Informing Stakeholder Strategy: Latent Class Analysis for Tailored, Whole-Venue HIV Interventions with Men in Zimbabwean Social Venues

Introduction Although not formally classified by World Health Organization, social venue-attending men are a diverse, high-risk group severely affected by HIV. Overlapping behavioural, structural, and psychosocial factors remain poorly addressed by current prevention efforts. We aimed to: (1) identify latent classes of men based on HIV acquisition/transmission factors; (2) examine associations between class membership and these factors; and (3) provide evidence-based typology to help inform stakeholder-led development of tailored HIV interventions for venue-going men. Methods We analysed data from 2,827 men recruited through time-location sampling in four Zimbabwean cities in the Priorities for Local AIDS Control Efforts surveys. Structured interviews captured socio-demographics, substance use, mental health, sexual behaviours, and biomedical outcomes. Latent class analysis of six dichotomous behavioural and psychosocial indicators identified subgroups, with the final model selected using Akaike Information Criteria, Bayesian Information Criteria, and interpretability. Age-adjusted weighted Poisson regression was used to estimate associations between class membership and six HIV acquisition or transmission-related outcomes. Results The three-class model fit best. Class 1 (33.5%) had men paying for sex (23%), drug use (42%), hazardous alcohol use (29%), and common mental health symptoms (11%). Class 2 (48.1%) showed the highest paid sex prevalence (72%), hazardous alcohol use (60%), drug use (62%), and mental disorders (46%). Class 3 (33.5%) had low paid sex (4%) but very high drug use (92%) and moderate hazardous alcohol use (24%). These profiles appeared consistently across all sites. Age also differed: younger men were more common in Class 1 (< 30 years), while older men dominated Classes 2 and 3 (≥ 30 years). Adjusted for age, Class 2 had higher STIs (PR = 2.01, 95% CI:1.06–3.84) and ≥ 2 partners in the last 4 weeks (PR = 3.50, 95% CI:2.33–5.26). Class 1 had more partners (PR = 1.51, 95% CI:1.08–2.10) but less condomless sex (PR = 0.84, 95% CI:0.75–0.94) compared to Class 3. Conclusions The identified HIV acquisition/transmission profiles should not serve as prescriptive blueprints but as inputs for qualitative work and stakeholder engagement to support context-specific interventions. Recognizing subgroups of venue-going men with overlapping vulnerabilities—common mental disorders, substance use, and paid sex—can help tailor whole-venue HIV prevention and care strategies. Word Limit: 346 words