Partner disclosure willingness and HIV advocacy engagement among men in rural Ghana: a Bayesian logistic regression analysis

Background Men in sub-Saharan Africa consistently report reduced willingness to engage publicly in HIV advocacy or to disclose HIV status to sexual partners. Thus, this study examined both public advocacy willingness and partner HIV status disclosure willingness as distinct outcomes in this population. Methods A community-based cross-sectional survey was conducted among 331 men aged ≥15 years in the Atwima Kwanwoma District, Ashanti Region, Ghana, using multistage random sampling. Two binary outcomes were assessed: (1) willingness to volunteer as an HIV advocate (0.9% prevalence; n  = 3); and (2) willingness to tell a partner about HIV status (90.0%; n  = 298). Given this fundamental disparity in event prevalence, Bayesian logistic regression with weakly informative Normal (0, 1.5) priors was applied exclusively to the tell-partner outcome, which had adequate event density (298 events, 9 predictors). Willingness to volunteer was described through stratified proportions and cross-tabulations only. An exploratory mediation analysis examined whether stigma perception mediates the pathway from HIV awareness to volunteer willingness. Internal validation used 200-replicate bootstrap optimism correction. Results Willingness to tell a partner was near-universal at 90.0% (95% CI: 86.3–93.0%). In the Bayesian model for partner disclosure willingness (all Rhat ≤ 1.001; minimum Bulk ESS = 5,221), community encouragement of facility patronage, followed by positive family attitude toward PLHIV, tertiary education versus JHS, HIV awareness per standardised unit, and increasing distance to the testing site as a negative predictor. The Bayesian model achieved AUC = 0.975; bootstrap-corrected AUC = 0.915 (mean optimism = 0.005). The exploratory mediation analysis indicated that approximately 42% of awareness’s association with volunteer willingness operated through stigma perception, though the indirect effect credible interval crossed zero (median = −0.29, 95% CrI: −1.52 to 0.59), precluding firm causal inference. Conclusion The near-zero prevalence of volunteer willingness versus near-universal partner disclosure willingness reveals a fundamental divide between men’s public and private HIV-related engagement. Reducing distance barriers and cultivating positive family climates surrounding HIV should be central to disclosure-promotion strategies. Also, community-based male advocacy programmes require stigma-reduction as a prerequisite rather than a parallel activity.