Exploring the mental health and psychosocial problems of Congolese refugees living in refugee settings in Rwanda and Uganda: a rapid qualitative study

Background Refugees fleeing conflict often experience poor mental health due to experiences in their country of origin, during displacement, and in new host environments. Conditions in refugee camps and settlements, and the wider socio-political and economic context of refugees’ lives, create structural conditions that compound the effects of previous adversity. Mental health and psychosocial support services must address the daily stressors and adversities refugees face by being grounded in the lived reality of refugee’s lives and addressing issues relevant to them. Methods We undertook a rapid qualitative study between March and May 2019 to understand the local prioritisation of problems facing Congolese refugees living in two refugee settings in Uganda and Rwanda. Thirty free list interviews were conducted in each setting, followed by 11 key informant interviews in Uganda and 12 in Rwanda. Results Results from all interviews were thematically analysed following a deductive process by the in-country research teams. Free list interview findings highlight priority problems of basic needs such as food, shelter, and healthcare access; alongside contextual social problems including discrimination/inequity and a lack of gender equality. Priority problems relating to mental and psychosocial health explored in key informant interviews include discrimination and inequity; alcohol and substance abuse; and violence and gender-based violence. Conclusions Our findings strongly resonate with models of mental health and psychosocial wellbeing that emphasise their socially determined and contextually embedded nature. Specifically, findings foreground the structural conditions of refugees’ lives such as the physical organisation of camp spaces or refugee policies that are stigmatising through restricting the right to work or pursue education. This structural environment can lead to disruptions in social relationships at the familial and community levels, giving rise to discrimination/inequity and gender-based violence. Therefore, our findings foreground that one consequence of living in situations of pervasive adversity caused by experiences of discrimination, inequity, and violence is poor mental health and psychosocial wellbeing. This understanding reinforces the relevance of feasible and acceptable intervention approaches that aim to strengthening familial and community-level social relationships, building upon existing community resources to promote positive mental health and psychosocial wellbeing among Congolese refugees in these settings.


Introduction
Globally more than 70 million people are displaced, with over 80% in countries neighbouring their country of origin [1], many in low-and middle-income countries (LMICs) with limited resources. These include protracted situations where refugees have been living in insecure settings for generations [2]. Mental health is impacted by a complex interplay of social determinants [3], which play out in unique ways in the social ecology of refugee settings. This includes structural conditions of adversity such as insecure asylum status, restricted opportunities to work or pursue education, and limited access to services, all of which can compound pre-displacement exposure to conflict or traumatic events and exacerbate psychological distress and mental health problems [4][5][6]. Displaced populations also face adversity due to loss of or separation from family, the erosion of social support mechanisms and community networks, and potential cultural or linguistic barriers to negotiating health services [7]. Refugee settings compound this adversity by establishing social, economic and political systems that may challenge the long established social and cultural norms that governed and shaped everyday life in their communities before they became refugees [8]. For example, in some settings refugees face restrictions on the right to work, to cultivate land, or pursue higher education. Such conditions diminish the agency of refugees and may create dependency on outside institutions. They alsoplace demands on social and cultural norms over generations, which can impact on the way refugees relate to one another, and can increase the risks of violence and discrimination [8,9].
Conceptual frameworks aid our understanding of the complex and cyclical relationships between the ecology of refugee settings -including familial, social, community, and structural context such as conditions of poverty and disadvantage -and mental health and psychosocial wellbeing. These include frameworks that foreground conditions of adversity that arise through daily stressors such as difficulties accessing basic necessities of food and shelter [10,11] and the impact of losing personal, familial, and social resources essential for mental wellbeing [12]. The ecological model [13] emphasise the micro-, meso-and macro-level settings in which humans are embedded, directly and indirectly influencing individuals through continuous individual, family, community, and social environment interactions. This model encourages attention to the multiple levels at which health and wellbeing are impacted, for example, a problem such as gender-based violence that manifests in the family may be an expression of gender relations at the community level. These frameworks and model emphasise creating supportive and enabling environments that allow displaced refugees to rebuild a sense of self and regain a sense of control to promote positive mental health. Acknowledging these connections, mental health and psychosocial support (MHPSS) services increasingly recognise the significance of social and cultural factors shaping experiences of mental health problems and appropriate responses; and the importance of integrating MHPSS services into multi-layered health, social and community systems [7]. Community-based MHPSS services are recommended to build on individual and collective capacities and resources [14][15][16][17][18], overcoming refugees' hesitancy to engage with formal services [19].
Consequently, to be acceptable and successful, approaches to MHPSS services must be rooted in the daily stressors and adversities refugees face, situating mental health and psychosocial problems in their socio-cultural context. To achieve this, research to understand local community needs and priorities is critical [18,20]. Gaining this understanding can help identify opportunities to empower existing social supports, and to inform the contextual adaptation of MHPSS programs to ensure their acceptability and relevance [19]. Recognising this, we applied the Design, Implementation, Monitoring and Evaluation (DIME) rapid qualitative approach [21] with Congolese refugees living in a refugee camp in Rwanda and a refugee settlement in Uganda. This study forms part of the COSTAR project [22] which will contextually adapt, implement, and evaluate a community-based group psychosocial intervention with Congolese refugees living in these settings.

