Examining implementation of an intervention to reduce psychological distress and intimate partner violence in a refugee camp setting

ABSTRACT

An integrated approach to reduce intimate partner violence and improve mental health in humanitarian settings requires coordination across health and protection services. We developed and tested the Nguvu intervention, which combined evidence-based interventions for psychological distress and intimate partner violence among Congolese refugee women in Nyarugusu refugee camp (Tanzania). We conducted 29 semi-structured interviews with Nguvu participants and stakeholders to explore the relevance, acceptability, feasibility, and impact of this intervention. Participants reported that the intervention aligned with needs and filled a gap in programming, yet further adaptations may improve the fit of the intervention. The Nguvu intervention was acceptable to participants, including group discussion of sensitive topics. Confidentiality was highly regarded among staff and participants, which improved safety and acceptability. It was feasible to train non-specialist refugee workers to deliver the intervention with adequate supervision. Facilitators noted contextual challenges that made it difficult to implement the intervention: limited infrastructure, competing priorities, and population mobility. The intervention was perceived to improve awareness of the association between violence and mental health, reduce self-blame, and build skills to improve wellbeing. Recommended adaptations reveal promising, yet challenging future directions for addressing social determinants of mental health and implementing multi-sectoral programmes in complex humanitarian settings.

KEYWORDS:

• Mental health
• intimate partner violence
• refugee health
• implementation science
• process evaluation

Introduction

Intimate partner violence, which involves physical, sexual, psychological, and/or controlling behaviours perpetrated by a current or former male partner, is the most common form of gender-based violence in humanitarian emergencies (Stark & Ager, 2011), and remains a leading risk factor for mental health problems among women (Tol et al., 2017). The prevalence of intimate partner violence is estimated to be greater during periods of conflict and displacement due to disruptions in traditional gender norms and a confluence of social and contextual adversities (Krug et al., 2002; Murphy et al., 2017; Wachter et al., 2018). Recent observational research suggests a cyclical and reinforcing relationship between intimate partner violence and mental health problems, such that intimate partner violence increases risk for depression and post-traumatic stress disorder, which in turn may make women more vulnerable to violence in the future (Krause et al., 2008; Perez & Johnson, 2008; Tsai et al., 2016). Intervention studies have found that greater reductions in symptoms of post-traumatic stress and depression resulting from psychological intervention were associated with lower intimate partner violence risk (Iverson et al., 2011). Intimate partner violence and related mental health sequelae have been targeted in isolated ways for several decades (Tol, Stavrou, Greene, Mergenthaler, Garcia-Moreno, et al., 2013). However, given the complex, bi-directional relationships between intimate partner violence and mental health, an integrated, multi-sectoral intervention that simultaneously targets intimate partner violence and psychological distress may be a superior strategy (Tol et al., 2017). In humanitarian settings, it is widely acknowledged that mental health and psychosocial support interventions constitute essential parts of the gender-based violence response (Gender-Based Violence Area of Responsibility, 2018; UNHCR, 2020). In practice of humanitarian settings, however, MHPSS interventions within protection services are often limited in scope, operate in relative isolation from (mental) health services, and may be financed through different funding streams with different financial donors (Schell et al., 2020; Ulleland, 2013; World Health Organization, 2020). Developing and testing a strategy for integrated, multi-sectoral delivery of mental health and psychosocial support may serve as a model for integration across various humanitarian sectors to improve mental, physical, and social wellbeing (Greene et al., 2018).

Setting

Nyarugusu refugee camp was established in 1996 in response to the arrival in Tanzania of over 150,000 Congolese fleeing conflict in the eastern provinces of the DRC. As of early 2015, there were over 60,000 refugees in Nyarugusu refugee camp. Beginning in April 2015 there was a large influx of refugees arriving from Burundi and continued arrivals from the DRC leading to a population of over 150,000 refugees in Nyarugusu as of 2018 (UNHCR, 2018).

