Development of a Systematic Framework for Cultural Adaptation and Contextualization of Evidence-Based Psychological Interventions

Manaswi Sangraula (  sangraulamanaswi@gmail.com ) Transcultural Psychosocial Organization (TPO) Nepal https://orcid.org/0000-0003-0241-7945 Brandon A. Kohrt George Washington University Renasha Ghimire Transcultural Psychosocial Organization (TPO) Nepal Pragya Shrestha Transcultural Psychosocial Organization (TPO) Nepal Nagendra P. Luitel Transcultural Psychosocial Organization (TPO) Nepal Edith Van't Hof Department of Mental Health and Substance Use, World Health Organization Katie Dawson University of New South Wales Mark JD Jordans King's College London


Background
Prevalence of psychological distress is high amongst populations in low and middle-income countries (LMICs) that are affected by con ict, poverty, and violence (1,2). LMICs are often unable to cope with such high rates of distress, due to fragmented health systems and limited number of mental health professionals (3). In such contexts, innovative interventions are especially important. Innovative interventions are de ned as interventions that can be scaled-up to reach larger populations, can be delivered through routine health care, and utilize concepts of task-sharing (4). Along with expanding the reach of innovative interventions, cultural compatibility must also be addressed. Literature has shown that translating interventions directly from one culture to another without any other changes may reduce the effectiveness of the intervention and could even cause harm (5,6). Therefore, a rigorous cultural adaptation process to t the cultural context is essential prior to intervention implementation.
Effective cultural adaptation methods maintain the core principles of the treatment and should be based on culturallygrounded evidence and a clear, thoughtful process. Adapting evidence-based psychological treatments (EBTs), to incorporate cultural concepts of distress, is an important step in increasing the acceptability and relevance of the treatment and the overall effectiveness of the program (7,8). It is promising that even with small adaptations, psychological interventions, such as problem-solving and behavioral activation for depression as de ned by Western criteria, have been successfully implemented with symptom reduction in diverse populations (9). This evidence highlights the importance of selecting a cultural adaptation framework that utilizes tried-and-tested approaches while clearly guiding change on the delivery and contextualization of the intervention material.
For an EBT to be considered effective in a cultural context, it is also necessary to modify the presentation of intervention content, delivery framework of the treatment and its scalability to increase the acceptability and uptake for the desired population (10). Adaptation is an iterative process and often occurs alongside trial and error during implementation. There is a need to develop a framework that both maintains the delity and the core mechanisms of action of the treatment but acknowledges that a certain amount of program drift is natural (10).

Overview of Cultural Adaptation Frameworks
Existing cultural adaptation frameworks are varied in their theoretical underpinnings and their implementation process for adaptation. Though they may differ in design, the main intention for all cultural adaptation frameworks is to increase the effectiveness of the EBT by making changes that align with the culture of the bene ciary population, while maintaining the components of the evidence-based research that supports the treatment. There have been many attempts to organize the theories and elements suggested by the existing frameworks (11)(12)(13)(14). Findings were summarized from numerous studies and condensed to three main distinctions that need to be balanced when adapting EBTs.
The rst distinction is surface vs. deep adaptations. Surface structure adaptations refer to modifying super cial characteristics to better t preferences of the bene ciary population (15). An example is translating treatment materials to match the native language (16). Deep structure adaptations target cultural values, norms, traditions, beliefs, and the bene ciary population's perceptions of the illness's treatment and etiology. Examples include integrating traditional healing practices in the interventions and deriving intervention activities from local conceptual models of illness. Some cultural adaptations of interventions have used this dichotomy as an explicit framework, while others have used surface vs. deep as an approach and adapted it further (17). Data shows that even surface level adaptations, such as pictorial material depicting participants ethnically similar to bene ciary populations, resulted in discernable improvements in participant retention (18,19) highlighting the importance of even "simple" cultural adaptations.
The second distinction is between the adaptation of core vs. peripheral aspects of the intervention. Core components are the main evidence-based ingredients of an intervention that are integral to the treatment (20). Peripheral components are related to the acceptability and feasibility of the intervention and exist to support the core components and the goals of the treatment. Examples include changes to language and methods of engagement and delivery. While promoting adaptations that are responsive to the needs of the bene ciary population, it is also important to follow the intervention as intended. This delity vs. t distinction must be balanced to promote cultural appeal to the intervention while also following tried and tested methods to increase effectiveness (21). Despite the clarity brought about by identifying three key distinctions in the literature, these aspects of cultural adaptations are not straightforward. For example, common factors in psychotherapy such as therapeutic alliance, changing expectations of personal effectiveness, and encouragement (22,23) could be classi ed as peripheral, or supportive components, but in fact may be core and integral to the success of the treatment.
Other challenges have also been reported in the adaptation literature. Adaptation frameworks are noted to be di cult to implement in real-world settings (20). Most original studies have not described their tools and process, which makes it di cult to replicate their adaptation methods (24). Therefore, a theoretically-grounded and explicit adaptation guide is necessary, especially one that can be applied to various populations. Many adaptation studies have been conducted with ethnic minorities in the United States (25). Though this is an overlooked and often marginalized population, these minorities are based in a high-income country (HIC) where resources such as time and personnel may not be as constrained as in LMICs. These constraints create unique needs, especially in humanitarian settings, where a transparent, thorough, and prescriptive framework is necessary to guide rapid and systematic adaptions.
There is a growing body of psychological interventions being adapted and developed for LMICs and humanitarian settings (26). Problem Management Plus (PM+) is a trans-diagnostic intervention employing empirically supported strategies that can be delivered by non-specialists to adults impaired by distress (27). Two recent randomized controlled trials in Pakistan and Kenya demonstrated the e cacy of individually delivered PM+ (28,29). To make PM + more scalable, cost-effective, and acceptable in different contexts, it is important to test a group version as well (30,31). Group PM + includes several evidence-based strategies that target common mental health disorders, such as depression and anxiety, over ve in-person group sessions.

