calm seas." "The crew threw the thick and heavy ropes. The fog insisted on hiding the contours of the bodies waiting for us. Would Ad be there? Would Ad have decided to stay home, not knowing what to do with my frustration? If my mission as commercial envoy of the Araken had gone well -as had the other one, seven years before -the same could not be said of my other mission, my mission. Once again, I had gone up the wide river that penetrates the jungle up from the capital of the Northern potentate that we traded with. Once again, I had, with relative ease, hired a large group of carriers and a trustworthy guide who was very knowledgeable in the native legends about the existence of what they called the agan-uruk, the dimorphic creatures. Once again, I had been able to reach the edge of the great mountain chain. That time I had even climbed all the way up to the Great Lakes region, where the creatures supposedly live, despite having lost a handful of carriers, taken by the strange disease that infests the woods and marshlands, and that I was lucky to avoid, some may say thanks to the amulets that the head-priest of the Araken had given me. During that journey I had even managed to obtain a piece of a creature in one of our traps -a foot, to be exact, similar in all aspects to our own feet, therefore from either a dimorph or any normal native of the land. I had not found anything remotely like a camp or a village of any sort. Nothing." "I remember the ship scratching the stone of the pier. I could already see the shapes of people. Agitated, running back and forth, shouting names, kin looking for kin. Ad wouldn't do that. Ad always reserved the expression of emotions for smiles, looks, body movements. Most people experienced a ship's arrival as a once in a lifetime event. Their lives unfolded almost entirely in our port city of Golmak, maybe in the outskirts too, or in some portion of territory at a horseride's distance. To travel is to adventure, to adventure is to travel. Although both expressions have a slight phonetic difference -armek and arm<unk>k -they mean the same to the common folk: risk, danger, daring, challenging the infinitely calm routine of life in Golmak. Ad never wanted to travel with me. Ad did share my interest and curiosity for research and knowledge. The same indomitable urge to know how things work, well beyond the explanations in the Holy Ala'm or the Laws of the Araken -the latter being no more than pragmatic transpositions of the former. Ad had other ways of searching for knowledge. Reading was a favorite activity, reading like there was not enough time in life to read everything that has ever been written, whereas I.... I would travel-adventure. That was why we had decided that it would be wiser that Ad, not me, should be the one to someday get pregnant, to be the uru, the genitor from inside." "I remember how suddenly the fog lifted. There was Ad, an unmistakable shape, behind the band that was playing off-key a repetitive saruetz, and the high dignitaries and lineage heads that I had to salute, one at a time, with deference and respect, and from whom I received congratulations and promises of rewards. Behind all of them, Ad was the true center of my attention: the same beauty as in my previous return seven years before, the same demeanor, the same attitude, staying calmly behind other people, not wanting to be a part of the excited crowd. I like

Ethnography-based Anthropology and Science Fiction can engage in a productive dialogue since both address what is proposed as "altertopias". Utopias, dystopias, and cultural alterity share the possibility of imagining social and cultural organizations different from both those of the authors and those of the readers. These imaginations are intrinsically creative/artistic and political at the same time, and they critique power structures, especially when approached through a feminist stance. Inspired by the literary work of Ursula Le Guin, the article takes this further by experimenting with the inclusion of a fictional piece of "Social science fiction" that itself plays on Le Guin's themes.

a'm or the Laws of the Araken -the latter being no more than pragmatic transpositions of the former. Ad had other ways of searching for knowledge. Reading was a favorite activity, reading like there was not enough time in life to read everything that has ever been written, whereas I.... I would travel-adventure. That was why we had decided that it would be wiser that Ad, not me, should be the one to someday get pregnant, to be the uru, the genitor from inside." "I remember how suddenly the fog lifted. There was Ad, an unmistakable shape, behind the band that was playing off-key a repetitive saruetz, and the high dignitaries and lineage heads that I had to salute, one at a time, with deference and respect, and from whom I received congratulations and promises of rewards. Behind all of them, Ad was the true center of my attention: the same beauty as in my previous return seven years before, the same demeanor, the same attitude, staying calmly behind other people, not wanting to be a part of the excited crowd. I like it that way: Ad's attention, eyes, and body were focused on me. And when I got close, it was relief that I felt. Ad's smile seemed to be saying, once more, 'You can always try again'." "How many times does one make love with the same person as if it were the first time? Few people can reply without irony: 'Once'. That day Ad and I became part of the group of happy few who can say 'three' -the very first, when we met, the second when I came back from my first expedition, and the third on that foggy day. Still at the harbor, Ad hugs me gently and for very long. And whispered: 'We'll talk about your trip tomorrow'. I recall feeling Ad's breasts harden. Ad must have felt the same with my penis. And immediately, in those brief seconds of desire's work, when one's desire pulls the other's desire and vice-versa, I felt the same happening to Ad's penis and Ad felt the same happening to my breasts. And both Ad and I felt, under our penises, the anxious moisture of our vaginas growing." "'Maybe now we should talk about the expedition'. Suggestions such as this, when coming from Ad, mean decisions already taken. Ad knew that I would have two important meeting the following day. The first, with the Araken, I would attend alone, and not a shadow of a problem was foreseeable. On the contrary: the metal booty that I had obtained from the Northern barbarians was such that I could almost expect to be granted a title of nobility, not to mention land and riches. My status in the Araken seemed to have one direction only: forward and upward. The second meeting, I would attend with Ad, since we were both full members of the Guild of Sages, a semi-secret group of fifteen people dedicated to the search of knowledge of the world, of a whole planet Earth to be discovered. We were all well versed in the exegesis of the Holy Ala'm; we had all learned how to read, write, and think in theological schools and we all had, in some moment or other in our youth, seriously considered the pursuit of a priestly career. But we all had decided instead to engage in the pursuit of knowledge and curiosity, to try to recover the 'Science' that our ancestors had practiced millennia before." "'What will you tell the Guild about the trip?' -and Ad's voice was now in the opposite register of the music of sex from a few hours before. 'Don't you think it's better that after this second attempt you abandon your pursuit?' 'But I won't abandon my pursuit, Ad, even you still support me on that, don't you?' 'You know I do. But you also know that I do it because I know that you believe that you are right. But if you insist on the search for the dimorphs, Ev, it can be....' 'Dangerous?', I completed. I knew the risk was real. Although the members of the Guild of Sages shared a code of honor that placed the search for knowledge above other loyalties, it was also true that several members of the Guild were simultaneously members of the Araken. The Araken had 101 members, representing all the lineages of Golmak. They all knew each other from the priestly schools -that was what united us as an elite and separated us from the common folk. In such circumstances it was easy, too easy, for any rumor or suspicion to leave the Guild and rapidly contaminate the Araken." "As Eva g Sirdi, member of the Sirdi lineage, one of the oldest in Golmak, I felt protected. Not having pursued the priestly career had not been a disadvantage, since that was not expected of all young people. Priestly teaching was expected for all, not the profession. My choice of a commercial career had been respected, and even more so after my successes with the Northern Barbarians. Even my semi-secret life in the Guild could be seen as an eccentricity of the kind allowed to the powerful. However, there was a rumor in town about how my commercial trips were a ruse for the satisfaction of my curiosity, a word that although not outrageous, had a titillating, provocative overtone, a taste of the forbidden. The true problem lay elsewhere: the nature of my search and curiosity, the question that I was asking, the answer that that question suggested." "Once, after my first expedition to the North, Ad ordered me, very harshly, to sit and pay attention. 'Listen, Ev, listen very carefully. Sometimes a priestly education too early in life leads those who have the knowledge of the sacred texts to be those who forget them sooner. Listen to this: Upon creating the Universe, Ala'm rejoiced in the beauty and perfection of Their work. The sky, deep blue when lit by the moon or light blue when warmed by the sun, was a thing of perfection and tranquility. The sea, reflecting the colors of the sky, leaped its waves, now gently, then furiously. The earth, green and flat, dark and mountainous, connected sky and sea. Such was A'lam's dwelling. To inhabit it with life did Ala'm create the plants that grow from earth, cast their seed and generate their own twins. Not satisfied with the plants, did Ala'm create the animals that move on it and did Ala'm make them more intricate than the plants they fed upon. He called some female, he called some male, and from the union of female and male new and different creatures were born. But seeing Ala'm that much was the dispute between females for males and between males for females, They made a new being, one that would feed on plants and animals. In the divine image did They create it, without female or male, and granted it with the gift of curiosity so that it could roam the world and possess it and, in the journey, find Ala'm Themselves, the one that is everywhere and nowhere. "'Why are you Reading me that, Ad? I know the Holy Ala'm. By heart. 'Because you can't take away from people that minor sentence. Minor, but of the greatest importance: in the divine image. At least you can't do it yet. And only Ala'm knows when you can...' 'But what I defend is somehow there, in the text', I argued, 'Ala'm was not satisfied with two-sex animals. One can say that Ala'm made them evolve. And since we do not descend from, say, a snake or a mockingbird, we must descend from people that still had the two-sex characteristic, such as monkeys do. My theory is not incompatible with the sacred texts. Nor with the ability that Ala'm gave me to... question. You know that is my standing, Ad'. Ad knew. Ad also knew the legends and myths that I had registered in the North and that told of a mythical people with two sexes." "Ad accepted my theory and as a curious person, as a member of the Guild, also had a practical mind and thought that it was impossible to demonstrate my hypothesis: 'You will never find your missing link, I'm afraid. I would give anything to help you achieve that goal and yet I would also give everything to have you give up your travels.' Differently from the tasteless stories told by the common folk, our arguments did not invariably finish with love making, with the perfect symmetry that the uneducated like, the two members of the couple taking turns in penetrating one another, time and time again in the endless pursuit of the perfect amorous equilibrium. Our arguments were... arguments, as if we were in the Guild of Sages, and always ended in an impasse, a deadend. 'So, Ad, your problem is that you think that there is no missing link. That there isn't a species like our own but in a previous evolutionary stage, a stage with two sexes?' 'No, Ev, that is not my problem'." "I didn't understand Ad's fear until, two weeks later, I received the summons from the Court of Inquiring of the Araken. I had never been present in a Court session. All that was known about that most rigid and secretive institution on Earth was the product of rumor, cautionary tales, scaring legends made to instill obedience and consensus. The most terrible tale had involved a relative of mine, Ca. My uru or genitor from inside -the person who gave birth to medid not share genitors with anyone, did not have ere, neither from the inside nor from the outside. My ulu, or genitor from the outside, however -the one who had inseminated my uru -shared genitors with Ca. They was, therefore, ere of my ulu. My alik. I was still an adolescent when my uru and my ulu embarrassingly announced that alik Ca had been banned to exile in a foreign land. They never wanted to explain to me why. Whenever they tried to do so they would give up, and I felt that they were embarrassed, that they could find or proffer the right words. Only years later did I come to know the truth and through indirect ways: Ca had been denounced to the Court of Inquiring for using only one of his genitals (the vagina) in the sexual act, and for refusing to use the other. How bigoted people can be when they impose their notion of'sexual plenitude'! So narrow-minded, a law that criminalizes that! Today, as someone who dedicates life to the pursuit of knowledge and the satisfaction of curiosity, I think that my alik Ca was probably just ill, the victim of some recessive disease from the time in evolution when we had not yet achieved our stage of perfect unimorphs." "I never knew who denounced me to the Court. The Court accepts anonymous complaints and guarantees anonymity. The Court just could not accept that a member of the Araken and of one of the most prestigious lineages, and with priestly training, would use the sacred texts to find in them a hint, an opening, for a, as the sentence said,'sinful confabulation' on the origins of humankind in two-sexed creatures. Ad's fear was confirmed. The love of my life watched the trial without flinching and when testifying supported courageously the merit of the work of the Guild and the right to be free to speculate and pursue curiosity. When asked about opinion on my theory, Ad replied honestly that we did not agree on that. But the Court used that against me: 'The fact that not even the trespasser's partner agrees with the sinful delirium confirms the enormity of the deviousness and heresy of the curiosity path that the trespasser has chosen to follow'." "I couldn't even say goodbye to Ad. And there won't be another return journey to Golmak. Nor ship, or long weeks at sea, or precious metals to transport, or festive receptions, titles or riches, fog in the harbor, Ad's smile, hair, hands, eyes, breasts, cock, cunt, sweat, voice, smell. They have only granted me one wish: that the place of exile be the Northern Barbarian potentate of my last trip. But, if in Golmak I had been banned for questioning the literal interpretation of the Holy Ala'm and the consequent symbiosis between ecclesiastical and political powers, in the Barbarian lands of the North I was seen, in the purest and simple, and therefore most brutal way as a perversion. I ended up being banished twice, from both home and the shabby barbarian encampment on the shores of the Great Lakes. I am now roaming deserted landscapes. I brought with me this copy of the Holy Ala'm, my only Reading, my only Writing paper. And a charcoal portrait of Ad. I live like a hermit and, too often, like a madperson. I talk to myself. I argue with myself. I rest motionless for long periods of time, until hunger makes me look for food. Slowly I give up." This seemed to be the last note by Eva g Sirdi. However, the mentioned charcoal portrait of Ad contained, in the back, a brief and final note, certainly written in two different moments. It is, for all intents and purposes, the note that made the member of the medieval Guild of Sages the pioneer of the modern science of human evolution: "Exhausted, I woke up to the vision of a camp. No idea how I got there. A campfire beside me. Shabby tents made of canes. Tiny creatures (thought they were infants) inspect me with curiosity. They laugh. They are disgusted too. They point at my genitals and breasts. They touch them with sticks. And they talked! Don't understand the language, but they talk! The work of Ala'm, praised be They! Then I watched them closer. 'Mutilation', I thought. 'Abnormality', too. Some of the creatures had a penis, a bag for exo-eggs, like lower animals, where the vagina should be. Some had just the vagina, uncovered by a hanging penis, although somewhat hidden by more flesh. These are the aganuruk, they have to be! Agan-uruk. The "divided people". Dimorphic. Dimorphs. Found what I've been looking for! [...]" "[...] Too late. Henceforth I, most advanced representative of the higher stages of evolution and Ala'm's creation, will be just an object of disgust and curiosity, the strange unimorph of the tribe." * This fiction was inspired by the Gethenian theme of The left hand of darkness. It was developed to include not only the aspects of sex, gender, and sexuality, but also the social organization of power and knowledge, and plays with classical portrayals of the tropes of expedition, discovery, otherness (taken farther than cultural difference, including, therefore, questions of evolution). It also plays with time and chronology, picturing Earth society in a future that is an apparent return to our current past; and placing contemporary human bodies as "predecessors" in the evolutionary scale. This allows for the imagination of an altertopia that is simultaneously a fantasy and a "system" that draws inspiration from the archives of ethnography. There is, of course, danger in moving back and forth between fact and fiction, real lives of real humans, and those of imaginary characters. But in imagining bodies that are apparently radically different one ends up realizing the familiarity of the components; desire, sex, and the social organization of them can be imagined as otherwise, as different from our own, especially in what concerns inequality and power. Furthermore, gender disappears, relationality and sociality do not. This kind of social science fiction is necessarily permeated by feminist and queer values and perspectives, not taking for granted neither heteronormativity, nor binarism or the body. Although some ethnographic archives do provide us with a variety of gender and sexuality orders and regimes, and even with the non-centrality of gender in the Western conceptual sense, they do not provide us with the absence of gender.

Ethnography-based Anthropology and Science Fiction can engage in a productive dialogue since both address what is proposed as "altertopias". Utopias, dystopias, and cultural alterity share the possibility of imagining social and cultural organizations different from both those of the authors and those of the readers. These imaginations are intrinsically creative/artistic and political at the same time, and they critique power structures, especially when approached through a feminist stance. Inspired by the literary work of Ursula Le Guin, the article takes this further by experimenting with the inclusion of a fictional piece of "Social science fiction" that itself plays on Le Guin's themes.

