Association for Public Policy Analysis & Management

Although the opioid epidemic has had a profound impact on many population subgroups in the U.S., little is known about the racial, ethnic and geographic disparities in prescription opioid utilization. This study explores community-level racial/ethnic composition as a risk factor for opioid use, in addition to individuals’ own characteristics.

Our main data source is Medical Expenditure Panel Survey (MEPS) from 2013 to 2016. We use confidential geographic identifiers to supplement MEPS with community information on racial/ethnic, socioeconomic, and industry composition at the census block group level from the American Community Survey. Our outcome is whether the individual had any outpatient opioid pain prescription fills per year. The detailed demographic, socioeconomic, and clinical characteristics of individuals are obtained from MEPS. We first estimate Ordinary Least Squares (OLS) models separately for each racial/ethnic group to identify determinants. Then, using the Oaxaca-Blinder decomposition method, we decompose the gaps in the mean of the outcome into (1) group differences in the means of the determinants of the outcome in question and (2) group differences in the effects of these determinants.

While non-Hispanic whites were more likely to have outpatient opioid prescription fills than those in other racial/ethnicity groups, they were less likely to be associated with risk factors, such as poor health, lower income, and uninsurance. Our future analysis exploring the differences in the effects of these risk factors across race/ethnicity and also the differences in community-level risk factors will shed lights on whether opioid epidemic has spread through different mechanisms depending on race/ethnicity.