Aim
Considering the mechanisms of action, delivery, scalability, and quality assurance, we will describe an explicit guide to adapting and contextualizing EBTs. We will then highlight the key adaptations made through this process in the context of Group PM + in Nepal and gather data on how the adapted and contextualized intervention was perceived by participants, community members, and families. By tracking the progress of cultural adaptation from the methodology, implementation, to feedback from the community, we aim to meet the following objectives: 1. Create explicit guidance for cultural adaptation and contextualization that can be applied to various populations with a focus on implementation, scalability, and quality monitoring. 2. Report on the cultural adaptation process of Group PM+.
3. Gather feedback from program stakeholders after implementation to analyze the effectiveness of speci c adaptations. 4. Suggest the most valuable steps in adapting an EBT under time and resource constraints.

Setting
Nepal is a low-income country with a history of internal con ict, political instability, and natural disasters. In 2015, an earthquake resulted in injuries, deaths and displacement. 34.3% and 33.8% of participants in an earthquake affected district scored above the validated cut-off scores for depression and anxiety respectively (32). The country reports that there are 0.52 psychologists and 0.36 psychiatrists for every 100,000 people, mostly working in larger cities and are not easily accessible to those in rural areas (33). The lack of trained specialists and their limited reach to populations around the country makes an intervention with non-specialists a practical and important solution to consider. To date, there have been minimal efforts to adapt EBTs in the Nepal (34). Group PM+ was considered an appropriate intervention to test in this setting due to its scalability and task-sharing approach.
The Group PM+ intervention's adaptation process and feasibility trial was conducted in Sindhuli district, which was heavily impacted by the 2015 earthquake (35,36). Two Village Development Committees (VDCs) were selected for the adaptation process and the feasibility trial to follow. There were no specialized mental health treatment facilities in the district and the closest psychiatric services were approximately six-hours away from the study site. For this reason, local health posts were the rst portal of care and the referral sites for program participants. The adaptation process and implementation of the trial was conducted by the staff of Transcultural Psychosocial Organization (TPO) Nepal based in Kathmandu, Nepal. TPO Nepal was established in 2005 and is a leading mental health service delivery, training, and research organization (37).

Study Methodology
A study was conducted to demonstrate the feasibility and acceptability of Group PM+ in Nepal (36). The pilot feasibility study was registered on ClinicalTrials.gov (NCT03359486). A ten-step adaptation model was followed in preparation for a feasibility and acceptability trial. After the feasibility trial, qualitative interviews were conducted with program stakeholders to gather data on the acceptability and scalability of the adaptations (35). After making improvements to the adaptations, a de nitive Group PM+ trial will be conducted in a ood-affected area of Nepal to test the effectiveness of the program and the adaptations made to support it (38).

Ecological Validity Model (EVM)
We used the Ecological Validity Model (EVM) to guide our adaptation of the Group PM+ intervention (39,40). The EVM framework was selected because it allows the treatment to keep its core principles and directs focus to the periphery, but equally important, aspects of the intervention (40). This framework is based on the view that individuals must be understood within their cultural, social, and political environment. Their personal motivations, attitudes, beliefs, and goals must also align with the intervention. The EVM framework serves to "culturally center" an intervention through eight dimensions that must be incorporated for an intervention to have ecological validity and be embedded within the cultural context (41). These dimensions include language, persons, metaphors, content, concepts, goals, methods, and context (see Table 1). [

