Background: Service user and caregiver involvement has become an expectation in all aspectsrnof the mental health system, including in policy, planning, service delivery, qualityrnimprovement, research and education around the world. Service user involvement is particularlyrnpromoted as a strategy to scale-up quality mental healthcare service in low-and-middle-incomerncountries (LMICs). However, little is known about actual involvement experiences and how bestrnto involve service users and caregiversin mental health systems. No study of involvement hasrnbeen conducted in Ethiopia. Hence, this study aimed to achieve the following objectives. rnGeneral objective: To develop, pilot and evaluate a model of service user/caregiverrninvolvement in mental health system strengthening in Ethiopia rnThe specific objectives of the study were to: rnï‚· explore the experiences, perceived barriers and facilitators regarding service user andrncaregiver involvement in mental health system strengthening. rnï‚· develop an acceptable, sustainable and feasible Theory of Change model of service userrnand caregiver involvement in the expanding mental health system. rnï‚· equip service users and health professionals for greater involvement in mental healthrnsystem strengthening. rnï‚· pilot the feasibility and acceptability of a co-developed model of service user andrncaregiver involvement in mental health system strengthening. rnMethods: A predominantly qualitative approach was used, which was conducted in three phases.rnThe study was situated within participatory approaches informed by a critical social theoreticalrnperspectives and social-ecological system theory. The thesis consists of 5 articles, which havernbeen published or are about to be submitted for publication.rnPhase 1 provides the context for the thesis, which was a formative qualitative study conducted tornexplore the experiences of involvement, barriers to, facilitators and capacity building needs forrngreater involvement from the perspective of key stakeholders (Paper 1). In the study, 39 semistructuredrninterviewsrnwererncarriedrnoutrnwithrnpurposivelyrnselectedrnservicernusersrn(n=13),rncaregiversrn(n=10), heads of primary care facilities (n=8) and policy makers/planners/service developersrn(n=8). Thematic analysis was applied. rnIn Phase 2, a Theory of Change(ToC) model was developed in iterative participatory workshopsrnconducted in (i) Addis Ababa with purposively selected psychiatrists (n=4) and multidisciplinaryrnresearchers (n=3), and (ii) a rural district in south-central Ethiopia (Sodo) with communityrnstakeholders (n=24). Information from the workshops (provisional ToC maps, minutes, audiornrecordings), and inputs from Paper1 were triangulated to develop the detailed ToC map. ThisrnToC map was further refined with written feedback and further consultative meetings with thernresearch team (n=6) and community stakeholders (n=35) (Paper 2). Based on findings fromrnpaper 1 and 2, an empowerment training programme was developed, delivered and evaluated forrnacceptability, feasibility and preliminary impacts using a mixed-methods design in Paper 3rn(Chapter 7). The training was delivered separately for service users (n=12) and caregiversrn(n=12), and health professionals (n=18). The quantitative data consisted of process data,rnsatisfaction questionnaires, and a retrospective pre-test survey. Qualitative data included exit andrnfollow-up in-depth interviews with service users and caregivers. Descriptive statistics werernperformed for quantitative data, and qualitative data were analysed using a thematic analysisrnapproach. The findings were integrated through triangulation for convergent themes followingrnanalysis. rnPhase 3 involved Participatory Action Research (PAR) implementation strategy development,rnpiloting the ToC model and a case study of the piloting process and participantsâ€˜ experiences ofrninvolvement. Accordingly, a PAR protocol was developed to guide the ToC modelrnimplementation process (Paper 4) and the ToC model was piloted using PAR in three stages,rneach with iterative activities of plan, act, observe and reflect. Two stakeholder groups, arnResearch Advisory Group (RAG) and Research Participant Group (RPG), were established andrncollaborated in all stages of the PAR process. Data collection involved process documentation ofrnmeetings and activities: attendances, workshop minutes, discussion outputs, and reflective notes,rnparticipatory observation of sessions, and in-depth interviews with service users (n=6),rncaregivers (n=4) and health professionals (n=2) involved throughout the PAR process.rnDescriptive analysis was used for process data, and thematic analysis was used for qualitativerndata. Triangulation and synthesis of findings was carried out to develop the case study (Paper 5Chapterrn8).rnResults: The formative qualitative study showed that the concept of service user involvement isrnnew within the Ethiopian mental health system and that service user participants had little directrninvolvement in mental health systems strengthening. Several potential benefits were identified,rnincluded were improved appropriateness and quality of services, and greater protection againstrnmistreatment and promotion of respect for service users. This study identified key multilevelrnbarriers and facilitators to effective involvement, and many areas of capacity building needs forrngreater involvement. Stigma was considered to be a pervasive barrier, operating within the healthrnsystem, the local community and individuals. Low recognition of the potential contribution ofrnservice users seemed linked to limited empowerment and mobilization of service users. Therernwas a strong need for awareness-raising and training to equip service users, caregivers, servicernproviders and local community for involvement. rnThe participatory ToC model co-produced described the causal pathways for greater service userrninvolvement with key components (community, health organisation, service user and caregiver),rninterventions, preconditions, assumptions and indicators for each component along the pathwayrnto the long-term outcomes and impact. The participatory nature of ToC process raised awarenessrnof the possibilities for servicer user and caregiver involvement, promoted co-working andrnstimulated immediate commitments to mobilise support for a grass roots service userrnorganization. The empowerment training content, delivery process and standard of the trainingrnprogram met participantsâ€˜ expectations, improved positive gains in understanding about metalrnillness, stigma, service-user involvement, and human rights. The training had positive impacts,rnincluding increased self-confidence, sense of empowerment, social benefits, and perceivedrntherapeutic benefits. Through active collaboration and involvement the stakeholder groupsrnidentified their top research priorities considered as significant problems in their specific settingsrn(particularly the need for public awareness-raising and addressing stigma and discrimination)rnand developed intervention programmes, action plans and initiated actions that could resolvernsome of these problems. Key mechanisms used for inclusive participation included capacityrnbuilding and bringing together diverse stakeholders, anchoring the study in established strongrncommunity involvement structures, and making use of participatory strategies and activitiesrnduring the PAR process. The case study about participantsâ€˜ experiences of involvement in PARrnidentified and explained about: (i) expectations and motivation, (ii) experiences of the dynamics rnof the PAR process, (iii) perceived impacts of involvement in the PAR process, and (iv)rnimplementation challenges and future directions. rnConclusions: Service user and caregiver involvement in mental health system strengthening is arncomplex process, which is affected by multilevel factors that demands interventions thatrnrecognize this complexity.This thesis shows that effective involvement could be achieved byrnempowering key stakeholders for greater involvement. PAR can be a feasible and acceptablernapproach for empowerment and collaborative engagement of a range of stakeholders in mentalrnhealth system strengthening. Combining ToC with a PAR approach can provide a more inclusivernactive involvement model of stakeholders including service users and caregivers. Embeddingrneffective service user and caregiver involvement can be hindered with lack of strategic enablingrnregulatory infrastructure within the mental health system. rnImplications: The findings of this thesis can be of practical relevance to researchers and thosernworking in health care settings to develop service user and caregiver involvement in mentalrnhealth system strengthening. It contributes theoretical understanding for future research andrnprovides practical knowledge for application of participatory action oriented approaches aboutrnhow to develop participatory models and actively involve key stakeholders in the contexts of lowrnresource setting.