The 130th Annual Meeting of APHA

Abstract #46455

Joseph P. Morrissey, PhD¹, Kathleen C. Thomas, PhD¹, T. Scott Stroup, MD, MPH¹, and Kathleen Holladay². (1) Cecil G. Sheps Center for Health Services Research, University of North Carolina, Chapel Hill, 101 Conner Dr. Ste.302, Willowcrest Bldg., CB#3386, Chapel Hill, NC 27599, 919-966-5829, joe_morrissey@unc.edu, (2) Cecil G. Sheps Center for Health Services Research, University of North Carolina Chapel Hill, 101 Conner Dr. Ste.302, Willowcrest Bldg., CB#3386, Chapel Hill, NC 27599

This study compares self-reported service use with Medicaid data for a population of Medicaid recipients with severe mental illness (SMI). These data are used to assess the impact of a capitated Medicaid plan with a 'care as usual' plan. Using a quasi-experimental design, initial and follow-up interviews were conducted with SMI clients in both plans. Client interview data were then merged with Medicaid claims. Medicaid claims matched 74 percent of the client data. The Medicaid data provide information about the number and cost of claims. They also indicate in which plans a client received care. The Medicaid data provide an external source with which to compare self-reported utilization by SMI clients. Self-reported utilization data can vary from actual use. Clients may over or under-report use. Interviewer variability introduces bias. The operational definitions of utilization may be unclear for clients who typically see a variety of professionals for a variety of interrelated services. Analyses of self-report data show that utilization of costly services was lower under the capitated plan. The analysis will be extended to compare utilization in the Medicaid claims. These data will either serve to validate the self-report data, or to suggest another pattern of use. If claims data fail to support the self-reported finding of lower utilization in the capitated plan, it may be that dissatisfaction with the new plan led to under reporting of use that was not realistic. These results have important implications for future evaluation using self-reported and claims data on mental health services use.

Learning Objectives:

• Apply a method of merging unidentified claims data with survey data on the people who generated those claims.
• Assess the similarities and differences in self-reported and claims data on mental health service use by the severely mentally ill.
• Describe the implications of using each type of data for health plan evaluation.

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.