A chart on death rates at local hospitals in yesterday's editions failed to list the figures for Southern Maryland Hospital Center. In 1986, 15 percent of 2,013 Medicare patients at the hospital died. Government statisticians had predicted that the death rate should be between 9 percent and 16 percent. (Published 12/19/87)

The 1986 death rate at D.C. General Hospital, the city's main health care facility for the poor, was the highest among hospitals in the Washington metropolitan area, according to an unprecedented federal study of the fate of Medicare patients across the country.

D.C. General was the only local hospital with an overall death rate higher than the expected range, according to government statisticians. Seventeen percent of 816 Medicare patients died during the year, while statisticians said the range calculated for the hospital should have been between 7 and 15 percent.

Several other area hospitals had death rates near or above the top of the predicted ranges for patients with specific illnesses, such as stroke, heart disease or pulmonary disease. These included Arlington, Fair Oaks, Capitol Hill, Greater Southeast Community, Holy Cross and Shady Grove Adventist hospitals.

These statistics were disclosed yesterday in the Department of Health and Human Services' first comprehensive survey of death rates for Medicare patients -- about 40 percent of all patients -- in 5,971 acute care hospitals. The figures were released after extensive controversy between the government and the nation's hospitals over the validity of the data and whether it should be made public.

Perhaps the main lesson of the survey is that death rates vary dramatically among hospitals and can give some clue about the quality of care that can be expected, several nongovernmental health care specialists said yesterday.

"Not all hospitals are the same. Saying that is a tremendous step forward," said Warren Greenberg, a health care expert at George Washington University. "This very fact will awaken consumers to the fact that there are indeed differences between the hospitals in the United States."

Government and hospital officials cautioned that the numbers cannot be used as a "report card" to compare one hospital with another and do not mean that any particular hospital is "good" or "bad." Even neighboring hospitals may get different kinds of patients, healthier or sicker or otherwise likely to do better or worse, they explained.

But, officials said, any suspiciously high figures should prompt patients to ask questions of their doctors before going to a hospital with either a high total death rate or high rates in 16 categories that the Medicare program measured.

As a patient, "I would go to my physician and say, 'You're associated with this hospital. Have you seen this? What has been your experience?' " said Thomas G. Morford, head of the HHS bureau that monitors the quality of care.

Officials of the American Medical Association, American Hospital Association and Federation of American Health Systems said they favor meaningful information for patients, but in opposing yesterday's release, said it unfairly seems to rank hospitals on the basis of unreliable statistics.

Local hospital officials reacted similarly, with many saying the information about specific hospitals does not take into account the severity of illness: how sick a patient is when admited to the hospital, and other complicating factors.

John Dandridge Jr., the executive director of D.C. General, said in a letter to Medicare chief William Roper that the hospital "emphatically disclaims the {HHS} report on mortality rate within this hospital, primarily because the statistical method used is not valid." He said the government's predicted range of mortality is unfair, considering the sicker type of patients the hospital treats.

Dr. Lawrence B. Johnson, the hospital's medical director, pointed out in an interview that even though the hospital's overall mortality rate was high, the hospital ranked above the expected range in only one of 16 categories of illness examined by the government: pulmonary disease.

"The numbers do not represent poor quality of care," Johnson said, though he added that the hospital will review its records to "see if there are any opportunities to improve care."

Dr. Reed V. Tuckson, the District's health commissioner, said that a major flaw in the government's numbers was their failure to take into account the differing socioeconomic status of Medicare patients at different hospitals. He noted that several of the city's hospitals take care of a high proportion of black and minority patients, who he said are often sicker than the general population because of a lack of quality health care.

"It would come as no surprise that these hospitals, particularly D.C. General . . . would have a higher mortality rate than the hospitals that are caring for people that are better off," Tuckson said.

In fact, D.C. General's numbers were little different than those at several other public hospitals in big cities with large minority populations and financial difficulties.

The federal government derived its figures from hospital billing records, adjusting them to account for differences in patients' ages, sex, diagnoses, other medical conditions, previous hospital admissions and whether the patient had to be transferred from another hospital. Still, the officials conceded, these are incomplete measures.

"We do not have a perfect measure of quality," said Roper, administrator of the Health Care Financing Administration, the agency that runs Medicare, the federal insurance program for the elderly. "But the perfect should not be the enemy of the good."

One of HHS' principal advisers in preparing the statistics, Dr. Robert Brook of the Rand Corp. and the University of California at Los Angeles, said in an interview yesterday, "On the average, if I knew nothing else about a hospital, I would probably be better off" going to one with low death rates -- rates in the lowest quarter of its expected range -- than one with rates in the highest quarter.

Another method of calculation increases the percentage of hospitals with suspiciously high death rates considerably by counting both the hospitals outside the expected range and those with death rates at the high end of the range. The overall percentage of hospitals with suspiciously high death rates then becomes 11 percent, or more than one hospital in every 10, rather than 2.5 percent. These higher figures were considered by HHS officials but not released.

Brook estimated that "maybe half, maybe 40 percent" of the hospitals with strikingly high death rates are so listed because their care is really inferior, and the rest are there because their patients are truly different. An academic medical center or trauma center or burn center, for some examples, may care for the sickest patients with the lowest chances of survival and, thus come up with poor statistics.

Roper predicted that the groundbreaking effort -- the first time hospitals and their medical staffs have been able to compare data so extensively -- would lead hospitals and doctors to look hard at their own performances. Already, he said, he is hearing "story after story" about hospital staffs seeking out their poorest-performing members and demanding better results.

Dr. Sidney Wolfe, head of the Public Citizen Health Research Group, called the release "a big step forward" and "an acknowledgement that the public has a right to know about the important differences between hospitals."

HHS officials predicted that yesterday's report would be the first of an annual series of increasingly detailed information releases, including not only hospital death rates but also individual doctors' performances.

Many hospitals were caught unprepared a year and a half ago, when the government first released mortality figures for some of the nation's hospitals. This year, hospitals were given the information for their own facilities several months prior to public release, and many prepared detailed responses that were released along with the government figures.

Although many local hospitals were sharply critical of the government data, several that did well under under the survey boasted about their results in their responses.

Several of the hospitals that did poorly in particular categories said they reviewed their medical records and turned up no evidence of poor or inappropriate care. Officials explained the discrepancies between the Medicare numbers and their own records as a keen illustration of the inadequacy of the Medicare data.

For instance, at Holy Cross Hospital in Silver Spring, the Medicare numbers showed that 10 percent of 199 patients in the low-risk heart disease category died, outside of the predicted range of 2 to 9 percent.

But Dr. Marvin Schneider, chairman of the department of medicine, said that further review of the records of 16 of the patients who died showed that all were far more severely ill than the initial diagnosis indicated and should have been in different disease categories. He said that rather than indicating a quality of care problem, the numbers show that the hospital needs to rethink its classification of incoming patients.

"We review this stuff constantly," Schneider said, speaking of Holy Cross' own internal quality assurance process. "If {bad care} was going on, we'd pick it up instantly. We're not just going to sit back and let people die."

Greater Southeast Community Hospital officials said they were surprised by the relatively high mortality rate in the stroke category. But a review of the cases, they said, showed that many of these patients were so sick they were going to die, no matter what is the hospital's quality of care.

The hospital's chairman of medicine, Dr. Raymon Noble, said he's been caring for a 79-year old woman for five years, and he has treated her recently for heart disease and a bleeding ulcer. Last week she was admitted to the hospital after suffering from a stroke, Noble said.

"We're going to take care of her and she's going to die," Noble said. "She's going to be one of those numbers next year . . . . Her death is going to be viewed in a negative light."