Improving mortality of coronary surgery over first four years of independent practice: retrospective examination of prospectively collected
http://www.100md.com
《英国医生杂志》
1 South Manchester University Hospital, Manchester M23 9LT, 2 Cardiothoracic Centre, Liverpool L14 3PE, 3 Manchester Royal Infirmary, Manchester M13 9WL, 4 Blackpool Victoria Hospital, Blackpool FY3 8NR, 5 Leeds General Infirmary, Leeds LS1 3EX
Correspondence to: B Bridgewater ben.bridgewater@smuht.nwest.nhs.uk
Abstract
In British medicine consultants are appointed to hospitals after a defined period of training, once they have satisfied national training bodies and been successful in open competition for an advertised post. NHS consultants are independent practitioners who function under clinical governance systems in the institutions in which they are employed. Increasingly, they are also regulated by national initiatives, such as the planned publication of surgeon specific mortality data for coronary artery surgery in the United Kingdom in 2004.1
In a recent study we looked at the outcomes of coronary artery surgery for individual surgeons in north west England.2 An incidental finding was a strong association between the volume of operations each surgeon had performed and mortality. This may be due to a "learning curve" effect, with higher mortality in patients operated on by newly appointed surgeons, starting late in the study period. We studied this in more detail by looking at changes in case mix and mortality after new consultant appointments in cardiac surgery in north west England.
Methods
A total of 18 913 patients underwent surgery in the region during the study period, 5678 of whom were included in the learning curve analysis. Table 1 shows details of the new consultant appointments and number of patients operated on. The number of patients pooled into the first, second, third, and fourth years were 1967, 1534, 1283, and 894 respectively. The average number of patients per surgeon was 379. The learning curve dataset included 30% of the total number of patients undergoing surgery.
Table 1 Number of cases dealt with by new consultants according to year of appointment
Overall, 374 patients (2.0%) undergoing isolated coronary artery surgery during the study period died. Mortality in patients operated on by surgeons in their first four years after appointment and that of established consultants was similar (109/5678 (1.9%, 95% confidence interval 1.6% to 2.3%) and 265/13 235 (2.0%, 1.8% to 2.2%), respectively, P = 0.71). The mean predicted mortality for patients operated on by consultants in the first four years after appointment (EuroSCORE mean 3.1, median 3, 25th and 75th centiles 1 and 4) was similar to that of those operated on by more established surgeons (mean 3, median 3, 25th and 75th centiles 1 and 4) (P = 0.79).
The ratio of observed to expected mortality is the best risk adjusted summary of performance. The ratios for the patients operated on by surgeons in the first four years of practice and more established surgeons were 0.61 and 0.66, respectively.
There was a progressive decrease in observed mortality during the first four years after appointment, from 2.2% (1.6% to 3.0%) to 1.2% (0.5% to 2.1%) in the first and fourth years, respectively (P = 0.05, table 2). The predicted mortality progressively increased over the four years from a mean EuroSCORE of 3.0 to 3.3 (P = 0.03). The ratio reduced by a factor of 50% from the first to the fourth year after appointment (P < 0.001). Exclusion of locum appointments from the dataset had no effect on the results. After we accounted for the effect of time and case mix by using the logistic model, the adjusted mortality decreased from 2.3% in the first year to 1.0% in the fourth year (P = 0.019, table 2).
Table 2 Observed mortality, predicted mortality, observed to expected mortality ratios, and adjusted mortality in patients operated on by newly appointed surgeons according to year after appointment
Table 3 shows changes in mortality in the low and high risk groups. There was a significant reduction in mortality in the low risk group with increasing length of time since appointment that was not seen in high risk patients. Table 4 shows that for established consultants there was a significant reduction in mortality over the time of the study (P = 0.01).
Table 3 Changes in observed mortality in low and high risk patients for newly appointed surgeons by year after appointment. Figures are percentage mortality
Table 4 Overall mortality for patients undergoing surgery by established consultants for each year of study period
Discussion
Dyer O. Heart surgeons are to be rated according to bypass surgery success. BMJ 2003;326: 1053.
