SAN DIEGO — So-called zero-mortality hospitals subsequently experience mortality rates that are similar to or higher than those of other hospitals, Dr. Justin B. Dimick reported at a congress sponsored by the Association for Academic Surgery and the Society of University Surgeons.
To determine whether zero-mortality hospitals actually achieved better results than did other hospitals “or were just lucky,” Dr. Dimick and his associates obtained national Medicare data for 1997–1999 on five procedures that are widely included in quality improvement measures: coronary artery bypass grafting; abdominal aortic aneurysm repair; and resections for colon, lung, and pancreatic cancers.
For each procedure, the researchers defined zero-mortality hospitals as those with no inpatient or 30-day deaths over the 3-year period.
The investigators then compared the mortality rates of the zero-mortality hospitals for the subsequent year (2000) with the mortality rates at other hospitals for that year.
No significant difference in mortality was observed between zero-mortality hospitals and the other hospitals for the following four procedures: coronary artery bypass surgery (4.0% zero-mortality hospitals vs. 5.0% other hospitals), abdominal aortic aneurysm repair (6.3% vs. 5.8%, respectively), colon cancer resection (6.0% vs. 6.6%, respectively), and lobectomy for lung cancer (5.1% vs. 5.3%, respectively).
In pancreatic cancer resection, however, the mortality rate was significantly worse for zero-mortality hospitals than it was for other hospitals (11.2% vs. 8.7%, respectively).
The researchers also observed that zero-mortality hospitals had fewer cases of all five operations than the other hospitals had.
“More attention should be paid to sample size in quality measurement,” said Dr. Dimick of the University of Michigan, Ann Arbor. He also called for hospital quality measures that “are more reliable and precise.”
The findings suggest that in deciding where to have surgery, patients “cannot consider a reported mortality of zero as a reliable indicator of future performance,” said Dr. Dimick.
Mortality rates are publicly reported and are commonly used to measure quality, he said. “For instance, cardiac surgery report cards are published in many states, including New York, Pennsylvania, California, and New Jersey. But what's even more troubling is the extension of this approach to many other operations. Right now the Agency for Healthcare Research and Quality is using operative mortality rates as quality measures. These are being published on Web sites, despite data showing [such measures] may not be useful,” he said.
Patients 'cannot consider a reported mortality of zero as a reliable indicator of future performance.' DR. DIMICK