The Maryland Hospitals Association, responding in detail to a consumer group's report that found wide differences in post-surgical mortality rates among the state's hospitals, has accused the authors of irresponsibility, faulty logic and causing "needless public alarm."
The report, based on computer data kept by the hospitals, was released last month by the Washington-based Public Citizen Health Research Group. The authors, Dr. Eve Bargmann and Cynthia Grove, found high rates of mortality among the patients of some doctors, who were identified only by code names, as well as wide variations among hospitals in the death rates after several common operations.
Officials of most of the hospitals identified in the report, who had not read it before its release, declined immediate comment. But in a detailed response issued this week the hospital association accused the consumer group of the "cruel use of a set of computer printouts to reach some questionable conclusions."
Noting the report's call for an inquiry into a surgeon who lost 25 percent of his prostate surgery patients at a Baltimore hospital, George M. Wright, director of research and data management for the association, wrote, "The clear implication is that this physician can be expected to kill a quarter of his or her patients. One would expect that these bold and potentially libelous statements are well-supported by the data . . . . In fact there is no such support."
Wright said it was not possible to draw such conclusions from the data used, because the report did not take into account that a surgeon could be operating at several different hospitals where he or she would be identified by several different code names.
"A physician . . . may choose to treat his or her highest risk patients at a large teaching hospital on the grounds that state-of-the-art support, experience and knowledge are available there. This same physician may treat low-risk patients at the most mutually convenient facility. It would not be surprising to find that such a physician experienced a high mortality rate in the former case," Wright said.
Dr. Sidney Wolfe, director of the Health Research Group, said that hospitals will have to identify physicians by name and look at their overall records to find out if Wright's criticism is valid.
"So far, Maryland hospitals have not done so, or, if they have done so, they have not reported what they found," he said.
Wright also sharply criticized the report's finding that surgeons who did fewer than 10 operations of a specific type each year tended to have higher-than-average mortality rates.
"It may be true that certain physicians have inexplicably high mortality rates," Wright said. "It may be true that physicians doing less than 10 procedures a year are less competent than others. It is certainly true that the data available to the researchers is not sufficient to test these two hypotheses."
Wright's critique dwelt heavily on the complex statistical formula used by the researchers, suggesting that they had overlooked "chance" by only examining what happened in the years 1979-1980. By using different mathematical methods, he said, "It was found that mortality rates appear to differ among hospitals only because of chance year-to-year variation. The study's conclusion to the contrary is based on a fundamental error: the data was taken to be a single sample of 50 observations, instead of 50 samples of a single observation."
The association said that the consumer report would encourage hospitals in Maryland and elsewhere to stop collecting and reporting the kind of data used to compile the report.
Wright also claimed that the report had damaged the credibility of Maryland hospitals, and may have damaged the professional reputations of some surgeons, who he said were recognizable to their colleagues despite the use of code names. "An alleged 'high-mortality' physician may be harmed professionally if his or her colleagues put any trust in the study's conclusions," Wright said.
In a separate criticism of the consumer report, Dr. Robert M. Heyssel, director of Johns Hopkins University Hospital, which had been singled out as having the highest mortality rate of all Maryland hospitals for gall bladder surgery, accused the HRG of "tagging a world-renowned institution with a 'high mortality' label."
Heyssel said that a computer error resulted in the hospital's mortality rate being reported as 3.3 percent, instead of 3 percent, and he contended that some elevation in the hospital's mortality rate could be a result of the hospital's patients being much older and sicker than the "average patient."
The authors of the report responded that a 3 percent mortality rate "is still twice the Maryland average," and they said they found that the university hospital "actually had fewer of these high-risk patients than the Maryland average."