Venue: The Fuqua School of Business, Duke University, 1 Towerview Drive, Durham, NC 27708-0120
Presentation
Decomposing the Effect of Medicaid on Emergency Department Use
Rationale: The uninsured are often assumed to be more likely than the insured to use the emergency department (ED). However, recent research suggests that Medicaid recipients are twice as likely as the uninsured or privately insured to have an ED visit. Little research has explored reasons for the discrepancies in visits between Medicaid recipients and the uninsured.
Objectives: This paper seeks to examine the effects of observable characteristics such as demographics, socioeconomic status, health status, access and attitudes towards health insurance on differences in ED utilization, as well as the contribution of behavioral differences between the two populations.
Methodology: Using 2004 Medical Expenditure Panel Survey (MEPS) data, we apply nonlinear Blinder-Oaxaca decomposition analysis to analyze ED utilization of two subgroups: individuals who were covered by Medicaid for the entire year and individuals who were uninsured for the entire year. This analysis allows researchers to quantify the effect that differences in observables between the groups have on the differences in ED utilization, as well as examine the magnitude of the effect of differences in behavior or treatment between the groups that might affect the disparity in utilization.
Results: 27 percent of individuals age 19-64 who were enrolled in Medicaid for all of 2004 had an ED visit, while 10 percent of those who were uninsured the entire year had an ED visit. Of the 17 point difference in likelihood of visit between the two groups, 9 percentage points can be explained by differences in observable characteristics. The remaining 8 points are thus attributable to unobservable characteristics. Although observable characteristics account for just over half of the difference in ED utilization between Medicaid recipients and the uninsured, there does not appear to be a major observable characteristic that contributes substantially to this difference.
Conclusions: Medicaid recipients are more likely to have an ED visit than those who are uninsured all year. Decomposition analysis results show that differences in observable characteristics or endowments between the two groups explain about half of the disparity in visits, suggesting that unobserved behavioral differences and care-seeking behaviors drive much of the differences in ED utilization. These unobserved differences may reflect the low or no co-pays required in most states upon ED visit, limited availability to providers willing to see Medicaid patients outside of the ED, and other behavioral differences that might influence Medicaid enrollees to seek care in the ED.