Overview of Cultural Adaptation and Contextualization Methodology
The adaptation process was an ongoing, iterative process and some of the steps overlapped with one another.
Questions addressed in each step were based on what was or was not answered in the prior steps and the iterative process of this methodology more easily allowed for nding a balance between delity versus t. The format of this methodology was participatory and involved a high level of engagement with the communities where the intervention was delivered. For this reason, the methodology was helpful in ensuring that adaptations were made not only to the clinical treatment material but also the scalability and implementation of the intervention.
A detailed data collection process and documentation system allowed us to ensure that each adaptation made was based on evidence. We created a matrix before the start of the adaptation process based on the FRAME approach for documentation (42): 1) the eight broad dimensions from EVM, 2) implementation strategy (what exactly should be changed in the intervention material), 3) rationale for change (description of why it should be changed and what it would accomplish for the intervention), and 4) evidence for change (which adaptation step(s) the change was a result of). All changes and adaptations were listed in the EVM matrix during the length of the process (see supplemental online material).
The following approach charts the cultural adaptation process from the preparation phase before the feasibility trial, through the trial, and any changes after the trial (see Table 2). [

Identify the key mechanisms of action
The mechanisms of action are theorized process by which a psychological intervention alleviates distress and supports behavioral change (43). Cognitive-behavioral therapies are built upon a theoretical framework that changes in cognition will precede emotional and behavioral changes. Four types of interpersonal distress are theorized as the cause of depression in interpersonal psychotherapy. Behavioral reinforcement cycles of avoidance and distress are targeted in behavioral activation therapy. The speci c mechanisms of action for Group PM+ were identi ed by reading literature of the same intervention conducted in different contexts and discussing with experts that have developed and implemented the intervention (27,28,31), (see Table 3).
[ A systematic review of existing literature on mental health interventions in Nepal was conducted to extract data on: mechanisms of action, delivery agents, trainings, supervision methods, process measures, outcome measures, psychoeducation method, integration into health systems, and cultural/ethno-psychology elements. Databases such as PubMed, PsychInfo and PsychiatryOnline were searched as well as grey literature from policy briefs and annual reports of local NGOs. Though not a formal part of the literature review, interviews were also conducted with staff from leading mental health organizations in Nepal to identify issues for community engagement and implementation related to mental health research and service delivery in Nepal. Data were summarized for key ndings, gaps in research and methodology, and recommendations for Group PM+ adaptation in Nepal.

Training of Trainers (ToT)
Clinical Supervisors and supporting counselors were given a 10-day training by a Group PM+ trainer from a previous study site. This training focused on how to train facilitators in delivering Group PM+. The participants of this ToT identi ed overlap of approach with preexisting practices in the program site and suggested culturally t adaptations to intervention content. The Group PM+ Clinical Supervisors gathered the adaptations suggested by the trainees and reviewed them before nalization. Suggestions that modi ed the mechanisms of action were rejected. Other suggestions that adapted peripheral aspects of the intervention, such as the language or metaphors, were documented, accepted and nalized into the manual by the Clinical Supervisors.

Translation of Manual
Clinical supervisors incorporated initial changes into the English manual which was then translated to Nepali by a professional translator. During the translation process, clinical supervisors regularly reviewed the translator's progress to ensure that the manual would be translated into lay Nepali and could be understood by non-specialist delivery agents. After completing the translation of the manual, the Clinical Supervisors reviewed it fully for any additional changes in language. This was an ongoing process and focused on language rather than the content of the manual. Study staff without a clinical background also reviewed the manual to ensure its comprehensibility for lay persons.

Expert read-through
Experienced bilingual Nepali psychosocial counselors read through the Nepali language intervention manual and suggested changes in language and content to t into the cultural context during a one-day workshop. The main objective of this step was to gain additional perspective from persons experienced in the program's mental health context on the intervention's content, language, and applicability.

Formative qualitative study
Based on gaps identi ed in prior steps, a formative qualitative study was conducted at the proposed study site to gather data on implementation, such as the community's awareness of mental health, identi cation of pre-existing community resources and identi cation of practical problems faced by community members. Local stakeholders such as female community health volunteers (FCHVs), key leaders, health workers, and community members were identi ed for Key Informant Interviews (KIIs) (n=18) and Focus Group Discussions (FGDs) (n=2). Interviews were then coded by two coders (MS and RG) using deductive analysis and key ndings related to program implementation and cultural ethno-psychological elements were summarized to be applied to the manual and program implementation strategy, which included community sensitization programs, recruitment, family meetings, referral pathways, and follow-up with participants.

