Frontiers in Operations: Equitable Data-Driven Facility Location and Resource Allocation to Fight the Opioid Epidemic.

This article focuses on developing an integrated approach combining a predictive epidemiological model and a prescriptive mixed-integer optimization problem (MIP) to improve equitable allocation of opioid treatment facilities and budgets across U.S. states. The epidemiological model is a state-level compartmental ordinary differential equation (ODE) system that captures opioid epidemic dynamics, including prescription and illicit opioid use, addiction, treatment, and overdose deaths, with parameters estimated via a neural ODE-inspired fitting process using state-specific data. The MIP uses these dynamics to optimize the location of medication-assisted treatment (MAT) facilities and budget distribution, prioritizing reductions in opioid use disorder (OUD) prevalence and overdose deaths while incorporating socioeconomic equity through the Centers for Disease Control's Social Vulnerability Index and opioid prescribing rates. Results indicate that the optimized allocations could decrease OUD prevalence by approximately 9%, increase treatment uptake by nearly 89%, and reduce opioid-related deaths by about 0.58% over two years compared to baseline projections, outperforming population- or vulnerability-based benchmarks. The framework is adaptable to other interventions and drug epidemics, offering policymakers actionable, data-driven strategies to address treatment access disparities in the opioid crisis.