Psychosocial and Mental Health Indicators and Viral Non-Suppression Among Adults Receiving Antiretroviral Therapy in Gaza Province, Mozambique: A Cross-Sectional Study

Background/Objectives: Mental health challenges may compromise long-term HIV treatment outcomes, yet evidence linking psychosocial and mental health indicators with viral suppression remains limited in Mozambique. We examined the association between psychosocial and mental health indicators and viral non-suppression among adults receiving antiretroviral therapy (ART) in Gaza Province. Methods: We conducted a cross-sectional analytical study among 343 adults receiving ART at seven public health facilities in Xai-Xai District, Mozambique, in 2023–2024. The primary exposure was endorsement of at least one of three psychosocial and mental health indicators: social isolation, sadness, or use of medication for mental health problems. The primary outcome was viral non-suppression. Associations were estimated using Firth penalized logistic regression because of the small number of non-suppression events. Results: Overall, 41/343 participants (12.0%) reported at least one psychosocial or mental health indicator. Viral load results were available for 338 participants, of whom 19 (5.6%) had viral non-suppression. Viral non-suppression was more frequent among participants reporting at least one indicator than among those reporting none (12.2% vs. 4.7%). Reporting at least one indicator was associated with higher odds of viral non-suppression in unadjusted analysis (OR 2.95, 95% CI 1.04–8.34) and after adjustment for age and sex (aOR 2.97, 95% CI 1.05–8.38). After additional adjustment for ART duration, the association was attenuated and statistically uncertain (aOR 2.12, 95% CI 0.71–6.37). Social isolation showed the largest estimated crude odds ratio (OR 8.17, 95% CI 1.70–39.29), although estimates were imprecise. Conclusions: In the primary model adjusting for age, sex, and ART duration, the association was attenuated and imprecise (aOR 2.12, 95% CI 0.71–6.37). The findings are hypothesis-generating and should be interpreted in light of non-validated exposure measurement, substantial missing exposure information, and few non-suppression events.