Trajectories of Medicaid Enrollment Stability Among Individuals with Serious Mental Illnesses

Background and Purpose: Medicaid is the largest payee of behavioral health services in the U.S., making it a critical link to needed care for individuals with serious mental illnesses (SMI). Research has found that, despite their disability status, individuals with SMI face frequent disruptions in Medicaid coverage which leads to increased acute care use and decreased outpatient engagement. Unfortunately, the Affordable Care Act, which expands and stabilizes Medicaid coverage based on income thresholds, has not stabilized Medicaid coverage for individuals who qualify based on disability. Prior studies of Medicaid stability among individuals with SMI have relied on dichotomous measures of enrollment disruption. The current study adds to this literature by (a) delineating the types of Medicaid enrollment trajectories that exist among people with SMI over time, and (b) examining whether mental health diagnosis predicts these trajectories.

Methods: County behavioral health service claims from 2006 were obtained from a state in the Midwest and used to develop a sample of individuals diagnosed with schizophrenia, bipolar disorder, or major depression. Medicaid enrollment data was then obtained for individuals who had any enrollment in Medicaid during the 2007-2010 study period. The sample for the current analysis (N = 2,965) includes those individuals who had (a) a new Medicaid enrollment during the study period, and (b) a first new enrollment that occurred at least 52 weeks prior to the end of the study period. Group-based trajectory models (GBTM) were used to analyze weekly Medicaid enrollment and disenrollment patterns for the 52 weeks following each individual’s first new enrollment. Optimal group-based trajectory models were selected based on group size and Bayesian Information Criterion. Logistic and multinomial regression models were used to examine the association between mental health diagnosis and trajectory-group memberships while controlling for demographic variables.

Results: Our final model reveals five Medicaid enrollment trajectories among individuals in our sample: stable enrollment (61.7%), partial instability (9.2%), immediate drop-off (11.8%), rapid drop-off (6.9%), and gradual drop-off (10.4%). Although stable enrollment was the largest group in our model, 38.3% of our total sample did not fall into this group. Additionally, mental health diagnosis was not associated with membership in the stable enrollment group. Individuals with bipolar disorder and major depression had greater odds of being assigned to the gradual drop-off group than did individuals with schizophrenia (OR = 1.61 and OR=1.65 respectively; p < .01).

Conclusion and Implications: Results from our GBTM reveal large amounts of heterogeneity in Medicaid coverage stability among individuals with SMI over a 52-week period. This suggests that many individuals with SMI struggle to maintain coverage even over relatively short periods of time and that those with the most impairing mental illnesses, such as schizophrenia, are not spared such instability. Notably, our trajectory models suggest that once Medicaid enrollees with SMI lose coverage, it is extremely difficult for them to subsequently gain enrollment stability. Future research should incorporate GBTM to capture the heterogeneity with which individuals with SMI experience Medicaid coverage disruptions and examine how these trajectories effect health and healthcare outcomes.