Harnessing Predictive Modeling to Identify Current and Emerging Hotspots within the HIV/SUD Syndemic - Abstract With over 1.2 million people living with HIV (PWH) and a high prevalence of substance use disorder (SUD) among PWH—6.5 times higher than in the general United States (U.S.) population—syndemic HIV/SUD presents a barrier to the nation’s goal of ending HIV in the U.S. by 2030 (i.e., Ending the HIV Epidemic: A Plan for America initiative). The prevalence of SUD among PWH varies significantly across local communities in the U.S., reflecting differences in social and economic conditions and public health policies. To our knowledge, excluding alcohol-related research, no study has systematically examined the prevalence and correlates of SUD among PWH in a representative population. Addressing this critical gap is essential for developing targeted, effective interventions and policies. Additionally, without a comprehensive understanding of the local context and its impact on the HIV/SUD syndemic with regard to emerging hotspots, the HIV epidemic will continue to intersect with and amplify challenges related to SUD, stymying progress toward achieving national health goals. Despite the recognized critical role of social determinants of health (SDoH) in shaping HIV/SUD trajectories, few studies have focused on the complex interplay between substance use, HIV, and SDoH. This gap persists due to several key challenges, including: 1) limited longitudinal data resources for PWH; 2) the challenge of integrating individual- and aggregated community-level data; 3) the need for advanced analytical methods to address the high dimensionality of SDoH amidst dynamic, evolving patterns of HIV and SUD; and 4) the lack of cohesive, multidisciplinary frameworks that bring together expertise across epidemiology, biostatistics, social science, and public health. Leveraging our multidisciplinary team and preliminary HIV and SUD work in South Carolina, we will examine the HIV/SUD syndemic in the state by integrating the PWH cohort with the Agency for Healthcare Research and Quality's database on SDoH using our innovative Big Data approach. We will develop a novel dynamic spatial model that accounts for multivariate outcomes to understand dynamic co-incidence patterns of HIV and SUD by substance type at the county level (Aim 1); apply advanced machine learning techniques to identify SDoH that contribute to HIV and SUD patterns and may be amenable to targeted interventions (Aim 2); and develop predictive models to identify future high-risk hotspots for co-occurring HIV and SUD (Aim 3), enabling evidence-based, proactive resource allocation and intervention planning. The project will provide critical insights into community-based factors linked to poorer treatment access and outcomes for HIV and SUD, guide targeted policy and intervention strategies, and create a scalable framework for addressing future public health challenges. By leveraging diverse data sources and innovative modeling, this research has the potential to transform the prevention and management of HIV and SUD not only in South Carolina but also nationally by creating a predictive modeling tool that can be applied in other states and public health systems.
