Personal Responsibility in Medicaid: Evidence from the Healthy Indiana Plan
Discussant: Thomas Buchmueller
The primary study method compares outcomes among targeted socio-demographic groups in Indiana, compared to a set of states that did not expand in 2015. However, since 2014 expansion states were still experiencing growth of enrollment in 2015, we limit our control group to never-expansion states. We use DD models that alternatively feature neighboring non expansion states, all non expansion states nationally, as well as synthetic control methods of Abadie, Diamond and Hainmueller (2010) for situations of one treated state compared to multiple control states. Within Indiana, we examine outcomes in counties with low vs high baseline uninsurance rates. Where appropriate, we also compare outcomes among 19-64 yr olds to other relatively less affected age groups. We also compare results in Indiana to states that carried out more traditional Medicaid expansions to understand the impact of Indiana’s personal responsibility features.
The outcomes analyzed include enrollment data from the state and from CMS, insurance coverage (American Community Survey: state level and PUMA level analysis, data through 2016), health insurance access (Behavioral Risk Factor Surveillance System: state level, data through 2016), prescription drugs (state by class level, through 2017, administrative totals for all states), early prenatal care (Natality data: state and county level, through 2015), and hospitalization outcomes (detailed data for certain reasons for visit, from Indiana, through 2017).
The results from these population-level analyses serve as a complement to evaluation reports by the state government and by CMS, and will help broaden the knowledge base regarding the impacts of the Healthy Indiana Plan and personal responsibility in Medicaid.