Type for PI and non-PI (Lempers & Clarklempers, 1992) (Levy, 2010). Of those children, approximately 1500 were adopted to the United States in 2009 (United States Department of State, Office of Children's Issues). Although the quality of institutions varies throughout the country, Baby Homes in St. Petersburg, where most of the children in the current study resided prior to adoption, are selectively social-emotionally depriving, or Gunnar's (2001) level three institutions, meeting all needs except for typical caregiver-child interactions and stable, long-term relationships with consistent caregivers. Such an environment can be expected to be associated with subsequent relationship problems. Specifically, children receive adequate medical care, sanitation, nutrition, toys, and equipment; however, they experience 60-100 different caregivers during their first 19 months in the institution, and children typically see no caregiver on two consecutive days. Children rarely, if ever, experience warm, sensitive, contingently responsive interactions, and there is little opportunity to form an attachment relationship (The St. Petersburg-USA Orphanage Research Team, 2005Team,, 2008) ). Indeed, many institutional children are classified as having disorganized attachments (Dobrova-Krol et al., 2010;The St. Petersburg-USA Orphanage Research Team, 2008;Vorria, et al., 2003;Zeanah, Smyke, Koga, & Carlson, 2005). Thus, research with these children can more specifically address the impact of early social-emotional deprivation, rather than other factors often associated with institutions, on later outcomes. --- THEORETICAL BASIS FOR INVESTIGATING POST-INSTITUTIONAL OUTCOMES: ATTACHMENT Attachment theory (Ainsworth, Blehar, Waters, & Wall, 1978;Bowlby, 1982) posits a probabilistic view of development, such that early experience, particularly in an attachment relationship, sets a child on a general path that leads to a set of possibilities rather than a predetermined outcome. The attachment relationship is the child's first experience of emotional closeness, and through consistent, responsive interactions with a few adults, children progressively develop "internal working models" of the purpose and structure of relationships that help them form expectations for the future (Bowlby, 1982;Sroufe, Carlson, Levy, & Egeland, 1999). Although children's models and expectations can change with experience, they enter each new relationship with learned dispositions that influence their interpretation of situations and interactions. Thus, the longer the child remains in a maladaptive situation, the more difficult it becomes to return to positive functioning (Sroufe et al., 1999). In the institution, high children-to-caregiver ratios (approximately 5-7: 1 during daytime and 9-14: 1 at night), age-based transitions (approximately two or three transitions by 2 years), high staff turnover, and inconsistent staff schedules (generally 10-24 hours 2-3 times per week, with 52-56 days of vacation per year), create an environment in which forming an attachment is all but impossible (Chisholm, 1998;The St. Petersburg-USA Orphanage Research Team, 2005). The inconsistency and emotional neglect from caregivers may contribute to problems with emotion regulation, including suppressing negative emotions and displaying positive emotions even in stressful situations (Cassidy, 1994;Bakermans-Kranenburg et al., in press). The lack of contingent interactions with a stable caregiver may be related to decreased emotional and behavioral control, including effortful attention regulation and inhibitory control (Bakermans-Kranenburg et al., in press;Dobrova-Krol, Bakermans-Kranenburg, van IJzendoorn, & Juffer, 2010). 1.1.1 Support for an attachment perspective. Attachment to a caregiver and responsive caregiving are theorized to support early behavioral, cognitive, and social development in non-PI children (Ainsworth, et al., 1978); and insecure attachment, especially when disorganized, is related to poorer outcomes in these areas (Carlson, 1998;Shaw, Owens, Vondra, Keenan, & Winslow, 1997;Sroufe et al., 1999). Research has also established a connection between extended institutional experience, as assessed by age at adoption, and behavior problems (e.g., Gunnar, Van Dulmen, & The International Adoption Project Team, 2007;Hawk & McCall, in press;Rutter et al., 2010), executive functioning problems (Bruce, Tarullo, & Gunnar, 2009;Merz & McCall, in press;Pollak et al., in press), and social difficulties (Ames, 1997;Gunnar et al., 2007;Julian, 2010;Rutter, Kreppner, & O'Connor, 2001) in childhood and adolescence. Deficient early social-emotional experience and lack of relationships in the institutions are hypothesized to be especially important to the quality of important relationships (e.g., mother, sibling, and best friend). One study (Vorria, Wolkind, Rutter, Pickles, & Hobsbaum, 1998) found that the lack of responsive, individualized care from a consistent caregiver, not physical deficiencies, resulted in more emotional disturbances and less harmonious, intimate relationships with peers in currently institutionalized children than children raised by their biological parents (non-PI). Studies have also found that longer time spent in the institution is associated with less secure attachment to adoptive mother (Bakermans-Kranenburg et al., in press;Chisholm, 1998;Juffer & Rosenboom, 1997;van den Dries, van IJzendoorn, & Bakermans-Kranenburg, 2009). Unfortunately, relationship quality in older PI children and adolescents has not been adequately addressed, even though relationship quality may be the domain of functioning most related to social-emotional deprivation and lack of attachment relationship. One relevant study found that later adoption was associated with poorer self-reported support from a best friend (Le Mare, Warford, & Fernyhough, 2001). --- IMPORTANCE OF RELATIONSHIPS WITH MOTHERS, SIBLINGS, AND FRIENDS --- Mothers. Although most research focuses on e arly mother-child relationships, the quality of this relationship remains important throughout adolescence. Non-PI children tend to rate their maternal relationships high in reliable alliance, affection, enhancement of worth, instrumental help, intimacy, and satisfaction, but quality decreases in adolescence (Furman & Buhrmester, 1985b, 1992;Lempers & Clarklempers, 1992). Adolescents' perceptions of high quality relationships with mothers have been associated with lower levels of adolescent depression (Aseltine, Gore, & Colten, 1998) and delinquent behaviors (Branstetter, Furman, & Cottrell, 2009;Hair, Moore, Garrett, Kinukawa, Lippman, & Michelson, 2005;Hair, Moore, Garrett, Ling, & Cleveland, 2008;Mason, Cauce, Gonzales, & Hiraga, 1994), higher levels of mental well-being (Armsden & Greenberg, 1987;Hair et al., 2008), and increased friendship quality (Kerns, Klepac, & Cole, 1996;Lieberman, et al., 1999). Very little research has examined mother-adolescent relationships in adoptees. Institutionalization has been associated with poorer mother-child relationships in early childhood, although most young PI children have good relationships with their adoptive mothers (Bakermans-Kranenburg et al., in press;Chisholm, 1998;Juffer & Rosenboom, 1997;van den Dries, et al., 2009). One study found that a similar proportion of PI and non-PI adolescents had positive relationships with their adoptive or biological mothers (Hodges & Tizard, 1989). Conversely, adoptive mothers of non-institutionalized domestic adoptees have reported more disagreements with their adolescents than biological mothers or stepmothers (Lansford, Ceballo, Abbey, & Stewart, 2001). --- Siblings. In studies of non-PI children, sibling relationships are exemplified by high levels of conflict, along with affection, alliance, companionship, intimacy, and nurturance (Furman & Buhrmester, 1985a,b, 1992;Lempers & Clarklempers, 1992). This environment of simultaneous closeness and conflict theoretically allows children to test and to develop social skills necessary for maintaining other relationships (Furman & Buhrmester, 1985a). Specifically, positive sibling relationships are linked to behaviors necessary for high quality friendships, including better conflict resolution strategies (Franco & Levitt, 1998;Recchia & Howe, 2009), higher levels of companionship and recreation with friends (Franco & Levitt, 1998), increased self-disclosure (Howe, Aquan-Assee, Bukowski, Lehoux, & Rinaldi, 2001), increased reciprocal interactions (e.g., companionship, emotional responsiveness, and prosocial activities: Karos, Howe, & Aquan-Assee, 2007), and, for girls, more emotional understanding (Howe et al., 2001). Further, positive sibling relationships have been directly related to increased friendship quality in middle childhood and early adolescence (Franco & Levitt, 1998;Yeh & Lempers, 2004). The only study of PI adolescents' sibling relationships reported more difficulties with siblings for PI children than non-PI comparisons (Hodges & Tizard, 1989). Among noninstitutionalized domestic adoptees, no difference was found in relationship quality between adoptees and their non-adopted siblings and two biological non-adopted siblings (McGue et al., 2007). Behaviorally, PI children adopted as sibling groups have higher ages at adoption but have lower or similar rates of behavior problems compared to single adoptees (Boer, Versluis-den Bieman, & Verhulst, 1994;Van den Oord, Boomsma, & Verhulst, 1994), and similar patterns are found for domestic adoptees (Erich & Leung, 2002;Hegar, 2005). This finding suggests a protective effect of having a sibling because older age at adoption is generally related to more, not fewer, behavior problems (Cederblad et al., 1999;Gunnar et al., 2007;Rutter et al., 2010). Further, social-emotionally deprived PI children may benefit more from additional interactions with siblings than non-PI children because of their limited earlier relationship experience. --- Friends. Approximately 78% of non-PI children in middle childhood have at least one best friend (Parker & Asher, 1993), and the best friend relationship is exemplified by companionship, intimacy, enhancement of worth, and some conflict (Furman & Buhrmester, 1985b;Lempers & Clarklempers, 1992). In adolescence, friendships tend to be high in support, intimacy, and affection, with some decreases in companionship and nurturance with age (Furman & Buhrmester, 1992;Lempers & Clarklempers, 1992). These findings may be qualified by sociometric status, such that high-and average-accepted children have more supportive and intimate friendships and are more satisfied with their friendships than low-accepted children (Parker & Asher, 1993). However, low-accepted children show the greatest variability in friendship quality (Parker & Asher, 1993), suggesting that some low-accepted children have very satisfying friendships, but perhaps with only one or two close friends. Poor friendship quality in non-PI children has been related to loneliness (Nangle, Erdley, Newman, Mason, & Carpenter, 2003), depression (Nangle et al., 2003), high school dropout rates (Parker & Asher, 1987), behavior problems (Hartup, 1995), and criminality (Parker & Asher, 1987). Conversely, high quality friendships are related to school involvement, higher grades, and social competency (Hartup, 1995;Gifford-Smith & Brownell, 2003). Further, children and adolescents who have at least one supportive friend are more involved in school (Berndt & Keefe, 1995), are more popular and socially competent, achieve higher grades, and have fewer behavior problems than those with no supportive friends (Hartup, 1995). PI adolescents who spent more than 24 m onths in a globally depriving Romanian institution reported less support from a close friend than non-PI children, early-adopted noninstitutionalized Romanian controls, and PI children adopted before 24 months (Le Mare et al., 2001). Hodges and Tizard (1989) also found PI children to be less popular and less likely to confide in friends than non-PI comparisons. Indirectly, several studies that use parent-reported measures of children's social problems, which are likely related to friendship quality, find that PI children have more reported social problems (Hoksbergen, Rijk, Van Dijkum, & Laak, 2004;Stams, Juffer, Rispens, & Hoksbergen, 2000;Verhulst, Althaus, & Versluis-Den Bieman, 1990) and disinhibited social behaviors (Bruce, et al., 2009) than non-PI children. Also, later adopted PI children (> 7-24 months at adoption) tend to have more parent-reported social problems than earlier adoptees (Cederblad, et al., 1999;Gunnar et al., 2007;Hawk & McCall, in press;Julian, 2010). These studies suggest that more exposure to institutions with deficient social-emotional relationships, few contingent-responsive interactions, and few opportunities to learn appropriate social skills, is related to more social and friendship problems. --- SIBLING CHARACTERISTICS AS POTENTIAL MODERATORS OF RELATIONSHIP QUALITY Sibling relationships may be especially important contributors to social skills and friendship quality (see above; Franco & Levitt, 1998;Yeh & Lempers, 2004). Certain characteristics of siblings have been associated with enhanced relationship quality in non-PI children; however, essentially nothing is known about these factors among PI children. --- Relative age. Among non-PI children, there are benefits of having an older rather than a younger sibling. Non-PI children with older siblings report greater admiration, intimacy, affection, and prosocial behavior in their sibling relationships than children with younger siblings (Buhrmester & Furman, 1990); however, companionship may be greater with closely spaced younger siblings (Furman & Buhrmester, 1985a). Similarly, although antagonism and conflict are generally high in sibling relationships, these negative interactions decrease with age for children with older siblings but not for those with younger siblings (Buhrmester & Furman, 1990). Karos et al. (2007) found that both the negative correlation between rivalry and reciprocal interactions and the positive correlation between reciprocal interactions and socio-emotional problem solving in the sibling relationship was seen only for children with older siblings. Further, children have reported that parents favor younger siblings (Furman & Buhrmester, 1985a), suggesting that relationship quality with mother may be higher for children with older than younger siblings. Similarly, because high quality sibling relationships are associated with positive friendships and social skills (Recchia et al., 2009;Yeh & Lempers, 2004) and having an older sibling is associated with better sibling relationship quality (Buhrmester & Furman, 1990;Furman & Burhmester, 1985a), it is expected that having an older sibling will be related to positive friendship quality. --- Gender composition. For non-PI children, same-gender siblings may be more beneficial than opposite-gender siblings. Same-gender siblings report higher levels of positivity (females only; Buhrmester & Furman, 1990), warmth and closeness (Furman & Buhrmester, 1985a), companionship (Furman & Buhrmester, 1985b), and intimacy (Furman & Buhrmester, 1985b) than opposite-gender siblings. In fact, Furman and Buhrmester (1985b) described same-sex sibling relationships as friendships, but with higher levels of conflict. The positive effects of same-gender sibling dyads may also be associated with friendship quality (see above), although the association with maternal relationship quality is unclear. --- Adoption status of sibling. The benefits of a sibling who is a biological child of the adoptive parents vs. an adoptee are unknown. For international adoptees (PI and non-PI), having more biological children in the adoptive family was associated with more behavior problems among adoptees, but not for those adopted with a biological sibling (Boer et al., 1994). Those adopted with biological siblings also had fewer problems than single adoptees, regardless of adoptive siblings (Boer et al., 1994;Erich & Leung, 2002). Conversely, biological children, who have not experienced institutional neglect, may be better models for appropriate relationship behaviors than other PI children. --- THE CURRENT STUDY The current study represented the first inquiry of PI children's self-reported relationship quality with mothers, siblings, and best same-sex friends. Children aged 10-17 years who had been adopted from primarily social-emotionally depriving Baby Homes in the Russian Federation completed a self-report relationship quality questionnaire to investigate the following questions: 1. What is the nature of the pattern of important relationships in PI children, and does this pattern differ from that of non-PI children? It was expected that PI children's pattern of relationships with mother, sibling, and best friend would be similar to that of non-PI children, but that PI children would report poorer relationship quality. 2. Does more prolonged social-emotional deprivation early in life relate to the quality of important relationships during later childhood and adolescence? Children adopted at older ages were expected to have poorer quality relationships with mother, sibling, and best friend (Le Mare et al., 2001). Attachment theory and past research on s ocial skills (Julian, 2011) and behavior problems (Hawk & McCall, in press) with this sample suggest relationship quality might be poorer in children adopted after 18 m onths of age, but relationship quality could continue to decrease with accumulated experience in the depriving institution (Sroufe et al., 1999). 3. Do sibling relationship characteristics moderate the association between age at adoption and relationship quality? Children with older siblings and those in same-sex sibling dyads were expected to show a s maller age-at-adoption effect than those with younger and opposite-sex siblings, respectively (Buhrmester & Furman, 1990;Furman & Buhrmester, 1985a). Predictions could not be made about sibling adoption status. --- METHODS AND EXPERIMENTAL DESIGN --- PARTICIPANTS Participants were 91 children between the ages of 10 and 17 years (M = 13.98, SD = 2.27) who were adopted from institutions in the Russian Federation into USA families. The Russian institutions have selective social-emotional deficiencies but adequate physical resources (see above; The St. Petersburg-USA Orphanage Research Team, 2005, 2 008). Because of this distinction, outcomes for these children can more specifically be attributed to early socialemotional deprivation. --- PROCEDURE Parents and children were made aware of the study through an article in the adoption agency's newsletter. They then received the packet of questionnaires containing a letter from the adoption agency director and informed consent materials. Some questionnaires were child self-report, and a separate informed assent for the child was included. Specific instructions were given for the parents not to look at the children's answers, and for the children to complete the questionnaire in private and to seal it for privacy. Participants were offered a modest payment ($15) for completion of the surveys (child response rate = 22%). --- Demographic information. Parents (78% mothers) reported the child's date of birth, date of adoption, and date of testing, which were used to calculate age at adoption and age at assessment. They also reported the child's gender, and the age, gender, and adoption status of all siblings. Children indicated their own age and gender, whether they had a sibling, the age and gender of their sibling, and whether they had a b est friend. Information was crosschecked between informants to ensure accurate data; no discrepancies were found. Because many parents did not report family demographic data, such as family income and parent's education (n = 27) or pre-adoptive child factors, such as birth weight (n = 71), they were not used as covariates in the analyses. Age at adoption was used as a surrogate for time in the institution (r =.69 using Baby Home records; Hawk & McCall, 2011). In previous analyses with this sample, rates of parentreported behavior problems, social skills, and executive functioning were minimal before but much higher after 18 months at adoption (Hawk & McCall, 2010;Julian, 2010;Merz & McCall, in press). Thus, age at adoption is used both as a continuous (M = 13.21, SD = 9.13) and a categorical (0-17 months: n = 71; 18-48 months: n = 20) variable in the current analyses. Due to the small sample size, characteristics of the sibling relationship were divided into three groups. Sibling relative age was defined as no sibling (n = 27), older sibling (n = 25; mean age difference = 5.91, SD = 6.39), or younger sibling (n = 36; mean age difference = 3.22, SD = 2.31). Gender composition was defined as no s ibling (n = 27), same gender (i.e., both boys or both girls; n = 26), or opposite gender (n = 36). Finally, sibling adoption status was defined as no sibling (n = 27), biological child of adoptive parents (n = 19), or adopted child (n = 35). Only four children had a genetically related sibling; therefore, the adopted child category included siblings both genetically related and unrelated to the focal child. --- Network of Relationships Inventory: Social Provisions Version. The Network of Relationships Inventory: Social Provisions Version (NRI: SPV; Furman & Buhrmester, 1985b) is a child self-report questionnaire that addresses relationship quality with significant others. It has previously been used to assess relationships with parents, grandparents, siblings, friends, teachers, and romantic partners, and has been given to children in 3 rd grade through college (Furman & Buhrmester, 1985b;Furman & Buhrmester, 1992). For the current study, children were asked to answer questions based on relationships with their mother, samesex best friend, and sibling. These nominations were self-selected, and if children had more than one best friend or sibling, they were asked to select the one who is most important to them. They could also indicate that they had no best friend (n = 3) or no sibling (n = 27). 1 Confirmatory factor analyses supported the use of these two composites.. Higher scores on these scales reflect poorer relationship quality. Relative Power was not included in analyses because is not included in either of the broader indexes. Although the NRI: SPV is a self-report measure that shows the children's subjective interpretation of the relationship, rather than its objective quality, the children's perception may be the most important aspect of the relationships because this shapes their own behaviors and their interpretation of others' behaviors (Furman, 1996). The NRI: SPV had adequate reliability in the current sample; Cronbach's alpha internal consistencies for all subscale scores were greater than.60 (M =.78), except mother instrumental aid (.30). The composite scores had Cronbach's alphas between.85 ( best friend negative interaction) and.94 (best friend social support). Validity had been examined by comparing scores across informants, with correlations of.34 for Support and.63 for Negative Interactions for scores between best friends (Furman, 1996). In the current sample, validity of the NRI was investigated by comparing composite scores to the parent-reported Internalizing and Externalizing broadband scales of the Child Behavior Checklist (CBCL, Achenbach & Rescorla, 2001). Using both extreme CBCL scores (T > 61) and mean scores, high Internalizing and Externalizing scores were associated with lower Social Support and higher Negative Interaction scores; however, these associations were only significant for mothers (Internalizing and Externalizing; most p <unk>.10) and siblings (Internalizing only; all p <unk>.09). Thus, poorer self-reported relationship quality did appear consistent with parent-reported behaviors. --- SELECTIVE RESPONDING To determine whether the current sample, which consisted only of children whose parents allowed them to complete the surveys, was selective, CBCL behavior problem scores were compared between children with self-report and parent-report data (n = 83) and children of the same age (10-17) with only parent-report data (n = 63). A multivariate analysis of variance (MANOVA) comparing the two groups on CBCL Internalizing, Externalizing, Social, Thought, and Attention problems was not significant, F(5, 128) = 1.39, p =.23, nor were any univariate tests (all p >.10). Thus, it appears that the sample was not selective by parents or children for problem behavior. --- RESULTS --- RELATIONSHIP QUALITY IN THE FULL SAMPLE It is possible that a relationship quality measure created for non-PI children may not be appropriate for PI children. To address this concern, the subscale scores for each relationship (within-subject) and their association with age at assessment (dichotomized 10-13 and 14-17) and gender (both between-subject) were examined in individual repeated measures analyses of variance (ANOVA) with post hoc comparisons between relationship sources. These analyses included only children with data for all three relationships (n = 64). Mean scores for each subscale for the current PI sample and a non-PI sample of Lempers and Clarklempers (1992) are presented in Table 1, and score profiles of PI children are presented in Figure 1. The relationship quality among sources, described below, was similar to those found in the non-PI literature, supporting the appropriateness of the NRI for PI children. analyses included only children with data for all three relationships. In comparisons between PI and Non-PI (Lempers & Clarklempers, 1992) children for each relationship, significance is indicated in non-PI relationship column; analyses included all children with data for each relationship. a p <unk>.05. b p <unk>.01. c p <unk>.001. --- Comparisons between relationships (main effects). Relative to the other referents, PI children rated their mothers as the highest source of overall 1 and Figure 1). --- Non-PI versus PI. Table 1 shows comparisons between the current PI sample and a sample of predominantly white, 11-19-year-old children reared by their biological parents in working-class or middle-class families in rural midwestern towns (Lempers & Clarklempers, 1992) 1). --- Age at assessment and gender. Research with non-PI children suggests that as children grow from middle childhood through adolescence, their perceived social support from mothers and siblings decreases, while their social support from friends increases. Conflict is highest in the sibling relationship, and conflict tends to increase with age for mother relationships but decrease in friendships. Girls also tend to report higher relationship quality scores than boys (Buhrmester & Furman, 1990;Furman & Buhrmester, 1992). To investigate age at assessment and gender effects (between-subject) and their interactions with relationship source (within-subject) in the current sample, repeated measures ANOVAs were calculated with post-hoc contrasts for relationship source interactions. 57) = 6.49, p =.003, a nd Support, F(2, 57) = 5.86, p =.005, g irls reported higher quality friendships than boys, but no di fferences or slightly lower quality in mother and sibling relationships. Girls also reported more Conflict, F(2, 57) = 5.49, p =.007, with mothers than boys (see Table 3). The three-way interactions were not significant. --- AGE AT ADOPTION The primary variable of interest was age at adoption, because it had the possibility of reflecting the influence of the institutional experience. Because age-at-assessment and gender effects were seen in the sample, these variables were included in the following analyses as covariates and possible moderators and, as such, are reported only if they qualify age-at-adoption effects. To investigate the association between age at adoption and relationship quality, regression analyses were performed separately for each relationship source predicting the NRI composite scores with gender, age at assessment, age at adoption, and their interactions. Composite scores were used rather than subscale scores to decrease the total number of analyses and because the subscale scores included only three items each. Variables were entered in blocks to determine explained variance; however, tables report only the final model for clarity. Blocks were entered as follows: 1) Gender, 2) Age at Assessment, 3) Age at Adoption, 4) Gender X Age at Assessment, 5) Gender X Age at Adoption, 6) Age at Adoption X Age at Assessment. --- Form of age at adoption. The first analyses addressed the form of the age at adoption function. Three competing models were tested: a linear model, a dichotomous model (cut-off = 18 months), and a spline regression model (knot = 18 months; testing whether the slope of the regression line jumps or changes at 18 months). The only significant age-at-adoption effect for any relationship, controlling for the other variables, was the interaction of the linear age at adoption variable and gender in the prediction of best friend Negative Interaction, R 2 change =.09, F(1, 77) = 8.33, p =.005. The overall regression was marginally significant, F(6, 78) = 1.99, p =.08, R 2 =.13. For girls, older age at adoption was related to more Negative Interaction, but for boys, older age at adoption was related to less Negative Interaction (see Table 4 and Figure 2). Neither the dichotomous model, F(6, 79) = 0.89, p =.51, nor the spline regression model, F(10, 75) = 1.33, p =.23, were significant. Subsequent analyses used only the linear age-at-adoption variable, and figures depict this linear association --- EXTREME SCORES PI children are often distinguished by extremes rather than average scores (e.g., Hawk & McCall, in press;Julian, 2010;Verhulst, et al., 1990), and extreme scores are more indicative of problematic behaviors rather than fluctuations within the normal range. Because no standardization sample exists that describes scores indicative of NRI "extreme problems" in non-PI children, "extreme good" and "extreme bad" scores were created in two ways using the current data. First, for Social Support, mean scores greater than 4 were considered "good" (M = 4.45) and mean scores less than 2 were considered "bad" (M = 1.46). For Negative Interaction, scores less than 2 w ere "good" (M = 1.56) and greater than 4 were "bad" (M = 4.60). This provided a description of the qualitatively highest and lowest scores. Second, z-scores were created, and, for Social Support, children with z-scores in the highest 15% were considered "good" (M = 2.47), whereas those in the lowest 15% were considered "bad" (M = 4.65). For Negative Interaction, the lowest 15% were "good" (M = 1.73) and the highest 15% were "bad" (M = 3.72). The 15% cut-off is similar to the definition of "extreme problems" on the CBCL. Analyses revealed similar outcomes for both types of extreme scores; thus, only the qualitative score outcomes are reported. Table 5 presents the percentage of children with "good" and "bad" scores based on age at adoption, age at assessment, and gender, although the latter two are not discussed in text. Chi-squared analyses compared earlier vs. later adoptees, younger vs. older children, and boys vs. girls for each relationship quality score. A higher percentage of later adoptees reported "good", <unk> 2 = 4.98, p =.03, and "bad", <unk> 2 = 5.28, p =.02, Social Support from their siblings than earlier adoptees. Later adoptees also had a higher percentage of "bad" Negative Interaction, <unk> 2 = 4.98, p =.03, with their siblings (see Table 5). Although age at adoption was dichotomized for simplicity, these percentages also increased linearly (e.g., percentage of children with extremely low sibling social support by age at adoption: 0-5 months = 0%; 6-11 months = 3%; 12-17 months = 10%; 18-23 months = 20%; 24-48 months = 25%). Tests of age at adoption X age at assessment and age at adoption X gender revealed no significant effects. Note. Comparisons are within variables (e.g., 0-17 vs. 18-48 months). a p <unk>.05. b p <unk>.01. --- SIBLING CHARACTERISTICS The following analyses used linear regression to examine whether characteristics of siblings (relative age, gender composition, adoption status) were associated with relationship quality both directly and through interactions with age at adoption. Thus, main effects of sibling characteristics are reported if there is no age-at-adoption interaction. Variables were entered in blocks as follows: 1) Gender, 2) Sibling Characteristic, 3) Age at Assessment, 4) Age at Adoption, 5) Age at Assessment X Sibling Characteristic, 6) Age at Adoption X Sibling Characteristic. Tables show only the final model for simplicity. --- Relative Age. Gender --- Mother. The overall regression predicting Social Support was significant, F(9, 81) = 3.45, p =.001, R 2 =.28. The R 2 change was significant for the interaction of age at adoption and sibling relative age