Nguvu intervention

The Nguvu intervention integrated elements of cognitive processing therapy and advocacy counselling. Cognitive processing therapy is a manualized, evidence-based psychotherapeutic intervention developed for survivors of assault that focuses on developing skills to manage distressing thoughts that lead to emotional problems (Gallagher & Resick, 2012; Resick & Schnicke, 1992; Tran et al., 2016). The intervention focuses on identifying and modifying ‘stuck points’, which translated in Kiswahili to ‘mawazo mgando’. Stuck points are beliefs about why gender-based violence happened and its impact on the individual and the world around them that often lead to self-blame (Resick & Schnicke, 1992). A 12-session protocol was previously found to be effective in reducing depressive, anxiety, and post-traumatic stress symptoms as well as improving functioning among survivors of gender-based violence in the eastern DRC (Bass et al., 2013). At the time of the study, this was the only randomized controlled trial of a psychological intervention that was found to be effective in reducing mental health problems for survivors of gender-based violence in a low-resource, conflict setting (Tol, Stavrou, Greene, Mergenthaler, van Ommeren, et al., 2013). Furthermore, the majority of refugees in Nyarugusu at the time the study was initiated were from the eastern DRC and thus we anticipated that this intervention would be appropriate and adaptable to this population and context. To improve feasibility for the refugee camp context, we built on a more recently developed 6-session protocol of cognitive processing therapy for the Nguvu intervention (Nixon, 2012; Nixon & Best, 2008).

In addition, we aimed to integrate an advocacy and empowerment component into the Nguvu intervention. Advocacy counselling focuses on increasing autonomy, empowerment, and strengthening linkages to community services by supporting survivors in exploring potential strategies that are supported by the facilitator through safety planning and goal setting (Tiwari et al., 2010; Trabold et al., 2020). At the time of the study, only one randomized controlled trial showed impacts of an advocacy intervention among survivors of intimate partner violence in a non-Western setting (Tiwari et al., 2010), which we combined with cognitive processing therapy to develop the Nguvu intervention. Elements from both treatments were combined into an eight-session intervention. The first session, focused on empowerment, a danger assessment, and developing a safety plan, was delivered in an individual format. Subsequent group sessions (with 6–13 women) focused on introducing the relationship between ‘stuck points and feelings’, learning safety and trust, improving self-esteem, and reviewing the safety plan and helpful coping strategies (Greene et al., 2019).

Implementation approach

In addition to shortening the intervention, several aspects of the delivery and implementation of Nguvu were designed to improve its feasibility in the context of Nyarugusu. This study employed a task-sharing model by training lay refugee incentive workers who were already working with the humanitarian partner organization’s women’s empowerment and protection programme in Nyarugusu to facilitate the intervention in pairs. They had varying levels of experience working with protection and more generic psychosocial support programmes in the camp, but lacked dedicated training in psychological interventions. The facilitators received training from experts in trauma-informed psychological and intimate partner violence interventions. Details of the intervention and training are provided elsewhere (Greene et al., 2019; Tol et al., 2017). The implementation approach was guided by input from a community advisory board that was assembled during the formative phase of the research. The community advisory board consisted of female refugee incentive workers who were involved in a range of programmes that were related to health and protection (e.g. SGBV, community mobilizers/outreach, community health workers, nurses/midwives, legal support), community and religious leaders, and traditional healers.

Women with past-year histories of intimate partner violence and currently elevated levels of psychological distress were recruited from women’s groups and randomized to receive Nguvu or usual care through the gender-based violence response programme (Greene et al., 2021; UNHCR, 2016). The gender-based violence response programme that existed at the time of the study consisted of case management (including basic counselling) and referrals to protection, medical, or legal services, as well as preventive interventions, including community awareness raising and dialogue sessions, establishing men’s accountability groups, among other activities(UNHCR, 2016). Women in the Nguvu study condition were able to access any other usual care services available in Nyarugusu refugee camp, including those for women seeking support for intimate partner violence. Nguvu was designed to be implemented within protection services offered by an existing humanitarian agency that had a long history of implementing protection services within Nyarugusu Refugee Camp. Following the last quantitative assessment, we conducted a qualitative process evaluation – the focus of this paper.

Objectives

The overall objective of this process evaluation is to examine the relevance, acceptability, feasibility, and impact of implementing the Nguvu intervention in a refugee setting from the perspective of intervention stakeholders.