Practice rounds
Clinical supervisors conducted Group PM+ practice rounds to gain experience delivering the 5-session intervention, gather feedback from the participants on their comprehension and relatability of the intervention, and apply any further changes to the manual and implementation strategy. Practice rounds were conducted with one female group from a nearby community organization and one male group. During these practice rounds, the facilitators (clinical supervisors) noted if adaptations already made in the manual were feasible and acceptable among participants and if further adaptations were necessary to improve the participants' engagement and understanding of the material. After each session, the participants were encouraged to give feedback to the facilitators on content, language, materials and methods used, and facilitation skills. This information was collected through informal interviews with the participants and noted down by the Clinical Supervisors.

Team adaptation workshop
A team workshop with the study Principal Investigators, Program and Research Coordinators, and Clinical Supervisors was conducted to summarize all intervention adaptations listed to date on the EVM matrix. Certain adaptations were accepted or rejected based on if and to what degree they brought changes to the core mechanisms of action. Once all adaptations were thoroughly discussed, Clinical Supervisors made nal changes to the manual before the start of the trial. Program staff also modi ed competency and quality monitoring procedures. Because of the iterative nature of this methodology, most large-scale changes, such as the delivery agents, delivery location, and target population, had been discussed with the team and integrated into the manual and into the delivery of the intervention prior to the workshop.

Implementation and Supervision
Lay Nepali community members were recruited to deliver Group PM+ to their communities as part of the feasibility trial (35,36). Regarding the trial procedures, Enhanced Usual Care (EUC) was compared to Group PM+ intervention. We employed a randomized control trial (RCT) design where the two chosen VDCs in Sindhuli district were randomly assigned to EUC or PM+. Approximately 60 participants, that were similar in population size, ethnic demographics, and access to health facilities, were recruited for each arm. During the Group PM+ training, the facilitators (n=4) were encouraged to suggest changes in the manual's language and feasibility and acceptability of the proposed implementation strategy. These changes were further incorporated into the manual and other program documentation. Cultural adaptation was an iterative process throughout the feasibility trial. All staff recorded notes about rst-hand experiences working on program recruitment, delivery, and engagement with the community, and shared these experiences with their supervisors. Some changes were made in real-time while others required further discussion at the end of the trial.

Review through Process Evaluation
After completing the intervention, qualitative interviews were conducted with eld staff, intervention participants, participants' families, and other key community stakeholders to gain perspective on the successes and challenges of varying adaptation and implementation strategies. Questions for these stakeholders were related to feasibility and acceptability of various cultural adaptions, feedback on the implementation strategies, program delity, and overall challenges. A total of 31 Key Informant Interviews (KIIs) and 6 Focus Group Discussions (FGDs) were conducted with stakeholders and then analyzed by the Program Coordinator and Clinical Supervisor. A deductive data analysis process was used; key themes were identi ed prior to analysis and a codebook with themes related to feasibility and acceptability of the manual content and implementation was developed (35). Interviews were coded using NVivo software and the two coders (MS and RG) established an acceptable inter-reliability rate (IRR) (IRR = 0.8) during the coding process. Once analyzed, key changes and their rationale were discussed within the core team and implemented further to the intervention before the de nitive trial.

Conceptualization of Stress and Tension
As part of the adaptation process, we aimed to create a conceptual model that linked each mechanism of action to how distress was viewed in this context. This model needed to be relatable enough for the lay-facilitators to use during their training in order to promote a comprehensive and culturally appropriate view of how the Group PM+ intervention could reduce distress. The literature review and qualitative study before the trial found that the word tension was used commonly in lay-Nepali language by community members of all ages, gender, and socioeconomic status as a nonstigmatizing term to refer to stress (44). Tension was used as an idiom of distress, as a proxy to depression complaints which is targeted by the intervention. The tension ethnopsychology model was conceptualized during the workshop as the team was nalizing the adaptations before the trial (see Figure 1).
[ FIGURE 1] According to the ethnopsychology model, adversity or practical problems lead to tension. This tension can have a physical manifestation and lead to somatic problems and/or emotional problems. Tension can also lead to a lack of energy, feeling unmotivated, and isolation from friends and family. Each Group PM+ session addresses managing the roots of tension or its effects, which are also integrally linked with one another. Because of the contextual t of the model, elements of the tension model were also used during the recruitment process to explain the effects of adversity on our lives to local community members.

Key Adaptations
Adaptations made before the feasibility trial were systematically documented in the EVM matrix (see Supplementary online material). Key adaptations are summarized in Table 4. [

Relatability of clinical content
Changes were made to the stories used during sessions based on the formative qualitative interviews, practice runs, and expert read-throughs. In these stories, the female character presented problems such as having too much work at home, relationship issues with her husband and mother-in-law and her own mother's death. The male character presented problems such as the stress of returning from a manual labor job abroad, nancial burden of raising a family, and relationship issues with his wife. Participants in the trial found these problems, the case story, and its pictorial representations on posters to be highly relatable and sympathized with the characters in the story (35).
Clinical supervisors and facilitators noted that physical health was a source of stress for the majority of program participants. This may be because most participants were of older age and lacked easy access to proper treatment. Though the intervention focused heavily on managing emotional and practical problems, the feedback received demonstrated that we needed to add more content on physical problems and train facilitators on how to work with participants who only discussed these types of problems. Some suggestions included adding characters in the case story that faced more physical ailments, and posters representing these problems to validate participants' distress. Facilitators should also be trained further in how somatic problems are connected to mental well-being, using the tension model as a guide, so that they can further address the participants' concerns.

