Profession
Volume-quality correlation not clear cut, study says
■ For diseases such as pneumonia, higher patient volume could mean worse physician performance.
By Kevin B. O’Reilly — Posted May 1, 2006
- WITH THIS STORY:
- » Discounting volume
Practice makes perfect? Not always. That's the finding of a recent study that sought to find out whether physicians who cared for a higher volume of pneumonia patients in a given year were more likely to follow quality guidelines and achieve superior outcomes.
Hundreds of studies have used patient volume as a proxy in quality measurement for high-risk surgical procedures, such as coronary bypass artery graft or care for complex diseases such as HIV/AIDS. Nearly 70% of studies examining physician performance in those areas found a statistically significant association between higher patient volume and superior outcomes, according to a 2002 medical literature review published in the Annals of Internal Medicine.
But the much-touted volume-outcome correlation appears to fall apart for care such as pneumonia, according to a study published in the Feb. 21 Annals of Internal Medicine.
The evaluation of 9,741 doctors who cared for 13,480 patients admitted to hospitals for pneumonia found little variation between physicians who cared for as few as two patients and those who cared for as many as 29 over two six-month periods. In fact, physicians with the highest volume of pneumonia patients delivered worse care on some measures. The results were similar at the hospital level.
"We were surprised that we didn't find an association in the direction we expected, and in fact, at least for several quality measures, we found just the opposite sort of association," said Peter K. Lindenauer, MD, lead author of the study and medical director of clinical information systems at Baystate Health Systems in Springfield, Mass.
He said referring patients to high-volume physicians and hospitals might make sense for procedures such as a hip replacement but that most patients still go to their local community hospital when they come down with pneumonia symptoms.
"Pneumonia care could be so straightforward that anybody can take care of pneumonia patients, and hence why would you expect to see a volume-outcome relationship?" Dr. Lindenauer said. "The hospital that takes care of one pneumonia patient could be as good as a hospital that takes care of 100 or 1,000. Maybe if you've seen one pneumonia patient, you've seen them all."
Few studies on nonsurgical situations
The study, "Volume, Quality of Care, and Outcome in Pneumonia," shows that the higher-volume-equals-better-outcomes premise is "not a slam dunk," said Robert Wachter, MD, chief of medicine at the University of California, San Francisco Medical Center.
"It flies a little bit in the face of prevailing wisdom that practice always makes perfect."
Dr. Wachter said caring for pneumonia could be like driving. "When you compare the incredibly proficient driver versus the competent driver, both are unlikely to get into a car accident," he said. "The baseline level of competency is good enough for certain things."
Complicated diseases are more analogous to risky activities like skiing, he said, where the difference between competence and proficiency can make a huge difference in outcome.
Organizations such as the Leapfrog Group, a coalition of health insurance purchasers, have sought to use data that show there is a quality-volume correlation to support what it calls evidence-based referrals to larger hospitals within a given region.
John Birkmeyer, MD, chair of the Leapfrog Group's expert panel on evidence-based hospital referrals, said that while scores of studies have found a positive volume-outcome relationship, when it comes to nonsurgical care "we don't know very much."
Dr. Birkmeyer, director of bariatric surgery at the University of Michigan in Ann Arbor, said volume has a two-pronged association with outcomes. The first is "the practice-makes-perfect model of volume, which really is no different from why the amount of practice would matter in golf or tennis." The second is the elements of care -- systemic and technological -- for which volume acts as a proxy.
While larger hospitals and physicians who specialize in a certain type of disease care are presumed to have the advantage for surgical and nonsurgical situations, things don't always play out that way.
"Big hospitals tend to be a little bit clumsy and slow to implement efficient processes with everyday stuff," Dr. Birkmeyer said. "We could postulate that with pneumonia, big hospitals have been slower to adopt evidence-based protocols."
This is the key, according to Bruce Bagley, MD, director of clinical quality improvement for the American Academy of Family Physicians.
"The authors assume that volume is the proxy for quality," Dr. Bagley said. "They didn't measure the quality. Whether you have one patient or 100 patients, if you have a system that makes sure everybody does the right thing every time, it's volume insensitive."
Some experts defended using volume to measure quality.
"I don't think this tells us that volume is not a good indicator of quality," said Arthur A. Levin, MPH, director of the Center for Medical Consumers in New York City. "It just tells us that it is limited in its predictive power. Volume may have a stronger relationship to some things, and it may have no relationship or a negative relationship to other things."
Levin, who has posted the number of different types of surgeries performed by New York physicians on his group's Web site, said that sophisticated quality measurement has been too slow in coming.
"That's why I've always pushed volume," he said. "It doesn't require risk adjustment, and it's there."