SAN FRANCISCO The generalization that the more procedures a hospital does, the better it is, may be an oversimplification at best and misleading at worst, according to two studies presented at the annual clinical congress of the American College of Surgeons.
In one study, Dr. Melissa A. Meyers and associates compared colectomy mortality in rural and urban hospitals using Medicare data on 279,385 patients who had surgery between 1994 and 1999.
Overall mortality was the same in the two groups. In small rural hospitals with a low volume of procedures, mortality was 6.7%. In urban hospitals, most of which had higher volume, the rate was 6.4%, said Dr. Meyers of the surgery department at Dartmouth-Hitchcock Medical Center, Lebanon, N.H.
Analysis of the data did show some evidence that the more colectomies a hospital performed the lower the mortality, but that observation held only for the urban hospitals. Rural hospitals had no such correlation between volume and mortality, though 90% of the rural hospitals had a low volume. Moreover, the mortality at the rural hospitals was not much different from that at the best urban hospitals, where the rate was 5.6%.
"Hospital procedure volume is a poor proxy for quality in a rural setting, and we need to develop better ways to gauge quality in hospitals overall," Dr. Meyers said.
In the second study, Dr. Dharam Kumbhani and colleagues looked at 30-day mortality for 10 different surgical procedures in the Veterans Affairs (VA) system. The study was a repeat of an earlier, highly controversial investigation that the investigators decided to revisit, with more recent data. Both studies used data from the VA National Surgical Quality Improvement Program.
The earlier study found no relationship in the VA system between surgical volume and outcome for eight different surgical procedures. The present study, which looked at procedures ranging from carotid endarterectomy and total hip arthroplasty to pancreaticoduodenectomy, again found no relationship between low volume and worse outcome, said Dr. Kumbhani of the VA Boston Healthcare System.
In 8 of the 10 surgical procedures, there was a statistically significant relationship between low volume and the observed-to-expected ratio of 30-day mortality. However, this difference was not clinically significant, Dr. Kumbhani said.
Moreover, when the data were analyzed using a hierarchical model that accounted for patient and hospital factors, no relationship was found between volume and 30-day mortality.
"We believe that systems of care are much more important than volume in determining the quality of surgical care," Dr. Kumbhani said. "A lot of high-volume centers have better risk-adjusted outcomes, not because they have higher volumes but because they have better systems in place."
The findings of his study are particularly robust and are probably more accurate than other studies of volume and surgical outcome, because the VA program collects all of its data prospectively and was designed for just this type of analysis, Dr. Kumbhani added.
Most of those who attended the presentations were gratified by the results. During the animated discussion period, it was suggested that the studies should serve as a cautionary note to efforts to measure quality solely in terms of volume, because the volume-quality equation perhaps only holds for very sophisticated procedures such as transplants.
However, Dr. Justin Dimick of the Veterans Affairs Medical Center in White River Junction, Vt., who was a designated discussant for the VA study, took issue with generalizing its results. The study's findings are at odds with an extensive body of research showing that the more a surgeon or a hospital does a particular procedure the better they are at it, he said.