Public insurance expansions and smoking cessation medications

Abstract

We study the effect of public insurance on smoking cessation medication prescriptions and financing. We leverage variation in insurance coverage generated by recent Affordable Care Act expansions to Medicaid. We estimate differences-in-differences models using administrative data on the universe of Medicaid-financed prescriptions sold in retail and online pharmacies 2011-2017. Our findings suggest that these expansions increased Medicaid-financed smoking cessation prescriptions by 34%. This increase reflects new medication use and a shift in payment from private insurers and self-paying patients to Medicaid. Adjusting our estimate for changes in financing implies that Medicaid expansion lead to a 24% increase in new medication use.

Keywords: I1; I13; I18; JEL classification; Smoking cessation; expenditures; healthcare; prescriptions; public insurance.

Figure 1.. Trends in smoking cessation medication prescription fills and refills per 100,000 non-elderly adults before and after Medicaid expansion: 2011–2017

Source: State Drug Utilization Data. Notes: Unit of observation is the expansion-period. All outcomes are converted to the rate per 100,000 non-elderly adults. States with substantial expansions before 2011 excluded from the analysis (see Appendix Table 1). Data points more than 12 periods in advance of or 14 periods following Medicaid expansion for expansion states and Q2 2014 for non-expansion states are excluded. N=1,207. Mean value in expansion states in pre-expansion period is 257.

Figure 2.. Event study of the effect of Medicaid expansions on smoking cessation prescription fills and refills per 100,000 non-elderly adults 2011–2017

Source: State Drug Utilization Data. Notes: Unit of observation is a state-year-quarter. All outcomes are converted to the rate per 100,000 non-elderly adults. All models estimated with OLS and control for state characteristics listed in Table 1, and state and period fixed effects. The omitted period is the quarter prior to expansion for expansion states. Non-expansion states are coded zero in all state-quarter lag and lead variables. Data points more than 12 periods in advance of or 14 periods following Medicaid expansion are excluded. 95% confidence intervals account for state-level clustering and are reported in vertical bars. States with substantial expansions before 2011 excluded from the analysis (see Appendix Table 1). N=1,207. Mean value in expansion states in pre-expansion period is 257. Coefficient estimates are average marginal effects.

Figure 3.. Event study including state-specific linear time trends of the effect of Medicaid expansions on smoking cessation prescription fills and refills per 100,000 non-elderly adults 2011–2017

Source: State Drug Utilization Data. Notes: Unit of observation is a state-year-quarter. All outcomes are converted to the rate per 100,000 non-elderly adults. All models estimated with OLS and control for state characteristics listed in Table 1, state and period fixed effects, and state-specific linear time trends. The omitted period is the quarter prior to expansion for expansion states. Non-expansion states are coded zero in all state-quarter lag and lead variables. Data points more than 12 periods in advance of or 14 periods following Medicaid expansion are excluded. 95% confidence intervals account for state-level clustering and are reported in vertical bars. States with substantial expansions before 2011 excluded from the analysis (see Appendix Table 1). N=1,207. Mean value in expansion states in pre-expansion period is 257. Coefficient estimates are average marginal effects.