Rates of mental health disorders in Cambodia are notably higher than in other low-or middle-income countries (1). Prevalence studies have estimated the probable rates of common mental health disorders in Cambodia stand at 16.7% for depression, 27.4% for anxiety and 7.6% for post-traumatic stress (PTSD) (2). Compared with global figures, this is approximately five times higher than the rest of the world (3). Following the Khmer Rouge period (1975–1979), when over one million people were killed, mental health services in Cambodia started to gradually re-emerge during the 1990s to tackle the crisis. Faced not only with a population living in the aftermath of genocide and trauma; mental health services were also confronted with a society struggling with inadequate access to food, financial stress, familial issues, and the fear of landmine injuries (4) However, the lack of resources and suitable infrastructure meant these mental health needs have largely gone unaddressed and unmet e.g., the number of psychiatrists and psychiatric nurses is approximately, 0.33 and 0.26 per 100,000 people (4). In addition, the total number of inpatient beds is 15, which is approximately 0.1 beds per 100,000 people (5). This is further compounded when one looks at the additional barriers to accessing community mental services for those with disabilities (6, 7). This is despite evidence showing elevated psychological distress and depression among Cambodia adults with physical disabilities (8).

Interventions for mental health disorders are clearly needed in Cambodia, however evidence suggests there is an absence of appropriate numbers of mental health professionals and mental health funding allocations (9). In recent years more Cambodian nationals are being trained as psychologists or mental health therapists, however the results of a qualitative study (n = 95) indicates that for these nationals, using Western-based therapy presents a myriad of challenges at individual, agency, family, cultural and community levels (10). It can be seen as infantilising, demeaning and ethnocentric when explaining westernised concepts of mental health. As such, there is a need for imported therapeutic models to be more culturally responsive to generate acceptance and impact in Cambodia (10). Innovative solutions are clearly required to address mental health issues in Cambodia. In particular, there is a definitive need for the Cambodian government to develop effective research and development strategies, and to devise a clear policy framework for health research to support projects which aim to train Cambodia nationals to deliver co-produced interventions to aid in psychosocial support (11–14). Given that international services addressing mental health issues have been criticized for cultural insensitivity and an inability to integrate with local communities (15), the best solutions are arguably those which are developed within the contexts in which they will operate (9). However, it remains unclear which mental health psychological and social interventions are the most effective in a Cambodian context augmented by the lack of cultural congruency and poor quality research studies (16). In a systematic review of psychological and social interventions for mental health issues and disorders in Southeast Asia, Maddock et al. (17), found some promising preliminary evidence that meditation and/or yoga could reduce the prevalence of symptoms of depression, anxiety and depressive disorders, and improve well-being in depressed clients (18, 19). These findings correlate with Cramer et al. (20) who conducted a systematic review of 12 Randomised Controlled Trials (RCTs) examining the effectiveness of mind–body yoga and relaxation interventions, however as Maddock et al. (17) state although the findings are encouraging more research and in particular RCTs of a higher quality are needed in a South East Asian context to consolidate the evidence base. Interestingly, Maddock et al. (17) also found that the highest quality, and most promising evidence came from the evaluations of lay delivered interventions. This evidence demonstrates the feasibility and potential sustainability of implementing such interventions in resource constrained contexts such as Cambodia (17). In light of the significant challenges facing mental health services in Cambodia, there is a need to develop low cost mental health interventions that could be delivered at scale within local communities by non-professionals (21). This process is known as “task shifting” and is recognised as a viable strategy for increasing access to mental health services for underserved populations (22–24).

Peer support in a mental health context can be described as involving people with personal experience of mental health problems in the delivery of support services (25). Peer support can include one to one support delivered by a person with experience of mental health problems, mutual support groups and peer-led mental health services (26). The uniqueness of peer support is that the individuals share similar characteristics or experiences as their peers within the group and thereby can create a connection which facilitates sharing of challenges and experiences as an “insider” (25). Several RCTs support the view that peer support programs can lead to improvements in empowerment, quality of life and depression and have a positive impact on hope and personal recovery (27, 28).