Methods
We applied the DIME rapid qualitative approach developed and applied extensively with adults in low resource settings [20,21], including with refugee populations [23][24][25]. The rapid qualitative assessment involves two stages: first conducting free-list interviews with local community members to identify problems the community face and a brief description of each. The results are analysed to identify priority problems relating to mental health and psychosocial wellbeing. Second, the priority problems are further explored in key informant interviews with community representatives knowledgeable about the selected topics. We conducted the study in each site (30 free list and 10-15 key informant interviews). Interviews were conducted in Kinyarwanda in Rwanda, and Congolese Kiswahili in Uganda.

Setting
Over the last two decades, the Democratic Republic of the Congo (DRC) has experienced three large-scale conflicts: the First Congo War (1996-7) and Second Congo War (1998)(1999)(2000)(2001)(2002)(2003), and the Kivu Conflicts in Eastern DRC which continue to date [26,27]. As of December 2019, Uganda and Rwanda host 397,638 and 76,266 DRC refugees respectively [28]. The two COSTAR study refugee settings are located 480 km apart: the Kyangwali refugee settlement in Uganda, and Gihembe refugee camp in Rwanda. Both settings are administered by the respective country government, with operational support from the United Nations High Commissioner for Refugees (UNHCR).
Gihembe refugee camp was established in 1997, and as of March 2015 hosts 14,774 refugees [29] in 3030 densely populated households. The camp was established to host survivors of the Mudende massacre [30]. Mudende was a refugee camp in Western Rwanda hosting Congolese refugees from Eastern DRC which in August and December 1997 suffered attacks by armed groups crossing the border from the DRC. There are an approximately equal number of male and female refugees with families averaging 4-5 people [29], and given their protracted stay, all speak Kinyarwanda. The majority of Gihembe refugees are Christian, with a minority Muslim population. The Kyangwali refugee settlement was established in the 1960s and currently hosts 109,207 people, of which 108,164 are refugees and 1403 are asylum seekers [31]. The majority of the refugees in Kyangwali are from the DRC (105,514), followed by South Sudan (3273), Rwanda (327), Burundi (71), Kenya (10), and Somalia (10,32). The Kyangwali settlement has approximately 39,846 households spread over a large geographical area, each with access to land to cultivate. The settlement has an approximately equal number of males and females [32] with families averaging 5-6 people, and the common language is Kiswahili. Both Gihembe and Kyangwali refugee settings host international and national refugee organisations, and have elected refugee representatives who represent community perspectives on administrative and governance matters.
Congolese refugees have experienced evolving Ugandan and Rwandan refugee policy and practice, with both countries recently committing to the United Nations Comprehensive Refugee Response Framework [28,33]. Uganda follows a self-reliance refugee model which includes the right to work, freedom of movement, and access to education and healthcare [34,35]. For example, 53.5% of refugees in Kyangwali report that they have an occupation [36]. Rwanda currently adopts a graduated refugee camp model with the aim of fostering greater refugee integration, and with support delivered through a cash and voucher mVisa system that allows families to make their own food choices [30]. The refugees in Gihembe currently do also have the right to establish businesses inside and outside the camp, and to move outside the camp to cultivate rented farmland or pursue employment.
Given the long and complex history of conflict in the DRC and the Great Lakes Region more broadly, refugees in these settings have been displaced at different times [28]. Gihembe refugee camp hosts a more established population who have been living there since 1997 [28], with some recent arrivals as a result of renewed hostilities in the DRC in 2012-13. Conversely, Kyangwali refugee settlement has experienced a pattern of continuous arrivals since its establishment, with a recent surge in arrivals as a result of the 2018-19 DRC Conflicts [28]. As a result, Kyangwali and Gihembe host a mix of more and less established Congolese refugees who may have experienced different levels of access to resources and support services since arrival. This is important as differences in resource access and support services can become a source of adversity through intra-community conflict over perceived inequity in resource allocation.

Participants
We interviewed male and female adult Congolese refugees living in Gihembe Camp, Rwanda; and Kyangwali Settlement, Uganda. Free list interviews were conducted with a convenience sample of 30 participants in each site (n = 60) purposively selected according to the inclusion criteria: Congolese refugee living in Gihembe/Kyangwali, over 18 years of age, fluent in Kinyarwanda or Congolese Kiswahili, and able and willing to talk to the research team. Participants were recruited from a selection of villages/neighbourhoods (umudugudu in Kinyarwanda) spread geographically across each site, with efforts made to achieve gender and age-range representation.
Key informant participants were purposively identified by free list interviewees as knowledgeable about the priority topics to be explored. Key informants were local community members including community and religious leaders, teachers, etc. The key informants did not have professional roles to respond to the mental health and psychosocial problems explored in interviews (i.e. they were not community counsellors, or health care or social workers). This approach aims to ensure that interview responses prioritise local community perceptions of problems and solutions, rather than views underpinned by professional training [21]. The demographic profile of free list and key informant interviewees are summarised in Table 1:

