WASHINGTON -- In a significant policy change, Clinton administration officials say they will not prod elderly Medicare patients to join health maintenance organizations, in part because they have discovered that the government loses money on people enrolled in such private health plans.
Bruce C. Vladeck, head of the federal Health Care Financing Administration, which runs Medicare, said
he would not aggressively promote HMOs for beneficiaries until he could guarantee consistent high-quality care and had a better way of paying for it.
"Our payment methodology is so primitive that it doesn't save money," Mr. Vladeck said. "We are losing money on HMO patients. Before we undertake aggressive promotion of HMOs for Medicare beneficiaries, we have to make progress on both issues: payment methodology and quality assurance."
In the Reagan and Bush administrations, Medicare officials insisted that HMOs would both raise the quality of care for elderly people, by coordinating services, and save money for the government.
But a new study by MathematicaPolicy Research, a private consulting concern, said the government paid 5.7 percent more for Medicare patients in HMOs than it would have paid if those people had been in the regular Medicare program. Medicare "is not achieving its goal to save money" through HMOs, said the study, done under contract to the government.
The problem, essentially, is this: While Medicare paid the HMOs a per-capita fee close to the cost of services used by the average Medicare beneficiary, the patients who enrolled in the HMOs tended to be healthier than the average, meaning that they would probably have used fewer services anyway.
The report does not discuss President Clinton's plan to remake the nation's health care system, which seeks to hold down health spending by inducing large numbers of consumers to enroll in HMOs and other forms of managed care. But the findings suggest that it may be difficult to achieve the large savings that his plan envisions.
"HMOs reduce utilization and don't seem to cut the quality of care," said Randall S. Brown, chief author of the Mathematica study. "But if you want to reap any savings, you better set the payments right."
Under Mr. Clinton's plan, organizations known as regional alliances would buy coverage for large numbers of people in a particular area.
Just as the government has not found a reliable way of adjusting its payments to HMOs to reflect the expected health needs of Medicare beneficiaries, so the alliances may have difficulty adjusting their payments to networks of insurers, doctors and hospitals to reflect the needs of alliance members.
Proponents of an alternative to the president's plan argue that instead of seeking elusive savings through managed care, all Americans should be lumped into a government-run "single payer" plan like the one in Canada.
But administration officials argue that the nation is not yet ready for a health care system financed entirely by taxes -- or the tax increases such a system would require -- and that the market forces that would be unleashed by the president's plan would succeed in holding down costs.
Health maintenance organizations normally charge a flat sum, fixed in advance, for each member, regardless of how much care the person uses.
By contrast, under traditional fee-for-service arrangements, which most Medicare beneficiaries use, doctors are paid separately for each service or procedure. Many health policy experts say these arrangements give doctors a financial incentive to perform extra services.
The four-year Mathematica study said that the Medicare payment rates to the HMOs did not fully reflect the makeup of the people who chose to enroll. Had these people obtained their care through fee-for-service, the government would have paid less than it paid the HMOs.
After more than a decade of experience with HMOs in the Medicare program, the study said, the government has not found a completely satisfactory way of adjusting payments to reflect the health status and medical needs of the people who join a particular plan.
Under Mr. Clinton's proposal, the government would devise a formula to make such adjustments, but actuaries and other experts say this will not be easy, and many HMOs complain that Medicare payments already are too low.
The study said that HMOs appeared to reduce the intensity and frequency of medical services used, especially hospital stays, without harming the quality of care.
But it found that Medicare paid an average of $413 a month for each person in an HMO, or $22 more than it would have paid for the same person in the regular Medicare program, even taking into account the additional services that patient would presumably have used in a fee-for-service arrangement.
Disclosure of the Mathematica study comes two months after the General Accounting Office, an investigative arm of Congress, said there was "little empirical evidence" of savings from HMOs and other forms of managed care used by employers for their workers.