Keywords: Health Insurance, Medicaid expansion, Emergency Department use, Propensity score weighting, boosted regression trees
A growing body of research shows that Medicaid expansion is not dramatically increasing the number of ED visits as had occurred in some state level expansions. However, whether those newly gaining Medicaid coverage utilize EDs differently is unknown. Using patient visit-level data from a national emergency medicine group, we examined changes in ED use by newly-covered Medicaid patients compared with those remaining uninsured in non-expansion states from 30 EDs (2013-15) across 7 states. Boosted regression trees were used to balance a large set of clinical and health care spending variables (twang). We ran propensity-score weighted models and found that in comparison to non-expansion states, patients gaining Medicaid coverage in expansion states showed statistically significant increases in the proportion of visits that resulted in overall and non-ambulatory care sensitive hospital admissions from the ED and ‘unavoidable emergent’ ED visits (by 3.1, 2.7, and 1.1 percentage points respectively, reflecting 21%, 23%, and 7% increases, p<0.05 for each). Obtaining Medicaid insurance through the ACA shifted patients’ ED use towards higher-acuity medical conditions.