County-Level Opioid Prescribing Rates, Socioeconomic Deprivation, and Opioid Overdose Mortality 2006-2016: Spatial Analysis and Mediator Decomposition in 3,084 U.S. Counties

This study investigates association between county-level opioid prescribing rates, socioeconomic deprivation, and opioid overdose mortality from 2006-2016 with spatial lag regression and mediation decomposition in 3,084 U.S. counties within the context of epidemiology and public health policy, an area of growing scientific importance given its implications for opioid prescribing policy targeting by deprivation index, spatial spillover naloxone distribution priority, and mediation-based intervention design for opioid epidemic response. Using spatial lag regression with queen-contiguity weights matrix, mediation analysis (prescribing rate as mediator of deprivation-mortality association), and panel fixed-effects regression with county and year fixed effects, we examine socioeconomic deprivation increasing prescribing through supply-side factors (pain clinic density, pill mill concentration in low-income rural areas) and demand-side vulnerability (unemployment, chronic pain, mental illness); prescribing creating opioid-dependent population whose overdose risk persists even after prescribing reduction due to transition to heroin/fentanyl in 3,084 counties x 11 years = 33,924 county-year observations; 284,000 opioid overdose deaths across study period; CDC WONDER suppression threshold applied (n<10 reported as censored) drawn from CDC WONDER multiple-cause death data, DEA ARCOS dispensing data, CMS prescriber-level data aggregated to county, and 2000/2010 Census ACS Area Deprivation Index. Results indicate that per 100 MME/person/year increase in prescribing rate associated with 1.84 additional overdose deaths/100k (spatial lag, p<0.001); ADI effect 38.4% mediated through prescribing; spatial autocorrelation Moran I=0.48 (p<0.001); Appalachian counties 2.84x national rate (p < 0.001), with +1.84 deaths per 100 MME; 38.4% mediated through prescribing; Moran I=0.48; Appalachian 2.84x as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to epidemiology and public health policy and carry actionable implications for the design of programs and policies targeting opioid prescribing policy targeting by deprivation index, spatial spillover naloxone distribution priority, and mediation-based intervention design for opioid epidemic response.