Monitoring surgical mortality
http://www.100md.com
《英国医生杂志》
Scottish scheme has worked well but may not be transferable to other settings
Should surgical mortality be routinely monitored? In this issue Thompson and Stonebridge present a compelling argument for systematic audits (p 1139)1 and Esmail, in the first part of a new series on the General Medical Council and revalidation, argues that doctors will have nothing to fear from the GMC's revised plans (p 1144).2
The Scottish Audit of Surgical Mortality is a voluntary, peer reviewed, critical event analysis that has become an established part of standard surgical practice in Scotland. Scottish surgeons have shown tremendous support for the programme—99% of surgeons participate and 91% of deaths under surgical care in Scotland are audited. They support the scheme perhaps because it seems to be effective. After errors in specific processes of care (failure to use intensive care units and failure to use prophylaxis for deep venous thrombosis) were identified by the scheme as contributing to surgical deaths, system-wide changes occurred and the frequency of such errors greatly declined.
The potential effectiveness of a programme that focuses on death as the only critical event, however, may be limited. Although errors occur often in medicine,3 errors contributing to death occur in only 6% of cases identified by Scottish Audit of Surgical Mortality. Errors that do not occur often or do not generally result in mortality are likely to be missed by such a programme. In addition, the focus of the programme on processes of care would indicate that feedback at the hospital level is at least as essential as feedback at the individual surgeon level.
The grassroots, clinician led model has worked well in Scotland but may not be easily transferable, particularly in settings where results of such a programme could have market influences. In the United States, most audits have taken the form of report cards where mortality (and in some cases morbidity) rates are calculated for a given procedure at the hospital or individual surgeon level, and the rates (generally adjusted for comorbidities) are compared between sites and surgeons. In such a system, attention is focused generally on the outliers who have poor results, with (in most cases) neither integrated analysis of the root cause nor any attempt to determine the processes of care that result in worse outcomes.
In New York state, adjusted mortality rates for cardiac surgery have been publicly disseminated since 1990. Although some evidence exists that this programme has resulted in a lower than expected cardiac surgery mortality rate for the state, what is not clear is if public dissemination of the information was necessary. Few patients who have bypass surgery are aware of the publicly available mortality rates of their surgeon or hospital.4 Even when they do know the rates, other factors may be more important. The hospital chosen by Bill Clinton for coronary bypass surgery, for example, had the highest mortality rate for this procedure in New York state in 2001, the most recent results available to Mr Clinton at the time of his surgery.5 For non-cardiac procedures, most hospitals do not have sufficient case loads to compare reliably mortality at the individual hospital (let alone surgeon) level.6
Publication of mortality audits in this setting serves little purpose—other than, perhaps, to create a false sense of doing something to improve quality. In fact, the underlying assumption of report card programmes may be misguided; clinically significant errors are committed at all institutions and by all surgeons, not just by the outliers with poor results. To build a framework for trust, the development and systematic adoption of effective methods to minimise errors for every patient must be a priority of the entire surgical community.
Nancy N Baxter, assistant professor
Division of Surgical Colon and Rectal Surgery, Department of Surgery, University of Minnesota, MMC 450, 420 Delaware Street SE, Minneapolis, MN 55455, USA (baxte025@umn.edu)
Education and debate pp 1139, 1144
Competing interests: None declared.
References
Thompson AM, Stonebridge PA. Building a framework for trust: critical event analysis of deaths in surgical care. BMJ 2005;330: 1139-43.
Esmail A. GMC and the future of revalidation. Failure to act on good intentions. BMJ 2005;330: 1144-7.
Committee on Quality of Health Care in America, Institute of Medicine. To err is human: building a safer health system. Washington, DC: National Academy Press, 2000.
Schneider EC, Epstein AM. Use of public performance reports. A survey of patients undergoing cardiac surgery. JAMA 1998;279: 1638-42.
Altman LK. Clinton surgery puts attention on death rate. New York Times 6 Sep 2004:section A: 1.
Dimick JB, Welch HG, Birkmeyer JD. Surgical mortality as an indicator of hospital quality. The problem with small sample size. JAMA 2004;292: 847-51.
