The Department of Health and Human Services took a major step yesterday toward providing the public with usable government ratings of the quality of hospitals.

HHS announced that in December it will disclose actual death rates for Medicare patients at each of the nation's 6,000 hospitals participating in the elderly care program, together with an estimate of the death rate that could be expected at each hospital based on the age, sex and certain other characteristics of the patients.

The information will allow a doctor or patient to compare the death rates at a specific hospital with those expected at that hospital, given its mix of patients.

The plan to publish overall death rates for Medicare patients plus death rates for 16 major categories of illness, together with the methods of deriving the statistics, was revealed in a proposed rule to be published in the Federal Register. The rule, signed by Dr. William L. Roper, administrator of Medicare, will be subject to public comment and may be revised before becoming final.

A year ago Medicare published a preliminary list of hospital death statistics that hospitals denounced as grossly misleading to the public. The list was drawn from raw statistics on deaths per Medicare patient, and hospitals argued that most people would assume that any hospital with a high death rate gave worse care than one with a low death rate.

The hospitals said the data made no adjustment for differences in age, sex, severity of illness on admission, economic status of the patient (poorer patients may be sicker because they could not afford prior medical care) and other factors.

The new proposal, by including an expected percentage of deaths for each hospital based on the age, sex, incidence of complicating illnesses and prior hospitalizations of the patients at that hospital, takes care of some of these objections, said Mike Bromberg, director of the national office of the Federation of American Health Systems, which represents about 1,200 for-profit hospitals.

But he said the published figures will "still be misleading because they don't adjust for severity of illness on admission or for economic status," both of which can make a big difference in whether the patient recovers.

The American Hospital Association said the new method would be better than last year's, but still not adequate to prevent misconceptions that mortality rates alone reflect quality of care.

Each hospital will be given the data a month before it is published and allowed to submit a response that will be published with the basic data when it is released to the public.

The figures will include the actual death rates and the expected death rates not only for all Medicare patients, but for separate illnesses that account for 80 percent of deaths, such as heart, lung and kidney disease, stroke, cancer, urologic disease and some other categories of illness.