Data collection
Research was conducted between March and April 2019. Twelve interviewers in Kyangwali (8 males, 4 females), and 8 in Gihembe (4 males, 4 females) who spoke the participants language were locally recruited. All interviewers received a 3-day training in the basic principles of qualitative interviewing, including how to conduct free list (FL) interviews, and the key principles of ethics including obtaining voluntary informed consent.
FL interviewers worked in pairs, one asking questions and probing responses, and the second making a verbatim written record of what was said in the local language. Each FL interview lasted between 20 and 45 min. In FL interviews participants were asked to list all the problems that adults living in Gihembe/Kyangwali face, and to provide a 1-3 line description of each. At the end of each interview the transcript was reviewed to ensure  clarity of the written record, and the interviewers reviewed the list of problems to identify those that might be related to mental health and psychosocial wellbeing (defined as those relating to thinking, feeling or relationships [21]). Interviewers then asked participants to recommend community members knowledgeable about each problem to be invited to key informant interviews. Following FL interview data analysis to select priority problems (described below), key informant (KI) interviews were conducted to gather more detailed data on the priority problems relating to mental health and psychosocial wellbeing. Interviewers were provided an additional 1 (Gihembe) or 1.5-day (Kyangwali) training for KI interviews, covering asking open-ended questions and probing skills, and refreshing key principles of research ethics and voluntary informed consent. In KI interviews the participants were asked to provide: a) a description of the problem, including symptoms and effects; b) perceived causes of the problem; c) effects of the problem on the individual The money we get is not enough 9 Housing 9 Food that is not enough 9 The problem of unwanted pregnancies among girls (Ikibazo cy'inda z'indaro mu bakobwa)

7
Aid is not enough 7 The problem of official papers 5  Killing one another 4 May lead to tribal disputes 3

Hatred within villages (machuki) 3
There is no peace and freedom (hakuna amani na uhuru) 2

Imprisonment 2
Lack of co-operation (between community members) hinders development 2 Early pregnancies associated with difficulties at delivery and reduce chances to lead a good life 1 Leads to poverty status 1 What do people currently do to address the problem

Number of respondents
Sensitizing community about effects of discrimination 6 Advice giving 5 Teaching people to live in harmony and love one another 3 Reporting problems to NGOs 2 Separating tribes that are fighting 1 and those close to them; d) what people currently do about the problem; and e) what people think could be done about the problem. Interviewers again worked in pairs with one asking questions and the other recording a written summary responses in the local language. Creating a written record is the recommended DIME approach, which seeks to facilitate rapid data collection and analysis in humanitarian settings. Written transcripts were reviewed with the KI at the end of each interview to ensure clarity of meaning and completeness. Each KI interview lasted on average 60 min, including obtaining consent. Should the participant have more to say or be knowledgeable about additional problems identified for KI interviews, interviewers would agree a subsequent date and time to continue the interview.

Ethics
Ethical approval for the study was provided by the University of Liverpool, the University of Makerere, Uganda National Council for Science and Technology, and the University of Rwanda. Additional administrative approvals were obtained to conduct research in the Kyangwali refugee settlement and the Gihembe refugee camp. All participants were approached by interviewers, provided a brief description of the study, and asked if they would consider taking part. If in agreement the interviewers would find a quiet location to discuss the participant information sheet, answer any questions, and complete a written informed consent form with the participant, including consent to reporting anonymised research results. For illiterate participants a line or mark was accepted in lieu of a signature, alongside witness confirmation (a participants' nominated family member or friend) of the voluntary nature of consent. Participants were free to pause or end the interview at any time without repercussions, and were provided refreshments during the interviews. Study trainers were available throughout data collection and analysis, and conducted daily team de-briefs. All data was anonymised through participant identification numbers. The only exception to this was the names of potential KIs recommended by FL participants which were recorded separately to FL data and destroyed once interviews had been conducted.

Data analysis
All data was analysed in the original languages at each site by the interviewers, supported by the trainers who were provided verbal translations where required. A thematic analysis approach was applied that was deductively driven by the interview questions, as recommended in the DIME manuals. For FL analysis interviewing pairs reviewed written FL interview transcripts and listed problems and their reporting frequency on a summary sheet. When problems were worded differently a consensus decision was made as to whether the items were the same or different to one another. When the same, the most appropriate wording was selected for the summary sheet by consensus. If no consensus could be reached, both items were listed as separate problems. Once complete, the problem lists were re-ordered by frequency alongside a single brief description using the participants wording that interviewers agreed best described the problem. Summary sheets were then reviewed with local community stakeholders Table 3 Problem of discrimination/ubaguzi (Uganda) (Continued)

Imprisonment of perpetrators of discrimination 1
Reporting to the village chairperson 1 Having community dialogues about discrimination 1 Government putting announcements on the radios about its effects 1 What should be done about the problem if resources were available Number of respondents Advising people to love one another 4 Establishing strict laws to fight discrimination 3 Sensitizing the communities about effects of discrimination 2 Games uniting people 2 All people should be taken abroad 1 Formation of community development groups 1 Joint business venture 1 Establishing more places where water can be got 1 Give them money to buy the food they need. 1 Government should pass rules against discrimination 1 Government should punish perpetrators 1 Table 4 Problem of inequity/inequality /ubusumbane (Rwanda)

