Spatiotemporal Analysis of Opioid Dispensing Rates on Opioid Overdose Mortality in Counties Across the United States

Keywords: opioid, opioid overdose, opioid dispensing, spatiotemporal analysis, multivariate regression
Abstract Type: Poster Abstract

Abstract

The opioid overdose death rate in the United States has increased over the past decade, rising from 7.3 per 100,000 in 2011 to 24.7 per 100,000 in 2021 [1]. Research indicates that the prevalence of opioid dependence and misuse at the individual level are directly correlated with county level dispensing rates [2]. We hypothesized that counties in the United States with higher opioid dispensing rates are associated with higher opioid related overdose death in those counties. We sourced county-level data for 2019 to 2022 opioid dispensing rates from the CDC Overdose Prevention database and county level opioid-related drug poisoning (overdose) incidents from CDC WONDER. We used ArcGIS Pro to visually display data. Salem County, Virginia had the highest opioid dispensing rate, while Lee County, Kentucky had the highest opioid overdose death rate. GeoDa was used to run an exploratory multivariate spatial regression on the association between county opioid dispensing rates (independent variable), poverty rates (covariate), and crude death rate due to opioid overdose (dependent variable). In the multivariate regression adjusting for poverty level, the relationship was statistically significant (β < 0.0001, p=0.00014), however our R-squared value was fairly low (0.15). These results indicate that there is significant multivariate association between crude overdose death rates, prescription opioid dispensing rates and poverty rates, but further models should be generated to understand what covariates, such as age, gender, and health care access are associated with the variation in opioid death rates.