Reinforcement of mechanisms of action
A combined competency and delity checklist was created based on both Group PM+ elements and common factors in psychological treatments, drawn from the ENhancing Assessment of Common Therapeutic factors (ENACT) tool (45). The delity checklist was used to measure the whether key activities were implemented and the competency with which they were completed in each session. Facilitators also used this checklist while conducting sessions to ensure that the mechanisms of action were addressed thoroughly during the session. Clinical supervisors attended at least two of the ve sessions per PM+ group and used the delity checklist as a tool to rate the skills of the four facilitators. Each session had 9-10 items and rated the facilitator's level of competency to the intervention manual on a scale of 1-3. The Reducing Tension Checklist (RTC) tool was also developed to assess whether participants applied the mechanisms of action learned in the sessions to their daily lives. This tool was used pre-and post-intervention (35).
The participants noted that of the different techniques they learned, the deep breathing technique was the most memorable and was used the most outside of the sessions (35). This technique was noted as the most tangible, accessible, and lead to the most immediate results. Facilitators found it di cult to track or gauge how often participants were practicing at home. For the next trial, participants and facilitators recommended making all mechanisms of action more memorable and easier to practice at home by creating a method to track time that participants spent practicing outside of sessions.
As a result of the process evaluations, the team decided to incorporate more imagery and memorable metaphors to each of the sessions. For example, the second session focused on effective problem solving, was considered to be one of the most di cult sessions for the facilitators to deliver and for the participants to practice outside the sessions. As a result of these ndings, we decided for the de nitive trial to incorporate an image of a hand with a step for problem solving written on each nger. A similar card was created for each of the ve sessions with visual imagery of each mechanism of action on one side and a space to plan when to practice on the other side. This set of cards was called the tension toolkit. We also decided to incorporate several concrete tools to support each of the sessions, such as a small pouch (thaili) after the third session for participants to store a rock, or kernel of corn each time they do a pleasurable activity at home. The objective of these tools is to provide physical items to help participants remember the content of the sessions and skills learned throughout the intervention.

Use of local idioms to increase acceptability and reduce stigma
Community sensitization events were a vital component of recruitment because stigma and lack of awareness of mental health issues in the program site (35). These events were led by the facilitators who invited community members to attend and discuss the causes, signs, and symptoms of tension. Facilitators prompted community members with questions about causes of tension and man ko samasya (heart-mind problems), and impact of adversity on personal behaviors and emotions. The participatory format of engaging with the community during sensitization activities while using de-stigmatizing terms was successful in encouraging community members to volunteer for study screening.
In further efforts to reduce stigma against the intervention, key adaptations were made to the metaphors, the concepts, and the "branding". To reduce stigmatization by the community for participating in the intervention and selfstigmatization by the participants themselves, a context appropriate name was voted on by the local staff that chose Khulla Man, meaning open-hearted. Some participants described their own heart-mind as being Khulla (open) or having a Khulla Man after the sessions. Similarly, the tension model conceptualized before the trial was successful in that it supported the facilitators' understanding of how each session supported in reducing stress. Importance of staff, training, and acceptable recruitment methods One of the key adaptations made for the feasibility was to recruit, train, and hire local lay workers, as a method of task shifting and to increase the capacity of local community members. A 20-day Community Psychosocial Worker (CPSW) Training, as is standard to certify this cadre of workers in Nepal, was delivered before 10 days of Group PM+ training. Because the four CPSWs who delivered the intervention were from different locations in the study catchment area, they conducted groups in areas that were close to their homes but not necessarily in their own neighborhood. As indicated by the qualitative interviews, participants found comfort in having their groups led by a facilitator who was from a familiar location but who the participants did not know very well. Similarly, helpers were hired to assist the facilitators during the sessions. Because concepts of time and punctuality were exible in this context, it was noted that having a helper remind participants to arrive on time helped increase attendance and reduce drop-out rates.
The Community Informant Detection Tool (CIDT) helped easily identify members of the community experiencing mental health symptoms through using vignettes and pictures. It has been previously successful in Nepal (46), and a general distress, man ko samasya, version was developed for this trial. Female Community Health Volunteers (FCHVs), mothers' group members, and other local community leaders received a one-day training on how to use this tool to refer those with general distress to the study. They were also trained on the severe mental health CIDT version to identify who not to recruit. However, this led to confusion amongst some trainees who referred those with severe mental illness to the study, since their signs and symptoms are clearer than those with general distress. For the next trial, it is recommended to train local community members on using the general distress version only and with regular on-site supervision.

