Abstract
Background
Mortality from overdoses involving opioids in the United States (U.S.) has reached epidemic proportions. More research is needed to examine the underlying factors contributing to opioid-related mortality regionally. This study’s objective was to identify and examine the county-level factors most closely associated with opioid-related overdose deaths across all counties in the U.S.
Methods: Using a national cross-sectional ecological study design, we analyzed the relationships between 17 county-level characteristics in four categories (i.e. socio-economic, availability of medical care, health-related concerns, and demographics) with opioid mortality. Data were extracted from the Robert Wood Johnson County Health Rankings aggregate database and Centers for Disease Control and Prevention (CDC)’s Wide-ranging Online Data for Epidemiological Research (WONDER) system.
Results: There were 1058 counties (33.67% of 3142 nationally) with reported opioid-related fatalities. Median opioid-related mortality was 15.61 per 100,000 persons. Multivariate regression results indicate that counties with the highest opioid-related mortality had increased rates of tobacco use, HIV, Non-Hispanic Caucasians, and females and were rural areas, but lower rates of food insecurity and uninsured adults. The rates of tobacco use and HIV had the strongest association with mortality. Availability of either mental health or primary care providers were not significantly associated with mortality. Severe housing problems, high school graduation rate, obesity, violent crime, and median household income also did not contribute to county-level differences in overdose mortality.
Conclusions: Future health policies should fund further investigations and ultimately address the most influential and significant underlying county-level factors associated with opioid-related mortality.
Acknowledgments
We would like to thank all of those who helped directly or indirectly in this study’s data collection, including Meredith Margaret O’Neal and Joshua Gross. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Author contributions
Research conception and design (JL, TCL), collection of data (JL and TCL), analysis (MLC and JL), interpretation of the results (JL, KC, and MLC), writing (JL, KC), revision (KC, JL, TCL, MLC).