High volume hospitals have lower death rates for many surgeries and HIV/AIDS, says UCSF study

Hospitals that handle a large volume of patients for several common surgeries
or for HIV/AIDS, have significantly lower death rates for those conditions than
lower volume hospitals, according to a recent University of California, San
Francisco study.  The researchers say these findings may be used by Medicare
and some employers to justify preferentially referring patients to high volume
hospitals, a trend that merits further consideration.

Many previous studies have shown lower mortality rates when patients with
certain conditions are admitted to high volume instead of low volume hospitals.
But the new study, led by R. Adams Dudley, MD, MBA, assistant professor of
medicine in UCSF’s Institute for Health Policy Studies, is the first to combine
these studies across all these conditions.  It is also the first to estimate
the annual number of hospital deaths in California that were potentially caused
by admission to a low volume hospital.

The study appears in the March 1 issue of the Journal of the American Medical
Association.

The researchers searched for the best study on several medical conditions, by
judging how well the study adjusted for confounding factors, such as the fact
that patients admitted to one hospital might be more sick than those admitted
to another.  They also favored studies that were conducted most recently, and
those that included the greatest number of hospitals.  They identified eleven
conditions and procedures for which high volume hospitals clearly had lower
death rates: ten procedures (such as coronary bypass surgery, coronary balloon
angioplasty, and removal of pancreatic cancer) and treatment for HIV/AIDS.

To estimate the number of deaths in California attributable to the use of low
volume hospitals, they applied the differences in death rates between high and
low volume hospitals from the best studies to California hospitalization and
death statistics.  Dudley estimated that in 1997, 602 deaths in low volume
California hospitals (26% of all deaths among patients in these studies) could
have been avoided if low volume hospitals had the same death rates as high
volume hospitals.

This is only an estimate, and not every low volume hospital has poor results,
Dudley said, but the tendency for high volume hospitals to have lower death
rates may convince some patients to reconsider where they go for complex
treatments.  “There are a lot of reasons to want to use your local hospital,”
he said, “but if you have one of these complicated problems and there is a
hospital with more patients like you across town or in the next community, you
may increase your chance of surviving by traveling a few miles.”
The researchers also calculated how far low volume hospital patients would have
had to travel to get to a high volume hospital. Surprisingly, 58% of low volume
hospital patients would have needed to travel 10 or fewer additional miles, and
25% actually traveled to a low volume hospital that was farther away than the
nearest high volume hospital. 

This finding reflects a little recognized fact: in California and other places
where most people live in cities, most low volume hospitals are in urban
areas.  For example, there are 42 hospitals in Los Angeles County that perform
cardiac bypass surgery, and 16 of those did fewer than 200 bypass operations-
the minimum number recommended by the American College of Cardiology to
maintain competency-in 1997.

One possible explanation for the reduced death rate at high volume hospitals is
that practice makes perfect. “High volume hospitals may be better at these
procedures because they do more,” Dudley said.  “However, the converse could
also be true - a hospital with better results may get more referrals, and so
become a high volume hospital.”

Whatever the reason for the difference between hospitals, studies such as
Dudley’s have convinced Medicare, the federal health program for senior
citizens, to use volume as a criterion for approving hospitals to perform organ
transplants or for designating hospitals as centers of excellence.  Major
employers in California are also considering the use of hospital volume as a
referral criteria.

This raises important social questions, Dudley said.  What if many patients are
referred to high volume hospitals and away from low volume hospitals?  Then
hospitals that no longer perform coronary bypass, for example, might not have
surgeons ready to do other kinds of heart procedures.  “These issues need
further public discussion and will require thoughtful approaches.  Medicare and
health plans may want to make sure their beneficiaries get the best possible
care, but they do not want some communities to lose hospital services because
patients are being referred away.  This may require, for example, focusing
referral efforts on cities in which several hospitals are providing the same
service.”

Co-authors are Kirsten Johansen, MD, UCSF assistant professor of medicine,
Richard Brand, PhD, UCSF professor of biostatistics, and Deborah Rennie, a
senior programmer at the Institute for Health Policy studies, and Arnold
Milstein, MD, MPH, from the Pacific Business Group on Health and William M.
Mercer, Inc.