Assessing layered HIV prevention programming: optimizing outcomes for adolescent girls and young women

Objective: To assess how exposure to multiple, layered interventions predicts HIV-related outcomes among adolescent girls (15-19 years) and young women (20-24 years) in Kenya. Design: Survey data from adolescent girls and young women (n = 736) with 14-16 months of engagement with DREAMS, a comprehensive HIV prevention program that provides a range of health education, life skills, social protection, and social and behaviour change interventions. Methods: Nonparametric recursive partitioning technique - classification and regression tree (CART) - to identify the best predictors (DREAMS interventions) for achieving the desired HIV-related outcomes (consistent condom use and no transactional sex or sexual violence). Results: Among adolescent girls, schooling support reduced the likelihood of engaging in transactional sex, whereas schooling support and exposure to parenting program reduced the likelihood experiencing sexual violence. Likelihood of consistent condom use increased among adolescent girls with exposure to preexposure prophylaxis (PrEP), schooling support, and the violence prevention programming. Among young women, multiple pathways reduced the likelihood of engaging in transactional sex: exposure to the male sexual partner program; exposure to the youth fund program; exposure to the violence prevention program; or exposure/engagement with schooling support, parenting programming, and the youth fund program. For young women, consistent condom use increased with schooling support and male partner engagement. Additionally, engagement in violence prevention program and male partner engagement increased the likelihood of not experiencing sexual violence among young women. Conclusion: Exposure to a combination of DREAMS interventions predicted outcomes that can reduce HIV risk among AGYW, though the pathways differed by outcome and age group.