Materials and methods

Participants and procedures

In December 2017, we conducted 29 semi-structured interviews with Nguvu participants, staff members, community advisory board members, and humanitarian practitioners to explore the implementation of the Nguvu intervention. Respondents were selected to maximize variation in perspectives on the implementation of Nguvu. We recruited ten Nguvu participants to complete the process evaluation interview at the end of programme implementation. Nguvu process evaluation participants were selected among those who were randomly assigned to receive the Nguvu intervention as part of the feasibility trial. Within this group we randomly selected one high attender (i.e. more than six sessions) and one low attender (i.e. less than four sessions) from each facilitator pair (k = 5 pairs). We enrolled all ten facilitators, the clinical supervisor of the facilitators, the research supervisor/project coordinator, four members of the community advisory board, and three representatives of the partner implementing agency to participate. Interviews were conducted in Kiswahili by trained research assistants with expertise in ethnographic interviews. Interview guides were initially translated from English into Kiswahili by a bilingual member of the research team with language adaptations made to improve comprehensibility during the research assistant training. All procedures were reviewed and approved by the institutional review boards at the Johns Hopkins Bloomberg School of Public Health (IRB0007219), the Muhimbili University of Health and Allied Sciences (2014-10-27/AEC/Vol.X/56), and the Tanzania National Institute for Medical Research (NIMR/HQ/R.8a/Vol.IX/2016).

Measures

The interview guide was designed to capture information about the barriers and facilitators to receipt, delivery, and implementation of the Nguvu intervention (Table 1) (Munodawafa et al., 2017; Murray et al., 2014). As suggested by the UK Medical Research Council’s framework for process evaluations, we inquired about factors related to context, implementation, and mechanisms of impact (Moore et al., 2015). The interview questions administered to Nguvu participants focused on the risks (perceived and experienced) associated with attending Nguvu activities, perceived impacts of the intervention, unexpected consequences and outcomes of participation in Nguvu, barriers and facilitators of attendance, and intervention acceptability. The interview questions administered to facilitators and clinical supervisors focused on the perceived impact of the intervention on participants, challenges associated with delivering the intervention, participant adherence and engagement, unexpected outcomes and consequences of implementing the Nguvu intervention, logistical challenges associated with scheduling and organizing intervention sessions, facilitator motivation, sufficiency of training and preparation, fidelity, and coordination. The interview questions administered to the project management team, which included the project coordinator, members of the community advisory board, and representatives from partner agencies, focused on implementation challenges, perceptions about acceptability and feasibility of this multi-sectoral approach, potential for sustainability, and recommendations for future programming.

Table 1. Summary of process evaluation interview guide aims and focal areas by participant type

Aim	Perspectives	UK Medical Research Council’s Framework for Process Evaluation Domains
1. To explore barriers/facilitators to participation in the Nguvu programme	− Participants (n = 10; 5 high and 5 low attenders; 2 participants from each facilitator pair)	Context − Barriers and enablers of session attendance − Acceptability and comprehensibility of the Nguvu programme for women in Nyarugusu Implementation − Organization of Nguvu sessions − Mechanism of impact − Risks (perceived or experienced) associated with Nguvu participation − Perceived mechanism of changes in wellbeing and violence − Unexpected outcomes or consequences of participation
2. To explore barriers/facilitators to delivering Nguvu sessions	− Facilitators (n = 10) − Clinical Supervisor (n = 1)	Context − External contextual factors affecting Nguvu sessions − Perceived acceptability, feasibility and potential sustainability of Nguvu Implementation − Implementation challenges − Facilitator motivation − Sufficiency of training and preparation − Fidelity to the Nguvu intervention − Coordination and management of Nguvu tasks with other IRC work − Mechanism of impact − Perceived mechanisms of change in participant wellbeing and violence − Participant adherence and engagement − Unexpected outcomes or consequences of delivering the Nguvu intervention
3. To explore barriers/facilitators to implementing Nguvu research and intervention activities in Nyarugusu	− Nguvu Project Coordinator (n = 1) − Implementing agency representatives (n = 3) − Community Advisory Board Members (n = 4)	Context − Structural challenges, particularly those unique to refugee camp setting − Impact of unforeseen challenges on implementation − Perceived acceptability, feasibility and potential sustainability of Nguvu programme − How external factors affected implementation of Nguvu Implementation − General challenges and successes − Coordination and communication within IRC and across partner agencies − Recommendations for improving implementation of future projects

Analysis

Data analysis involved both inductive and deductive approaches. Using a thematic analysis approach for the inductive component of the analysis, two independent coders (MCG, TS) read all interview transcripts and developed preliminary codes and themes describing factors that related to intervention implementation. The codes were harmonized through discussion. After developing the preliminary codebook, the coders reviewed the transcripts and applied the final codes independently to all study interviews. Discrepancies in coding were resolved through discussion. Then, these themes were mapped onto the process evaluation outcomes (deductive approach): relevance, acceptability, feasibility, and impact. Determinants of implementation were organized within these outcomes according to the UK Medical Research Council Framework and included context, implementation, and mechanisms of impact. All analyses were conducted using NVivo.