Orphanages in the Russian Federation are deficient primarily in social-emotional relationships (St. Petersburg-USA Orphanage Research Team, 2005). Children who spend their first months or years of life in orphanages have limited opportunities to form relationships and to develop proper social skills. This early experience has been associated with many problematic behavioral outcomes (Gunnar et al., 2007;Rutter et al., 2010). However, researchers have not frequently examined relationship quality of post-institutional (PI) adoptees, nor have they examined aspects of the adoptive family that might moderate institutionalization effects. The purpose of this study was to examine the self-reported relationship quality (using the Network of Relationships Inventory: Social Provisions Version; Furman & Buhrmester, 1985) of 10-17-yearold children adopted into the USA from Russian orphanages and to determine whether sibling characteristics (relative age, gender composition, sibling adoption status) moderate institutionalization effects. Older age at adoption was related to poorer friendship and sibling, but not mother, relationship quality. Older siblings and same-sex siblings buffered children from this negative age-at-adoption association.

gender revealed no significant effects. Note. Comparisons are within variables (e.g., 0-17 vs. 18-48 months). a p <unk>.05. b p <unk>.01. --- SIBLING CHARACTERISTICS The following analyses used linear regression to examine whether characteristics of siblings (relative age, gender composition, adoption status) were associated with relationship quality both directly and through interactions with age at adoption. Thus, main effects of sibling characteristics are reported if there is no age-at-adoption interaction. Variables were entered in blocks as follows: 1) Gender, 2) Sibling Characteristic, 3) Age at Assessment, 4) Age at Adoption, 5) Age at Assessment X Sibling Characteristic, 6) Age at Adoption X Sibling Characteristic. Tables show only the final model for simplicity. --- Relative Age. Gender --- Mother. The overall regression predicting Social Support was significant, F(9, 81) = 3.45, p =.001, R 2 =.28. The R 2 change was significant for the interaction of age at adoption and sibling relative age, R 2 =.08, F(2, 81) = 4.41, p =.02. F or children with no siblings or younger siblings, age at adoption was not associated with Social Support; but for those with older siblings, older ages at adoption were associated with more mother Social Support, holding all else constant (see Table 6 and Figure 3). The overall regression for Negative Interaction was not significant, F(9, 81) = 1.05, p =.41, R 2 =.10. --- Sibling Adoption Status. Regression analyses were conducted predicting relationship quality with gender, age at assessment, age at adoption, and sibling adoption status (no sibling, biological child of adoptive parents, adopted child; reference = no sibling). Although the regressions predicting mother, F(9, 71) = 2.29, p =.03, R 2 =.23, sibling, F(6, 49) = 2.3, p =.049, R 2 =.22, and friend, F(9, 68) = 2.03, p =.049, R 2 =.21, Social Support were significant, neither age at adoption nor sibling adoption status were significant predictors (see Table 7). The regressions predicting Negative Interaction were not significant for mother, F(9, 71) = 0.98, p =.46, sibling, F(6, 48) = 0.38, p =.89, or friend, F(9, 68) = 0.72, p =.69 (see Table 7). --- Mother. The overall regression predicting mother Social Support was significant, F(9, 79) = 2.96, p =.004, R 2 =.25. H owever, neither age at adoption nor gender composition was a significant predictor (see Table 8). The overall regression was not significant for mother Negative Interaction, F(9, 79) = 0.93, p =.51 (see Table 8). --- Sibling. The regression equation predicting sibling Social Support was significant, F(6, 57) = 4.14, p =.002, R 2 =.30. The change in R 2 increased significantly for sibling gender composition, R 2 change =.10, F(1, 61) = 6.79, p =.01, and the interaction of sibling gender composition and age at adoption, R 2 change =.10, F(1, 57) = 7.88, p =.007. For children with same-gender siblings, age at adoption was not associated with Social Support; but for those with an opposite-gender sibling, older age at adoption was associated with less Social Support (see Table 8 and Figure 4). Although the overall regression predicting sibling Negative Interaction was not significant, F(6, 56) = 1.15, p =.35, R 2 =.11, the R 2 change was significant for the interaction of sibling gender composition and age at adoption, R 2 change =.10, F(1, 56) = 6.50, p =.01. When the sibling was the opposite gender, older age at adoption was associated with more Negative Interaction, As a potential explanation, PI children tend to be conforming and accommodating, at least at younger ages, which may support the creation of relationships; and they may have a greater desire for relationships, given their early social-emotional deprivation. The combination of these factors may allow them to develop higher quality relationships than non-PI children, although this may be limited to earlier adoptees (see below). Because high quality relationships are related to positive social, academic, and psychological outcomes in non-PI children (e.g., Hartup, 1995;NICHD Early Child Care Research Network, 2009;Yeh & Lempers, 2004), these positive relationships may similarly buffer PI children from more problems. Although PI children exhibit higher than expected rates of problems, most are within normal limits (Gunnar et al., 2007). Possibly, being reared in an advantaged home with parents and siblings with good social skills and having friends may minimize problems in some PI children that otherwise might occur. Given these findings, future research on PI children should consider relationship quality as a potential moderator of long-term developmental outcomes. --- AGE AT ADOPTION Contrary to findings with parent-reported executive function, behavior problems, and social skills in this sample (Hawk & McCall, in press;Julian, 2010;Merz & McCall, in press), an ageat-adoption step-function was not found for relationship quality. Instead, increasing age at adoption was linearly related to increasingly poorer relationship quality (e.g., percentage of children with extremely low sibling social support by age at adoption: 0-5 months = 0%; 6-11 months = 3%; 12-17 months = 10%; 18-23 months = 20%; 24-48 months = 25%). Attachment dyads. Same-gender siblings may spend more time together, especially in younger childhood (Furman & Buhrmester, 1985a), resulting directly in better relationships. Finally, sibling adoption status was not associated with relationship quality either as a main effect or as a moderator of age at adoption. This finding bodes well for PI children because many adoptive parents are unable to conceive. This suggests that the beneficial effects of having a sibling, especially an older and/or same-gender sibling, are not dependent on t he adoption status of that sibling. --- LIMITATIONS The current findings are important because they are the first examination of perceived relationship quality in PI children and the moderating effects of sibling characteristics; however, some limitations should be acknowledged. First, the response rate (22%) was smaller than expected, resulting in a smaller sample and decreased power. This is a subsample of one of the largest populations of adoptive parents of PI children in the literature (total N = 1117; current wave N = 417), but the children's response rate could be smaller than parents' because both parent and child needed to cooperate. However, the children who did participate were not significantly different than those who did not, at least on m easures of behavior problems, suggesting that the current sample was not biased toward "good" children or children without behavior problems. Second, relationship quality consisted of PI children's perceptions, which could have reflected desired quality rather than objective quality; the perspectives of the relationship partner were not collected. The higher relationship quality may reflect a tendency for PI children to perceive their relationships as more idealized than non-PI children. However, the children's perception may be the best predictor of outcomes because children act and interpret others' actions based on perceptions (Furman, 1996). Further, the self-report measure was associated with parent-reported behavior problems. Finally, the current study is cross-sectional, so inferences cannot be made about age changes. Similarly, it could not be determined whether sibling characteristics were causal factors in relationship outcomes or the result of a third variable. Important events and characteristics of the children's lives between adoption and the current data collection were unknown, including aspects of the home life, how much time they spent with siblings when younger, familial social skills, and stressful events (e.g., death of a family member), all of which could have impacted relationship quality. Although these factors will be important to examine in future studies, the current study is unique in that it was one of the first to examine PI relationship quality and the first to suggest that sibling characteristics may moderate institutionalization effects. --- IMPLICATIONS The current findings suggest that resiliency following early deprivation is more possible than previously believed, at least in the domain of relationship quality. Given the nurturing and advantaged adoptive family environment, children with little to no r elationship experience or socio-emotional interactions are able to form high quality relationships, especially with adoptive mothers, even when they were adopted later than 18 months. Thus, the window for learning how to have relationships seems to be larger than originally believed, at least according to child selfreport. --- Sibling. The overall regression predicting sibling Social Support was significant, F(6, 57) = 3.38, p =.002, R 2 =.25. B ecause only children with siblings were included in these analyses, sibling relative age was converted to a dichotomous variable (older = 1). The change in R 2 was significant for the interaction of age at adoption and sibling relative age, R 2 change =.08, F(1, 57) = 5.70, p =.02. For children with younger siblings, age at adoption was not related to sibling Social Support; but for children with older siblings, older age at adoption was associated with more Social Support (see Table 6 a nd Figure 3). The overall regression for sibling Negative Interaction was not significant, F(6, 56) = 0.91, p =.49, R 2 =.09; however, the R 2 change was significant for sibling relative age, R 2 change=.08, F(1, 60) = 4.78, p =.03. Children with older siblings reported less Negative Interaction than those with younger siblings, t(56) = -2.25, p =.03 (see Table 6). --- Best Friend. The overall regression predicting friend Social Support was significant, F(9, 76) = 2.27, p =.03, R 2 =.21. T he interaction between age at adoption and sibling relative age, specifically older sibling, was significant; however, the R 2 change was not, F(2, 76) = 2.49, p =.09 (see Table 6). The overall regression predicting friend Negative Interaction was not significant, F(9, 76) = 1.50, p =.17, R 2 =.15, but the R 2 change for the interaction between age at adoption and sibling relative age was significant, R 2 change =.11, F(2, 76) = 4.68, p =.01. F or children with no sibling, older age at adoption was associated with more friend Negative Interaction; however, for children with younger or older siblings older age at adoption was associated with less Negative Interaction (see Table 6 and Figure 3). whereas when the sibling was the same gender, age at adoption was not associated with Negative Interaction (see Table 8 and Figure 4). --- Best Friend. The overall regression predicting friend Social Support was significant, F(9, 76) = 3.10, p =.003, R 2 =.27. The change in R 2 increased significantly for the interaction of gender composition and age at adoption, R 2 change =.06, F(2, 76) = 3.08, p =.05. H olding all else constant, when children had no sibling or a same-gender sibling, age at adoption was not associated with friend Social Support; however, when they had an opposite-gender sibling, older age at adoption was associated with more Social Support (see Table 8 andFigure 4). For friend Negative Interaction, the overall regression equation, F(9, 76) = 1.14, p =.35, R 2 =.12, was not significant (see Table 8). --- DISCUSSION PI children reported high levels of Social Support and low levels of Negative Interactions in their relationships with mothers, siblings, and best friends. Older ages at adoption were associated with poorer relationship quality with siblings and best friends, but not mothers. Older siblings and same-gender siblings buffered PI children from these negative age-at-adoption effects, especially for relationships with siblings. --- PI RELATIONSHIP QUALITY In the non-PI literature, 22% of children (aged 9-13) report having no best friend (Parker & Asher, 1993); however, in the current PI sample, only 3 children (3%) endorsed having no best friend. Further, PI children reported a similar pattern of relationships for subscale scores across the three referents as non-PI children, except that, contrary to expectations, they generally reported better, not worse, relationship quality than non-PI children (Lempers & Clarklempers, 1992). Spending time in a Baby Home, especially the lack of response-contingent interactions with a stable attachment figure and inadequate opportunities to develop working models of relationships, was expected to be associated with poorer, not better, relationship quality for PI children. theory suggests a rather specific age period during which most parent-reared infants develop a relationship with their caregiver (Bowlby, 1982), but the current data do not support such a specific age period. In contrast, Sroufe et al. (1999) Eventually, prolonged institutional exposure does seem to have a corroding effect on relationships, but the effects were not pervasive or consistent, and they were primarily with siblings and best friends, not with mothers. The mother relationship, which was not associated with age at adoption, was expected to have the strongest age-at-adoption effect because of the lack of maternal attachments in the Baby Homes. A potential explanation is that adoptive mothers are more invested in their children than siblings or friends, such that they are more willing to work on having a high quality relationship despite possible difficulties. Similarly, PI children may put more emphasis on the mother relationship because of their lack of attachment in the institution, and there may be a biological drive for a mother-child relationship over siblings and friends. --- SIBLING CHARACTERISTICS: HOW SIBLINGS BUFFER INSTITUTIONALIZATION EFFECTS The presence of a sibling buffered children from the negative effects of age at adoption on best friend Negative Interactions. Older ages at adoption were related to more friend Negative Interactions for children with no sibling, but not for those with a sibling. As the non-PI literature suggests, children with siblings may learn the skills necessary for high quality friendships, especially in terms of conflict resolution, within their sibling relationships (Franco & Levitt, 1998;Furman & Buhrmester, 1985a;Recchia & Howe, 2009). This practice may be particularly important for later adoptees, who spent a longer time without adequate relationship experiences and who had fewer opportunities to learn social skills early in life. Older siblings also muted the association between older ages at adoption and poorer relationship quality, consistent with the non-PI literature (Buhrmester & Furman, 1990;Furman & Buhrmester, 1985a;Karos et al., 2007). Older siblings may provide a more nurturing environment and be better models than younger siblings. PI children may learn more from older siblings, who are mentally and socially advanced and who may set the ground rules and tone of the sibling relationship, which could be more important for later adoptees, who are more likely to have difficulties. Such learning can occur sooner after adoption when the sibling is older rather than younger than the PI child. Similarly, PI children with same-gender siblings had higher quality sibling relationships than those with opposite-gender siblings, consistent with the non-PI literature (Buhrmester & Furman, 1990;Furman & Buhrmester, 1985a, 1985b). In the current sample, this was a buffering effect, in which children with opposite-gender siblings reported an association between older age at adoption and poorer sibling relationships, but this was not the case for same-gender sibling Further, the sibling relationship seems to promote resiliency in PI children, a topic that had not been examined previously. PI children with siblings presumably have more experience with interactions with children similar to their own age than those without siblings. Older siblings and same-sex siblings may spend even more time together, resulting in more opportunities to practice relationship skills and to develop an understanding of relationships.

