Low-income populations are disproportionately affected by smoking, and Medicaid coverage for smoking cessation treatments has expanded considerably over the past couple of decades. The 2009 Affordable Care Act further increases cessation services in Medicaid by both increasing the size of the population eligible for such services and increasing the generosity of coverage. While coverage and higher generosity of coverage can encourage the use of cessation treatments, their impact on actual quitting is not clearly established; evidence from Medicaid is scarce, primarily correlational, and inconclusive. While case studies from Massachusetts (Land et al 2010) and California (Schauffler et al 2001) recorded reductions in smoking rates after the introduction of cessation coverage, an early study using cross-state data to evaluate the link between quitting and Medicaid cessation coverage (Liu 2009) showed conflicting results that were not robust to controlling for unobserved state characteristics. A subsequent study of pregnant women (Adams et al 2013) concluded that smoking in expectant mothers enrolled in Medicaid before pregnancy was lower in states with cessation coverage, but a number of limitations prevented identification of causal effects. Most recently, Greene et al (2014) extended Adams et al (2013) by examining the relationship between quitting and cessation coverage in the general (not pregnancy-restricted) Medicaid population. Using both cross-state and within-state variation in coverage, they found a positive association between quitting and increased coverage, but, as in previous studies, identification of effects was limited by the absence of a comparison group. Thus, existing evidence on the effects of Medicaid cessation coverage is based on trends among Medicaid beneficiaries without a counterfactual, potentially biasing the estimates by not accounting for underlying trends in quitting behavior that may not be specific to Medicaid yet may coincide with increased Medicaid generosity in cessation coverage.
This study expands the literature by estimating the effects of Medicaid cessation coverage on past-year quitting in a difference-in-differences framework, using changes in Medicaid benefits within states and employing a counterfactual group of low-income adults not on Medicaid. We evaluate two aspects of smoking cessation treatment: pharmaceutical (nicotine-replacement therapy and prescription medication), and non-pharmaceutical (counseling), each with two levels of generosity (copayment and no copayment required). We obtain individual-level data with restricted state identifiers from the 2002-2014 National Health Interview Survey, using 9 states that introduced pharmaceutical coverage and 14 states that introduced counseling coverage during the study period. Preliminary results suggest that introducing Medicaid coverage for pharmaceuticals increases the likelihood of quitting, provided that this coverage does not require copayments, and that this effect may be stronger among lower-income beneficiaries. We find relatively larger effects for counseling coverage, which appears to be effective regardless of cost-sharing. Identifying the effects of expanded cessation coverage on quitting can inform the process of weighing the associated benefits and costs, and can be useful for identifying optimal levels of generosity in coverage of different interventions.