Outputs and Applications
This study provides a model that incorporates changes to the clinical content, scalability and implementation in a cultural adaptation method. Ultimately, this study further integrates cultural adaptation into implementation science research (47). The scalability and implementation aspects were shown to be just as important as the clinical content when adapting an intervention. As a result, a tension ethnopsychology model was developed as a conceptual foundation to key adaptations. Adaptations were made to case stories, visuals, and materials to re ect the context. Using a context appropriate intervention name, utilizing the CIDT for recruitment, and conducting community sensitization events supported in reducing the stigma. We also adapted the facilitator trainings, and created the delity checklist, RTC, and tension toolkit to reinforce the mechanisms of action and ensure quality of care. As part of this study, we have also designed a clear and detailed documentation process that will assist in conducting evidencebased adaptations to future interventions. This work builds on prior research in LMICs and can be integrated for use into approaches such as the Replicating Effective Programs (REP) framework (48).
The outcomes of the pilot Group PM+ trial demonstrates that the speci c adaptations made to the intervention as well as the contextualization of the implementation process was feasible and acceptable. The intervention had a high retention rate with 75% of the participants completed 4 -5 sessions (35). All facilitators (n=4) scored above a 75% on the delity checklist developed to measure the competency and overall delity to the intervention. Though the study was not powered, the intervention group showed a greater change in outcomes, especially in general psychological distress. We hypothesize that the positive quantitative and qualitative results of the pilot trial can in part be attributed to the cultural adaptations and contextualization that supported the overall implementation of the Group PM+ intervention.
We have created one cultural adaptation guides with two applications; the rst model to be used in contexts without time or resource constraints and the second in contexts such as LMICs with various constraints (see Figure 2).

[FIGURE 2]
The longer process includes a thorough literature review and a read-through by experts. Though we conducted a formative qualitative study as part of the adaptation process, we found that an assessment of the intervention site could have su ced for the purpose of adaptation. This assessment should be tailored to the needs of the intervention and should be approached as a method to gather information from and to engage the local community, which was proven to play a large role in the success of the intervention. This could include, amongst other activities, meeting with local stakeholders, potential service users, their families, and other services being offered in the area.
In reality, adaptation processes are often conducted under staff and time constraints in low-resource settings. Throughout the adaptation process, we took note of which steps are absolutely necessary versus those that could be combined or eliminated altogether, while not compromising quality in favor of time. Using this "time constraints" method, it is still possible to create deep rather than surface structure changes. The literature review step has been modi ed from in-depth to an overall review. The expert read-through has been eliminated since we found that the training of trainers and other steps that involved study clinical staff who developed in-depth knowledge of the intervention when conducting the other adaptation steps were best t to suggest the most meaningful changes. The practice rounds were found to be the most helpful and the greatest number of adaptations were made from this step as compared to the others. The practice rounds were an important step in addressing the logistical aspects of the intervention, such as time management, venue location and also to gather feedback from participants within the targeted population. Therefore, we recommend shortening other steps and focus mainly on conducting several practice rounds or sessions as part of the adaptation process, if under extreme time constraints.
The eight dimensions of cultural adaptation, as presented by Bernal and colleagues (11), were helpful in conceptualizing the types of changes that could be made, while preserving the treatment's core mechanisms of actions. However, we found that a few of the dimensions, such as content and context, are similar in their de nitions and overlapped with one another. These dimensions also seemed to apply to the majority of the adaptations compared to a dimension such as persons that applied to very few adaptations. As a result, we found it best to focus less on which dimension an adaptation would be categorized as and instead allow the eight dimensions to serve as a framework to better understand that the treatment and implementation needs to be adapted within the cultural context.

Limitations
Limitations to the study must be accounted for before using this adaptation model for other interventions. This paper documents the adaptation process till the end of the feasibility trial. It does not incorporate ndings from the de nitive trial to understand if all the adaptations made after the feasibility trial were successful. It is di cult to identify which adaptations were the most important and which had the highest impact on the intervention because there are many potential factors contributing to its success. Though manual content and implementation complement one another, it is also di cult to isolate which aspects were most successful and to speci cally recommend it for future interventions. The current results are also limited to qualitative work and the participants', staff, and stakeholders' perceptions of what were and were not effective. Regardless of compressing the adaptation process to t resource limited contexts, the nature of the cultural adaptation itself is lengthy, iterative and requires depth and heavy documentation. However, these are compromises that must be made for a successful adaptation process.