Description of the problem Number of respondents
We have limited job opportunities because of having refugees status (papers) 3 Those who have a chance to find a job outside the camp, you realize that their salaries are not equal to those of nationals 3 When you finish secondary school you don't have financial support to go to the University like Rwandans 2 Poverty status 2 We cannot get bank loans because we don't have guarantee to give and also the fact that we are not Rwandans 2 In the sector of shelter within the camp, there is inequality (ubusumbane); they do not permit us to construct new house as we wish nor do repairs 2 There is gender inequity and gender inequality (ubusumbane bujyanye n' igitsina). 1 As refugees are not respected (Gusuzugurwa kw' impunzi) 1 Refugees who are teachers in schools outside the camp do not earn the same salaries as Rwandans. That causes teachers from the camp feel they are not like others, they feel less value and are ashamed (ipfunwe) like during the meeting. (representatives of organisations in the refugee settings) to identify problems potentially related to mental health and psychosocial problems -defined as those relating to thinking, feelings and relationships -to explore in KI interviews. Additional considerations in the selection of priority problems included that they were mentioned by multiple respondents and contained descriptions that appeared to be severe in terms of impact. A similar process was followed for KI interviews analysis with the interviewers and trainers reviewing written transcripts and recording responses to: a) description; b) perceived causes; c) perceived effects on the individual and those close to them; d) what people currently do; and e) what people think could be done. These responses were compiled into summary tables alongside the codes of participants. Where participants had reported what interviewers agreed was referring to the same thing the interviewers would agree the wording of the item and record both participant codes alongside it.
For the purpose of reporting the data has been translated into English, with original language terms -in Kiswahili for Uganda, or Kinyarwanda in Rwanda -retained where relevant. Analysis was conducted independently at each site, before the results were compared across the two sites. This approach sought to ensure analytical attention to the specificities of the data at each site, before considering the data sets alongside one another to identify potential commonalities and differences in the experiences of Congolese refugees in two settings, following previous DIME studies that have adopted this approach [37]. Have the opportunity to get loans from the banks 1 Youths who completed secondary school be given a financial opportunities to go to university 1 Those in old age be given financial support on different occasions 1 Increase vocation training opportunities among the youth to prevent them from turning to substance use 1 Introduce a literacy school for adults so that they can learn how to read 1 To set up a library with enough books so that educated people continue to acquire various knowledge 1 Increasing the money we are given every month 1 Those of us who are yet to be resettled to a third country, they can also help us to go abroad because in the camp we feel like people living in isolation 1

Results
The Uganda and Rwanda FL results are provided alongside one another. We then identify the priority mental health and psychosocial problems selected for KI interviews, and present KI interview results for thematically similar problems explored across both sites.

Free-list interviews
In Table 2 we list the problems reported by the 60 FL respondents in Uganda (n = 30) and Rwanda (n = 30) in decreasing order of frequency. Across both settings the most commonly reported problems relate to daily living such as poverty, unemployment, and lack of healthcare access and education opportunities. Alongside these participants reported social problems including alcohol and drug abuse, sexual and gender-based violence, teenage pregnancy, and discrimination/segregation. Problems related to their refugee status were also highlighted, including difficulties obtaining official papers or in Rwanda obtaining employment if one's status is identified as "refugee", and a lack of resettlement opportunities. This problem was mentioned more often by refugees in Rwanda as compared to Uganda, which may reflect that the majority of study respondents had been living in Gihembe for over 21 years.
Problems potentially related to mental health and psychosocial wellbeing (i.e. relating to thinking, feeling and relationships) selected for KI interviews in Uganda include discrimination (ubaguzi) (n = 8), no intervention for torture and trauma victims (hakuna musada kwa watu ambao walinyanyaswa na kihihi) (n = 4), domestic violence (ukatili wa nyumbani) (n = 3), and alcohol and substance abuse (kutumia pombe na madawa ya ulevi mubaya) reported as "drunkardness/alcoholism" (ulevi) (n = 2). After discussion between the Ugandan research team and local stakeholders the problem of alcohol abuse was expanded to incorporate substance abuse because these were felt to be interconnected problems that should be considered holistically. In Rwanda KI interviews explored unwanted pregnancies amongst girls (n = 7), inequity/inequality (ubusumbane) (n = 4), sexual and gender-based violence (ihohoterwa rishingiye kugitsina) (n = 3), and drug abuse (kwishora mu biyobyabwenge) (n = 3). We chose further explore the problem of discrimination (Uganda) and inequity/inequality (Rwanda) because as a problem related to social relationships and cohesion this was felt demand require further unpacking to consider its relationship to mental wellbeing.

