Volume versus outcomes. How should the orthopaedic
surgeon and the medical community at large react? This issue of The
Journal contains an important article by Katz et al. dealing
with the association between hospital and surgeon volume and outcomes
of total hip surgery in the United States Medicare population1. It appears logical to assume that
the more experience that a surgeon has with a particular procedure,
the better the result and that a so-called learning curve may be
associated with many surgical procedures. One can hardly argue with
the goal of improving our care of patients by providing the best
possible outcome. This is particularly true of high-volume procedures
such as total hip replacements, which are performed at the rate
of more than 200,000 per year2. The majority of both primary and
revision total hip arthroplasties in patients in the Medicare population
are performed by surgeons who carry out fewer than ten of these
procedures per year.
Higher surgeon and/or hospital volume has been associated with
better outcomes in several areas, including cardiac surgery3,4;
cancer surgery5; cataract surgery6; and medical management of various
conditions, including myocardial infarction7. There have been similar
reports involving orthopaedic procedures, including total knee8-11
and total hip12 arthroplasty.
Kreder et al. noted significant differences in the case mix of low-volume
providers compared with that of high-volume providers12. Surgeons
and hospitals with a volume below the 40th percentile managed patients
who had a more adverse risk profile in terms of age, comorbidity,
and diagnosis. However, even after adjustment for this case mix
there was a significant relationship between surgeons who averaged
fewer than two hip replacements annually (which the authors defined
as low-volume surgeons) and a worse outcome. Kreder et al. found that
the patients managed by these low-volume surgeons tended to have
higher mortality rates, more infections, higher rates of revision
procedures, and more serious complications. Furthermore, they found
that the duration of hospitalization was inversely related to surgeon
volume and directly associated with hospital volume.
When one looks at volume versus outcomes, there are several questions
that should be considered: What do the data mean? What are their
limitations? How should the data be used? What are the unanswered
questions?
Katz et al. analyzed Medicare claims data for more than 58,000
elective primary total hip replacements and almost 13,000 elective
revision total hip replacements performed between July 1995 and
June 19961. They assessed the relationship between surgeon and hospital
volume and mortality, dislocation, deep infection, and pulmonary
embolism occurring in the first three postoperative months. They
found that patients treated with primary hip replacement in hospitals
in which more than 100 of those procedures were performed per year
had a lower risk of death than those operated on in hospitals in
which ten or fewer were performed per year. The rate of dislocation
was significantly lower for patients treated with primary hip replacement
by surgeons who performed more than fifty of the procedures per
year than it was for those operated on by surgeons who performed
five or fewer per year. Patients treated with a revision hip replacement
by surgeons who performed more than ten of those operations per year
had lower mortality rates than those treated by surgeons who performed
three or fewer such operations.
We believe that it is important to carefully review the study limitations
described by Katz et al., including the fact that analyses of Medicare
claims are limited by the lack of key clinical information such
as operative details, psychological and physical functional status,
and pain. Furthermore, the complexity of the operative procedure,
which is particularly important in the case of revisions, is not
captured in the claims data. In addition, both Katz et al. and Kreder
et al. identified key areas for additional research.
Data such as these may be used as a reason to concentrate procedures
or medical care in regional, high-volume centers in an attempt to
reduce mortality and morbidity13,14; however, as Katz et al. pointed
out, the trade-off between the comfort of having surgery at a community
center and the better outcomes in larger, referral centers should
be examined explicitly. Furthermore, since the Medicare data lack
information on pain relief and physical as well as psychological
function, the effects of procedure volume on pain relief and functional improvement
as well as on the durability of the implant should be examined in
order to provide a more complete picture of the influence of volume
upon outcomes.
A final question to be addressed is: "How good is good enough?" One
can hardly argue with the contention that the results of total hip
replacements are extremely good. Katz et al. reported that the overall
mortality rate remains low, <2% for primary hip
replacement, even in the lowest-volume centers. Furthermore, neither
Kreder et al. nor Katz et al. identified a clear-cut point at which
the complication rate changes dramatically; rather, there appears
to be a continuum (that is, a linear relationship) between volume
and outcome, making it difficult to identify a specific cutoff point.
What is the orthopaedic surgeon to do with this information? The
orthopaedist must take a proactive role and first of all be aware
of the data and understand their limitations. In addition, more
extensive research is necessary to refine and clarify the data before
they can be used to justify changes in the delivery of total joint
services in this country. The findings of this study should serve
as a foundation for thoughtful dialogue about further improvement
of the already remarkable success of total hip arthroplasty, and
we must continue to pursue answers to the unanswered questions.