Orphanages in the Russian Federation are deficient primarily in social-emotional relationships (St. Petersburg-USA Orphanage Research Team, 2005). Children who spend their first months or years of life in orphanages have limited opportunities to form relationships and to develop proper social skills. This early experience has been associated with many problematic behavioral outcomes (Gunnar et al., 2007;Rutter et al., 2010). However, researchers have not frequently examined relationship quality of post-institutional (PI) adoptees, nor have they examined aspects of the adoptive family that might moderate institutionalization effects. The purpose of this study was to examine the self-reported relationship quality (using the Network of Relationships Inventory: Social Provisions Version; Furman & Buhrmester, 1985) of 10-17-yearold children adopted into the USA from Russian orphanages and to determine whether sibling characteristics (relative age, gender composition, sibling adoption status) moderate institutionalization effects. Older age at adoption was related to poorer friendship and sibling, but not mother, relationship quality. Older siblings and same-sex siblings buffered children from this negative age-at-adoption association.

Background Worldwide, more than 11 million people are imprisoned at any given time, and more than 30 million people move through the prison system annually [1,2]. In Canada, there are more than 250,000 adult admissions and 14,000 youth admissions each year to correctional facilities [3,4]. On any given day, there are about 40,000 adults and youths in correctional facilities [5][6][7]. In Canada, jurisdiction over correctional facilities is shared between the federal, provincial, and territorial governments. Admission to a correctional facility prior to sentencing is called remand, and persons in remand are considered detained rather than incarcerated. Persons in remand and persons who are sentenced to less than 2 years are detained or incarcerated in provincial or territorial facilities, and those who are sentenced to 2 years or longer are incarcerated in federal facilities. International data suggest that people who experience detention or incarceration have poor health compared with the general population, as indicated by data on the prevalence of mental illness, infectious diseases, chronic diseases, and mortality [8]. Detention and incarceration may serve as a unique opportunity to provide health care, to initiate programs to improve health, and to link persons with appropriate services on release. Such interventions could improve the health of people who experience detention and incarceration, and also decrease health care costs [9], improve health in the general population [9][10][11][12][13][14], improve public safety [9], and decrease re-incarceration [9,15,16]. Decisions regarding priorities for research, programs and policies should be informed by Canadian data on the burden of disease and interventions in this population [17]. We conducted a scoping review to describe the extent and type of quantitative health status research conducted between 1993 and 2014 on people who have experienced detention or incarceration in correctional facilities in Canada. In summarizing these data, we aimed to identify areas that have been well defined and gaps in evidence that we can use to inform future research. --- Methods We conducted the scoping review [18,19] according to a protocol that we defined a priori (available from authors on request). --- Search strategy We searched Medline, PsycINFO, Embase, the Cochrane Library, Social Sciences Abstracts, Social Services Abstracts, Sociological Abstracts, CINAHL, Criminal Justice Abstracts, ERIC, Proquest Criminal Justice, Proquest Dissertations and Theses, Proquest Dissertations and Theses: UK and Ireland, Web of Science, and Scopus in April 2014 (see Additional file 1 for search strategy). We reviewed reference lists of included studies and relevant reviews. We did not use any language restrictions, though we used only English language search terms. We searched websites of relevant organizations, specifically the Correctional Service of Canada, Statistics Canada, the Office of the Correctional Investigator, Public Safety Canada, the provincial and territorial Ministries responsible for correctional facilities, PASAN, The John Howard Society of Canada, the Canadian Association of Elizabeth Fry Societies, and the Canadian HIV/AIDS Legal Network. We also consulted with knowledgeable persons in some of these organizations. --- Study selection and data extraction Population We included studies of adults and adolescents who had been detained or incarcerated in a prison or jail in Canada, whether they were remanded or sentenced, and whether the study focused on the period prior to, during, or subsequent to detention or incarceration. We included studies that included other populations if the studies presented stratified results for persons who met this population criterion. We excluded studies that did not specify that participants had been in detention or incarcerated, e.g. studies of "offenders" or "forensic" populations that did not specify a history of detention or incarceration. --- Study period Studies were eligible if they reported data on health from 1993 to 2014. We chose 1993 as the start date for our period of eligibility because we wanted to capture recent data and we hypothesized that a reasonable number of studies would have been conducted after this date. In addition, this date follows the enactment of the federal Corrections and Conditional Release Act in 1992 [20], which may have affected the health care services provided in federal correctional facilities. We included studies that did not specify the dates on which they were conducted if they were published in or after 1997, which assumes a maximum four-year lag time from conducting a study to publication. --- Study types We included experimental studies (i.e. randomized controlled trials, quasi-randomized controlled trials, and non-randomized controlled trials), quasi-experimental studies (e.g. controlled before-after studies, interrupted time series studies), and observational studies (i.e. cohort studies, case-control studies, cross-sectional studies). We included reports of administrative data as well as studies that collected primary data. --- Outcomes of interest We included indicators of health as defined by the Canadian Institutes for Health Information [21], including indicators of health status and health system performance, and social determinants of health as per the Public Health Agency of Canada [22]. Since we were interested in defining health status, we included studies that provided absolute measures of health (e.g. percent or number of persons with a certain condition or behaviour) and not studies that specified only relative measures of health (e.g. relative risks, odds ratios, etc.). For some determinants of health, such as gender, race/ethnicity, employment and education status, we included studies only if they also specified other health status data or if they summarized these data for the whole source population of interest. We included only studies that reported individual-level data (e.g. a study that reported only the total proportion of urinalysis tests that were positive in an institution but did not specify the number or percent of persons who had positive tests would be excluded). --- Review procedures Two reviewers (FGK and AS) independently reviewed titles and abstracts for eligibility, and any disagreements were resolved by discussion. For full article review, we conducted a pilot to ensure a high level of agreement regarding eligibility. One reviewer (FGK or AS) then reviewed each full article to assess eligibility, with discussions regarding any decisions that were not clear. --- Data extraction For eligible articles, one reviewer (FGK or AS) extracted data using a data extraction form that we developed, piloted, and modified. We extracted data on study context, populations included, design, outcomes, and results. --- Analysis We planned a priori to summarize data based on study design, whether each study was conducted in the federal or provincial/territorial system, with females or males, and with youth or adults, in which provinces or territories the study was conducted, and the types of outcomes included. We classified observational studies as cross-sectional if the data were collected at a single point in time and no intervention occurred, and longitudinal if the data were collected over a period of time and no intervention occurred. We categorized interventional studies as per the Cochrane Handbook classification (Table 13.2) [23]. We classified a study as involving youth if the record specified that youth or adolescents were involved or that the study was conducted in a facility for youth or adolescents or with persons younger than 20, and as involving adults if the record specified that the study was conducted with adults or in a facility for adults, or if persons included were aged 17 and older. We categorized outcomes and behaviours into one of the following categories: death, chronic disease, communicable disease, mental health excluding substance use outcomes (using DSM-V categories when appropriate [24]), substance use, injury, sexual and reproductive health, health system, and social determinants of health. To identify data from studies that were reported in multiple publications, we sorted and compared extracted data on the basis of author name, study location, dates, and sample size. For the purposes of summarizing data, we considered data described in multiple publications as a single study if the sample data matched. --- Results As shown in Figure 1, we identified 2560 records: 2419 through database searches, 34 from reference lists and sources known to the authors, and 107 on websites. After eliminating duplicates, there were 2239 records remaining. Of these, 515 were eligible for full review. We were unable to retrieve 1 article [25]. Of the remaining 514 full articles, 8 were duplicates that had not been identified previously, and 219 articles were eligible for inclusion. These 219 articles represent 194 unique studies (see Additional file 1). Key characteristics of included studies are summarized in Table 1. Seventy-five studies were conducted in multiple provinces, of which 92.0% (n = 69) were conducted in federal correctional facilities only. One third of studies were conducted in British Columbia, Ontario, or Quebec, and in almost ten percent of studies, the geographical location was not specified. More than 60% of studies were conducted in only federal facilities. Over ninety percent of studies were conducted with persons while detained or incarcerated, and collected data only from the period while they were detained or incarcerated. Eighty-six studies included only men, and in another 35 studies, more than two thirds of participants were men. More than three quarters of studies included only adults. Seventy-one point six percent of studies were cross-sectional, 20.6% were longitudinal, and 7.7% were interventional. The majority of studies (N = 119) were conducted with the general inmate population, with some exclusions based on concerns about language, literacy, safety, or mental health. Other populations studied include persons who were participating in specific programs in correctional facilities (N = 17), persons convicted of sexual offenses (N = 12), persons who used drugs (N = 5), Aboriginal persons (N = 5), and persons with a history of self-injury (N = 5). One quarter of studies (N = 49) included less than 100 participants, 17 of which were conducted in the general population. Sixty-four studies used only data from administrative databases or file reviews, i.e. the researchers did not collect primary data, most of which (n = 58) were conducted only in federal facilities. Administrative For 65 studies, the dates when the study was conducted were not clearly specified. Dividing the period under study (1993 to 2014) into two equal intervals, it was possible to ascertain the period when the study was conducted for 154 studies based on the dates of publication as well as the dates when the study was conducted as reported in articles; data were collected before 2004 in 80 studies, between 2004 and 2014 in 51 studies, and spanning both periods for 23 studies. Table 2 shows the number of studies that reported outcomes in different health status categories. A large number of studies reported mental health (N = 99), substance use (N = 86), and social determinants of health (N = 80) outcomes, while comparatively few studies reported communicable disease (N = 35), injury (N = 30), sexual and reproductive health (N = 29), and in particular chronic disease (N = 6) outcomes. Regarding specific outcomes, a large number of studies reported data on --- Discussion This review identified 219 publications representing 194 studies that were conducted from 1993 to 2014 with people who experienced detention or incarceration in correctional facilities in Canada. The majority of studies were conducted with persons during detention or incarceration, with persons in federal facilities, and with only men. The greatest number of studies presented mental health, substance use and social determinant of health outcomes, and few studies reported chronic disease, sexual and reproductive health, and injury outcomes. Notably, the great majority of people who experience detention or incarceration in Canada have short sentences or are in remand and therefore serve their time in provincial or territorial facilities, with less than 5% of admissions to federal facilities [3,4,26]. This review found that most research on health status conducted from 1993 to 2014 has focused on persons in federal facilities. This may be due to longer periods of incarceration in federal facilities that provide more time to assess health and conduct research, a greater focus on rehabilitation (including health) in the federal system, greater accessibility of administrative and health data in the federal system compared to provincial facilities, or the challenges of following persons post-release who are not under community supervision. More clearly defining health and intervening to improve health in persons who experience detention or incarceration in provincial and territorial facilities could have a relatively large health impact on this population, with potential ripple effects on family and community health, public safety, and costs to society of health care and criminal justice system involvement [9]. Considering the large number of people who experience detention and incarceration each year in Canada and the fact that the State has a clear obligation to provide health care during detention and incarceration, the number of studies identified in this review is small. There is a particular lack of data in this population for important outcomes such as chronic diseases, sexual and reproductive health, and injury, some of which may be amenable to primary, secondary, and tertiary prevention interventions [27]. The paucity of evidence identified could reflect a lack of collection of health data (whether routinely or for specific projects), a failure to analyse collected data or to disseminate collected data, or limitations in our search strategy. It may also be, in part, a downstream effect of the lack of dedicated funding in Canada for research focused on prison populations, in contrast with the USA [28]. Population health status data should inform decisions about how to focus limited resources in correctional facilities and after release, and data on the general population of persons who experience detention and incarceration could be used to determine the population burden and to estimate and compare the impact of proposed interventions. Instead of multiple independent small studies of specific outcomes, researchers and health administrators should use strategies that are more efficient and provide a more complete picture. Existing processes at intake or routine evaluation could be optimized by identifying which data should be routinely collected [29], standardizing questions or measurements across jurisdictions including levels of government, and implementing or improving the use of electronic databases [30]. A periodic population --- Childhood abuse 25 Other adverse childhood experiences 16 Child welfare involvement 9 Out of home placement 4 --- Residential school involvement 4 Witnessed domestic violence 9 Family and social connectedness 18 Coping skills 4 Locus of control 3 Self-esteem 6 Intelligence 2 Literacy 1 health survey could provide a cross-section of health data across persons in detention in Canada, potentially including physical measurements and biological sampling [31]; similar comprehensive health surveys have been conducted across correctional facilities in the USA [32,33] and in federal facilities in Canada [29,[34][35][36][37][38][39][40]. Further, with appropriate measures in place to ensure informed consent and privacy, data from administrative sources and from periodic surveys could be linked to external administrative data sources to look at health status in the community before and after detention or incarceration, including vital statistics registries, health services utilization data, and social services utilization data [41,42]. There are several potential limitations to this review. We included only studies that reported absolute data on health status, given our interest in understanding the quantitative burden of disease, which means that we have excluded some studies that reported health outcomes, and we did not include qualitative data. While we endeavoured to optimize our search strategy, we may have missed relevant studies, as noted above, especially studies that were not published in the peer-reviewed literature. We have described our search strategy in detail and provided our search terms in an Additional file 1 for transparency and reproducibility. Though we aimed to capture studies conducted from 1993 through 2014, our study would not have captured most studies conducted in the past few years, given the typical lag in time from data collection to study publication. Regarding study procedures, only one author reviewed most full articles and extracted data from eligible articles, which may have led to errors in determination of eligibility or in the data presented. We attempted to minimize errors by discussing and defining in detail the eligibility criteria, piloting our review and data extraction process to achieve a high level of consistency and accuracy, and checking extracted data. More work is required to improve our knowledge about the health of persons who experience detention and incarceration in Canada and to facilitate the application of health status data. Together with key stakeholders, including persons with a history of detention or incarceration [43], provincial, territorial and federal governments should consider which health status data are required for action to improve health, health care and security, and then ensure that these data needs are reflected in data collection at intake and other routine evaluations, as well as in surveillance programs, surveys, and other research initiatives. Consideration should be given to the analysis and dissemination of collected data to optimize their reach and impact. Persons conducting research and making decisions about health initiatives need to review both published and gray literature to inform their work and to minimize the duplication of research efforts [17]. Finally, as data emerge on various aspects of health in this population and on effective interventions, we should iteratively assess and define research priorities for improving health. --- Conclusions Health status data are limited for persons who experience detention and incarceration in Canada. Data are lacking on chronic disease, injury and sexual and reproductive health outcomes, and for persons in provincial facilities and after release. Further research should be done to elucidate health status in this population, and research should be streamlined to improve efficiency. Consideration should be given to which data are required for action to improve health and health care in this population, and efforts should be made to ensure that knowledge is disseminated to decision makers and other key stakeholders. --- Additional file Additional file 1: Search strategy and reference list for identified studies. --- Competing interests The authors declare that they have no competing interests. --- Authors' contributions FGK led the study, developed the protocol, conducted the searches, reviewed records and articles for eligibility, extracted data, and wrote the article. AS reviewed records and articles for eligibility and extracted data. All authors contributed to the study design and to revising the manuscript, and approved the final manuscript.

