FIGURE 1. Difference-in-differences (DID) regression results estimating the effect of Medicaid expansion on the number of opioid treatment programs (OTPs) in a state per 100,000 population^a.

^a All models were conducted with the xtreg command in Stata, version 15, with standard errors clustered at the state level and with control for optional and mandatory-access prescription drug monitoring programs, pain clinic laws, state median income, state age distribution, state race-ethnicity, state poverty and unemployment rates, and state and year fixed effects.

^b Extended-release injectable naltrexone.

^c DID coefficients (change in number of programs per 100,000 population after Medicaid expansion). Shaded areas indicate 95% confidence intervals.