Causal pathways and cost-effectiveness of DREAMS' impact on psycho-social mediators, social norms, and adolescent health outcomes in rural and urban K

Persistently higher rates of HIV infection among adolescent girls and young women (AGYW) compared to their male peers, especially in East and Southern Africa has led to large investments in targeted HIV prevention. Barriers to reducing HIV incidence are multi-faceted, and funding agencies are increasingly supporting delivery of complex interventions. Comprehensive investments in adolescent development, including HIV prevention, are crucial in achieving the Sustainable Development Goals, and can yield short as well as long-term benefits for health, education and social-economic wellbeing. Understanding the magnitude and mechanisms of impact for comprehensive interventions can provide insights on how to modify them to improve their effectiveness, or adapt them to other contexts.

To enhance our knowledge and optimise multi-sectoral HIV programming among AGYW, this PhD aims to learn from an independent evaluation of a scaled-up complex intervention ("DREAMS"), by analysing large, observational quantitative datasets from prospective cohorts of AGYW in two settings in Kenya: informal urban settlements in Nairobi and Siaya (western Kenya, rural). The DREAMS (Determined, Resilient, Empowered, AIDS-free, Mentored, Safe lives) Partnership was implemented across ~60 districts in 10 African countries with high HIV burden since 2016, and implementation is continuing during 2021. DREAMS was not randomised, and aimed to target the most vulnerable AGYW. The DREAMS core package provides comprehensive HIV prevention programs to decrease risk among AGYW and enhance supports at the individual and contextual levels. Interventions at the individual level aim to empower AGYW, and include HIV testing services, social asset building, or sexual and reproductive health services. Contextual-level interventions include social protection programs, community-wide efforts to address social norms for violence prevention, and strategies to reduce HIV risk among the male sexual partners of AGYW.

Current research from the independent evaluation conducted between 2016 and 2019 shows that uptake and layering of multiple interventions improved over time and reached high proportions of AGYW in both settings. Participation in DREAMS had positive and consistent impacts on some outcomes like social support and uptake of HIV testing, but impacts on other outcomes (e.g., recent pregnancy) differed by setting, and between younger AGYW (15-17 years) vs older (18-22 years).

Building on this earlier work, this project seeks to explain these findings in greater detail, and explore the causal pathways from DREAMS interventions, through potential psycho-social mediators to important developmental health outcomes relating to social protection behavioural protection, and biological protection. By analysing observational data using causal inference approaches, I will conduct my research guided by emerging theoretical methodological frameworks such as emulation of hypothetical trials, and causal mediation analysis, and apply advanced statistical methods for analysing these kinds of data. This will allow me to advance my understanding of how to obtain reliable causal estimates of effect from non-randomised studies, and contribute knowledge that other researchers can apply in their work. Through further training facilitated by the studentship, I will enhance my interdisciplinary skills in social research methods and health economics, which are crucial for answering some of the research questions alongside the quantitative statistical analysis. Insights from this PhD will help strengthen continued programming of DREAMS, and more broadly will make a valuable contribution to applied research aimed at understanding how to improve and optimise complex, multi-sectoral intervention delivery for HIV prevention among AGYW.