Background: We conducted a scoping review to define the extent and type of quantitative health status research conducted from 1993 to 2014 with people who have experienced detention or incarceration in correctional facilities in Canada. Methods: We searched 15 databases, reviewed reference lists and relevant websites, and consulted with key stakeholders to identify eligible studies. We reviewed records for eligibility and extracted relevant data from eligible articles. Results: We identified 194 studies that were eligible for inclusion. Most studies were conducted with males and with persons in federal facilities, and focused on mental health, substance use, and social determinant of health outcomes. Conclusions: Health status data are limited for several outcomes, such as chronic disease, injury and sexual and reproductive health, and for persons in provincial facilities and post-release. Efforts should be made to improve data collection and knowledge dissemination, so that relevant data can be used more effectively to improve health and health care in this population.

processes (21st Century Nanotechnology Research and Development Act 2003;European Commission 2004, 2007;Netherlands Organisation for Scientific Research 2008). Political concern for the societal impact of science and technology is in itself nothing new (cf. Roosevelt 1936). What distinguishes recent policies is a widespread interest in socio-technical integration at the'midstream': 'co-operative' or 'interdisciplinary' research that targets early stage R&D decisions, as opposed to 'upstream' funding or 'downstream' regulatory decisions (Fisher et al. 2006). The European Commission for instance aims to: ''encourage actors in their own disciplines and fields to participate in developing Science in Society perspectives from the very beginning of the conception of their activities'' (European Commission 2007, p. 6). While these mandates mark a political interest in interdisciplinary research efforts to integrate social and ethical concerns at early stages of R&D, the appropriate means by which such integration is to occur is still open to experimentation. The recently developed framework of midstream modulation (MM) opens one potential avenue for interdisciplinary collaboration in the research laboratory.1 Two 'laboratory engagement studies' (Fisher 2007) have applied this framework to address the question of social responsibility in research practices, focusing on researchers' critical reflections on the broader socio-ethical context of their work. These studies sought to gauge to what extent MM could help render more visible the broader context of laboratory research, and whether research participants considered critical reflection on this broader context to be relevant. --- Engaging Researchers with the Socio-Ethical Context of Their Work Contrary to the 'neutrality view' of social responsibility-the notion that the social responsibility of researchers is exhausted by the disinterested pursuit of scientific knowledge-scholars have argued that the social responsibility of researchers should include critical reflection on the socio-ethical context of their work (Verhoog 1980). This normative stance reflects recent observations in ethical and normative scholarship (cf. Douglas 2009), including engineering ethics and the ethics of science and technology. Several engineering ethicists have argued for the early assessment of moral issues in technological design by direct involvement of scientists and engineers. Van de Poel and Van Gorp have similarly argued that ''designing engineers have a moral duty to reflect on the ethically relevant choices they make during the design process''(2006, p. 335). While laboratory science differs in many ways from engineering, similar challenges have been voiced in relation to laboratory science. According to Ziman, ''the transformation of science into a new type of social institution'' requires that the ethical dimensions of research should become part of the 'ethos' of science (1998, p. 1813). Accordingly, various scholars have suggested new multidisciplinary engagements in light of the radical ethical challenges posed by new and emerging science and technology (Herkert 2009, personal communication;Khushf 2006;Moor 2005;Schuurbiers et al. 2009b). If ethical and normative scholarship has established a moral imperative for, and a general vision towards, integrating such reflection into research, it has been less clear on how to implement this vision. Theoretically established claims that scientists and engineers should reflect on the normative dimensions of their work do not in themselves enforce or encourage such reflection. Indeed, policy calls for ethical reflection may have at best a tangential effect on research practices because researchers generally perceive the broader socio-ethical context of research as peripheral to their work (Guston 2000;Rappert 2007;Schuurbiers et al. 2009a). The question of implementation can thus stymie broad normative commitments to ethical reflection in research practice. The studies presented here sought to tackle this challenge by supplementing the descriptive techniques of MM with the explicit normative commitment of an 'embedded ethicist'. While MM is more attuned to raising'reflexive awareness' among R&D practitioners (Fisher et al. 2006), could it offer possibilities for defining a context-sensitive form of ethics, using ethnographic methods that would open up the 'black box of science and technology' to normative inquiry (Van de Poel and Verbeek 2006)? --- Midstream Modulation Midstream modulation is a framework for guiding intervention-oriented activities in the laboratory that aims to elucidate and enhance the'responsive capacity' of laboratories to the broader societal dimensions of their work (Fisher et al. 2006).2 Developed by Erik Fisher during a three-year laboratory engagement study, MM has been applied in a range of laboratories around the world as a form of'sociotechnical integration research', or STIR (Fisher and Guston 2008). 3 MM extends more traditional laboratory ethnographies by augmenting participant observation methods with distinct engagement tools that allow for feedback, discussion and exploration of research decisions in light of their societal and ethical dimensions. An 'embedded' social or human scientist interacts with laboratory practitioners by closely following and documenting their research, attending laboratory meetings, holding regular interviews and collaboratively articulating decisions as they occur through the use of a protocol 4 that maps the evolution of research and helps feed back observation and analysis into the laboratory context itself (Fisher 2007). Regular use of the protocol allows for collaborative exploration of the nature of research decisions, with the ultimate aim of shaping technological trajectories by rethinking the processes that help characterize them (Fisher et al. 2006). Since the general possibility and utility of MM was tested in an earlier pilot study (Fisher 2007), the studies presented here aimed to explore the extent to which MM could be applied to enhance lab-based critical reflections on the broader socioethical context of research. As such they attempted to bring together the normative approaches of the ethics of science and technology with the descriptive richness of science and technology studies (STS) (Radder 1998;Van de Poel and Verbeek 2006;Zuiderent-Jerak and Jensen 2007). These research studies asked two questions: (1) How can broader social and ethical dimensions of research be rendered visible in the laboratory? and (2) Do laboratory practitioners perceive critical reflection on the broader socio-ethical context of their work to be relevant? --- First-and Second-Order Reflective Learning To assess the research findings in light of these questions, I distinguish between first-and second-order reflective learning (Van de Poel and Zwart 2009;cf. Sclove 1995;Wynne 1995;Schot and Rip 1997;Grin and van der Graaf 1996). First-order reflective learning is an iterative process by which a professional experimentally finds solutions to problems using several lines of inquiry. This process ''takes place within the boundaries of a value system and background theories'' (Van de Poel and Zwart 2009, p. 7). First-order reflective learning thus concerns ''improvement of the technology and the improved achievement of one's own interests in the network.'' Second-order reflective learning, on the other hand, ''requires a person to reflect on his or her background theories and value system'' ( Van de Poel and Zwart 2009, p. 7). In second-order learning, value systems become the object of learning while in first-order learning these are taken for granted. This distinction can be applied to the social responsibility of researchers: firstorder reflective learning is reflection 'within' the research system. Van de Poel and Zwart note, ''In first-order reflective learning, moral issues are dealt with within the bounds of the background theories and are approached from within the value system of the actor'' ( Van de Poel and Zwart 2009, p. 7). In terms of responsibility, such forms of reflection involve compliance to one's internal responsibilities towards the research community such as the responsible conduct of research and environmental health and safety. Second-order reflective learning involves reflection 'on' the research system, including the value-based socio-ethical premises that drive research, the methodological norms of the research culture, and the epistemological and ontological assumptions upon which science is founded (Verhoog 1980): the background theories and values of the research system itself become the object of learning. The value of MM with respect to the challenge for the ethics of science and technology lies in its ability to support second-order reflective learning. In addition to several instances of first-order learning that occurred as a result of the interdisciplinary interactions, MM served to enhance critical reflection on the socioethical context of lab work. Note that the studies did not assume that laboratory practitioners have a general'reflexive deficit', or that scholars from the humanities and social sciences are somehow more reflexive. Rather, they sought to test the hypothesis that social scientific and humanistic practitioner knowledge could complement, through interdisciplinary collaboration, natural scientific practitioner knowledge. --- Midstream Modulation in Delft and Tempe The STIR studies described here consisted of two consecutive laboratory engagement studies: in the Department of Biotechnology at Delft University of Technology, The Netherlands (Fall 2008) and in the School of Life Sciences at Arizona State University, Tempe, USA (Spring 2009). A total of eight laboratory researchers participated in the studies. I had regular interactions during a period of 12 weeks with four of these researchers. The other four participants acted as 'controls', doing only the pre-and post-interviews at the beginning and end of the study (see Table 1). The participants were all PhD students in molecular biology. Researchers in the Delft Department of Biotechnology focused on the use of microorganisms for industrial production of chemicals from renewable resources and as diagnostic systems, while those in the Tempe Photosynthesis Group applied genomic and molecular biological techniques to elucidate physiological processes in cyanobacteria with a view to bioenergy generation. --- Data Collection Following the MM pilot study (Fisher and Mahajan 2006), interactions with research participants consisted of pre-and post interviews, participant observation, regular application of the protocol and collaborative drafting of visual representations of the research process. The pre-and post-interviews enquired into the research objectives, decision-making structures, implicit and explicit references to societal goals in the project description and changes in participants' awareness of and attitude towards ethical and societal dimensions of the research. The preinterviews marked the beginning of a period of participant observation in which I followed the 'high interaction' participants, spending 8-12 h per week in the lab and participating in regular lab meetings whenever possible. During the research phase, the STIR protocol was applied (Fisher and Mahajan 2006;Fisher 2007;Schuurbiers and Fisher 2009). Reconstructing decisions by way of the protocol allows for reflection on how the interplay of various decision components leads to decision outcomes, constituting a collaborative process in which both observed and reported information is reflected back to the practitioner over time. The embedded scholar thus becomes ''part of the convergence of goals, strategies, and socio-material configurations'' (Fisher and Mahajan 2010). Given the normative background that motivated these studies, my engagements attempted, in addition to bringing out latent MM 'de facto' considerations, to examine how issues in the ethics of science and technology as such could be brought to bear on the research process with the goal of 'deliberately' expanding what researchers took into account (Fisher et al. 2006). Since the goal of STIR was to explore the extent to which interdisciplinary interactions may serve to bring out a range of potential latent and implicit broader issues, I tried not to determine in advance which issues were to be considered as relevant. Indeed, a wide range of issues emerged as a result of the interactions-and were classified only in retrospect (see Table 2). Schematic overviews of the research progress indicate the links between the interrelated series of decision processes mapped over the twelve-week period (e.g., see Fig. 1). As with the protocols, the initial drafts of these overviews were based on earlier conversations, and were discussed regularly with participants, and adapted on the basis of the feedback provided. New drafts were discussed at the following meeting, and the iterative process was repeated. These overviews, and the regular discussion of them, confirmed my understanding of the unfolding research project, built my 'interactional expertise' (Collins and Evans 2002), and identified relationships between the research and the broader discussions held during the protocol meetings. --- Objects of Reflection The iterative process of observation and feedback by means of the protocol and research overviews served to render normative issues that were directly related to the research at hand more visible to myself and my collaborators. Observation and feedback predominantly focused on research goals (knockout or overexpression of protein production pathways followed by phenotypic characterization) and molecular biological techniques (plasmid insertion, the polymerase chain reaction [PCR], separation gels, high performance liquid chromatography [HPLC], and so forth). Still, reconstructing 'technical' decisions by way of the protocol quite naturally brought out'microethics'-normative issues concerning ''individuals and internal relations of the engineering profession'' (Herkert 2005, p. 373). Unpacking a decision not to repeat a gel run for instance could bring out financial and time considerations, but also more overtly normative issues such as concerns about the expectations of a supervisor or the epistemic norms of the research community (verifiability, impartiality, scrupulousness). Asking why research participants took protective measures against harmful effects of carcinogens brought out personal health and safety and environmental considerations, but could also invite a research participant to comment on how colleagues ought to behave, or lead into a discussion about the appropriateness of safety regulations. In addition to the kinds of microethical discussions-lab practices, responsible conduct of research and environmental health and safety concerns-emanating directly from the laboratory work, the feedback processes also occasioned discussion of macro-ethical issues, normative issues that apply ''to the collective social responsibility of the profession and to societal decisions about technology'' (Herkert 2005, p. 373). Enquiring into the impact of a confidentiality agreement on the freedom to publish research results could lead us to examine intellectual property, confidentiality and the influence of private investors on research. A question on the relationship between expectations raised in a research proposal and the actual work done could serve to explore the role of promises and expectations in research, science-policy interfaces and hype-disillusionment cycles in research. Ultimately, repeated questions like ''How do you know that the results you have just obtained are actually a result of your transformations?'' led to discussions on philosophical topics like reductionism and the problem of underdeterminacy of scientific data. Table 2 categorizes the range of topics discussed and provides indicative questions that initiated such discussions, showing how implicit value judgments were rendered explicit by asking 'broader' questions. Most of these topics were addressed in each of the interactions, given that their discussion was dependent on the nature and stage of the research projects as well as the particular experiments performed at the time of study. These findings suggest that researchers frequently deal with normative and social issues but without necessarily labeling them as such, as the notion of de facto modulation (Fisher and Mahajan 2006) posits. Researchers are not accustomed to viewing their decisions from a normative perspective or discussing the normative aspects of decisions explicitly. Such broader issues were brought into focus by routinely asking different kinds of questions than those usually encountered in the midst of laboratory research: questions about the normative dimensions of lab practices, about researchers' personal moral concerns, about the possible longer term ethical, legal and social implications of research, and so forth (see also Table 2). Thus, the methods and techniques of MM can help render ethical and societal dimensions of research more visible to practitioners within the context of the laboratory. In addition to these kinds of discussions brought about by applying MM methods and techniques, several kinds of learning occurred as a result of the interactions. This speaks to the question of whether research participants perceived critical reflection on the broader socio-ethical context of their work to be relevant. --- Reflection 'Within' the System In several ways the iterative observation and feedback processes occasioned instances of first-order reflective learning, i.e., learning related to technological improvement and the improved achievement of the researcher's own interests. The regular occurrence of 'efficiency' discussions, probing for possible overlooked considerations or alternatives of a technical nature, on several occasions led to improvement of the technology or the improved achievement of the research participant's interests in the situation in which he or she was working. For instance, after observing R1A repeatedly preparing small amounts of stock solution for a gel, I asked whether making a bigger batch could save time. Efficiency discussions were a matter of trial and error: participants appreciated my effort, but had often thought about possible alternatives already. In other cases, my questions suggested new alternatives. Applying the protocol to a particular experiment that R2A was performing, we determined that there was an opportunity to identify a specific chemical compound involved in cell-to-cell communication. R2A was searching for the compound in a bottom-up fashion, by measuring cell reactivity to different candidate compounds. When I proposed a top-down experiment, determining the presence of the compound in a sample where the anticipated cell communication was already occurring, R2A replied: My supervisor decided to do it this way. Probably the current experiment was easiest.... But that might be the way to go, now that this doesn't work. Such efficiency discussions thus served a threefold purpose: they elucidated the details of the experiments; probed whether an outsider's perspective could occasion new research opportunities; and built trust, enhancing a sense of co-labor. When I asked R1D at some point whether our interactions led him to perceive new research opportunities, he said: [It happened] just now. Well, I have to look back, I have to think about what I've done every now and then, to