Beyond Access: How Are Communities Engaged in Digital Mental Health Interventions in Low- and Middle-Income Countries?

Introduction - Objectives: Community engagement may improve the relevance, acceptability, and sustainability of digital mental health interventions, yet how communities participate in their development and implementation remains unclear. We examined levels and modalities of community participation in digital mental health interventions in LMICs and their variation across intervention phases and study characteristics. Methodology: We conducted an exploratory secondary analysis of 110 studies from an international scoping review. Participation was classified by the highest documented level: none, consultation, user adaptation, co-design, co-implementation, or non-classifiable, and further grouped as consultative/adaptive or collaborative. Associations with study and intervention characteristics were explored using Fisher’s exact tests with Monte Carlo simulation (100,000 replicates) among 106 classifiable studies. Study selection followed a structured screening process, from 5,045 identified records to 110 studies included in the secondary analysis. Two reviewers independently verified the selection process and the reliability of the extracted data. Results and Discussion: Of 110 studies, 40 (36.4%) documented no participation and 4 (3.6%) were non-classifiable. Consultation was identified in 8 (7.3%), user adaptation in 28 (25.5%), co-design in 28 (25.5%), and co-implementation in only 2 (1.8%); 36 (32.7%) were consultative/adaptive and 30 (27.3%) collaborative. The reported phase was mainly design (54/110; 49.1%), compared with evaluation (20/110; 18.2%) and implementation (8/110; 7.3%), while 28 (25.5%) did not report a phase. All co-design studies were reported at the design phase (28/28), whereas both co-implementation studies were reported at implementation (2/2). Among user-adaptation studies, 20/28 (71.4%) were reported at design and 8/28 (28.6%) at evaluation. Peer or community-actor involvement was reported in 50/110 studies (45.5%) and cultural adaptation in 72/110 (65.5%). Collaborative participation was more frequent in interactive applications/technologies (41.9%) than in web/platform interventions (12.5%). Participation level was exploratorily associated with digital modality (p=0.029) and implementation scale (p=0.038). Conclusions: Community participation in digital mental health interventions in LMICs remains heterogeneous and appears more frequently documented at the design than implementation stage. Despite frequent cultural adaptation and peer or community-actor involvement in nearly half of studies, co-implementation was rare. Extending participation beyond design to implementation may strengthen community engagement throughout the intervention lifecycle.