Results

Relevance of the Nguvu intervention

Nguvu aligns with local needs and fills a gap in programming

The Nguvu intervention was described as relevant to the context given that it aligned with the needs of the target population and filled a gap in programming within Nyarugusu refugee camp. Respondents described the context of extreme poverty, inequitable gender norms, a history of traumatic events and forced displacement, and significant suffering and hopelessness, which worsened family functioning, violence, and related psychological and physical sequelae. The ubiquity of intimate partner violence in Nyarugusu was surprising to some Nguvu facilitators, while others acknowledged it as a common reality for many women due to the socio-historical context.

One indication of the perceived relevance of the intervention were frequent statements with regard to the need to expand the reach of the intervention beyond the feasibility trial. Many also described the need for Nguvu to extend to nearby refugee camps and communities given the shared needs and relevance. There was a strong sense of collective benefit of the intervention being present in the camp even among women who did not participate. For example, one facilitator described: ‘It is just the support that people are crying for. They say, if you support a woman, you have supported all the women here’. Participants, facilitators, research staff, and members of the community advisory board described the desire for continued delivery of Nguvu sessions to current participants in addition to extending the reach to new beneficiaries. Respondents noted that recognizing the experiences of other women, receiving support and a sense of solidarity, and the appropriateness of a cognitive processing and psychoeducational approach were described as furthering the relevance of the Nguvu intervention.

They were just sitting at home thinking that they are alone, but once they are there they become one. – Facilitator

Suggested adaptations to improve the relevance of Nguvu

There were several suggestions for adaptations to improve the relevance of the Nguvu intervention. First, all respondent types (community advisory board, implementing agency staff, facilitators, research staff, and participants) described the importance of including and involving men in efforts to reduce gender-based violence. These respondents acknowledged the utility of supporting women who have survived intimate partner violence, but also stressed the importance of intervening with men to change their behaviour and reduce the perpetration of gender-based violence . Respondents described the need to create awareness and provide education to men about the impact of intimate partner violence, change social norms that justify violence, and strengthen family relationships and functioning through sessions targeting men individually or through village meetings. Another common suggestion to improve the relevance of the Nguvu intervention was to integrate economic empowerment, microfinance, and skills training components. Respondents often described financial dependence as being a factor that make them more vulnerable to violence. They recommended introducing loan programmes or income-generating activities as strategies to financially empower women, improve social support and community involvement, while also reducing risks of abuse due to economic stress within the household.

Because of the violence here in the camp [and the limited] resources, you find that the economic status of women’s families, husbands, or partners is low. You find that the flour that they have inside is little. Then [the family/husband/partner] comes home, you have not cooked! Now it is a fight. The fight is not to beat the person, it is to mistreat that person while there is no food at home. I would ask Nguvu to empower these women … I would ask Nguvu to support these women so they can provide for themselves financially little by little. – Facilitator

Acceptability of the Nguvu intervention

Acceptability of discussing intimate partner violence with Nguvu staff and in a group setting

The format and content of the Nguvu intervention was generally considered acceptable. Most women were comfortable discussing sensitive topics in a group setting and the environment was described as collaborative, open, and supportive. The acceptability of the format was surprising to some facilitators and implementing agency staff:

Before they were not free to talk about sexual abuse, but now they have gained trust and friendship with the [facilitators], they can talk about it now. It is not Africans’ way to talk about sexual life, so we didn’t expect them to express that way. – Implementing Agency Staff

One facilitator identified that some women during the screening process were uncomfortable reporting their experiences of violence to research staff and, even though they were likely eligible, did not disclose these experiences and were therefore not selected to participate in the intervention. Among participants who were enrolled, the perceived impacts of the intervention, becoming more aware about gender-based violence and mental health, feeling safe and comfortable, and realizing the client-centered nature of the intervention made it more acceptable.