tell you what I did, so to say. So that forces me to some kind of realization.... At the same time I've been working on a presentation for a work meeting. At that moment I also realize that knocking out those genes could well have more consequences than we think.... And then I started reading back, like what is the capacity of that transporter, and then I came across a calculating error.... So, on the one hand, you force me to think, and on the other hand a work meeting forces me to think. So... it comes from both sides so to say. These examples indicate that regular application of the protocol facilitated firstorder learning, although it is difficult to pinpoint precisely what triggers the learning process. R1D found his calculating error as a result of being ''forced to some kind of realization.'' Perhaps my questions instigated this realization process, or perhaps it emerged from thought processes developing in the researchers' minds as they explained their work to me. In any case, the collaborative process stimulated mutual learning. There were other instances of this kind of learning, such as when I was discussing one of the draft research overviews with R2A. Looking at the number of research lines he was simultaneously pursuing, he realized how much he had taken on, leading him to the conclusion that he needed to make decisions about which research lines to pursue and which ones to drop:... it's a good following of the process.... I think you can pretty much see how the thinking evolves, right? I mean, the first insertion, that was my supervisor's idea, and then I came up with other stuff, and we get to the point where I'm thinking about stuff that is not even cyanobacteria genes, but something else. When I enquired later about the relevance of our discussion, he commented that he had never given research planning much thought, but saw the value of it now: For me that was the most important point, that I see how much I have to do, or have done, or how sometimes stuff gets entangled with other stuff if you never realize that things are related. Then you end up with a contest, and entrepreneurship, and things which you never thought about, and then... It's also fun to see how you have four lanes, or forks, and then one of them stops, because you're trying to advance the other one, and try to keep all of them running at the same time. Apart from efficiency discussions, considerations of a more explicitly normative nature in some cases led to changes in lab practice. For instance, several research participants who wore two plastic gloves to prevent getting acrylamide on their skin, would subsequently open a cupboard without first removing one of the gloves. When invited to present my findings to the research group at the final lab meeting I attended, I noted this lack of compliance with environmental health and safety regulations, feeding back my observation. The example sparked a hefty debate. Some researchers in the group felt strongly about complying with such regulations, particularly with regard to wearing lab coats, even though no one seemed to ever wear them. A few days later I received unsolicited news that several lab members had now started wearing lab coats again: Apparently, the presence of an outsider in the lab enabled a change in laboratory practice, as a result of rendering explicit and discussing the latent moral considerations of lab practitioners, particularly the'recognition' (quite literally) of personal safety and well-being as a moral value. As this behavioral change illustrates, laboratory-based, collaborative work that was structured by MM was able to accomplish what regulations up to that point could not. Along with the other examples cited, it also confirms that MM can encourage first-order reflective learning by elucidating and enhancing laboratory decisions, whether aimed at improving the technology (a more efficient experimental setup, less time-consuming procedures) or achieving one's own interests (better research planning, compliance to existing regulations). Such reflection 'within' the system of course has value, but more encompassing reflection and learning, such as called for in the ethics of science and technology, would go beyond issues of compliance and improvement and would enhance the capacity of scientists and engineers to reflect on the broader socio-ethical context of their work and the reasons for the regulations in the first place. It would require 'broad and deep' learning (Schot and Rip 1997, p. 257), including second-order reflection on the background theories and value systems of the research context in which researchers operate. --- Reflecting 'on' the System In addition to microethical considerations, broader social and ethical dimensions of research were also regularly discussed during protocol meetings. One example of second-order reflective learning relates to the moral dimensions of genetic engineering. R1D at one point considered integrating a heterologous gene in the micro-organism with which he was working. He faced a choice between integrating a human gene and a mouse gene, both of which fulfilled the required characteristics. Discussing the choice with his supervisors, he invoked a range of technical considerations such as substrate specificity, affinity, capacity, availability of a plasmid and scientific novelty. The question of whether integrating a human gene would be morally acceptable was not discussed. Still, R1D expressed his moral reservations during one of the protocol meetings: R1D: I'm cloning a mouse gene, because... I decided like I'm not going to do a human gene. At least, there was a choice between human and mouse, well, then I'll go for mouse, that's a bit... safer. I subsequently probed R1D for the moral arguments he might have: Me: Why would that matter? A gene is a gene, right? A sequence of base pairs that you can reproduce synthetically. R1D: It's an image-thing. Practically, pieces of DNA from one organism work better than others, and synthetic genes don't always work optimally, probably because of interaction with the genome. Where it comes from is important, it's a bit... ethical. The DNA is still from that person. You put a piece of human in a micro-organism. I would have less difficulty if we would synthesize the DNA based on the sequence of a human fragment of DNA. R1D's response included some morally relevant dimensions. Beyond the practical consideration that ''pieces of DNA from one organism work better than others, and synthetic genes don't always work optimally'', he showed awareness of possible issues in relation to public concern by saying that ''it's an image thing''. He also expressed a moral value with respect to the integrity of the human genome: ''You put a piece of human in a micro-organism.'' His response led us to explore each of these dimensions further. The 'practical consideration' prompted discussion about reductionism: if genes are nothing more than strings of nucleotides, then why would synthetic genes not work optimally? In addition to further practical considerations (synthetically produced genes may have overlooked point mutations for example), we considered the background assumptions behind genetic engineering (the assumption that genes express proteins may turn out to be more complicated than expected due to unknown gene-gene interactions in the living system). The potential for public concern led to a discussion on how to address public concerns about genetic modification. From the possible moral values involved in the acceptability of using genomic material of human origin came discussion of deontological and utilitarian views in ethical decision making and the question of normative pluralism. Evaluating the relevance of these discussions at a later stage, R1D commented: R1D: I had given it some thought subconsciously, but I never really gave it careful thought.... Ethics can be very boring, until you reach dangerous territory, and then it becomes fun. This response suggests that the perceived relevance of ethical issues for researchers increases when discussed in relation to concrete situations and, furthermore, that their discussion in close proximity to the research activities that occasioned them may expand the kinds of considerations that researchers invoke when making morally relevant decisions. These are moments when the embedded ethicist can introduce broader perspectives and invoke theories from other ways of knowing while maintaining a direct bearing on the research at hand. There were numerous occasions for bringing a broader normative perspective to bear on the work done in the laboratory during the lab studies, for example on the regulation of research on genetically modified organisms, intellectual property and the ethics of promising. Another example of second-order learning occurred when discussing synthetic biology. While regularly ordering synthetic genes from chemical suppliers, research participants did not see their own work as being related to synthetic biology, nor to the ongoing debates on synthetic biology in ethics and the social sciences. Upon learning that R1D had ordered a synthetic gene I asked: Me: Would you call this synthetic biology? R1D: That depends. What is synthetic biology? Much of what is now called synthetic biology resembles what we do: putting a piece of synthetic DNA in a host. But I think synthetic biology is making all components synthetically.... Really to develop a cell from scratch might take another twenty years. R1D did not consider normative questions on the desirability of building cells from scratch to be relevant because of the practical complexities involved and the long time span before that vision might become a reality, whereupon I invited him to take a historical perspective. I referred to the progress that was made in molecular biology in recent decades, and how we probably would not have predicted 20 years ago that ordering a synthetic gene would be a standard procedure by 2010. I invited him to reflect on recent developments from this broader perspective, where 20 years is just around the corner. R1D: Then you would need to think about the use, or the goal. If you can build a cell, then you can build other things as well. We shouldn't go in the direction of synthetic higher organisms. There's always a risk that others move in the wrong direction. You shouldn't be using it for other purposes. It's like a knife: you can use it for good or for bad.... That's why we should maybe think about these things. Then there has to be extra regulation. Taking the longer-term perspective that ethicists and social scientists may take when reflecting on new developments such as synthetic biology, R1D started to think about his research in a markedly different way. By contemplating the longterm impacts of his work, he started to reflect on the broader purpose and potential outcomes of the developments of which his own work was a part, acknowledging the relevance of broader reflection. A third example of second-order learning concerns the social relevance of research. Questions concerning the future use of research outcomes were regularly discussed in each of the studies. Responses from all eight of the research participants to the two questions on social relevance featured in the pre-interviews shared a similar ambiguity. All participants responded positively to the first question: does society benefit from research? C1A: One of the main goals is that society benefits, from any research. It's not just a fun thing we're doing here. R2A: I wouldn't see what would be the point otherwise. If it would not help the rest, if that's the reason, than usually... Society should benefit; what would be the point otherwise? While being convinced of the general societal benefits flowing from scientific research, participants had more difficulty in predicting the possible benefits of their own research projects in response to the more concrete follow-up question: does society benefit from your research? C1A: I hope so. It's not my immediate goal; I haven't thought much about it. What I'm doing is basic research; this is probably a little bit far away from... What I'm doing is too far away. R1A: Honestly, I don't see any significant contribution, no. Maybe there is very slightly, slightly, indirectly, related to contributing ideas, maybe there is some technology... But otherwise, the result, for us researchers, we're excited but for other people, who cares? Wanting to pursue this perceived discrepancy between the general benefits of research and the specific benefits of individual research projects, I revisited the question of social relevance throughout each of the studies. Research participants responded in a similar fashion: a general picture emerged in which the ultimate benefits of research cannot and should not be accurately predicted. Participants gave several historical examples of knowledge flowing from basic research that only much later turned out to have practical use like the invention of the light bulb, penicillin or X-radiation, and concluded that unrestrained basic academic research is ultimately more likely to increase the possibility of socially relevant applications than directly demanding social relevance. Increasing calls for social relevance were therefore seen to pose a danger to scientific progress, and ultimately to societal progress, by stifling the innovative power of research: R1D: If you invest more in society-improvement, then the learning curve of science will become less steep. So... in the end it's less good for science... And in the end maybe also for society... in the long term. Interestingly, most of the research projects under study relied predominantly on funding from private organizations and were strongly driven by the need to deliver practical applications. When I questioned the amount of freedom involved in privately funded research, research participants readily acknowledged that their freedom is limited because of the expectations of the private investor. They saw this as the inevitable result of decreases in government funding: the only way for a research group to survive is by strengthening links with private industry. But while acknowledging that this shift in funding mechanisms limited their academic freedom, they continued to invoke the principle of unrestrained academic research to argue against calls for social relevance. Their background assumptions and value systems were in tension with recent changes in funding mechanisms. I subsequently tried to challenge their assumptions by first assuming them: supposing that one cannot predict the societal benefits flowing from research, and therefore academic research should be unrestrained, then how should a private investor determine which types of research to fund, given that funding sources are necessarily limited? Me: The question is: how do you make the decisions whether I should fund genetic modification of cyanobacteria, or whether I should maybe fund your colleagues who do evolutionary growth of cyanobacteria? R2A: That's why the, well the way that I thought is that politicians are the voice of the people, and those are the ones that automatically decide who gets the money, because they should have, they should know, what people want. So if people want cleaner fuels, then they give money to cleaner fuel. If people wanted better dogs, than they would find someone else. I think it's driven like that. To press the question, I would ask how the research participants would decide which research to authorize if they were a policy maker. R2A took recourse in a process of democratic decision making: R2A: Right, I guess the policy has to be made, [based] on the average of what people think.... [T]he policy [should not] be made on the thinking of one person only, but on what most people think. Me: But how about if big masses of people, like in Europe, say we don't want any genetic modification? Would you say, well, that's the majority vote, I'll just quit my job and find another? R2A: Probably not like that. But... I tend to be objective on those sorts of issues, so... Someone who can prove to me that that was the best decision, I would follow it. If someone would have a good argument I probably would... not quit my job, but find a different approach. I guess, I don't know. Such discussions thus problematized the unquestioned assumption that the demand for societal relevance hampers societal benefit. Research participants realized that some kind of demarcation criterion was needed to determine which research to fund, only to realize that this would involve measuring the value of knowledge as a function of some kind of external relevance, contradicting their original assumption that the utility of research cannot be predicted. The MM feedback mechanisms allowed for attending to broader questions as they impinge on the daily work of researchers, and pointing to possible tensions and ambiguities in research participants' responses. The value of these'second-order' discussions lies not so much in having motivated directly observable changes in practice, but in the fact that participants engaged in critical reflection on the broader socio-ethical context of their work. Participants observed the ambiguity in their initial responses, realized that some criterion of relevance is needed 'in the real world' to determine what projects to authorize, and showed interest in reflecting on it in more nuanced ways: R1D: Yeah, you pull... away from the science a little, you put [the science] in a somewhat different perspective, more like... You look at science as a society so to say, where all kinds of things happen. R2D: What I think is useful is that one can indeed think about what kind of societal interest is involved when someone does this kind of research.... I think it's really interesting that people will start thinking about the use much more. These findings suggest that participants began to reflect in new ways on the underlying background theories and value systems operative in research. By challenging unquestioned assumptions, discussing what future applications could come out of the research, and sharing different visions on the role of science in society, the socio-ethical context came to life within the context of researchsomething that participants indicated not having experienced before, neither through their 'ethics and society' curriculum nor ethically-oriented funding requirements. Research participants indicated that the ongoing discussions during and alongside the actual conduct of research did not hamper, but instead added value to the research process in several ways. In the words of R1D,'stepping into the helicopter' could serve as a guide to research planning, to identify overlooked opportunities, to relate lab research to its broader policy contexts, and to uncover latent normative issues. When during the post-interview I asked R1D whether he thought the study was useful to him, he replied:... everybody should perhaps reserve free space in their agendas every now and then, stop all experiments... and think.... Maybe you could... Should one integrate this in each and every PhD project? That someone from outside the faculty comes along, and you need to account for your actions towards that person. And the guy

In response to widespread policy prescriptions for responsible innovation, social scientists and engineering ethicists, among others, have sought to engage natural scientists and engineers at the 'midstream': building interdisciplinary collaborations to integrate social and ethical considerations with research and development processes. Two 'laboratory engagement studies' have explored how applying the framework of midstream modulation could enhance the reflections of natural scientists on the socio-ethical context of their work. The results of these interdisciplinary collaborations confirm the utility of midstream modulation in encouraging both first-and second-order reflective learning. The potential for second-order reflective learning, in which underlying value systems become the object of reflection, is particularly significant with respect to addressing social responsibility in research practices. Midstream modulation served to render the socio-ethical context of research visible in the laboratory and helped enable research participants to more critically reflect on this broader context. While lab-based collaborations would benefit from being carried out in concert with activities at institutional and policy levels, midstream modulation could prove a valuable asset in the toolbox of interdisciplinary methods aimed at responsible innovation.

