Gambling with Health: A Re-Evaluation of the Oregon Medicaid Lottery

In Baicker et al's “The Oregon Experiment – Effects of Medicaid on Clinical Outcomes,” the author’s utilize a “natural experiment” in which low-income adults were randomly selected to apply for Medicaid to estimate the causal effect of Medicaid on clinical outcomes, healthcare utilization, and health expenditures (Baicker et al, 2013). The 2008 Oregon Medicaid lottery provided something akin to a randomized clinical trial for Medicaid coverage across the state; however, the authors merely look at the effect of having any Medicaid coverage on these outcomes and not how long a person was covered by Medicaid. In our analysis, we look at the effect of 1-6 months, 7-12 months, 13-18 months, 19-24 months, and greater than 25 months of Medicaid coverage and we find significant health impacts from Medicaid coverage do not start prior to 18 months. This suggests that with an average follow up time of only 25 months, Baicker et al’s study ended too soon to adequately measure the impacts of the Medicaid expansion.