At first it was lack of education, so we did not see the importance of the program, but as we were receiving education things changed and there were no more challenges. But it was hard at first. – Participant

Failing to fulfil financial expectations among facilitators and participants undermined the acceptability of the Nguvu intervention

Some contextual factors and expectations compromised acceptability of the implementation of Nguvu. First, policies in Nyarugusu refugee camp restrict refugees’ right to work. Refugees are allowed to work for small incentives (i.e. refugee incentive workers). To promote sustainability, Nguvu was designed to be integrated into an existing implementing agency’s programmes and incentive worker payments were cost-shared between the Nguvu programme and the partner implementing agency. However, facilitators saw these roles as distinct and expected for Nguvu to supplement their pre-existing incentives. The low allowable incentive worker rates produced inequities and tensions between national staff and refugee workers.

Please gather for us a little bit of incentive, that will help us also. Even if we are refugees, we are humans as you are … it is not that we, the refugees, do not deserve it. – Facilitator

Similarly, there were expectations across stakeholder groups that they would be provided with material or financial support related to their different roles and participation. Most commonly these requests were related to providing participants with small items (e.g. soap, food) or financial support during Nguvu sessions similar to the compensation they received when completing research interviews. The material and financial support was primarily seen as something that would motivate attendance and participation, but also compensating for their time away from family, their homes, and other opportunities to generate income and resources for their family.

Safety, privacy, and confidentiality

The intervention implementation was designed to protect the participant’s privacy and ensure that their partners and community were not aware that they were attending an intervention related to their experience of intimate partner violence given concerns that this may be unacceptable to their partners and may put women at increased risk of violence. Participants and facilitators reported complying with safety procedures, which included referring to the intervention as one related to women’s health that would improve their wellbeing and strengthen their families, which was considered acceptable to their partners. Some women nevertheless struggled with getting permission from their partners to attend sessions. Selecting facilitators that were members of the community meant that they were more trusted by male partners and able to reach women and arrange sessions without producing suspicion or problems in the home. In general, confidentiality was highly regarded among participants and facilitators, and they were not aware of any adverse events amongst themselves or other participants that were related to their participation in Nguvu.

Feasibility of the Nguvu intervention

The refugee camp context: environment, coordination, population characteristics & infrastructure

Contextual factors that challenged the implementation of Nguvu included characteristics of the refugee camp environment, limited infrastructure, competing priorities, population mobility, and cultural factors. The large size of the camp and limited transportation options made it difficult for facilitators and participants to meet for sessions. Transportation and distance barriers were amplified during the rainy season when individuals were often unable to travel by foot, especially without sufficient rain gear. Another challenge related to the inter-institutional environment and coordination of activities. Facilitators and implementing agency staff described feeling as though Nguvu was ‘in competition’ with other programmes. Some essential activities, such as distribution of food and household items, often conflicted with Nguvu sessions making scheduling and coordination of activities difficult.

Another challenge was the issue of distribution. You schedule a day to do your session and then it turns out there is a distribution. That was also a challenge because the times here were not predictable. You find that today is the day for flour, another distribution is for nets, the other distribution is for taking gas, and another distribution is for soap. It was interfering with one another. You schedule [Nguvu] today and it fails. And you find that they need you the next day. – Facilitator

Scheduling challenges were compounded by other coordination and communication difficulties. Participants often had different preferences regarding the timing of intervention sessions. The flexibility of facilitators and late arrivals often resulted in sessions starting and ending late, which was frustrating for participants and often a source of conflict with their families. Furthermore, communication with participants especially those who did not have continued access to cell phones made scheduling challenging. Nguvu groups utilized different strategies to manage communications and scheduling. One group nominated a group member who would be the primary point of contact with facilitators and went door-to-door to remind women in her group of sessions. Communication with other stakeholders, including the community advisory board and implementing agency staff was also challenged by high turnover and communication across different levels of management.

Another barrier to implementation and retention in Nguvu was population mobility. Nguvu was initially implemented during an active resettlement period when many refugees were being resettled to Europe or North America. Simultaneously, there was a large influx of refugees from Burundi and continued arrival of Congolese refugees, which altered activities and resulted in re-allocation of resources to meet the needs of the growing population in Nyarugusu refugee camp.

Limited infrastructure within the camp produced ongoing challenges to implementation. The Nguvu programme did not have an independent office or intervention space and the implementing partner’s spaces were often not available for use due to other activities.

