Surgeon Volume and Operative Mortality
http://www.100md.com
《新英格兰医药杂志》
To the Editor: Birkmeyer et al. (Nov. 27 issue)1 demonstrated that an individual surgeon's volume was a better predictor of lower mortality for certain high-risk operations than hospital volume. Kizer, in an accompanying editorial,2 interpreted these data to conclude that patients and health plans should preferentially choose high-volume surgeons for certain procedures or should at least avoid low-volume surgeons.
As a surgeon and surgical educator, I do not doubt the validity of the data reported by Birkmeyer et al., nor do I disagree with Kizer's recommendations. However, we must remember that every high-volume surgeon began practice as a low-volume surgeon, who subjected some number of patients to the increased risk associated with undergoing an operation performed by a low-volume surgeon. So, as Kizer states, we are faced with an ethical conundrum. An educated, informed patient with the means to choose will prefer a high-volume surgeon. But on whom will the low-volume surgeons operate in order to become high-volume surgeons? Patients of modest knowledge and means? And if there is no opportunity to practice on less choosy patients, where will future high-volume surgeons come from?
Andrew P. Gutow, M.D.
University of Michigan Medical School
Ann Arbor, MI 48109
agutow@umich.edu
References
Birkmeyer JD, Stukel TA, Siewers AE, Goodney PP, Wennberg DE, Lucas FL. Surgeon volume and operative mortality in the United States. N Engl J Med 2003;349:2117-2127.
Kizer KW. The volume-outcome conundrum. N Engl J Med 2003;349:2159-2161.
To the Editor: Birkmeyer et al. infer from Medicare statistics that a higher surgical-case volume accounts for improved outcomes of various surgical procedures. Is it possible that the cart has been placed before the horse?
From my view as a referring physician, a somewhat different perspective may be obtained. There is a well-developed grassroots information network at every institution, which provides dynamic information on surgical skills, experience, and unexpected outcomes. This network drives referrals to surgeons with good outcomes and is the basis of the relationship between outcome and volume. Birkmeyer et al. have not established that volume begets quality, rather than the other way around.
Gifted, high-volume surgeons must at some point evolve from recently trained, meticulous, caring, low-volume surgeons, and older surgeons who reduce their volumes of procedures later in their careers do not grow additional thumbs. Although volume is a convenient metric, I fear that it could be misused to disenfranchise high-quality surgeons.
Jay A. Erlebacher, M.D.
Englewood Hospital and Medical Center
Englewood, NJ 07631
drerle@mac.com
To the Editor: Birkmeyer et al. verify that, on average, higher-volume hospitals have lower procedural mortality. They conclude, "Patients can often improve their chances of survival substantially, even at high-volume hospitals, by selecting surgeons who perform the operations frequently." The authors base this conclusion on data that they consider to be valid.
If the data on outcomes are available, why use volume as a surrogate for quality? Why not publicize and use the outcome data themselves? This would avert the referral of patients to high-volume surgeons who perform less well, and it would not punish lower-volume surgeons who perform well. Volume-based referral may be appealing to large purchasers of health care, but is it really in the best interest of patients and health care providers?
Steve G. Hubbard, M.D.
East Alabama Medical Center
Opelika, AL 36803
steve_hubbard@eamc.org
Dr. Birkmeyer replies: Given the current interest in "evidence-based referral," Dr. Gutow wonders how low-volume surgeons can best become high-volume surgeons. Rather than have established surgeons "practice" on the less informed, I believe surgeons should gain experience under supervision during their residencies or fellowships at high-volume teaching hospitals. Once training is complete, surgeons interested in performing high-risk operations should develop practices specializing in those areas. As Dr. Gutow suggests, even surgeons who follow such guidelines might be labeled low-volume as they build their practices. However, public reporting systems could potentially avert this problem by flagging or omitting recent surgical graduates.