Ibrahim et al. (29) highlight that peer support significantly contributes to promoting recovery and reducing stigma through shared lived experiences, for example, Lyons et al. (30) conducted a systematic review and meta-analysis, including data from 2,131 participants across eight trials, showing that group peer support interventions lead to small but significant improvements in personal recovery outcomes, such as hope and empowerment. Further, Cooper et al. (31) conducted an umbrella review summarizing evidence from 35 reviews on peer support approaches. Despite mixed effectiveness results, their findings indicated that peer support might improve depression symptoms, self-efficacy, and recovery. They also identified factors promoting successful implementation, such as adequate training, supervision, a recovery-oriented workplace, and effective collaboration. As such, given the paucity of high-quality research and a heterogeneity between studies and types of peer support delivered it is challenging to conclude on the effectiveness of peer support (26, 32–34). It is also unclear as to whether these interventions will address depression or anxiety symptoms because of trauma. Nevertheless, given the findings of a recent systematic review and realist synthesis, peer support may be a viable option for low-middle income countries such as Cambodia as it is an economically viable way to engage underserved populations at a community level given the impact of historical traumas on community mental health (35).

This paper reports on the results of a mental health intervention development study using a mixed methods for Cambodian adults with physical disabilities using both face-to-face and online delivery methods. The study was designed by a multi-disciplinary and multinational collaboration between physical and mental health academics/practitioners across social work, psychology, physical rehabilitation (prosthetics and orthotics) and staff at the Cambodian School of Prosthetics and Orthotics (Exceed Worldwide). Exceed Worldwide trains Prosthetist Orthotists and Technicians at three separate clinics across Cambodia and provides free physical rehabilitation services, to those most in need.

Using a mixed methods design, the research team developed and implemented an 8-week peer-led intervention (known as a Friendship Group (FG) for Cambodian adults with physical disabilities using both face-to-face and online delivery methods (Trial no: Trial Registration: NCT05725707 - retrospectively registered, 23/02/2023). Included within study objectives are a number of questions to explore whether a larger scale trial is feasible. As such, the design of this study is straddled between early intervention development and feasibility trial. Ethical approval was granted by the University's ethics committee (REF: 028_2021). Where relevant, the authors followed the CONSORT checklist for feasibility trials (36) as well as the guidance on complex intervention development by O'Cathain and colleagues (37) (see Table 1). Scores across both delivery methods were compared at baseline and post-intervention and included measures of psychological distress and PTSD as outcomes, and facts of mindfulness, worry and rumination as proposed mediators of change.

The selection of outcome measures for this study was carefully considered to ensure relevance and appropriateness for the Cambodian context, as well as to comprehensively capture the impact of the intervention on mental health outcomes. Each measure was chosen based on its validity, reliability, and cultural relevance to the target population. As indicated above, all participants in the study completed a screening survey to determine suitability. This included some basic socio-demographic information as well a series of validated psychometric scales described below.

The data were screened for missing values and any error cases, such as extreme outliers. There were no missing values or error cases on any of the outcomes. The Wilcoxon Signed-Rank test was used to assess changes in pre-post scores for psychological distress, PTSD, worry, rumination and facets of mindfulness. With relatively small group sample sizes, differences between group allocations could confound intervention outcome measurement and it is therefore important in such circumstances to compare baseline differences between groups during the analysis (53). In order to do so, this study used analysis of covariance (ANCOVA), as this method increases the power of analyses of group allocations by reducing any unintentional baseline differences due to the allocation process, which increases a study's capacity to obtain a valid estimation of the intervention effect between groups (54, 55). In order to support balance in prognosis, intention-to-treat analysis was also employed (56). This allowed a pragmatic estimate of the benefit of the friendship group, rather than of its potential benefit in patients who receive treatment exactly as planned to be attained (56). No p-value adjustment was be made for multiple comparisons, as controlling for Type 1 error in this manner is likely to increase the chances of Type 2 error (57).

Qualitative data was used by the research team to understand the feasibility and acceptability of FG's from the perspective of those who engaged in the intervention. Five focus group interviews were conducted 1–2 weeks post-intervention - four with FG participants (n = 21) and one with FG facilitators (n = 8). Audio recording took place during each interview, and this was transcribed verbatim.

Given the small sample size and non-randomized design of this feasibility study, the statistical analysis is presented tentatively. Descriptive statistics were used to summarise baseline participant demographics and psychometric scores. Sum scores were determined for each construct within the screening survey. Wilcoxon Ranked Signed tests and two-way ANCOVAs were conducted to compare baseline and post-intervention scores for primary and secondary outcome measures.