Key informant interviews
We summarise KI interview responses for mental health and psychosocial problems across the two sites that explore similar themes of social inequalities, sexual and gender-based violence (SGBV), and alcohol and drug abuse. The results are drawn from 11 KI interviews in Uganda, with 7 respondents exploring all three problems, and 4 addressing one or two; and 16 KI interviews in Rwanda, with 4 respondents exploring each problem.
First we present the results for the problems of discrimination (ubaguzi, Uganda, Table 3) and inequity/inequality (ubusumbane, Rwanda, Table 4): These results contain similarities, including identifying gender as a key site of discrimination, particularly noting the impact of discrimination and inequity/inequality on the education of girls. Another similarity is the connection between experiencing discrimination or inequity/inequality and emotional impacts such as "jealousy", "lack of love for one another", "sadness/feeling upset" and experiencing "shame". Recommendations for responding to these problems from both settings highlight the role of dialogue and learning.
The results also identify key differences in the problem description, with participants in Uganda making frequent reference to discrimination by tribe, religious sect, and language. In contrast, in Rwanda inequity/inequality is identified as arising from participants' refugee status in contrast to Rwandan nationals (although this is also identified in the results from Uganda, it is less frequently mentioned). These differences potentially reflect the refugee communities in each setting, with Uganda hosting a greater diversity of nationalities, languages, and ethnic origin.
Second, we present results for the problem of domestic violence (ukatili wa nyumbani, Uganda, Table 5), and sexual and gender-based violence (ihohoterwa rishingiye kugitsina, Rwanda, Table 6).  Similarities are notable, including that descriptions of violence are centred on the familynotably the husband and wife relationshipand foreground impacts upon children. Poverty is highlighted as a cause of interpersonal violence, with the role of alcoholism also noted. Both sets of data foreground the health impacts of violence including physical injuries and rape, with a particular emphasis on children as suffering adversely from the effects of violence such as experiencing "child neglect" and "mistreatment". Recommendations for addressing these problems include awareness raising, counselling, enforcing laws against perpetrating violence, and providing education and vocational opportunities.
There are also some notable differences, primarily the role of polygamy as both a cause and consequence of violence in Uganda which was not identified in Rwanda. In Rwanda the role of physical proximity to others in the community is identified a cause of violence which is not present in Uganda, reflecting the role of the physical organisation of the refugee settings in perpetuating violencewith Gihembe densely populated whilst Kyangwali is a large settlement allowing refugees to live in small compounds that are spread out.
Finally, we present the results of the problem of alcohol and substance abuse (kutumia pombe na madawa ya ulevi mubaya, Uganda, Table 7) and drug and alcohol abuse (kwishora mu biyobyabwenge, Rwanda, Table 8).
Results on substance and alcohol abuse across the two settings show significant similarity, including the impact of past experiences causing emotional distress that leads to substance abuse. Another similarity is the impact of alcohol and drug abuse at both the community and family levels, including adding to community insecurity and physical violence within marriages and towards children. Overlap is also evident in recommendations to respond to this problem, with a notable role for giving advice or counselling and enforcing laws, as well as providing employment and vocational opportunities.

Discussion
This study contributes to understanding the problems faced by Congolese refugees living in two refugee settings in Uganda and Rwanda. Findings identify priority mental health and psychosocial problems that relate to social cohesion and social relationships. Social cohesion is a complex concept that has been conceptualised to contain multiple dimensions, including common values, social order, social solidarity, social networks, and place attachment (see e.g. [38]). We adopt a broad understanding of social cohesion as entailing connectedness, a sense of belonging, and solidarity among refugees [39]. A lack of social cohesion is evident in the problem of discrimination and inequity; and poor social relationships are described in the impact of alcohol and drug abuse and sexual and gender-based violence which affect relationships at the family and community levels. To be acceptable MHPSS services must be rooted in the social ecology of refugees everyday lives [18], and recognise the impact of daily stressors [10,11]. To achieve this, study findings will inform the implementation of a community-based group psychosocial intervention that seeks to rebuild intra-community connections, and the adaptation of instruments to evaluate the interventions effectiveness [21].
Problem FL findings reflect dimensions of human development as conceptualised by the United Nations, which builds on Sen's work on human capabilities [40]. Briefly, Sen claims that there is moral value in a person's freedom to achieve wellbeing, which is understood in relation to people's capabilitiestheir opportunities to do and be what they have reason to value. Notably participants' responses reflect foundational aspects of human development such as access to basic needs including food, shelter, and health care to maintain a decent standard of living; followed by contextual social problems including systemic discrimination, inequality, and gender inequity. This problem prioritisation is important for understanding the daily living conditions of Congolese refugees in Gihembe and Kyangwali, and consequently identifying the impediments to achieving positive mental wellbeing [40]. Prioritising meeting basic needs such as shelter and food is commonly reflected in studies exploring the MHPSS needs of refugees [10]. For example, research with South Sudanese refugees in Table 6 The problem of sexual and gender-based violence/ihohoterwa rishingiye kugitsina (Rwanda) (Continued) like tailoring and hair dressing and cooking If they could financially support us, we will all find what to do 1 Those in old age would be given jobs 1 To ban those illegal drinks the way they banned Kanyanga (local illegal liquor) 1 We think that they should find another refuge for us 1 They would install for us public lights on the streets and houses 1 When someone is sick and they realize they cannot treat him/her they would immediately refer him/her to main hospital 1 They can bring back firewood or give us gas 1 Table 7 The problem of alcohol and substance abuse/kutumia pombe na madawa ya ulevi mubaya (Uganda)

