Medicaid managed care: are academic medical centers penalized by attracting patients with high-cost conditions?

OBJECTIVE: To determine whether case-mix and health utilization disparities exist between Medicaid enrollees within a Michigan managed care organization (MCO) who selected primary care providers (PCPs) affiliated with a major academic medical center (AMC) and enrollees who selected community providers.

STUDY DESIGN: A retrospective cohort study using cost estimates obtained from claims data and based on a standardized Medicaid fee schedule.

METHODS: We established the prevalence of 25 high-cost chronic medical conditions from the claims data for capitated Medicaid enrollees from January 1, 1997, through October 31, 1999. We assessed differences in healthcare cost estimates per member for Medicaid enrollees at AMC primary care sites versus other community sites using t tests and linear regressions, including analyses stratified for Temporary Assistance for Needy Families (TANF) and Aid to Blind and Disabled (ABAD) programs.

RESULTS: Enrollees with AMC providers had a much higher cumulative prevalence of the 25 high-cost chronic medical conditions (95.6 per 1000 enrollees versus 65.6 per 1000; P < .001), and virtually all of this difference was confined to ABAD enrollees. Estimated total costs were also higher for ABAD Medicaid enrollees at the AMC sites than for those at community sites. The average total services and pharmacy cost estimates per ABAD member were $1219 higher per member per year at the AMC sites (P < .001), primarily from costs of inpatient hospitalizations. Regression analyses demonstrated that differences in the prevalence of the 25 high-cost chronic medical conditions accounted for about 50% of the cost differences observed between sites. These analyses suggest that at least half of the observed cost disparity was due to adverse selection.

CONCLUSIONS: This study found both significant case-mix and cost disparities for ABAD patients, suggesting that AMC primary care sites experienced substantial adverse selection. Unless approaches to account for adverse selection are put in place, this phenomenon could jeopardize ABAD Medicaid recipients' ongoing access to needed medical care.