Focus group data were analysed thematically (58) and followed a semi structured format that included core topics concerning – (1) general acceptability of the friendship group; (2) barriers to engagement (practical, material or psychological); (3) attitudinal or behavioural change during or post engagement; (4) facilitators or inhibitors of intervention effectiveness. Initially, transcripts from focus group interviews were read multiple times for familiarisation. Open coding was conducted on each transcript and labels that shared similarities were sorted into clusters and themes by two members of the research theme before being discussed among the wider research team. Through collaborative discussions, these codes were refined and grouped into overarching themes. The themes were then reviewed and defined to ensure they accurately represented participants' experiences. This systematic approach ensured a comprehensive analysis of the qualitative data.

RO1: To compare uptake and retention between face-to-face and online delivery, aiming for at least 70% enrollment of eligible participants and a dropout rate not exceeding 30%.

Following screening, 52 met inclusion and were invited to attend weekly FG's. Of those who had consented and agreed to take part, 22 were allocated across the two online groups and 19 participants were allocated across the two face-to-face groups (see Figure 3). Post-intervention, data was available for 18 participants in the online group (8 females and 10 males) and 14 participants in the face-to-face group (8 females and 6 males). This resulted in an uptake of 78% and an overall attrition rate of 22% with 26% dropout in face-to-face groups and 18% in online groups (Table 5).

RO2: To explore the potential benefit of weekly friendship groups in improving psychological distress.

A Wilcoxon signed-ranks test showed a statistically significant decrease in psychological distress scores [Z = −3.808, p < .001] from pre [Mdn = 20, IQR = 16.5–25.5] to post [Mdn = 16, IQR = 14–18.5] Friendship group intervention. A Wilcoxon signed-ranks test also showed a statistically significant decrease in PTSD scores [Z = −2.239, p < .025] from pre [Mdn = 4, IQR = 3–5] to post [Mdn = 3, IQR = 2.75–4] Friendship group intervention. There was also a statistically significant decrease in worry scores [Z = −3.904, p < .001] from pre [Mdn = 5, IQR = 3.5–6.5] to post [Mdn = 3, IQR = 3–4] intervention. There were no statistically significant changes in facets of mindfulness [Z = −1.71, p < .09] or rumination [Z = −1.9, p < .06] from pre to post Friendship group.

RO3: To compare pre to post outcome data from those who attended the face-to-face group and those who attended the online group.

There were no significant group differences between the face to face and online groups in psychological distress [F (1, 39) = 2.38, p = 0.13, η² = .06], PTSD [F (1, 13) = .04, p = 0.84, η² = .003], worry [F (1, 39) = .15, p = 0.7, η² = .004], rumination [F (1, 39) = .23, p = 0.63, η² = .006] and facets of mindfulness [F (1, 39) = .66, p = 0.42, η² = .017] from pre to post Friendship programme.

RO5: To train (task-shift) local Cambodia staff that work in the Cambodian School of Prosthetics and Orthotics to deliver a weekly “friendship group” intervention for stress, low mood, and anxiety.

RO6: To gather qualitative feedback from participants and facilitators to understand the acceptability and potential barriers to the intervention.

Five focus group interviews took place involving a total of 29 participants. Four focus groups took place with FG members only (n = 21) and one focus group was for facilitators only (n = 8). Several interconnected themes emerged during these interviews. These included the mental health benefits of Friendship Groups as conceptualised through knowledge acquisition (psycho-education and learning from others), skill development (mindfulness practices and learning to talk about feelings) and peer support. Issues of intervention acceptability and comparison between online and face-to-face support were also highlighted (Figure 4).

This intervention development study has shown promise in relation to a number of key areas. Initial findings show reduced psychological distress, levels of worry and PC-PTSD-5 scores for those attending the 8-week FG intervention (see logic model – Figure 5). There were no significant differences in primary outcome scores across online and face-to-face groups. Rumination and levels of facets of mindfulness showed a mean reduction however, these were not statistically significant. Overall, the intervention achieved an uptake rate of 78%, with an attrition rate of 22% across all four groups. Interestingly, the dropout rate was higher in the face-to-face groups (26%) compared to the online groups (18%). The lower attrition rate in the online Friendship Groups (FGs) was surprising given the technological barriers; however, this may reflect the scarcity of local in-person mental health services. As a result, the study's uptake rate of 78% exceeds the acceptable threshold of 70% for feasibility studies (36, 62). Additionally, the retention rate of 78%, with a dropout rate of 22%, falls within the acceptable range of up to 30% (63, 64). These metrics indicate that our intervention is both feasible and acceptable to participants.

