About 10 percent of Medicare patients who enter hospitals each year under the new fixed-payment-per-case system don't need hospital care, the inspector general of the Department of Health and Human Services reported yesterday.

However, Richard P. Kusserow said another study by his office contains better news: Only one Medicare hospital patient in 100 is being discharged "quicker and sicker" than medically appropriate.

A third study showed that 20.8 percent of Medicare hospital patients are placed in the wrong billing category, and in three-fifths of those cases the error favored the hospital. The billing errors, projected to all Medicare patients being treated under the new system, indicate a net overpayment of $308 million in the year starting Oct. 1, 1984, Kusserow said.

The studies were undertaken to determine how hospitals handle Medicare patients under the new payment system that is applicable to more than four-fifths of all Medicare patients. The studies are based on 239 representative hospitals admission and discharge practices in the year starting Oct. 1, 1984.

Kusserow's office estimated that about 8 million Medicare patients were admitted to hospitals under the new system that year, so a 10 percent rate of unnecessary admissions would mean 800,000 Medicare patients go to the hospital needlessly who could be treated elsewhere at less cost.

He declined to give details, saying the full report would be available in a few weeks. He said, however, that the 10 percent figure is lower than some earlier estimates.

Medicare officials said Kusserow's finding that very few patients are being discharged too soon agrees with their studies. But they said their studies showed that billing changes favoring hospitals have fallen off sharply since the year covered by Kusserow's study. They had no comment on the 10-percent excess-admissions finding.

The studies shed new light on the operations of Medicare's new prospective payment system, which changed the financial rules in 1983 for hospitals with Medicare patients, in an effort to hold down costs.

Previously, hospitals were paid actual costs, with extra payments for each day in the hospital and for each added procedure. A hospital collected more by keeping patients longer, and the suspicion was that many hospitals were keeping patients longer than medically necessary.

Under the new rules, the hospital receives a fixed payment per case for each Medicare patient in an illness category, regardless of the time the patient stays in the hospital or the amount of resources expended during his or her stay. The faster the hospital gets the patient out, the less its costs and the more it can make on the fixed payment.

This led to fears that some hospitals, in excessive cost-cutting, would push out some patients before it was medically appropriate. But the Kusserow study concluded that only 0.8 percent of Medicare hospital patients were discharged inappropriately.

Jack Owen, executive vice president of the American Hospital Association, said, "The one-in-10 estimate for inappropriate admissions probably indicates an improvement in hospital performance in recent years as compared to earlier, when general professional standards about who needs to be admitted were more permissive."

Owen said the billing system was entirely new in 1984 and 1985, and 30,000 medical record personnel were just being trained in how to classify illness for billing purposes. "Twenty percent is actually less than I would have anticipated," he said, noting that the Medicare administrator has said that in the past few years, billing errors of this type have declined sharply.

On the "sicker and quicker" findings, Owen said, "They prove what the hospitals have been saying . . . that despite fears and allegations, we are not pushing sick people out of the hospitals just to make money."