Description of the problem Number of respondents
Abrupt fights at home (mapigano nyumbani) 5 Drunkards staggering (kuanguka) on the road in broad day light 3 Drinking alcohol and smoking (kuvuta) 3 Beating up children 2 Picking fights with people for no reason (kupigana na watu bila kuwa na shida nao) 2 Getting lost on the road 2 Falling on the road 2 Being dirty everyday 1 Stealing people's property 1 Chewing marijuana and selling it (watu wanakula mairungi na kuchuruza) 1 Raping women (ubakagi wa wanawake) 1 Spending many hours in bars drinking alcohol 1 Youth have created groups for smoking marijuana 1 Increased number of bars which sell alcohol 1 Being nude in front of children 1 Hurling insults at people 1 Stealing food from gardens 1 Urinating in own trousers 1 Reeking of alcohol and marijuana 1

Perceived causes of the problem Number of respondents
Having no peace at home (kutokuwa na amani ndani ya nymba) 4 Being in bad company 4 Wanting to forget many bad thoughts (kutaka kusahau mafikiri mabaya) 4 Being jobless 3 When s/he remembers things he went through when in DRC (kama anakumbukavinavyo alipitia ndani ya DRC) 3 Loosing hope because they have no parents nor job opportunities (kukata tamaa juu hawana wazazi walakazi) 2 Seeking pleasure (kutafuta raha) 1 To replicate what youth watch in movies 1 Received little money which is no good for them and they are not happy because of the hardship in the settlement 1 Having no say in community matters and no friends (kutokuwa na neno kwa komunote wala kukua bila rafiki) 1 Lack of self-worth/respect.(ukosefu wa maana wa kibinafsi) 1 It is a learned habit (born of parents who use alcohol) (ni tabia nazaliwa na bazazi banakunywa pombe) 1 Alcohol is very cheap 1 Too many bars open during daytime 1 Effects on those with the problem and those close to them Number of respondents When intoxicated they get into fights and are jailed 5 Poverty within families 4 Excess use may cause death 4 Separation/divorce 3 Diseases in the home 3

Rape women 2
Mental disorders (magonjwa ya kichwani) 2 Uganda report similar problems securing basic needs, and poor social cohesion including ethnic tensions, gender-based violence, and child protection concerns [41]; findings echoed in studies with refugees and refugee service providers in Tanzania, Rwanda and Burundi [42]. This problem prioritisation is complemented by theoretical exploration of the changed social context and systems in refugee settings that can create conditions that facilitate gender inequity and SGBV [9], such as the loss of traditional gender roles, poverty, and limited education opportunities. The layering of problems therefore highlights the dynamic interplay between the multi-level structural organisation of refugee settings and the multiple and intersecting social relationships at the community and family levels that correspond to ecological models of human development [13]. Building on the FL findings, KI interviewees' exploration of priority mental health and psychosocial problems bring out their common roots in problems of social cohesion. The relationship between alcohol and drug abuse and mental health problems has been explored in research with Burundian refugees [43], including relating the use of drugs and alcohol to a loss of social control (p.225). Furthermore, Tankink, Ventevogel, et al. [42] highlight the contextually embedded consequences of substance abuse both on individual mental health and wellbeing, and as a factor underpinning disruptions to social cohesion such as gender-based violence. Loss of Table 7 The problem of alcohol and substance abuse/kutumia pombe na madawa ya ulevi mubaya (Uganda) (Continued) Marital rape (ungono kwa kingufu) 2 Quarrels with neighbors (kusema mubaya na wajirani) 1

Motor accidents 1
Being sacked from jobs 1 Falling out with friends 1 People who drink too much lose respect (ukosefu wa heshima) of family members and other people 1 After receiving family food ration, the husband wants to sell the food to buy alcohol. This infuriates the wife who then stabs him with a knife.
1 Domestic violence (ukatiri wa nyumbani) 1 What do people currently do to address the problem Number of respondents Arresting abusers of alcohol and other substances 7 Community health talks about alcohol and substance abuse. 5 Religious leaders preaching against alcohol and substance abuse 2 Laws are in place to prohibit the sale and consumption of alcohol by minors 2 Advising alcohol abusers to reduce alcohol intake 2 Engage people in sports so as to spend less time in bars 1

Relocating bars 1
Prohibiting children from working in bars and consuming alcohol 1 What should be done about the problem if resources were available

Number of respondents
Restricting bars to open during daytime 5 To employ the alcohol abusers 4 To teach them how to drink responsibly 4 To enroll alcohol and substance abusers in self-help groups (kufanya vikumbi vya watu wanaokunywa pombe) 3 To retain children in school 1 Banning alcohol sale in towns 1 Gazette bars in residential places 1 Police to patrol bars during daytime 1 Organizing sport activities for children and youth 1 Conduct abrupt searches in bars to arrest people who employ minors 1 Proper disposable of alcohol sachets so that children don't lick them 1 Table 8 The problem of drug abuse/kwishora mu biyobyabwenge (Rwanda)

Description of the problem Number of respondents
They involve themselves in violence and there are those who stab one another (Bakora urugomo harimo nabateranaga ibyuma)

