The author ran Linear Probability Model and Logistic Model regressions on data from the 2017 Medical Expenditures Panel Survey to determine demographics of opioid prescriptions. The dependent variable was opioid prescription received (Y/N) and independent variables included age, education level, relative income, race, insurance status, etc. Results of the regressions showed that people with public or private insurance are more likely to be prescribed an opioid, that the problem mainly affects white people, and that as educational attainment increases, likelihood of opioid prescription decreases.
Linton, I., & Condliffe, S. (2020). Demographics of Opioid Prescriptions. Retrieved from https://digitalcommons.wcupa.edu/econ_stuwork/1