Should surgical mortality be routinely monitored? In this issue Thompson and Stonebridge present a compelling argument for systematic audits (p 1139)1 and Esmail, in the first part of a new series on the General Medical Council and revalidation, argues that doctors will have nothing to fear from the GMC's revised plans (p 1144).2
The Scottish Audit of Surgical Mortality is a voluntary, peer reviewed, critical event analysis that has become an established part of standard surgical practice in Scotland. Scottish surgeons have shown tremendous support for the programme—99% of surgeons participate and 91% of deaths under surgical care in Scotland are audited. They support the scheme perhaps because it seems to be effective. After errors in specific processes of care (failure to use intensive care units and failure to use prophylaxis for deep venous thrombosis) were identified by the scheme as contributing to surgical deaths, system-wide changes occurred and the frequency of such errors greatly declined.
The potential effectiveness of a programme that focuses on death as the only critical event, however, may be limited. Although errors occur often in medicine,3 errors contributing to death occur in only 6% of cases identified by Scottish Audit of Surgical Mortality. Errors that do not occur often or do not generally result in mortality are likely to be missed by such a programme. In addition, the focus of the programme on processes of care would indicate that feedback at the hospital level is at least as essential as feedback at the individual surgeon level.
The grassroots, clinician led model has worked well in Scotland but may not be easily transferable, particularly in settings where results of such a programme could have market influences. In the United States, most audits have taken the form of report cards where mortality (and in some cases morbidity) rates are calculated for a given procedure at the hospital or individual surgeon level, and the rates (generally adjusted for comorbidities) are compared between sites and surgeons. In such a system, attention is focused generally on the outliers who have poor results, with (in most cases) neither integrated analysis of the root cause nor any attempt to determine the processes of care that result in worse outcomes.
In New York state, adjusted mortality rates for cardiac surgery have been publicly disseminated since 1990. Although some evidence exists that this programme has resulted in a lower than expected cardiac surgery mortality rate for the state, what is not clear is if public dissemination of the information was necessary. Few patients who have bypass surgery are aware of the publicly available mortality rates of their surgeon or hospital.4 Even when they do know the rates, other factors may be more important. The hospital chosen by Bill Clinton for coronary bypass surgery, for example, had the highest mortality rate for this procedure in New York state in 2001, the most recent results available to Mr Clinton at the time of his surgery.5 For non-cardiac procedures, most hospitals do not have sufficient case loads to compare reliably mortality at the individual hospital (let alone surgeon) level.6
Publication of mortality audits in this setting serves little purpose—other than, perhaps, to create a false sense of doing something to improve quality. In fact, the underlying assumption of report card programmes may be misguided; clinically significant errors are committed at all institutions and by all surgeons, not just by the outliers with poor results. To build a framework for trust, the development and systematic adoption of effective methods to minimise errors for every patient must be a priority of the entire surgical community.
Nancy N Baxter, assistant professor
Division of Surgical Colon and Rectal Surgery, Department of Surgery, University of Minnesota, MMC 450, 420 Delaware Street SE, Minneapolis, MN 55455, USA (baxte025@umn.edu)
Education and debate pp 1139, 1144
Competing interests: None declared.
References
Thompson AM, Stonebridge PA. Building a framework for trust: critical event analysis of deaths in surgical care. BMJ 2005;330: 1139-43.
Esmail A. GMC and the future of revalidation. Failure to act on good intentions. BMJ 2005;330: 1144-7.
Committee on Quality of Health Care in America, Institute of Medicine. To err is human: building a safer health system. Washington, DC: National Academy Press, 2000.
Schneider EC, Epstein AM. Use of public performance reports. A survey of patients undergoing cardiac surgery. JAMA 1998;279: 1638-42.
Altman LK. Clinton surgery puts attention on death rate. New York Times 6 Sep 2004:section A: 1.
Dimick JB, Welch HG, Birkmeyer JD. Surgical mortality as an indicator of hospital quality. The problem with small sample size. JAMA 2004;292: 847-51.