Restoring Medicaid Coverage to Persons with Mental Illness Released from Prison

Thousands of offenders with serious mental illness are released from state prisons each year. Among the many barriers to a successful reintroduction to community living is the often considerable lag experienced in obtaining health care coverage post incarceration. Medicaid enrollment can be difficult to navigate even in the best circumstances, but can be even more difficult to obtain for persons with serious mental illness (SMI), resulting in delays in service and medication use. Several states have implemented expedited Medicaid restoration (EMR) policies to facilitate their re-entry, improve their access to needed community services, and reduce recidivism. However, little is currently known about effectiveness of these efforts and whether benefits, if any, accrue to the medical, mental health, or criminal justice systems. States participating in the Medicaid expansion under the Affordable Care Act (ACA) will see automatic enrollment of many offenders with mental illness released from state prisons.

Administrative data from corrections, Medicaid, and public mental health authorities in Washington State were linked for persons released from prison over a two-year period (2007-08). Persons were followed for 36 months post-release. Capacity constraints in both states prevented all eligible persons with SMI from receiving expedited Medicaid benefits, thus creating a natural comparison group of persons with SMI who were released from prison over the same time period, but who did not receive EMR (646 cases and 2450 controls).

Because of concerns over selection bias, we used propensity score weighting to obtain better balance on baseline risk factors, using a rich set of characteristics on offenders with SMI during the 3 years prior to incarceration. Study outcomes include the use and costs of health and mental health services in the community, subsequent criminal recidivism, and overall days in the community (defined as days not incarcerated and not in a hospital).  Cost models were run using generalized linear models with a gamma distribution and log link. Models were compared to OLS models, which yielded very similar results.

Persons referred for EMR had a 36% point higher probability of having any Medicaid coverage during the 12 months post release (p<0.01). We find a higher rate of use of antipsychotic and antidepressant medication by those receiving EMR, slightly greater use of outpatient mental health ($28; p<0.01) and outpatient medical care ($516; p<0.01) by EMR recipients. Costs of alcohol and drug treatment were also higher in the EMR group ($158; p<0.05). We find no difference in the cost of local hospital or state psychiatric hospitalizations, the rate of recidivism or the number of community days between the two groups.

These results indicate that expediting Medicaid increases access to services, thus increasing costs in the short run (1 year). These investments in health services do not seem to generate spillovers to the criminal justice sector during the 12 month window; results 36 months post-release are pending and will be available for conference presentation. These results provide important information on the effects of providing insurance coverage during a critical period of community reintroduction for persons with severe mental illness.