systems operative in research. By challenging unquestioned assumptions, discussing what future applications could come out of the research, and sharing different visions on the role of science in society, the socio-ethical context came to life within the context of researchsomething that participants indicated not having experienced before, neither through their 'ethics and society' curriculum nor ethically-oriented funding requirements. Research participants indicated that the ongoing discussions during and alongside the actual conduct of research did not hamper, but instead added value to the research process in several ways. In the words of R1D,'stepping into the helicopter' could serve as a guide to research planning, to identify overlooked opportunities, to relate lab research to its broader policy contexts, and to uncover latent normative issues. When during the post-interview I asked R1D whether he thought the study was useful to him, he replied:... everybody should perhaps reserve free space in their agendas every now and then, stop all experiments... and think.... Maybe you could... Should one integrate this in each and every PhD project? That someone from outside the faculty comes along, and you need to account for your actions towards that person. And the guy sitting in front of you would only have to ask: why? Why this? Why that? Couldn't you do that differently? And how does it work? --- Discussion These experiences suggest that the broader socio-ethical dimensions of research were rendered more visible within the research context and that research participants perceived such broader reflection to be relevant. MM served to encourage researchers to address the socio-ethical context of their work through collaboration and in real time. The lab studies aligned with the objective of realtime technology assessment to ''provide an explicit mechanism for observing, critiquing, and influencing social values as they become embedded in innovations'' (Guston and Sarewitz 2002, p. 94) while adopting the overtly normative standpoint that researchers should engage in critical reflection. Like the MM/STIR pilot study (Fisher 2007), these studies helped bring out latent ethical and societal dimensions of research, rendering explicit considerations that hitherto remained implicit, at a time when they could influence researchers' decision-making. Unlike the pilot study, they also aimed to introduce relevant socio-ethical knowledge and perspectives, and initiate discussion of specific moral questions as they arise in the laboratory context. As Ibo Van de Poel and Peter-Paul Verbeek note: Synergy between engineering ethics and STS... could result in an empirical and reflexive research, which is empirically informed and critically contextualizes the moral questions it is asking but at the same time does not shy away from the effort to actually answer them. (Van de Poel and Verbeek 2006, p. 234) The approach I adopted in these studies is not morally agnostic. It invokes the procedural norm that researchers have a moral obligation to critically reflect on their research. Yet a commitment to such 'deliberative modulation' does not require the embedded humanist to enter the laboratory with a predetermined set of substantive norms; as the laboratory engagement experiences made abundantly clear, the content of critical reflection can only emerge as a result of situated interactions over time. Such collaborative, situated critical reflection combines different ways of thinking and knowing: those of the laboratory researcher and those of the embedded social researcher (Gorman et al. 2009). It instilled a sense of urgency, concreteness and relevance to research participants that differs essentially from reading about them in a textbook, for example. It also supports early detection and warning signals of the ethical valence of research outcomes that may otherwise go unnoticed. Additionally, MM can take a more focused (and less speculative) approach towards ethical reflection that could lead to more meaningful interactions between scientists and ethicists (cf. Nordmann and Rip 2009). Note however that the sample size of these lab studies cautions against overgeneralizing: the results need to be compared with other findings to confirm or refute these observations. The perceived value of second-order reflective learning proceeds by way of the perceived value of first-order learning, of improving the achievement of one's own interests. During each study's duration, initial reticence from research participants turned into enthusiasm for discussing both the progress and the broader aspects of their research. Given that'rethinking' knowledge production in research systems depends on the willingness of research communities to rethink their own practices, such collaborative approaches could be more effective than external forms of critique. Of course, this dependency on research participants' willingness to engage implies certain limitations too. While the 'voluntaristic' approach towards collaborative engagement can enhance researchers' critical reflection, it also builds an asymmetrical relation between the researchers and the embedded scholar. As a guest in the research group, the latter is dependent on the acceptance and endorsement of the hosts, and critical views cannot be allowed to disrupt good relationships. This may not be a problem if the collaboration is seen by research participants to be conducive to first-order learning, but could become a problem when there is strong normative disagreement. In those cases, the embedded ethicist has no 'jurisdiction' (Anthony Stavrianakis, personal communication). The need to respect operative conditions and dynamics within the laboratory inevitably limits the range of possible critiques. Furthermore, the collaborations are constrained by their social and institutional environment. Existing, internal responsibilities often take precedence over a researcher's broader social responsibilities. That said, MM has been found to enhance the critical reflection of research participants on the socio-ethical context of their work. Such reflection is arguably needed if other social and ethical programs-upstream engagement, technology assessment, codes of conduct, etc.-are to be successful. The reflective learning documented here provides modest indications of Webster's vision of STS, that is, helping to set the terms on which science might be accorded a socially warranted status that in important ways is distinct from, critical of and supersedes the conventional (scientistic) sense in which science has been legitimated (Webster 2007, p. 460). This vision must be tempered by the danger of the STS practitioner becoming an ''integral co-productionist element of the very structures of power and culture which might be just what STS should be challenging'' (Wynne 2007, p. 494). This is the real challenge for the embedded researcher: becoming part of the convergence of goals, strategies and configurations of the laboratory insofar as it provides access to different registers of justification (Arie Rip, personal communication), while not losing sight of the original intentions behind one's entrance into the laboratory. Walking the fine line between co-labor and critique may allow different voices to be heard at the heart of the R&D enterprise, tapping potentials for learning and change that could prove significant. --- Conclusion The laboratory engagement studies described here provide an indication of the potential for interdisciplinary collaborations to enhance the critical reflection of scientists and engineers, albeit in a relatively small sample size. They demonstrate that broader socio-ethical dimensions can be productively engaged during laboratory research. Midstream modulation was found to engender fruitful and meaningful collaborations between social and natural scientists, encouraging second-order reflective learning while respecting the lived morality of research practitioners. Not only did it help make broader socio-ethical issues more visible in the lab, it encouraged research participants to critically reflect on these broader issues. Contrary to their initial claims, participants came to acknowledge that broader socio-ethical dimensions permeated their research. Importantly, first-order learning seems to be a prerequisite for the possibility of second-order learning: research participants' willingness to engage in critical reflection on the broader socio-ethical context of research was seen to be dependent on their perception that the collaboration also improved the achievement of their own (research) interests. The ongoing observation-based feedback of the midstream modulation framework and STIR protocol allowed the laboratory researchers and embedded ethicist to build collaborative capacities and establish conditions for productive reflection on ethical and social considerations. While what counts as an ethical issue is to some extent a negotiation between the individual collaborators, the procedural norm of reflective learning can guide both practitioners as they deliberatively integrate socio-ethical assessment with ongoing and future research directions.

In response to widespread policy prescriptions for responsible innovation, social scientists and engineering ethicists, among others, have sought to engage natural scientists and engineers at the 'midstream': building interdisciplinary collaborations to integrate social and ethical considerations with research and development processes. Two 'laboratory engagement studies' have explored how applying the framework of midstream modulation could enhance the reflections of natural scientists on the socio-ethical context of their work. The results of these interdisciplinary collaborations confirm the utility of midstream modulation in encouraging both first-and second-order reflective learning. The potential for second-order reflective learning, in which underlying value systems become the object of reflection, is particularly significant with respect to addressing social responsibility in research practices. Midstream modulation served to render the socio-ethical context of research visible in the laboratory and helped enable research participants to more critically reflect on this broader context. While lab-based collaborations would benefit from being carried out in concert with activities at institutional and policy levels, midstream modulation could prove a valuable asset in the toolbox of interdisciplinary methods aimed at responsible innovation.

