Mental Health Treatment Study: Replication and Extended Impacts on Services Use

University Of Maryland Baltimore County, Baltimore MD

Investigators

Abstract

? DESCRIPTION (provided by applicant): The Mental Health Treatment Study (MHTS), a national randomized trial, funded by the Social Security Administration (SSA), provided access to evidence-based treatment and rehabilitation interventions for SSDI beneficiaries, including dual eligibles, with serious mental illness (SMI). Key MHTS interventions were systematic medication management, care coordination, and supported employment. The MHTS enrolled 2,238 persons during 2006-2008, randomized to treatment or control, and assessed impacts over a 2-year follow-up. Evaluation results, based on self-report data, revealed potentially important positive outcomes on (1) mental health status, (2) earnings, and (3) reduced use of inpatient and crisis outpatient services. Evidence indicated these positive outcomes increased over time in the 2-year period. We propose research to replicate and extend previous MHTS evaluation studies of service use impacts by analyzing a large database of CMS administrative data on MHTS trial enrollees. Our replications studies will broaden the scope of previous studies by including longer follow-up periods, up to 4 additional years, and a much wider range of behavioral health services covered by Medicare or Medicaid, including both inpatient and all outpatient services, as well as use of prescription pharmaceuticals. We will also develop separate service use impact estimates for major sub- groups of MHTS participants, including persons with schizophrenia. Major data sources for the analyses will include: Medicare claims and enrollment files, Medicaid MAX data, all MHTS interview data, and all SSA administrative data previously used in the MHTS study. Our proposal is a collaboration of the researchers from UMBC, Westat, and SSA involved in earlier MHTS studies. We will estimate MHTS impacts on volumes and patterns of service use, on major sub-categories of services, and sub-group/interaction impacts of MHTS using regression methods. Methods common in health services utilization studies will be used, including hurdle, zero- inflated, and other multi-part models. We will also test for attrition bias in our results using administrative data. Our analyse have the potential for significant public health impacts for both substantive and methodological reasons. The group of persons from which MHTS enrollees were drawn - adults with SMI who are receiving public income support and coverage through public health insurance programs - is large and accounts for an important share of total services provided through these public health insurance programs. Thus, full-scale implementation of the MHTS interventions could have an important influence on the volume and mix of services provided under these programs. From a methodological perspective, our findings will be based on more reliable administrative data, a longer follow-up period, a broader scope of service use outcomes, and assessment of attrition bias. These findings can serve to illustrate the public health importance, in future services intervention trials, of incorporating these same methods and leveraging Federal data sources across agencies, rather than relying solely on narrower self-reported data over a fairly brief follow-up period.