A cluster-randomized trial of interventions for adolescent mental disorders in Zimbabwe

Purpose: Young people have low uptake of mental health. We compared two task-shifted mental health care models, i.e., adult Friendship Bench (FB) delivered by community health workers and Youth Friendship Bench (YouFB) delivered by trained university students in Harare, Zimbabwe. We hypothesised that the peer-delivered YouFB would have greater uptake and effectiveness in managing common mental disorders (CMDs) in 16–19-year-olds compared to the standard FB model. We also aimed to evaluate the reach, fidelity, acceptability and cost of the YouFB compared to standard FB. Methods: We conducted an open-label cluster-randomised, hybrid type-2 implementation trial with cost analysis in 26 primary care clinics and their surrounding communities. Facilities were randomised 1:1 to FB or YouFB. The primary implementation outcome was uptake, defined as the proportion of adolescents aged 16–19 offered FB sessions for treatment of CMD who completed at least one FB session. Secondary implementation outcomes included reach, fidelity, and acceptability. The main clinical outcome was the clinical effectiveness of YouFB vs. FB at six months, assessed by changes in Shona Symptom Questionnaire (SSQ-14) scores. We also carried out a cost analysis from a societal perspective. Acceptability was evaluated qualitatively using in-depth interviews. Reach was calculated as the number of adolescents receiving FB sessions per clinic day. Results: Uptake in the FB and YouFB arms was 86.6% (187/216) and 95.6% (220/230), respectively (primary outcome). The number of completed FB sessions (feasibility) was higher in the YouFB arm than the FB arm (cluster-level mean prevalence 96.7% vs. 85.8%, prevalence ratio = 1.13; 95% CI:0.98–1.30). Among 528 trial participants, adjusting for baseline score, gender, education, marital status, employment and HIV status, the proportion of participants with SSQ-14 score ≥ 8 was similar by arm after six months, adjusted odds ratio = 0.65 (95% CI: 0.36–1.17). Total program costs were higher in the YouFB arm. Process evaluation found the YouFB to be highly acceptable. Conclusions: A youth-focused Friendship Bench intervention is feasible and acceptable. Recipients highly valued the ability to connect with a same-age peer and its easy accessibility. However, further intervention optimisation is needed to improve its clinical and cost-effectiveness. Trial registration: This trial was prospectively registered on 21/08/2018 with the Pan African Clinical Trial Registry database. Registration no PACTR201808181810124.