2
Coming home drunk, quarrelling (agatongna) and you realized that the family is not tranquil ntumererwe neza) 2 He/she changes in his/her talking style, walking style, it means that he/she shows another face (agaragaza indi sura) 2 There used to be a time they left schools and you realized that they were no longer studying 1 Falling down while walking and talking obscene (atukana) and also fighting 1 He/she beats another person with no reason 1 He/she can drink and then go to 'gare' (meeting place at the center of the Gihembe camp) and strip naked. 1 He/she throws stones at a car yet he/she is not mad (umusazi) normally 1 A person may be under influence of drugs and then rape a minor 1 He/she is not afraid, is not obedient (ntiyubaha), and won't be afraid of his/her leaders 1 It leads some to become bandits (abarara) 1 When the night falls it is not only the phone he/she can also snatch a handbag from you 1 The camp management team come and arrest those who sell them and punish them 1 It [actions to address the problem] is often done by the government of Rwanda and they also urge us to fight against them 1 social control is echoed in participants' descriptions of the behaviours of those with alcohol and drug abuse problems that emphasise outward public perception, for example identifying getting lost on the road, being dirty everyday and falling down while walking and talking obscene and also fighting. These descriptions echo another aspect of the COSTAR study which explored what FL participants consider to be a 'good life'. Findings identified the role of outward appearance and dress which are linked with self-esteem and how a person is viewed in the community, including the avoidance of shame (Robinson J, Chiumento A, Kasujja R, Rutayisire R, White R: When you are dirty, you have confidence in you: an exploration of the 'good life', personal appearance, and mental health with Congolese refugees in Rwanda and Uganda, under review). These connections are echoed in KI findings relating to discrimination/inequity where participants described a lack of love for one another, back-stabbing among different tribes and not respected because you are a refugee. Whilst in Uganda intra-community tensions were structured by gender, tribal, linguistic or religious affiliations; in Rwanda, inter-community inequities were more commonly reported, arising between the treatment of refugees compared to Rwandan nationals. The difference in the internal/external conceptualisation of these problems is potentially related to the structural environment in which refugees are embedded, emphasising the role of the social, economic and political systems for how refugees relate to one another [8], and their host communities. This finding is also echoed in our results on SGBV which highlight the adverse effects of violence on children and young people who experience neglect and mistreatment. These echo studies exploring violence and protection risks for adolescent refugees living in refugee settings in Uganda that highlight the complex interplay of societal norms towards violence [44]; and contextual factors such as divisions within refugee communities as a result of new arrivals which heightened problems of food insecurity, in turn giving rise to psychosocial impacts as a result of hunger [45]. Therefore, our findings suggest complex, inter-linked, and multifaceted relationships between the structural and relational problems experienced by Congolese refugee communities in Rwanda and Uganda that impact on mental health and psychosocial wellbeing.
Specifically, our findings foreground that problems of social cohesion in family and community relations negatively impact the mental health and psychosocial wellbeing of Congolese refugees in Gihembe and Kyangwali. This is evident in participant's descriptions of the effects of problems, for example for the problem of inequity/ discrimination participants' identify this leads to: fighting, hatred within villages, jealousy among people and to lose hope. These effects indicate both low levels of social cohesion seen in a lack of connectedness or sense of belonging, as well as feelings of hopelessness, indicating low levels of mental wellbeing. The interconnected relationship between low levels of social cohesion and poor mental health outcomes has been identified in a recent study with Congolese refugees in Rwanda that highlights the relationship between mental health problems and Table 8 The problem of drug abuse/kwishora mu biyobyabwenge (Rwanda) (Continued) [drugs] Projects that came and helped our boys and girls and take them to study at Gahogo those who had dropped out of schools 1 'Inshuti z' Umwali' (Friends of young lady) they also came and helped girls and boys and gave them what to do instead of sitting idle 1 Security organ from the village (umudugudu) to the top is always at hand, when information is shared on time, we quickly intervene [to address problems] 1 What should be done about the problem if resources were available Number of respondents Maybe if possible, all of them should find jobs 2 If security is back they would go back home (country of origin) 2 There is a good thing the government did of taking to a rehabilitation centre, but after 1 year they should not abandon them like this, they can find away to bring them together in vocations They can make the procedure of going to America easy 1 suicidal ideation, and social cohesion including a low sense of connectedness and belonging [46]. Our findings therefore confirm the relevance of conceptual frameworks that emphasise the complex and cyclical relationships between the ecology of refugee settings, encompassing daily stressors, experiences of loss, grief, and continuous social interactions at the community, family and individual levels, and mental health and wellbeing [10][11][12][13]. What is absent from the data is the spontaneous identification of problems that directly reference psychological or emotional states commonly recognised in the US and Europe, such as depression or anxiety. Whilst the data does reveal descriptions that may capture culturally bound idioms of distress [47] such as feeling despair, self-isolate and disowning oneself, sadness, lack of self-worth/respect and wanting to forget many bad thoughts, reflecting previous research with conflictaffected populations that identified a common syndrome with core features of sadness and social withdrawal that bore similarities to classifications of mental states such as depression or anxiety [19]. In this study the core conceptualisation of problems with potential mental health and psychosocial impacts were socialincluding behaviours considered antisocial such as alcohol and drug abuse, violence and gender-based violence; and the systemic problem of discrimination/inequity within and between communities. This problem prioritisation reflects research highlighting the role of social functioningunderstood as interpersonal interactions in daily life, including access to strong familial and community resources -as an important component of mental health and psychosocial wellbeing [48]. It is also in line with a systematic review that suggests social support seeking from family, friends and community groups is an important component of effective coping strategies in waraffected populations in LMICs [49], with reciprocal networks suggested to be the most beneficial [50]. Our findings reflect multiple qualitative studies with refugees in East Africa and globally that report contextually embedded experiences of interconnected structural and social constellations of refugees' lives that create conditions where social problems, and resulting poor mental health and psychosocial wellbeing, arise [40,51,52].
Consequently, the problems impacting upon mental health and psychosocial wellbeing in this study have their roots in disruptions to social cohesion. As such, proposed interventions should build on interpersonal, familial and community-level supports [9] to support the establishment of social norms that challenge inequity, injustice and abuse [18]. Such an approach offers opportunities to positively re-frame the social conditions that negatively impact mental health and psychosocial wellbeing. This approach resonates with the repeated suggestion of FL and KI participants to provide advice and sensitisation to communities about how to respond to social problems, and to build on community resources to establish educational, vocational and employment opportunities in an effort to reduce negative mental health impacts. This is a central component of the COSTAR project that is working to implement and evaluate a community-based group psychosocial intervention that seeks to promote interactions between individuals and their social environment to reestablish values, norms and relationships; whilst also exploring shared experiences and contextually embedded coping mechanisms that build on the collective strengths and resources of communities to promote mental wellbeing [16]. Alongside such interventions it will also be important for the global community to continue to address structural factors that are recognised to impact upon refugee mental wellbeing, such as refugee integration policies that grant or refuse refugees the right to education and employment opportunities, and the conflict and environmental factors that continue to force people to flee their homes [53].
A unique contribution from this study is the comparison of findings collected with the same methodology in two sites hosting Congolese refugee populations. This is important for understanding the commonalities and differences between the experiences of Congolese refugee communities living in two distinct settings, and for identifying potential opportunities for group-based psychosocial programs to help address the identified problems. Potential study limitations include the number of key informant responses for each problem, notably in Rwanda where adverse weather led to limited access to Gihembe camp to collect data within timelines. However, the high level of congruence of themes, and their confirmation at meetings with community stakeholders including organisations supporting refugees and refugee community representatives, suggest findings are broadly representative of the experiences of Congolese refugees living in Gihembe camp and Kyangwali settlement. Finally, whilst efforts were made to ensure a broad sample including mixed gender, ages, and length of time in the refugee setting to capture the potential diversity of experiences resulting from participant characteristics that affect their daily living, it is possible that the gender affiliations of our research assistants may have affected the engagement of some potential participants.

