The Effect of Reduced Coverage for Inpatient Psychiatric Care on Emergency Department Utilization

This paper is the first to study the effect of reduced insurance coverage for inpatient psychiatric care on emergency department utilization. I examine this relationship by exploiting Medicaid's Institutions for Mental Disease exclusion, which states that for beneficiaries between the age of 21 and 64, inclusive, Medicaid will not reimburse for medically necessary inpatient psychiatric care administered at an institution with more than sixteen beds. I employ a "fuzzy" regression discontinuity framework using the near universe of emergency department and inpatient discharge records from 2005 through 2011 for two states. For Medicaid-eligible men, I find that at age 21 reduced coverage for inpatient psychiatric care increases mental health emergency department visits by 11 percent. For Medicaid-eligible women, I tend to find statistically insignificant effects. To address the concern that the minimum legal drinking age has a confounding effect at age 21, I define my outcome variable to exclude all alcohol-related visits. I also conduct a placebo test to check that there is no discontinuity in this outcome for private insurance beneficiaries (i.e., individuals that do not experience a change in insurance coverage at age 21). Consistent with the hypothesis that my outcome variable is not confounded by alcohol use at age 21, I find no discontinuity in mental health emergency department visits for this group. We can, therefore, interpret the increase in mental health emergency department visits as the unintended cost of reduced coverage for inpatient psychiatric care.