Group interpersonal psychotherapy for depression in Uganda: intervention effects and the influence of social desirability bias in a pilot cluster randomized trial

Introduction: Group interpersonal psychotherapy (IPT-G) can expand access to depression care, but assessment conditions may influence outcomes. This pilot cluster-randomized trial was designed to refine outcome-measurement procedures and generate preliminary effect estimates to inform a larger trial. We evaluated a six-week model, examining associations between follow-up assessment conditions and depressive symptoms. Methods: In Mayuge, Uganda, female participants aged [&ge;]13 years with Patient Health Questionnaire-9 (PHQ-9) scores [&ge;]10 were randomized by village to IPT-G or enhanced treatment as usual. Primary outcome was the PHQ-9 score. Assessments occurred at baseline, two weeks and three months post-treatment. During early two-week follow-up, therapy facilitators remained visible near interview locations while mobilizing participants; thereafter, they left before subsequent interviews. At three months, interviews were conducted independently and participants were randomly assigned to receive a prompt emphasizing confidentiality, accurate reporting and that incentives were not contingent on responses. We examined associations between these two conditions and PHQ-9 scores. Results: Of 292 randomized participants, 263 provided three-month data. Mean PHQ-9 scores were lower in intervention than control participants at two weeks (1.73 versus 15.14; mean difference [MD] -13.41, 95% confidence interval [CI] -16.10 to -10.71) and three months (3.19 versus 10.97; MD -7.78, 95% CI -9.66 to -5.90; both p<0.001). Among intervention participants, prior assessment without facilitators and receipt of the confidentiality and honest-reporting prompt were jointly associated with 4.41-point higher PHQ-9 scores versus the reference condition (95% CI 2.37 to 6.45; p<0.001). In exploratory sensitivity analysis, an intervention subgroup exposed to both conditions had lower three-month PHQ-9 scores than controls (MD -5.19, 95% CI -7.39 to -2.99; p<0.001). Conclusions: Six-week IPT-G was associated with significant symptom reductions, but follow-up procedures influenced the apparent benefit, highlighting the need for independent outcome assessment in community-delivered psychological intervention trials. Trial registration: Pan African Clinical Trials Registry, PACTR202606549854263; retrospectively registered.