Although our analysis clearly demonstrates an association between surgeon volume and operative mortality for many procedures, it does not address the issue of causality. Higher volume may lead directly to better outcomes — the so-called practice-makes-perfect hypothesis. The possibility that surgeons who perform a specific complex operation frequently are more proficient would not surprise anyone who plays golf or a musical instrument. However, Dr. Erlebacher suggests the alternative hypothesis — that better outcomes may lead to higher volumes. Selective referral to surgeons with good outcomes requires that referring physicians can reliably judge surgeons' performance with selected procedures, an untested and uncertain proposition. A better understanding of the mechanisms underlying observed associations between volume and outcome may ultimately be useful for improving quality in all settings. In the meantime, surgical patients can often improve their odds of survival substantially by selecting high-volume surgeons, even if the reasons for their better performance are unknown.
Like Dr. Hubbard, many believe that evidence-based referral should be based on direct outcome data, not on proxy measures such as procedure volume. Outcome data are clearly superior for some procedures. With coronary-artery bypass surgery (CABG), for example, risk-adjusted mortality rates are tracked widely and, in many states, are made publicly available. For most other procedures, however, outcome data are not available. Even if they were, constraints on sample size would often limit their usefulness. With the exception of CABG, few high-risk procedures are performed frequently enough for it to be possible to assess provider-specific mortality with adequate precision. Consider a hospital performing six pancreatic resections in a period of three years (the U.S. median). A 0 percent mortality rate (no deaths) would not reliably indicate good performance. A 17 percent mortality rate (one death) or even a 33 percent mortality rate (two deaths) could reflect bad luck as much as poor performance. For some procedures, hospital or surgeon volume is better than no information about surgical quality. It may also be better than direct outcome data.
John D. Birkmeyer, M.D.
University of Michigan Medical School
Ann Arbor, MI 48104
The editorialist replies: I appreciate the comments of Drs. Gutow, Erlebacher, and Hubbard. Their comments underscore points made in my editorial and reinforce the points that further research is needed to determine the mechanism of the relation between volume and outcome, that actual performance data are preferable to volume as a surrogate measure, and that this relation creates a true policy conundrum. I hope the relevant surgical specialty boards and societies are wrestling with the important questions posed by Dr. Gutow on how surgical trainees can best acquire the necessary experience to be high-volume surgeons.
Kenneth W. Kizer, M.D., M.P.H.
National Quality Forum
Washington, DC 20005
As a surgeon and surgical educator, I do not doubt the validity of the data reported by Birkmeyer et al., nor do I disagree with Kizer's recommendations. However, we must remember that every high-volume surgeon began practice as a low-volume surgeon, who subjected some number of patients to the increased risk associated with undergoing an operation performed by a low-volume surgeon. So, as Kizer states, we are faced with an ethical conundrum. An educated, informed patient with the means to choose will prefer a high-volume surgeon. But on whom will the low-volume surgeons operate in order to become high-volume surgeons? Patients of modest knowledge and means? And if there is no opportunity to practice on less choosy patients, where will future high-volume surgeons come from?
Andrew P. Gutow, M.D.
University of Michigan Medical School
Ann Arbor, MI 48109
agutow@umich.edu
References
Birkmeyer JD, Stukel TA, Siewers AE, Goodney PP, Wennberg DE, Lucas FL. Surgeon volume and operative mortality in the United States. N Engl J Med 2003;349:2117-2127.
Kizer KW. The volume-outcome conundrum. N Engl J Med 2003;349:2159-2161.
To the Editor: Birkmeyer et al. infer from Medicare statistics that a higher surgical-case volume accounts for improved outcomes of various surgical procedures. Is it possible that the cart has been placed before the horse?
From my view as a referring physician, a somewhat different perspective may be obtained. There is a well-developed grassroots information network at every institution, which provides dynamic information on surgical skills, experience, and unexpected outcomes. This network drives referrals to surgeons with good outcomes and is the basis of the relationship between outcome and volume. Birkmeyer et al. have not established that volume begets quality, rather than the other way around.
Gifted, high-volume surgeons must at some point evolve from recently trained, meticulous, caring, low-volume surgeons, and older surgeons who reduce their volumes of procedures later in their careers do not grow additional thumbs. Although volume is a convenient metric, I fear that it could be misused to disenfranchise high-quality surgeons.
