Poster Paper: Does Medicare Increase Opioid Drug Usage: A Regression Discontinuity Study

Saturday, November 10, 2018
Exhibit Hall C - Exhibit Level (Marriott Wardman Park)

*Names in bold indicate Presenter

James W. Henderson, Scott Cunningham and Alexander Law, Baylor University

The overuse of opioids in the United States is a well-documented problem with annual deaths from overuse reaching upwards of 60,000. Policy makers and opinion leaders alike are calling it a national epidemic. The problem of moral hazard predicts that services covered by insurance will increase the likelihood of overuse because of the individual coverage risk and financial burden are removed. Medicare Part D provides prescription drug coverage for outpatient medications including opioid drugs for all individuals over the age of 65.

The primary source of data is the Integrated Public Use Microdata Series Medical Expenditure Panel Survey (IPUMS/MEPS). We append data on medical conditions from the Household Component Full-Year files and the Prescribed Medicines files for the Household Component Event files. The usable data set includes 590,233 observations.

The simple observational approach seems to indicate at lease a six percent statistically significant increase in opioid usage after an individual becomes Medicare eligible. Because age and Medicare eligibility are endogenous with several chronic conditions whose likelihood increases with age, we use age as the running variable for a sharp regression discontinuity design. In it, we use a dummy variable for the relevant threshold to indicate Medicare eligibility. Using this approach, we find no statistically significant increase in opioid usage from the age threshold for Medicare eligibility. Robustness checks find no statistically significant results when the sample is limited to specific medical conditions, dosages, or type of opiate. Narrowing age bandwidths from 15 to 10 to 5 years around age 65, the results do not change.