Mediation Analysis of the Efficacy of the Eban HIV/STD Risk-Reduction Intervention for African American HIV Serodiscordant Couples.

Targeting couples is a promising behavioral HIV risk-reduction strategy, but the mechanisms underlying the effects of such interventions are unknown. We report secondary analyses testing whether Social-Cognitive-Theory variables mediated the Eban HIV-risk-reduction intervention's effects on condom-use outcomes. In a multisite randomized controlled trial conducted in four US cities, 535 African American HIV-serodiscordant couples were randomized to the Eban HIV risk-reduction intervention or attention-matched control intervention. Outcomes were proportion condom-protected sex, consistent condom use, and frequency of unprotected sex measured pre-, immediately post-, and 6 and 12 months post-intervention. Potential mediators included Social-Cognitive-Theory variables: outcome expectancies and self-efficacy. Mediation analyses using the product-of-coefficients approach in a generalized-estimating-equations framework revealed that condom-use outcome expectancy, partner-reaction outcome expectancy, intention, self-efficacy, and safer-sex communication improved post-intervention and mediated intervention-induced improvements in condom-use outcomes. These findings underscore the importance of targeting outcome expectancies, self-efficacy, and safer-sex communication in couples-level HIV risk-reduction interventions.