Conclusion
This study proposes not only a clear adaptation process but also documentation tools to guide adaptations of future interventions. Through this process, adaptations have been made to not only the intervention but the EVM framework itself, with a focus on scalability, and quality monitoring. Though this process was used to adapt a mental health intervention in a developing country, it can be used to adapt interventions for various populations, such as ethnic minorities in high-income countries or in other LMICs. With the increase in interventions that employ the concept of task-sharing (49), this process also serves as an example for other interventions in humanitarian or developing country settings. Though this process was used to adapt a group intervention, it is exible enough to be used to adapt an individual intervention or even a treatment beyond mental health.

Authors' contributions
The study was designed by BAK, MJDJ, NPL, EvO, and MS. The manuscript was drafted by MS, with support from BAK and MJDJ. MS, RG, and PG conducted the implementation of the adaptation process. MS and RG conducted qualitative interviews, coding, and analysis. All authors read and approved the manuscript.

Competing interests
The authors declare that they have no competing interests.
Ethics approval and consent to participate Ethical approval has been received from the Nepal Health Research Council (NHRC) and the World Health Organization.

Consent for Publication
All participants provided written informed consent before participating in the study, which included consent to publish anonymous quotes from individual participants.

Methods
Methods are the procedures followed to achieve treatment goals. These methods and procedures should be congruent with the participants' culture and use of language.

Goals
Goals are the agreement between participants and facilitator in what participants would like to achieve during the course of the treatment. These goals must be realistic and t with the participants' values and personal motivations.

Context
Context refers to the participants' economic, social, political and cultural environment. This should look beyond just the participant as an individual and focus on outside factors, such as socialization, discrimination and family history, that could in uence the treatment.

Content
The knowledge, values, customs, and traditions shared by the participants should be integrated into all elements of the treatment. This can be seen as a starting point for culturally adapting the recruitment process, assessments, and the treatment itself.

Metaphors
Culturally appropriate symbols or concepts should be embedded within the intervention that support participants in absorbing the treatment's core mechanisms of action. Metaphors used may be pictorial, idioms, commonly used phrases or item and symbols.

Persons
A culturally appropriate intervention must consider the role of ethnicity, race, gender, class and other relevant social constructs in the relationship between the participants and facilitators. This relationship should respect expectations and limitations that are re ective of the local culture.

Language
Language is inherently attached to culture and is related to the expression of emotional experiences. The intervention should be in the language most comfortable and accessible to the participants and should also use appropriate terminology based on the education levels of the facilitators and participants. Code interviews using deductive analysis and summarize key ndings related to main dimensions [ Table 5].
Apply ndings from process evaluation to the manual and program implementation strategy for the future program implementation [ Table 5].

Managing Problems
Participants learn which of their problems are solvable and which are unsolvable. One solvable problem is chosen and participants brainstorm solutions, then identify manageable steps to implement their solutions and accomplish their goals.

Session 2-5
Behavioral Activation Participants review the inactivity cycle. They choose a small activity that they enjoy doing (e.g. making and drinking tea, meeting a friend etc.) or a task they need to complete and create a detailed plan about when and how to conduct this activity as a rst step in breaking the inactivity cycle.

Session 3-5
Strengthening Social Support Participants learn to recognize who amongst their family and friends are existing and potential sources of support and how best to strengthen connections with them. Participants could also identify broader community and organizational forms of support. Social network mapping activities are incorporated in this mechanism.
Session 4-5 Note: The rst four sessions of PM+ each address a speci c mechanism of action. The fth and last session is a review of the mechanisms of actions learned in the intervention. Four key mechanisms of action were identi ed; stress management, managing problems, behavioral activation, and strengthening social support 4,6 .
Because these mechanisms of action cannot be changed, adaptations in the next steps will be made to support the bene ciary population's comprehension of these techniques.

2.In-depth Literature Review
The literature review suggested that mental illness is deeply stigmatized in Nepal and idioms of distress such as man ko samasya or tension can be used in rural communities to refer more openly to general distress. This suggests a need for developing a non-stigmatizing conceptual framework for this intervention to be used in the Nepal context 1,3,4,5,8 .
Task-sharing trainings are common in Nepal, due to the lack of trained mental health professionals. Trainings are followed by frequent on-site and off-site supervision for the nonspecialists. We decided that for Group PM+, Clinical Supervisors will conduct weekly o ce supervision and will observe two sessions per group 2, 7 .