Introduction This article aims to analyze the impact of a documentary showing the daily routine of ten mothers in caring for their children with disabilities, exhibited in cine debates over two years in Brazil, Colombia and Japan to audiences of health and education professionals and students. The development and production of this film were based on two national and international research projects, involving science and cinema professionals, family members and the professional staff of institutions in the health and disability area, with the support of the Japan International Cooperation Agency -JICA and the Research Support Foundation of Rio de Janeiro (Fundaç<unk>o de Amparo à Pesquisa do Estado do Rio de Janeiro -Faperj). The film takes a new look at aspects of the mothers' lives and how they face up to the challenges posed by the disabilities. As a contemporary documentary, through various daily routines and thematic blocks, it joins narratives, reflections, photos and scenes, past and present, forming a lively and dynamic mosaic that encourages new outlooks. In this analysis, we will show how the communicative and expressive force of audiovisual language becomes a powerful social technology for developing a relevant and strategic theme for public health in Brazil. According to Nóvoa et al. 1, there is a close link between cinema and history, animated image and reality. One can sense the weight of the past and be attracted by the novelty the story evokes. Pires and Silva 2 speak of cinema as a cultural artifact that projects itself as a kind of discourse and contributes to the construction of social meanings, helping to disseminate new and inter-subjective knowledge. Cinema as art is capable of shaping opinions, eliciting reflection, and can be used as a strategy for demystifying prejudice and taboos 3. Codato 4 considers that the main purpose of image in the cinema is to seduce the eye and seek the interpretation, the sense and the meaning in the spectator. Here, cinema is treated simultaneously as an object and instrument of research. As a produced object, it provides an approximation to the real world, to the concerns and stories of anonymous lives, helping to depict the daily routine and life in societies. As an instrument, it becomes a method, a way of perceiving and understanding the world, dealing with reality, revealing a society's way of thinking and its representations. Observed and analyzed, it evokes ways of understanding, communicating and being in touch with oneself and others 5. Thus the use of cine debates constitutes a creative methodology for applied research in social and human sciences. According to Fróis 6, in a documentary film, the filmmaker winds up by recreating or reinterpreting the reality he portrays, because the genre allows him to act like a creative artist. Farina 7 maintains that a documentary should be understood as an original production, with conceptual choices, theoretical and technical, of framework, characters and organization of the final production. The author and director play an important role in the way in which the facts are portrayed, hovering on the thin line between fiction and reality. As Mascarello 8 sees it, if, on the one side, the contemporary documentary "constructs a flight line from the surfeit of reality that overwhelms us, on the other, it turns towards a'reality' that escapes us and challenges us in its inextricable exteriority". In this respect, Farina 7 concluded, a documentary production is not a "reflection of the actual", but a collection of material that stimulates social debates based on a slice of reality. This film, produced and handled as a social technology strategy, aims to challenge stereotyped and reified views of the families, the mothers and the disabilities, focusing on the differences of gender and the role the woman plays in this life path full of stumbling blocks and challenges. We start from the assumption that cinema as art, in particular the film in question, affects the social imaginary, informs about the daily lives of the mothers and children, and is also a powerful tool for tackling prejudice and discrimination. According to Pires and Silva 2, and based on Walter Benjamin 9, imagistically, film facilitates work on processes of sociocultural alienation and enables the building of new knowledge, initiation of new discourse and, through dialog, reaches subjects and promotes intersubjective processes. From recollections of the pregnancy, the initial and later difficulties of the child and communication of the diagnosis to the challenges of infanto-juvenile development, the families come up against uncertainties, the couple faces emotional difficulties in assimilating the news that initiates a difference, the mother tends to assume the daily tasks of intensive care, and the professionals have difficulty in giving the diagnosis, guiding and providing support in the most difficult stages of the life 10. The coldness and silence, the impersonal and distant contact, the omission of information reveal in the expressions of the health or education professionals that the problem is worrying and serious, the child is not go-ing to develop "as expected" and that the school will have difficulties in receiving it 11,12. On account of the disability, expectations concerning the child's future start to unravel and a fatalistic and prejudiced tendency to appear on the part of professions in the face of the parents' difficulties 11, shot through with feelings of anguish, despair, fear, sadness and shame, particularly affecting the mothers, not always free to cry and whose pain is barely understood, with many seeing family members and close friends drift away 13. In Western and Eastern societies, domestic tasks and child care are women's responsibility. Thus, as the main carers, mothers are at risk of stress on account of the demanding daily routine in which they sacrifice their love life, social and professional lives, as a woman and wife 14. Motherhood of a child with a disability is unique, imposes complex and challenging experiences, involves uncertainty about the future, unknown and unpredictable, requires specialized help, a support network and social support 12,15. There is much we need to understand in our civilization about the necessity for family, technical, social and religious support when a child's limitations are substantial. Based on the idea that society in general is unaware of the challenges and barriers faced by families and children with disabilities in the course of their development and social inclusion, the feature-length film "A special day" emphases the mothers' experience of caring and their relationship with their special needs children, the family, school and society. The documentary gives an account of children, young people and adults with various types of disability, neurological syndromes and autism, to a severe, moderate or slight degree. The circuit of cine debates, with exhibitions and reflections on the film, recorded the words and emotions of the audiences; and, in the words of Ramos 5, like a "magic mirror", the film revealed in body language, gestures and feelings, verbal and non-verbal communication what it is like to "care for a child with a disability, from morning to night", through the eyes of the mother, woman and wife. Our objective in this article is to analyze the impact of the film and its distribution to audiences of health and education professionals. --- Methodology First phase, with financing by JICA (2006-2008) -Amorim et al. 16 16 produced a first documentary, "Above the Waves" (28 minutes), based on interviews with mothers and children with intellectual disability and autism in Brazil, Colombia, Malaysia and Thailand. In the video, they draw attention to the similarity of the women's experience, despite cultural differences. On showing the video to mothers of "special" children in Brazil, we were surprised by how they felt free to speak, exchange, share and recall their experiences. Second phase with JICA financing (2011-2013) -This result lead us to invest in a second Brazilian documentary for use as a social technology tool for tackling disability in collective health. Through the "Very Special Women Network" project, sponsored by JICA, the filmmaker Yuri Amorim, assisted by Annibal Amorim 17, produced the documentary A Special Day (80 minutes), in partnership with the Instituto Educateur, Kinera Produç<unk>es, the Centro de Estimulaç<unk>o e Psicopedagogia Criart and their families. The interviews and the daily routines were filmed after definition of the guidelines for the script in six meetings between the network of women, the professionals and the filmmaker. Third phase with Faperj financing (2013-2016) -Cavalcante 18 set up the project Gênero e Deficiência retratados no cinema: biografias em debate ("Gender and Disability portrayed in the cinema: biographies under debate"), sent to the Office of the Vice-President for Environment, Attention to and Promotion of Health/Oswaldo Cruz Foundation, supported by the Instituto Educateur, Kinera Produç<unk>es and Criart, as well as the Veiga Almeida University/Integrated Social Practices Laboratory, of the Federal University of Rio de Janeiro/National Museum. Thus funds were raised for post-production of the film A special day, distribution and organization of up to ten cine debates in schools, universities and congresses, with a view to raising awareness among health and education professionals and the public in general. Content of the film A special day -The feature-length film introduces different ways of looking at gender and disability. It accords visibility to the plurality of experiences that accompany the arrival of a child with a disability, communication of the diagnosis, the impact of the news on the family and the disability in the lives of the women, mothers, wives and professionals, the adjustments they go through, their pain and achievements, what they learn, what the children with disabilities teach them and what they have to say to the world about their experience. It starts with the question: "If your life was a film, what could not be left out?"; and ends with the thought: "What would be a special day for you?", when the notion "special" used by the mothers, on referring to their children, acquires singular meanings and concepts in plays on words. Organization and systematization of cine debates -Exchanges were organized with events that permitted exhibition of the film in eight Brazilian towns and in Bogota/Colombia and Tokyo/Japan. A total of 22 cine debates were held, and nine meetings with health and education professionals and students from primary to post-graduate level were recorded in full. Debates were held after each showing, with the participation of the mothers who took part in the film, the filmmaker, social scientists or professionals and educators, who enlivened the exchange with the audience. The public's reactions were documented in films and audio recordings, transcribed and organized in an analytical table showing the subjects, what was said and the relevance attributed to the film. Analysis of the data -The transcriptions were grouped by events and exhibitions linked to health and education and the material was organized in narrative units, by subject and associated ideas. A cross reading enabled an analysis of relevant categories, taking into account the audiences and their particularities. Field impressions, theoretical reflections on documentary cinema, disability and family, gender and motherhood, as well as concepts that guided the interpretation of narrative and social contexts: Daniel Bertaux's 19 idea of a "life story", Philippe Lejeune's 20 "biographical pact", the "social imaginary", approximating cinema and history, according to Nóvoa et al. 1, and the notion of "social representation" for Codato 4. We started with Daniel Bertaux's 19 notion of "life story", which considers the story of a life as told by the person who lived it. What appears to be the basis is merely one facet of the person. As the documentary combined a number of "life stories", with different profiles, the resulting diversity was such that the contrast between the reports brought about a saturation point that renders the selection representative. Even faced with a "biographical pact" or an "authenticity pact", as Philippe Lejeune 20 points out, we will be heedful of the relativity of the events and the adventure of the art of speaking about oneself. As Nóvoa et al. 1 puts it, in the close relationship between cinema and history we perceive the "social imaginary", confirming the socio-histories as representations and their anchoring in a sociocultural context. In Codato's 4 view, film enters into a dialog with the "social representations" in their historic, social and aesthetic bias, in the power of the cinematographic image, revealing mechanisms that lie behind the impression of reality. The representations stimulate reflections based on the images and the relationships that one human being establishes with another. The limitations of the study lie in what it was possible to record and compile of a wide universe, in the analysis of words taken out of their conversational environments and in the risk of reducing plural and polysemic interpretations attributed by social players. Taking these points into account, the advantages of the analysis reside in the opportunity to compile and analyze content evoked by exhibition of the film and the debate, differentiating by target audience and the most expressive themes and emphasis, and as far as possible understanding the differences of discursive contexts and the atmosphere noted. --- Results and Discussions From the initial aim of distributing the film A special day and holding eight to ten cine debates, noting the film's impact on health and education professionals and students, it's fair to say that we have progressed further than expected, in that the film seems to have "grown its own legs" and is blazing the trail. One showing aroused interest in further showings, and the congresses and events provided opportunities for the protagonist mothers, production team and researchers to participate. Schools and universities, professionals and students became interested in seeing the documentary and helping to organize the debates. The work involved with the film exceeded expectations and it was therefore necessary to establish a profile for the events that would be closely monitored, filmed, transcribed and analyzed. We will first present the circuit of cine debates in Rio de Janeiro, in Brazilian states and in two countries that have partnered us in various stages of previous projects, Japan and Colombia. We analyzed cine debates held at two scientific events, two health education graduation and post-graduation events, two primary school and two high school meetings and one focusing on public policies. We will now present reflections on the impact of the cine debates by group analyzed. (1) Health professionals at scientific events; (2) Health professionals and graduate and post-graduate students; (3) Primary education teachers and students; (4) High school teachers and students. Between August 2013 and March 2016, as shown in Table 1, the A special day film circuit included one or more showings in eight Brazilian towns (Teresina, Jo<unk>o Pessoa, Manaus, Belo Horizonte, Rio de Janeiro, S<unk>o Paulo, Bras<unk>lia, Santa Catarina), in Tokyo, Japan and Bogota, Colombia. Specifically in Rio de Janeiro, where we were able to monitor it more closely, it was shown in 13 different venues. The film was discussed in seven congresses, seminars and forums, two international, three national and two regional, an estimated audience of 790 people. The major events that exhibited it considered the theme of attention to health, mental health, disability, autism, occupational therapy and humanization, with the emphasis on human rights, citizenship and diversity. Included in the 2nd ABRASME Film Festival Vida em Foco (Life in Focus), the documentary A special day followed an itinerant exhibition circuit through the Amazonian riverside communities in 2014, at the end of the 4th Brazilian Mental Health Congress, giving visibility to the various productions there. The film's initial landmark was its first showing, at the 1st International Healthcare Congress, in Teresina (PI), when it still had a running time of 90 minutes. In response to audience feedback, the filmmaker and director fine-tuned and cut it to achieve a final running time of 80 minutes, and this version competed and was selected in its first festival. In the Assim Vivemos (How we Live) 6th Disability Film Festival, held at the Banco do Brasil Cultural Center between 2013 and 2014, A special day was exhibited six times in Rio de Janeiro, three in S<unk>o Paulo and three in Brasilia, with a debate in each city. The documentary was considered one of the five best of 26 films from 17 countries (Australia, the United Kingdom, Canada, Israel, Russia, Brazil, Holland, France, Norway, Spain, Scotland, Iran, Belarus, Bulgaria, Poland, Germany and the United States), awarded the prize Mulheres Guerreiras ("Warrior Women") by the jury, and "best film" by public vote at the end of the circuit. As shown in Table 1, the film was exhibited at eight festivals and events, in institutions and companies (OAB, SESC, SESI, Petrobras) and film festivals (Psicodália Festival), to an estimated audience of 1,110 people. Finally, the film was shown in cine debates in universities and public and private schools in Rio, to an audience of 1,470 people. In all, between congresses, seminars, festivals, events, universities and schools, the majority in Rio, the film was exhibited to some 3,370 people -health and rehabilitation professionals, family members, people with disabilities, graduate and post-graduate students and professors, primary and high school teachers and students. The following analysis includes reflections on the film in Brazil and Colombia. We draw attention to the interest of a professor from the Social Service department of Japan Women's University, who included the film in her classroom methodology. --- Cine debates between health and education professionals (1) Health professionals at scientific events: For public health professionals, the film produced an overall view of those who look from the outside, a new way of seeing after an empathetic approach and a change of view concerning the professional way of acting. It presents an overall and plural view of what it means "to be the mother of a child with a disability", casts a sensitive eye on the mothers and carers, gives a voice to those who undergo the experience. The film gives depth, by reproducing words taken from daily lives. There were mothers in the audience during the debate, who gave their views: "It was extraordinary, the film portrayed what we experience, it showed our expectations and sorrows"; and siblings: "My mother had to learn to be a nurse, a teacher, and to fulfill various social roles to help my disabled brother". It emphatically brought some people closer to the reality of these mothers of "special" children, in comparison with other mothers, generated a recognition of the investment in the child and their resilient attitudes. One aspect of the film that caught a social scientist's attention was "the role of affection", something that requires a major psychic effort in extreme situations. He therefore concluded that these families are "special" in that they sustain affection that is demanding on a daily basis and is often given without support, without other structures, with no chance to catch a breath and recover vitality. It was striking that over half of the stories analyzed stressed that the film caused a change of view in relation to professional practice, making professionals rethink their way of seeing, hearing, perceiving, feeling and acting: "Professional people need to know how to share and hear, and to shed the fear of being creative". They acknowledged the importance of the exchange of views between mothers and professionals, the need to act in a more integrated way with the subject and their family, the importance of a support network for mothers and carers: "As a nurse, I feel guilty, because I never saw this side of things, I shall provide more help and see how the mother feels". This proves that, as stated by Naujorks et al. 21, cinema is a powerful pedagogical tool: "The actual representations of people with disabilities can be redefined, repeated, or brought up for discussion, expanding the possibilities of the individual and collective imaginary of people with and without disabilities in relation to those with disabilities". --- Target public --- Audience (2) Health professionals and graduate and post-graduate students: The importance of listening to families building strategies that meet their needs, the importance of a "breathing space" to enable a mother to be a woman, as well as the mother of a disabled child: "The film showed us a carer's strength and energy. Their sense of life has changed, but they have to find strength to cope with the fate the world has reserved for them in order to live". The importance of empowering mothers was emphasized: "We have to strengthen the women, because an empowered woman cares for the family, will track down resources, talk to the husband and mobilize the community, so we will be reinforcing the network". The specific function of gender is considered, the woman's role in the family as the one who provides daily help for the more vulnerable members and has a strategic caring role: "The idea is not to cut out the father, the siblings or even the grandfather, but to understand that the mother has a specific role because she's a woman, in most cases, the one who will provide care or abandon. So, to prepare for this situation, who should we strengthen? The woman?" A woman who manages to find a space to care for herself, as well as caring for others, will be contributing towards fulfilling her social role well and fighting a source of chronic stress. In the words of Neves and Cabral 22 : "Once empowered, a woman can demand her rights, fully exercise her citizenship, ensure quality care for the child, share decision-making with the professional who attend her sons and daughters and be effectively included in the caring process". The film presents situations in Brazil similar to those observed by professionals in Colombia -the similar incidence of abandonment by the fathers, the fear of dying and the worry about who will care for the child, the need to find strength to promote inclusion. For these reasons, the film was considered a model and an example of a way of acting with people with disabilities. As the film "Above the waves" shows, thematic similarities of the challenges for mothers and families are to be found in different cultures and societies. The advantage of film as art and social technology is to retrieve the subjects' experiences, the interaction between mothers and children, the carer's suffering, and to give visibility to the matter. "The documentary is beautiful, it shows the family life that we don't see", said one professional. The film shows one thing over and above the daily events of professional practice and research: "It's one thing for you, a researcher, to transcribe the mother's voice, and then interpreting it, but however much you effect the phenomenological epoché and reinterpret the words, it's you speaking, it's not her". The conclusion is therefore that: "when we stop to hear what she has to say, we realize how much we have to learn from these families, we're not the only experts, but we also have to learn and they need to act more as partners, we have to spend more time with them". On this point, the film gave rise to a very important consideration, a change of view and an understanding of the need to recognize the richness of the strategies these mothers and families create for facing up to the disability, the way in which they produce everyday knowledge that needs to be seen and better understood: "There are many more tools that are developed and that the health or rehabilitation professionals don't see because they only recognize their own tools, not those built by the families". This discovery of families as builders of a stock of knowledge, alongside the traditional scientific knowledge, could result in advances: "We need to open up a field of investigation to find out more about the tools built by these families; and how they can suggest tools, think of tools, in other words, involve them in this construction of knowledge and the state of knowing". In the book Pessoas Muito Especiais (Very Special People), Cavalcante 12 reveals a wide variety of experiences, strategies, practices, tools and knowledge developed by these families. However, the documentary is a quicker way of evoking new viewpoints, confirming the art's importance for the production of knowledge: "The film makes us reflect (...) invites us to assume this social responsibility as health professionals, as public health workers". The cinema, as an art that gives wings to the social imaginary and stimulates transformation in social representations is highlighted by Silva, Nascimento and Silva 23 : "we understand cinema as an art that expresses a certain view of the world, with philosophical, aesthetic, ethical and educative content, that enables and favors different apprehensions of this reality situated in a historic-social context". (3) Primary education teachers and students: A school that stands out for its social and political participation and engagement, in both regular teaching and scholastic inclusion, was chosen to participate in this debate. There were two meetings, which also included some parents and guardians, held in a festive atmosphere. Among the relevant themes highlighted were differences of gender and motherhood, questions of school and professional practice, discrimination and prejudice and public policies. Attention focused on the impact on the life of the couples, the separations, the absent fathers, the mothers left alone and accompanied. The film shows the 'viewpoint of the mother, the woman, in interaction with the family; and learning about this feminine experience helps us to understand them. The mothers' lack of time for themselves was noted -how much they need the time to be a woman, how greatly they feel they are losing various rights of life and coexistence, of going to the doctor, sleeping, enjoying themselves and even the "right to die". As Falkenbach et al. 24 put it, "talking about and studying the matter of the disabilities can provide useful information, however it is living and coexistence with these people that can promote in-depth learning and new views". A number of educators who spoke recognized the importance of the school's and the families' participation in the children's development and inclusive education. On the other hand, the school's commitment to socialization, to fighting prejudice and respecting differences was also discussed. "The film give us a jolt. The school starts to think of itself as a challenge, that of not only receiving the child (with a disability), but of feeling itself called upon to go a step further -what strategies for developing this individual can the school propose to the family?". The film also lead educators to think about a change of attitude, seeking to find ways of giving the mothers more support: "I used to think that it was good to listen, but it's not enough, we need to develop strategies". There was also a change of outlook and posture, a search for another way of listening, more attention and care in building re-lationships, better sharing of the responsibilities, working together, teaching appreciation of life, expanding the debate in schools: "What does it mean to be in a relationship? What is the school's responsibility? What is the family's responsibility? (...) when we offer special needs teaching, we have to encourage full development, there has to be a pedagogical route". There was recognition of how the theme of discrimination and prejudice is constructed socially, of what the school so often contributes towards the social construction of the preconception of gender and of other social groups; and of how it needs to assume a role in the fight against discrimination, and to act on behalf of all. Schools need to reinforce the policies for fighting prejudice and respecting differences. (4) Secondary education teachers and students: A school with a normal curriculum visited the Raul Cortez theater twice to see the film A special day, taking 400 pupils, parents and educators each time. It was a special moment. As in every session, crying and sobbing were heard in the course of the film. Six of the ten mothers who participated in the film were present at the meeting, which lead to an enthusiastic debate and a euphoric experience in which the audience, mainly comprised of young people from the primary teacher training course, was able to put questions directly to the characters in the film. The students were advised to present their questions in writing and some used the microphone to ask questions directly. The questions didn't stop coming and it was impossible to keep up with even a tenth of what was asked. We compiled the questions, removing the repetitions and arriving at a total of 79 questions, 64.5% (49) of which were put directly to each of the mothers, personalizing the exchanges and the themes. The remaining 35.5% (27) were questions directed to the mothers as a group, with only three to the mediator. Three main themes interested the students and guided their queries: questions of gender, family and maternal care; reactions, adjustments and coexistence after diagnosis of the disability; school and prejudice. The mothers lead the debate and entered into direct exchanges with the audience. We highlight the words of one mother, whose daughter has a serious and progressive degenerative disease. Her words are a good illustration of the impact of a severe disability on a woman's life: "Everything is very hard, because when you are given a special child, they don't come with an instruction manual. You don't know how to deal with it or who to appeal to. And everything changes in your life, right? Your family life, your personal life, your working life (...)". As Fiamenghi J<unk>nior and Messa 25 put it, "The parents of disabled children live with worry all their lives, from the baby's birth to old age, mainly due to the lack of institutions that can care for disabled people as they grow old". The depth and extent of this change and the impact on the woman's life, starting with the serious condition of a child with a disability 12,26,27, are clear from this narrative: "Our vocabulary changes (...) But from that moment, the vocabulary becomes one of therapy, convulsion, medication and new treatment. And suddenly you stop being yourself, you're no longer R. You're A's mother. And that's how the health professionals call you for the treatment sessions: -A's mother and there comes a time when you no longer know yourself, you don't recognize yourself any more". This clearly shows the change in the life of a woman, mother, wife, professional person and the importance of a support network. "The quality of the support provided by the support networks renders the experience of motherhood less arduous; and the more effective the assistance to these women, the more confident they will be in caring for the child" 15. --- Conclusion We saw from the words of professionals, educators and students that the film A special day impacts the public through the sensitive look at the little-known life experience of the mothers and the difficulties of the daily care, producing empathy for their hard reality and that of the disabled person and resulting in changes of perception. There are reflections on differences of gender, resilience, love, integral well-being, co-existence, respect, family, school, society and support network, from the carer's viewpoint. The feature-length film contrasts ten life stories, articulates various themes, perceptions and points of view, presents day-to-day situations based on life experiences and practices built, which not only clarify problems, difficulties, pain and suffering, but also point to ways of facing them, support, subjective and social tools that have been developed in the life cycle by women and families challenged by their child's disability. By giving visibility to the voices and experiences of the mothers, in the drama of their lives, it was possible to bring to light unknown information and the emotional, social and cultural strength of the warrior women who acquire life wisdom The film has been referred to as a model and example of how to act with disability, due to the similarity of experiences in different cultures and societies, such as Colombia and Japan; and also for revealing aspects of family life that escape the professional eye. More than showing the richness of experience of the mothers and their children, the film reveals that these families build tools for dealing with the disability, other than the professional tools, and this constitutes an expertise to be investigated and studied. Hearing the voice of these women, mothers, wives and professionals led the post-graduates to conclude that much is to be learnt from them, and that these mothers need to be more involved as partners in health care and the building of knowledge. Educators and primary and high school students were more struck by the family life experience, seeing the effects of the disability on the couple's life and the overload assumed by the woman and mother. The mothers' lack of time for themselves and how they need to receive more support called attention. The school party saw how the family could be more of a partner in the challenge of monitoring and teaching pupils with disabilities. The film "gives a nudge to schools" insofar as it shows that it is not enough to take in a special student, it is also necessary to welcome the family and support the mothers. The film prompts a change of view and posture, inviting an investment in the quality of the relations between family and school. It also give more visibility to the experiences of prejudice and discrimination, helping the school community to recognize itself as a social space that tends to reproduce stereotypes and that could work in a more conscious way to deconstruct prejudices and stigmas. Cinema as art stimulates faster transformations in the imaginary and produces new social representations, showing itself to be a high-impact technology. It evokes in-depth debates, produces a greater approximation of social realities and life experiences and also shows the differences of gender and promotes empathy and respect for the mothers and their children. The documentary A special day gave visibility to diverse experiences in the face of deficiency, showing that as a result of the extreme care, a woman profoundly changes her life and no longer recognizes herself, and it is thus crucial to expand support and increase the network of very special women. --- Collaborations FG Cavalcante, LF Lau, GF Barbosa, DLG Berlim, NC Menezes and DC Braga participated in both the cine debates and structuring of the article. FG Cavalcante, AC Amorim and YC Amorim cooperated in production of the film, the cine debates and revision of the article.

The article analyzes the impact of a documentary on the daily lives of mothers and children with disabilities, exhibited at conferences, festivals, universities and schools in Brazil, Colombia and Japan. Produced by film and health professionals, social scientists and a women's network, the feature film "A special day" was awarded at Assim Vivemos VI International Disability Film Festival (2013-2014). The objective is to analyze the impact of the film and its diffusion among professionals and students from health and education areas. We adopted a methodology of cine debates that were recorded, transcribed and coded. The circuit toured 22 locations with one or more exhibitions for 3.370 people, of which nine presentations were analyzed. Our interpretations were guided by the notions as life account, biographical pact, imaginary and social representation. We analyzed gender issues, family and maternal care; reactions, adaptations and life after the diagnosis of disabilities; discrimination and prejudice; school and inclusion. The study showed that the communicative and expressive power of film becomes a powerful social technology to address gender, maternity and disability as an important and strategic issue for public health in the country.