Nguvu intervention facilitators, content & procedures

Other barriers and facilitators to implementation included the recruitment and enrolment process, the complexity of the intervention, fidelity and adherence, facilitator characteristics and motivation, coordination and scheduling, limited resources, and the multi-sectoral integrated implementation model. Facilitators reported that participants were sometimes misclassified during screening due to misreporting of eligibility criteria. Some participants who were uncomfortable responding to questions, specifically about sexual violence, were misclassified as ineligible. A more commonly reported misclassification related to participants falsely reporting experiences of violence and distress in order to be considered eligible and be enrolled in the programme. Facilitators reported that these women, once they found that they would not receive material benefits from participation and did not find the content of the intervention relevant, would often drop out.

Participants who missed sessions or did not understand some of the content often requested that the facilitator repeat information during subsequent sessions so they were able to advance through the Nguvu intervention, which caused the facilitators to ‘end up repeating instead of moving forward’. There were also disruptive group dynamics, including participants leaving sessions to answer phone calls or bringing children to sessions. Nguvu staff arranged for childcare during sessions and facilitators introduced strategies to set expectations in groups to reduce these disruptions. In general, participants were patient with others who required that material be repeated and the facilitators reported that most participants were attentive and engaged. Facilitators developed strategies for assessing comprehension among participants:

Participants were very attentive during the sessions. The challenge was their ability to understand fast as they took longer to understand the sessions, but I was trying to explain to them using different examples for them to understand fast. I tested their understanding during sessions by asking the participants to give examples during our discussion. – Facilitator

Some facilitators also struggled with the more difficult concepts and required additional support and supervision by the clinical supervisor and other facilitators. The confidence of facilitators grew over time as they practiced the material and started working with women. Working with a co-facilitator and the peer supervision process were helpful sources of support and guidance that improved their confidence. Throughout Nguvu, clinical trainers provided refresher trainings, but there remained a desire for ongoing and extended training from experts to bolster their skills at the end of the project. The facilitators were very motivated to master the material. They described both intrinsic and extrinsic motivations for serving as Nguvu facilitators. Many of the facilitators also identified as survivors of intimate partner violence who had experienced distress and reported experiencing benefits of the intervention during training that motivated them to share that healing with other women in their community. Noticing improvements in participants was a very rewarding aspect of being a facilitator. They also noted that facilitating Nguvu sessions allowed them to build relationships and a sense of community with women in their sessions. Facilitator motivation was reinforced by the respect, appreciation, and recognition they received through their new role within their community.

Impact of the Nguvu intervention

Nguvu was consistently described by participants as an intervention that was beneficial for women affected by intimate partner violence and elevated psychological distress in Nyarugusu. Norms related to violence and gender, including justificatory attitudes toward violence and discouraging women from seeking services or supports outside of the family to address gender-based violence, were seen as cultural and contextual factors that hampered effects of the intervention. Facilitators felt that the ‘Nguvu project is ideal to decrease violence even if it will not decrease all of it, but it lowers it and one feels better’.

Perceived impacts of Nguvu extended beyond participants. Almost all facilitators noted the impact that the Nguvu trainings had on them personally. These impacts ranged from changes in their mood, feeling empowered with the new knowledge and skills they possessed, and improving their family functioning. Participants and facilitators also described the social benefits they experienced.

Nguvu was perceived as achieving these benefits through several mechanisms. First, simply building awareness of the relationship between intimate partner violence and psychological distress made women feel more comfortable discussing their experience, helped women recognize mental health problems and violence, provided tools to protect themselves and improve their wellbeing, and reduced self-blame.

After seeing the difficulties facing us as women, we realize that an event might occur, but it is not our mistake. I used to blame myself and thought that maybe that event happens because of me. I used to put those ideas in my brain. I blamed myself, but after I have been healed through the training, I found I was not guilty for what happened to me. Why do I put static ideas in my mind and blame myself? All of those things helped me a lot … realizing what static ideas are, removing static ideas, and understanding why you blame yourself for an event which already occurred. That cured me a lot. - Facilitator

Practical skills such as managing stuck points, breathing, and relaxation techniques, and understanding the relationship between thoughts and feelings were described as mechanisms by which Nguvu improved mental health. Nguvu was described by participants as ‘molding the brain’, ‘bringing peace in our homes’, bringing ‘changes to our health and mind’, and that it ‘changes us and makes us learn how to live’.

