The Effect of Medicare Health Maintenance Organizations on Hospitalization Rates for Ambulatory Care-Sensitive Conditions
Published In: Medical Care, v. 44, no. 10, Oct. 2006, p. 900-907
Posted on RAND.org on January 01, 2006
OBJECTIVE: The objective of this study was to estimate the effect of Medicare Health Maintenance Organization (HMO) enrollment on hospitalization rates and total inpatient days for ambulatory care-sensitive conditions (ACSCs) after controlling for selection. RESEARCH DESIGN: Simultaneous equations using a discrete factor selection model are used to estimate the probability of HMO enrollment, hospitalization rates, and total inpatient days for ACSCs. SUBJECTS: Enrollment data on Medicare beneficiaries in California were linked to hospital discharge data from the California Office of Statewide Health Planning and Development for January through December 1996. The following beneficiaries were excluded: 1) end-stage renal disease, 2) under 65 years of age, 3) not covered by both Medicare Part A and Part B, 4) switched between HMOs and fee-for-service (FFS), and 5) switched between HMOs. The sample was stratified by age, gender, race, county, disability, Medicaid eligibility, HMO status, and death. A 2% random sample from the 4 California counties with the largest Medicare enrollment yielded 10,448 HMO enrollees and 11,803 FFS beneficiaries. RESULTS: Using a discrete factor selection model, we estimated the rate of ACSC hospitalizations among FFS beneficiaries would decline from 51.2 to 44.2 per 1000 if all FFS beneficiaries joined an HMO. Similarly, the mean total inpatient days for ACSC hospitalizations would be reduced from 7.5 days to 5.1 days if all FFS beneficiaries joined an HMO. CONCLUSIONS: After controlling for selection, Medicare HMO enrollees have lower hospitalization rates and fewer total inpatient days for 15 ACSCs than Medicare FFS beneficiaries. These findings suggest selection of healthier beneficiaries into HMOs does not completely explain their lower rates of ACSC hospitalization.