Vogel B, Johnson C, Radcliff T, Duncan P, DeWald L; AcademyHealth. Meeting (2003 : Nashville, Tenn.).
Abstr AcademyHealth Meet. 2003; 20: abstract no. 626.
University of Florida, Department of Health Policy and Epidemiology, Box 100177 Health Science Center, Gainesville, FL 32610 Tel. (352) 265-0680 x86541 Fax (352) 265-8047
RESEARCH OBJECTIVE: This research compares Medicaid payments for those Medicaid recipients who made the transition from MediPass, Florida's primary care case management system, to Florida's new Provider-Sponsored Network (PSN), a new incentive payment system designed for a large cohort of Medicaid managed care recipients in South Florida. STUDY DESIGN: We used a pre-post, quasi-experimental design, obtaining recipient, provider, and claims data on all the patients of any provider associated with the PSN over the period March 1999 through February 2001. This time period was chosen to bracket the implementation of the PSN in March 2000. Our goal was to compare the payment levels of those Medicaid recipients who had a MediPass enrollment span followed immediately by a PSN enrollment span. In so doing, we sought to isolate the impact of this new organizational form on Medicaid payments while holding constant the impact of recipient and provider characteristics on utilization and payment. POPULATION STUDIED: The 16,751 individuals who made an immediate transition from MediPass to the PSN over the period March-October 2000. PRINCIPAL FINDINGS: In comparing payment levels under MediPass to payment levels under the PSN, we found that expenditures were considerably lower for the PSN than for MediPass. For example, total expenditures per member month under MediPass were $405 and $288 under the PSN, for a difference of $117 per member month. These results strongly suggested that the PSN was incurring lower Medicaid expenditures than MediPass.Upon closer inspection, however, we discovered selection bias in the timing of MediPass transitions to the PSN, with lower cost enrollees transitions first and higher cost enrollees transitioning later. As a result, during the full eight-month transition between MediPass and the PSN, MediPass expenditures were higher and PSN expenditures lower than would have been expected if the entire cohort had been simultaneously transitioned.To adjust for this, we eliminated all data from the eight-month transition period. Using these more appropriate data, we found fewer and smaller differences in utilization and payment between MediPass and the PSN. For example, total expenditures per member month under MediPass fell to $367 (compared to the previous $405). Total expenditures per member month for the PSN increased to $349 (compared to the previous $288). This difference of $18 is considerably smaller than the original difference of $117. When administrative costs, withholds, and shared savings are examined, they add approximately $44 per member per month to PSN payments over March-December 2000, suggesting that the PSN had higher total Medicaid payments per member month than MediPass. CONCLUSIONS: We find evidence that during the initial months of operation, the PSN incurred higher Medicaid payment per member month than MediPass. IMPLICATIONS FOR POLICY, DELIVERY OR PRACTICE: For policymakers, our work demonstrates the importance of careful incentive design in innovative Medicaid programs. For researchers, our work emphasizes the importance of the nature of the transition in pre-post quasi-experiments in avoiding bias in the results.
Publication Types:
Keywords:
- Florida
- Health Expenditures
- Humans
- Managed Care Programs
- Medicaid
- economics
- hsrmtgs
Other ID:
UI: 102275913
From Meeting Abstracts