Quantitative data indicated positive outcomes in relation to worry and PTSD symptoms. This was explored within focus group data which indicated a potential combination of factors, such as knowledge acquisition, skills development, and peer support. Our findings align with those of Ibrahim et al. (29), who identified that organizational culture, adequate training, and clear role definitions are critical for the successful implementation of peer support work. Gaining new knowledge regarding the impact of mental health issues as well as learning skills to support self-care (e.g., mindfulness practice and communicating how they feel to others) appeared to help to build confidence among group members. The peer support component provided a useful context in which these experiences and new knowledge could be shared and validated. As such, the focus on shared learning and practical advice created an environment more akin to a training course than a mental health intervention (“after I have this training” and “so I took this mental health course”). This may illustrate how change is influenced through social learning theory (SLT) (65) considering that support networks, the environment and context within which support is provided have proven to be vitally important in the success of this intervention. In turn, this may have increased the acceptability of the FG intervention by reducing perceived stigma.

Common emotions cited by group members following the intervention were increased feelings of calm and reduced feelings of anger. This supports some of the quantitative data that showed reductions in stress as well as PTSD, anger being a common emotion (66, 67). Cooper et al. (31) highlighted the effectiveness of peer support in improving depression symptoms, self-efficacy, and recovery, which parallels our observation of reduced psychological distress and worry among participants. In addition, while the intervention did not target PTSD symptoms specifically, the shared experience of a physical disability may have enabled some participants to re-evaluate their own experience and the impact it has had. This new information may have reassured them that are “not alone” and bring new meaning to the event – a key step during PTSD treatment (68).

Quantitative data showed no significant differences in psychological distress scores across both delivery formats, with both showing positive effects. Qualitative data did reveal some additional barriers in regard to online delivery with disruptions in internet connection and background noise meaning that groups sometimes felt rushed and less opportunity for in-depth discussion. In most cases however, solutions to connection issues where solved over time as members got more familiar with the technology and facilitators took a more active role in encouraging attendance in advance of meeting start time. Interestingly, these initial issues did not seem to affect retention and attendance, with attrition in online groups (14%) lower than face-to-face groups (32%). An additional positive of online groups was the ability to access services that were unable available locally and this was acknowledged by several members.

There appeared a high level of acceptability across group members and facilitators during the study indicating potential feasibility of a larger scale research project. Those who attended online and face-to-face FG's reported positive outcomes and some indicated a willingness to expand the service or continue on after the intervention period. Lyons et al. (30) found that group peer support interventions lead to significant improvements in personal recovery outcomes, such as hope and empowerment, which supports our findings of increased participant confidence and reduced psychological distress. The label of “Friendship Group” rather than “mental health support group” appeared culturally appropriate and the incorporation of techniques, such as mindfulness, behavioural activation and peer support appear to translate well within the local context. FG facilitators grew in confidence as the intervention progressed and appeared to find it a beneficial experience. There were no reports of unnecessary or additional burden or harm and this included both the intervention itself as well as the survey questionnaire participants were asked to complete.

In conclusion, the friendship group intervention delivered in both online and face-to-face formats appears feasible and acceptable for adults with physical disabilities within the Cambodian context. Initial data revealed positive findings in terms of reduction in psychological distress, worry and PTSD symptoms as well increased feeling as calm. The combination of new knowledge and skills within the context of peer support appeared as the main drivers behind the benefits that members experienced. While the role of wider societal stigma wasn't explicitly acknowledged by group members, in relation to both physical and mental health disability. The interconnected nature of new knowledge and peer support appear to reduce its impact. This may have implications for future recruitment and retention. By adopting a peer support model and training community-based workers, the research team was able to deliver this mental health intervention at a much lower cost. Future studies would do well to build upon this work by using a more robust (controlled) experimental research design and recruiting a larger sample. Moreover, careful consideration of online platforms for connectivity and familiarity among participants is necessary, alongside ensuring a quiet space for both face-to-face and online group delivery. Expanding the length of the initial orientation session for online members and framing Friendship Groups as training courses may lessen associated stigma and improve engagement. Including additional content around skills development and minimising gaps between initial screening and intervention delivery are also important considerations.