3.Training of Trainers (ToT)
Trainers suggested adding pre-existing counseling techniques to the Group PM+ intervention in order to strengthen the mechanisms of action. These techniques were previously proven to be effective in the Nepal context. Examples include grounding techniques (where participants are brought to the present moment by using senses to identify what is around them) and memapping (an activity where participants identify their close relationships in pictorial form) 4,5,7 .
As part of the training, trainers read the manual and reviewed other implementation material thoroughly and suggested further language, such as man ko samasya, to decrease self-stigma. This aligned with the results of the literature review 1,5,7,8 .
Trainers suggested adding multiple culturally appropriate ice breakers and energizers throughout the group sessions. This was to ensure that participants would stay on task throughout each session 4,5,7 .

4.Translation of manual
Translation of the manual was an iterative process where the translator met often with study staff to review translations and to ensure that the language was simple and accessible for lay facilitators 1,4 .
While the translator was making progress on the manual, the Clinical Supervisors felt the need to create a delity checklist for the facilitators to follow during sessions and to use alongside the manual 7 .

5.Expert readthrough
Experts with experience counseling in Nepal stressed the importance of gender matching facilitators and participants for the intervention 2,7,8 .
Family engagement was identi ed as a key component in prior mental health interventions in Nepal. A family meeting, to involve participants' families in the clinical process, was added as a component to Group PM+ 2,4,5,6,7,8 .

6.Formative qualitative study
Local community members identi ed that it is deemed acceptable for facilitators of lower caste to work with participants of higher caste in the local area. Therefore, facilitators from all castes can be hired to deliver the intervention 2,8 .
Nearby resources such as health posts, mothers' groups, domestic violence NGOs, and safe homes were identi ed as referral services for intervention participants if needed 7,8 .
Gender issues and social discrimination were identi ed as sources of stress for community members. It was suggested to add these issues in the manual and other clinical material 4,8 .

7.Practice rounds
Participants in the practice rounds suggested adding more posters and visuals to the clinical content. They also suggested personal problems that characters may have for these materials. Examples included an unemployed man returning home from working a labor job abroad, a daughter-in-law having an argument with her mother-in-law, and a woman unable to concentrate on her work in the farm due to stress 3,4,8 .
Clinical supervisors found that when running the practice sessions, participants would often arrive late and some would forget the day of the week the sessions were held on. This highlighted the importance of having a helper who could call the participants a day before the session as a reminder or gather them from their homes before the start of each session. Participants also suggested the distribution of a calendar during the rst session as a reminder of when to attend the following sessions 4,7,8 .
Clinical supervisors noted that some participants were dominant while others were quieter. It was suggested to increase training in group facilitation skills and managing dominant and quiet participants. Facilitators must also take special note of gender, socioeconomic status, ethnicity/caste and other social factors and their possible effects on group participation 2,4,7 .
Some participants in the practice rounds expected monetary bene ts from the intervention. Therefore, clinical supervisors found it necessary to clarify that the program is not for those who need support with their economic situation but is for those with man ko samasya and tension 4,8 .

8.Team Adaptation Workshop
The role of the helper was further de ned during the workshop. From experience gathered from each of the steps, the team agreed that the helper would also be responsible for administrative tasks, such as hanging posters and writing on the board, that would allow the facilitators to place all of their focus on the participants and session material 2 .
Study staff found that it would be necessary to include a method to evaluate participants' progress and suicidal tendencies at the start of each session and to conduct referrals as necessary. It was decided during the workshop that the Psychological Outcomes Pro le (PSYCHLOPS) would be administered during snack time at the start of each session 6,7 .
A general distress version of the Community Informant Detection Tool (CIDT) will be created and used for recruitment of participants 4,5,6,7,8 . The tension conceptual model was created during the workshop as program staff brought together various conceptual adaptations informed by the adaptation process. The conceptual model relates each session to an aspect of general distress 4,5 .

Implementation
During the recruitment of program participants, the facilitators voted to change the Group PM+ program title to Khulla Man, meaning open-hearted, a more culturally appropriate and de-and Supervision stigmatizing program name 3,4,8 .
Community sensitization events were led by facilitators and program staff to raise awareness about mental health and recruit participants into the program. Over time, these events went from lecture heavy to discussions about the man (heart-mind) and the causes and symptoms of man ko samasya. This method of engagement was found to be especially helpful with recruitment and de-stigmatizing mental health issues 3,4,5,7,8 .

Review through Process Evaluation
Facilitators mentioned that conducting practice groups supported them in feeling prepared before the start of the trial and recommended that the facilitators conduct more practice groups before the next trial 3 .
Participants noted that though they enjoyed the sessions and practiced what they had learned at home, they sometimes had di culty in remembering all the techniques learned. As part of the review after the pilot trial, a "tension toolkit" was developed by study staff to help participants remember techniques learned in each session. The toolkit included cards for each session with pictures of the techniques learned, and a space to track how many times participants practiced each technique. A certi cate of completion was also included to encourage participants to continue utilizing what they learned after the nal session 3,4,6,7 .   CASupplementaryMaterial.docx