Study Design, Methods, and Modeling in Networks to Inform HIV Interventions and Policy in Marginalized Populations

The Networks and Causal Inference for Public Health Research (NCIPHER) Lab at the University of Rhode Island (URI) was established in 2018, with an initial pilot project supported by Advance Rhode Island Clinical Translational Research (RI-CTR), to develop methodological and computational approaches for evaluating interventions in real-world settings where individuals are connected in ways that impact their health. Leveraging bioinformatics and high-performance computing resources through Advance RI-CTR, we integrate empirical and simulation methods to estimate causal intervention effects (i.e., change in an outcome caused by a specific intervention) beyond treated individuals, including spillover through social and healthcare networks and clusters. Our work demonstrates that accounting for spillover improves understanding of the effectiveness of HIV and opioid use disorder (OUD) interventions. In the Transmission Reduction Intervention Project Athens, Greece, participants not exposed themselves to community alerts about HIV risk, with 50% of their immediate contacts exposed, reported three fewer unsafe injection behaviors (per 100 participants), compared to those who had 20% of their contacts exposed (95% CI: -7, 0). We also found 26 fewer reports (per 100 persons) at six months under individual treatment with medication for opioid use disorder with 60% of other individuals in the network treated, compared to no treatment with 20% of others treated (95% CI: -38, -13). Understanding effects in networks can improve translation from intervention delivery to population-level impact, supporting evidence-based policy.