December 1993

Authors: Randall S. Brown; Jeanette W. Bergeron; Dolores Gurnick Clement; Jerrold W. Hill; Sheldon Retchin

Executive Summary

Since the early 1980s, the Health Care Financing Administration (HCFA) has been encouraging health maintenance organizations (HMOs) to provide Medicare coverage to enrolled beneficiaries in return for fixed prepaid premiums. A five-year evaluation of the Medicare risk program conducted by Mathematica Policy Research, Inc., (MPR) shows that the program is achieving some of its goals but not fulfilling its promise in other areas.

In April 1985, HCFA implemented the Medicare risk program for HMOs and competitive medical plans (CMPs) as a means of reducing costs to the Medicare program while increasing beneficiaries choice of health care delivery systems. By paying HMOs and CMPs a fixed monthly amount in advance for providing all of the Medicare-covered services an enrollee might need, HCFA created incentives for these providers to deliver care more cost-effectively than providers under Medicares conventional fee-for-service (FFS) method of reimbursement. In order to share in the expected savings, HCFA set the payment rate at 95 percent of the actuarial estimate of what it would typically spend in reimbursements for these beneficiaries, had they not joined an HMO. To assess whether the risk program achieved these objectives and delivered care of comparable quality to that provided by FFS providers, MPR and the Medical College of Virginia conducted an extensive evaluation of the risk program for HCFA.

Findings In Brief

The risk program does not save money for HCFA in fact, costs are higher than they would have been had the enrollees not joined the HMOs. Costs under the risk program were 5.7 percent higher than they would have been under FFS because beneficiaries with chronic health problems were less likely than healthy beneficiaries to enroll in HMOs, and the payment (capitation) rates failed to reflect this favorable selection fully. Although payment rates to HMOs were set at 95 percent of HCFAs projected FFS cost for enrollees, these projections were too high, by about 11 percent, on average. HCFAs simple method of basing the payment rate for individuals on their age, gender, and a few other readily available characteristics fails to account fully for the healthier-than-average mix of beneficiaries who choose to enroll in HMOs. Thus, instead of saving 5 percent as intended, HCFA spent nearly 6 percent more than it would have for enrollees had they not joined the HMOs.

Although the payment rates were set too high to result in savings to HCFA, HMOs delivered care more efficiently than FFS providers without reducing quality of care. Even when the favorable selection of enrollees is considered, HMOs used fewer hospital days. They also utilized other expensive resources (for example, home health visits and skilled nursing facility days) less intensively. HMO members were more likely than nonmembers to have had some recent contact with a physician, but again HMO care was less resource intensive fewer patients had a large number of visits, and patients with chronic ambulatory conditions used specialists less and had fewer follow-up visits.

HMOs and FFS providers delivered care of comparable quality for both inpatient and ambulatory care. The percentage of hospitalized patients who died and the percentage readmitted were similar in the HMO and FFS sectors for both conditions examined (stroke and colon cancer). Similarly, for each of the three ambulatory health problems examined, HMO enrollees who had the condition were just as likely as their FFS counterparts to be symptom-free by the time of our interview. Lower resource intensity appears to be the result of more efficient use of services, not evidence of poorer quality care.

Beneficiaries enrolled in risk plans generally feel that HMOs lower premiums and more extensive benefits are adequate compensation for a lower level of satisfaction with their health plan. Although few HMO members are dissatisfied with their quality of care, access to care, and the amount of personal attention they receive, the proportion rating their care as excellent on any of these dimensions is significantly lower than the proportion of non enrolled beneficiaries. However, enrollees are much more satisfied than non enrollees with their costs, and 14 out of 15 enrollees would recommend their plan to a friend or family member.

HMOs appeared to have mixed experiences in the risk program, with nearly half dropping out of the program by 1990. Among active plans in 1988-1990, about half earned profits in a given year.

The most successful risk plans tended to manage utilization aggressively and to have large enough enrollments to hold down average costs per member. HMOs more efficient use of resources suggests that, if the payment mechanism were changed to better account for favorable selection into HMOs, the risk program could yield savings to HCFA while still allowing HMOs to prosper. The results of our study suggest that basing the capitation payment for a beneficiary on whether he or she has a history of serious illness, in addition to the demographic characteristics already incorporated into the payment formula, would eliminate the increased costs to HCFA.