Discussion

Following a feasibility trial, we conducted 29 semi-structured interviews to understand the perspectives of stakeholders with regard to relevance, feasibility, acceptability, and impact of a novel intervention that integrated elements of cognitive processing therapy and advocacy counselling (Nguvu) amongst Congolese refugee women facing IPV in a Tanzanian refugee camp. Qualitative interviews with refugees, practitioners, Nguvu project staff, and community advisory board members reinforced the complexity and pervasiveness of IPV in refugee settings. Similar to previous research, respondents highlighted the role of armed conflict, trauma, instability, changing gender roles and norms, economic stress, and changing social dynamics in exacerbating gender inequality and perpetration of gender-based violence (Fry et al., 2019; Horn et al., 2014; Wachter et al., 2018). Nguvu was found to be a relevant and acceptable approach to addressing these related health and protection challenges in a refugee camp context. The Nguvu intervention was perceived to improve mental health by reducing self-blame through identifying and modifying ‘static ideas’ (i.e. stuck points), increasing awareness and solidarity, strengthening social support, and providing skills to manage risky situations as well as negative thoughts, and feelings. Despite these perceived benefits and the fit of the approach within the refugee camp context, we identified significant implementation challenges that compromised the feasibility of the intervention and should be considered before future delivery and testing.

Respondents provided several suggested adaptations to Nguvu primarily related to the components focused on intimate partner violence. First, participants noted a need violence prevention programming in Nyarugusu that targets and/or includes men. As part of the initial design of the intervention, we purposefully worked in an area where IPV prevention and other response programming was already ongoing – according to a detailed mapping exercise (Greene et al., 2019). Some of the prevention activities that were separately implemented, but available while Nguvu was being delivered included men’s discussion groups. Given that these activities were already ongoing, and we were concerned that involving men in the Nguvu group activities may carry risks for retaliation, we opted not to include the partners of participants. However, participants in this qualitative study still encouraged additional initiatives be focused on men. For future programme development therefore it is important to explore (1) whether implementation of Nguvu alongside existing prevention activities sufficiently covers the prevention needs of Nguvu participants or whether prevention activities should be incorporated within the Nguvu intervention; and (2) whether it is feasible to include safe and survivor-centered ways to induce behaviour change in IPV perpetrating partners, which remains controversial and insufficiently researched (Travers et al., 2021). Recent evidence from Zambia showed that providing psychological intervention activities to address mental health of men (e.g. alcohol misuse and posttraumatic stress) can reduce IPV perpetration (Murray et al., 2020) and have sustained effects (Kane et al., 2021).

Another consistent recommendation was to integrate economic empowerment initiatives into gender-based violence programming. Existing research on the impacts of economic empowerment on intimate partner violence is mixed. Some studies have reported increased risk of intimate partner violence by challenging traditional gender roles and making men feel economically vulnerable (Slegh et al., 2013; Tappis et al., 2016). However, reflections from economic interventions in Rwanda and Uganda have found that complementing economic empowerment programmes for women with strategies to engage men may mitigate these risks (Green et al., 2015; Slegh et al., 2013). Integrating economic components and providing material or financial incentives was also suggested as a strategy to promote engagement and acceptability of Nguvu, which is consistent with previous process evaluations of psychosocial interventions in refugee settings (Tol et al., 2018). Given national refugee policy prohibiting work for refugees, there are limited opportunities for sustainable economic empowerment.

This recommendation from participants raises the broader question of how many elements can feasibly be integrated into one intervention package. Our starting point was the development of an intervention that addressed both IPV and mental health, issues traditionally targeted separately in the humanitarian protection and health sectors respectively. IPV in low-resource refugee settings is a complex issue with multiple determinants and mutually reinforcing relationships across various modifiable risk factors (Logie et al., 2019; Rubenstein et al., 2020; Timshel et al., 2017; Tol, 2020). Actors in various humanitarian sectors have recognized the need for multi-sectoral, integrated interventions (James et al., 2020; Logie et al., 2019; Rubenstein et al., 2020; Timshel et al., 2017; Tol et al., 2020; Welton-Mitchell et al., 2018). However, multi-sectoral, integrated interventions often face logistical limitations – an observation highlighted by the finding from this study. One potentially promising future direction for efforts aiming to address mental health and its social determinants is to identify more specific psychological targets that are at the core of vicious cycles (e.g. self-blame in this study). This goal can be furthered both by participatory techniques that map the full complexity of violence and other social determinants (e.g. jointly developed theories of change, or community-based system dynamics modelling) as well as more basic research into the specific mechanisms that tie social determinants and mental health (e.g. particular emotion regulation processes or cognitive biases).