Jay A. Erlebacher, M.D.
Englewood Hospital and Medical Center
Englewood, NJ 07631
drerle@mac.com
To the Editor: Birkmeyer et al. verify that, on average, higher-volume hospitals have lower procedural mortality. They conclude, "Patients can often improve their chances of survival substantially, even at high-volume hospitals, by selecting surgeons who perform the operations frequently." The authors base this conclusion on data that they consider to be valid.
If the data on outcomes are available, why use volume as a surrogate for quality? Why not publicize and use the outcome data themselves? This would avert the referral of patients to high-volume surgeons who perform less well, and it would not punish lower-volume surgeons who perform well. Volume-based referral may be appealing to large purchasers of health care, but is it really in the best interest of patients and health care providers?
Steve G. Hubbard, M.D.
East Alabama Medical Center
Opelika, AL 36803
steve_hubbard@eamc.org
Dr. Birkmeyer replies: Given the current interest in "evidence-based referral," Dr. Gutow wonders how low-volume surgeons can best become high-volume surgeons. Rather than have established surgeons "practice" on the less informed, I believe surgeons should gain experience under supervision during their residencies or fellowships at high-volume teaching hospitals. Once training is complete, surgeons interested in performing high-risk operations should develop practices specializing in those areas. As Dr. Gutow suggests, even surgeons who follow such guidelines might be labeled low-volume as they build their practices. However, public reporting systems could potentially avert this problem by flagging or omitting recent surgical graduates.
Although our analysis clearly demonstrates an association between surgeon volume and operative mortality for many procedures, it does not address the issue of causality. Higher volume may lead directly to better outcomes — the so-called practice-makes-perfect hypothesis. The possibility that surgeons who perform a specific complex operation frequently are more proficient would not surprise anyone who plays golf or a musical instrument. However, Dr. Erlebacher suggests the alternative hypothesis — that better outcomes may lead to higher volumes. Selective referral to surgeons with good outcomes requires that referring physicians can reliably judge surgeons' performance with selected procedures, an untested and uncertain proposition. A better understanding of the mechanisms underlying observed associations between volume and outcome may ultimately be useful for improving quality in all settings. In the meantime, surgical patients can often improve their odds of survival substantially by selecting high-volume surgeons, even if the reasons for their better performance are unknown.
Like Dr. Hubbard, many believe that evidence-based referral should be based on direct outcome data, not on proxy measures such as procedure volume. Outcome data are clearly superior for some procedures. With coronary-artery bypass surgery (CABG), for example, risk-adjusted mortality rates are tracked widely and, in many states, are made publicly available. For most other procedures, however, outcome data are not available. Even if they were, constraints on sample size would often limit their usefulness. With the exception of CABG, few high-risk procedures are performed frequently enough for it to be possible to assess provider-specific mortality with adequate precision. Consider a hospital performing six pancreatic resections in a period of three years (the U.S. median). A 0 percent mortality rate (no deaths) would not reliably indicate good performance. A 17 percent mortality rate (one death) or even a 33 percent mortality rate (two deaths) could reflect bad luck as much as poor performance. For some procedures, hospital or surgeon volume is better than no information about surgical quality. It may also be better than direct outcome data.
John D. Birkmeyer, M.D.
University of Michigan Medical School
Ann Arbor, MI 48104
The editorialist replies: I appreciate the comments of Drs. Gutow, Erlebacher, and Hubbard. Their comments underscore points made in my editorial and reinforce the points that further research is needed to determine the mechanism of the relation between volume and outcome, that actual performance data are preferable to volume as a surrogate measure, and that this relation creates a true policy conundrum. I hope the relevant surgical specialty boards and societies are wrestling with the important questions posed by Dr. Gutow on how surgical trainees can best acquire the necessary experience to be high-volume surgeons.
Kenneth W. Kizer, M.D., M.P.H.
National Quality Forum
Washington, DC 20005