Background: Opioid overdose related deaths have increased dramatically in recent years. Combating the opioid epidemic requires better understanding of the epidemiology of opioid poisoning (OP) and opioid use disorder (OUD).
Objective: We aimed to discover geospatial patterns in problematic opioid use and its correlations with demographic features related to despair and economic hardship, most notably the US presidential voting patterns in 2016 at census tract level in New York State.
Methods: This cross-sectional analysis used data from New York Statewide Planning and Research Cooperative System claims data and the presidential voting results of 2016 in New York State from the Harvard Election Data Archive. We included 63,958 patients who had at least one opioid use disorder (OUD) diagnosis between 2010 and 2016, and 36,004 patients with at least one opioid poisoning (OP) diagnosis between 2012 and 2016. A logistic regression model was used to determine the associations between patient level characteristics (sex, age group, race, and payment type) and OUD and OP patient rates at census tract level.
Results: Several areas shared similar patterns of OUD rates and Republican vote: census tracts in Western New York, Central New York, and Suffolk County. The Spearman rank correlation between OUD rates and the Republican vote was 0.38 (P < 0.0001). A multiple regression model of census tract level demographic and socioeconomic factors explains 29% of the variance in OUD rates, with disability and republican vote the biggest predictors.
Conclusions: At the census tract level, opioid use disorder rates were positively correlated with Republican support in the 2016 presidential election, disability, unemployment, and unmarried status. Socioeconomic and demographic features explain a large portion of the association between the Republican vote and opioid use disorder. Together, these findings underscore the importance of socioeconomic interventions in combatting the opioid epidemic.