The process evaluation revealed several critical implementation challenges related to human resources, coordination, and contextual barriers. This study employed a task-sharing approach by training refugee women from the community without prior training in psychosocial interventions to deliver Nguvu, which was integrated within existing protection services. This implementation approach was designed to promote feasibility and longer-term sustainability within a context lacking specialists. However, the low (national policy-mandated) incentive rate for refugee workers produced tensions between refugee and national staff, competition between implementing agencies, and facilitators struggled to justify absorbing new responsibilities into an existing role for which they received minimal compensation. Sustainable and equitable models of compensating the refugee workforce that are rights-based, fair, build self-reliance, and can be adapted to diverse contexts and policy landscapes are needed (Morris & Voon, 2014). Competing priorities posed a major challenge to delivering the Nguvu group sessions, which has also been reported as a barrier to implementation in previous studies of psychosocial interventions in refugee settings (Tol et al., 2018). The distribution of essential goods and services, such as food or bed nets, was largely unpredictable. Strategies to improve system-level coordination and the flexibility of intervention delivery to overcome these challenges without compromising fidelity are needed to improve the feasibility of multi-session, group psychosocial interventions in refugee contexts. Contextual factors relating to a challenging environmental context, poor infrastructure, and limited resources made it difficult to implement the Nguvu intervention (McAlpine et al., 2020). Addressing these challenges often defaulted to facilitators who were resourceful, but often had to assume responsibilities outside of their scope of work to ensure that sessions could be delivered safely and reliably. The motivation and ownership of the intervention by facilitators were essential to the implementation of Nguvu. However, greater buy-in and ownership across stakeholder levels is needed if Nguvu is to be replicated and scaled in other contexts.

These implementation challenges add to a growing literature on psychosocial intervention implementation challenges in low-resource and humanitarian settings. Similar to our study, previous process evaluations that have applied the UK Medical Research Council Framework have noted the important role of external factors and the environmental context, cultural norms, the characteristics of intervention personnel, including lay health workers, the inclusion of community members and advisors in implementation decision-making, and the importance of the fit of the intervention to the culture and context (Munodawafa et al., 2018). While further examination of these determinants of implementation is needed, including quantitative measurement and triangulation of implementation outcomes, it is also imperative that researchers work with a range of stakeholders, including refugee communities, practitioners, and policymakers, to design and test strategies to overcome these challenges and improve the relevance, acceptability, and feasibility of mental health and protection interventions in humanitarian settings.

Conclusions

Generally, stakeholders saw the value in the Nguvu integrated approach to address IPV and mental health problems and felt that with adaptations, the intervention could have an important impact on the wellbeing of refugee women and their families. This process evaluation revealed opportunities to explore improvements to the content and implementation of the intervention, including further integration of activities focused on male perpetrators, integration of economic activities, and a number of implementation challenges related to the challenging (low-resource and dynamic) context and multi-sectoral, multi-stakeholder nature of the project. Future research is needed to further tailor strategies for increasing the reach, adoption, scalability, and sustainability of Nguvu in a humanitarian setting. These implementation outcomes should be considered alongside efforts to evaluate the effectiveness of Nguvu adapted according to recommendations presented in this process evaluation and other evidence for improving mental health and reducing IPV in humanitarian contexts.

Acknowledgements

We would like to thank UNHCR and IRC Kasulu field office staff for their assistance in field operations. We are very grateful to the refugee incentive workers who served as research assistants and intervention facilitators, without whom this research would not have been possible. We would also like to thank the women who participated in this study for generously sharing their time and experiences with us.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statements

Data will be made available upon reasonable request to the corresponding author ([email protected]).

Additional information

Funding

The research was funded by Elrha's (Enhancing Learning and Research for Humanitarian Assistance) Research for Health in Humanitarian Crises (R2HC) Programme, which aims to improve health outcomes by strengthening the evidence base for public health interventions in humanitarian crises. R2HC is funded by the Foreign, Commonwealth & Development Office (FCDO), Wellcome, and the UK National Institute for Health Research (NIHR). Visit elrha.org for more information about Elrha's work to improve humanitarian outcomes through research, innovation, and partnership.