Reflections on DIME
The DIME approach applied in this study has two interrelated objectives [1] to elicit an indication of priority problems that Congolese refugees in Gihembe and Kyangwali face to inform the contextual adaptation of a group psychosocial intervention; and [2] to provide local language terminology for the translation and adaptation of mental health instruments. We found the first objective to be partially met as study findings provide a understanding of the complex and dynamic interplay of factors impacting upon mental health and psychosocial wellbeing from the perspective of local community participants whose voices may not otherwise be heard. These indicate the sort of problems and proposed solutions participants may bring to a group psychosocial intervention, providing contextual information to situate the intervention in the lives of Congolese refugees. Some of this understanding could also have been obtained through "practice-based evidence" [54] that enters into a dynamic relationship with knowledge, practice and context. This could include for example gaining the knowledgeable insights of local researchers, the practicebased insights of stakeholders such as refugee organisations, and the contextual understanding of refugee community representatives and members. These approaches should continue to prioritise and value diverse community participation and perspectivesincluding through purposive inclusion of local community members not represented through formal channels -to identify MHPSS priorities and shape interventions, valuing the lived-through experiences of refugees that plays a role beyond research findings [55,56]. For the second objective the data obtained has emphasised commonly used linguistic terms in Kiswahili and Kinyarwanda, and encouraged critical reflection on terminology and translation into English [57] in an effort to accurately reflect the precise concerns being voiced. This understanding of language has furthermore been directly incorporated in the translation and contextual adaption of instruments to assess mental health and wellbeing, suggesting this may be where the DIME approach is strongest. Given these reflections further critical reflection on the DIME approach, and potential alternative methods, is recommended.

Conclusion
Recognising the importance of anchoring MHPSS interventions in the context of refugees' everyday lives, this rapid qualitative study has explored the priority problems of Congolese refugees living in two settings in Uganda and Rwanda. Our study makes an important contribution to the literature on the position of Congolese refugees by exploring the commonalities of experiences across two distinct refugee settings. Notably, findings reveal significant commonalities in the priority problems related to mental health and psychosocial wellbeing that are rooted in problems of social cohesion within refugee communities and between refugee and host communities. This understanding reinforces the importance of proposed intervention approaches that strengthen familial and community-level social relationships to promote positive mental health and psychosocial wellbeing, alongside addressing structural conditions that lead to forced migration and conditions of daily adversity that precipitate mental health problems [10][11][12][13]53]. Our findings confirm that implementing a community-based group psychosocial support intervention that seeks to foster communities of support by harnessing the collective strengths and resources of refugee communities to promote mental health and psychosocial wellbeing [16,17] offers a potentially appropriate response to addressing the problems identified by participants in this study.