Racial Inequities in Opioid Overdose Prevention: The role of local context in the effectiveness of state-level overdose prevention policies

PROJECT SUMMARY Deaths due to opioid overdose are a pressing public health crisis and the number of deaths per year is increasing. Rates of fatal overdose are rising faster for Black people than for any other racial/ethnic group, largely driven by the increased prominence of synthetic opioids. Black people have the highest percentage of overdose mortality attributable to synthetic opioids and have experienced the greatest increase in related mortality rates. Interventions are critically needed that can effectively reduce overdose mortality and do so equitably for Black people. Overdose prevention policies (OPPs) (i.e., Good Samaritan laws and naloxone access laws), a class of state-level policy interventions intended to reduce overdose mortality, may be able to address rising racial inequities in fatal overdose. However, there are critical gaps in the literature regarding the effectiveness of OPPs, including a lack of prior research into which provisions may better reduce overdose deaths for Black people, and a lack of prior research into the extent to which local contextual factors modify the effects of state-level OPP provisions. Additionally, the common practice of enacting OPP provisions in packages creates a significant methodological challenge for standard causal inference approaches of assessing the effectiveness of individual OPP provisions. The scientific objective of this research plan is to assess the effectiveness of state-level OPP provisions to equitably reduce overdose mortality and identify local-level factors that may produce racialized effectiveness of provisions. This project uses novel causal machine learning methods in conjunction with a combination of restricted mortality data from the National Vital Statistics System and multiple publicly available data sources to address the methodological challenge and fill the critical gaps outlined above. This innovative data-driven approach will be complemented by taxonomies of hypothesized OPP provision effectiveness produced by a panel of opioid policy experts using the Delphi method. By doing so, this project will empirically evaluate which sets of OPP provisions are most effective at reducing overdose mortality overall, and specifically among Black people, and estimate the role of local contextual factors (e.g., access to harm reduction services or local law enforcement practices) in producing varied effects of OPP provisions. This research plan is complemented by a career development plan that builds upon the applicant’s background in epidemiology and biostatistics and includes new training in (1) development and implementation of state-level drug policies; (2) measurement and evaluation of policy intervention effectiveness; and (3) machine learning methods to identify salient causal measures from high- dimensional data. The combined research and training plan will prepare the applicant to successfully transition to an independent research career aimed at using novel statistical and computational methods to identify and evaluate interventions to reduce racial inequities in substance use related harms.