Bridgewater B, Grayson AD, Jackson M, Brooks N, Grotte GJ, Keenan DJ, et al. Surgeon specific mortality in adult cardiac surgery: comparison between crude and risk stratified data. BMJ 2003;327: 13-7.
North west quality improvement programme in cardiac interventions. www.nwheartaudit.nhs.uk (accessed 12 Jan 2004).
Roques F, Nashef SA, Michel P, Guaducheau E, de Vincentiis C, Baudet E, et al. Risk factors and outcome in European cardiac surgery; analysis of the EuroSCORE multinational database of 19 030 patients. Eur J Cardiothorac Surg 1999;15: 816-23.
Hosmer D, Lemeshow S. Applied logistic regression. New York, NY: John Wiley, 1989.
Hannan EL, O'Donnell JF, Kilburn H Jr, Bernard HR, Yazici A. Investigation of the relationship between volume and mortality for surgical procedures performed in New York state hospitals. JAMA 1989;262: 503-10.
Keogh B, Kinsman R. National adult cardiac surgical database report 2000-2001. London: Society of Cardiothoracic Surgeon of Great Britain and Ireland, 2002.
Roques F, Michel P, Goldstone AR, Nashef SAM. The logistic EuroSCORE. Eur Heart J 2003;24: 1-2.
Michel P, Roques F, Nashef SA; EuroSCORE Project Group. Logistic or additive EuroSCORE for high-risk patients? Eur J Cardiothorac Surg 2003;23: 684-7.
Hassan A, Pozzi M, Hamilton L. New surgical procedures: can we minimise the learning curve? BMJ 2000;320: 171-3.
Bull C, Yates R, Sarkar D, Deanfield J, de Leval M. Scientific, ethical, and logistical considerations in introducing a new operation: a retrospective cohort study from paediatric cardiac surgery. BMJ 2000;320: 1168-73.
Orr RK, Hoehn JL, Col NF. The learning curve for sentinel node biopsy in breast cancer: practical considerations. Arch Surg 1999;134: 764-9.
Gates EA. New surgical procedures: can our patients benefit while we learn? Am J Obstet Gynecol 1997;176: 1293-9.
Dincler S, Koller MT, Steurer J, Bachmann LM, Christen D, Buchmann P. Multidimensional analysis of learning curves in laparoscopic sigmoid resection: eight-year results. Dis Colon Rectum 2003;46: 1371.
Soot SJ, Eshraghi N, Farahmand M, Sheppard BC, Deveney CW. Transition from open to laparascopic fundoplication: the learning curve. Arch Surg 1999;134: 278-81.
Ford WD, Crameri JA, Holland AJ. The learning curve for laparoscopic pyloromyotomy. J Paediatr Surg 1997;32: 552-4.
Ikhena SE, Oni M, Naftalin NJ, Konje JC. The effect of the learning curve on the duration and peri-operative complications of laparoscopically assisted vaginal hysterectomy. Acta Obstet Gynecol Scand 1999;78: 632-5.
Waldman DJ, Steven A, Ypurstine SA, Smith HL. Learning curves in health care. Health Care Manage Rev 2003;28: 41-54.(Ben Bridgewater, consulta)
Correspondence to: B Bridgewater ben.bridgewater@smuht.nwest.nhs.uk
Abstract
In British medicine consultants are appointed to hospitals after a defined period of training, once they have satisfied national training bodies and been successful in open competition for an advertised post. NHS consultants are independent practitioners who function under clinical governance systems in the institutions in which they are employed. Increasingly, they are also regulated by national initiatives, such as the planned publication of surgeon specific mortality data for coronary artery surgery in the United Kingdom in 2004.1
In a recent study we looked at the outcomes of coronary artery surgery for individual surgeons in north west England.2 An incidental finding was a strong association between the volume of operations each surgeon had performed and mortality. This may be due to a "learning curve" effect, with higher mortality in patients operated on by newly appointed surgeons, starting late in the study period. We studied this in more detail by looking at changes in case mix and mortality after new consultant appointments in cardiac surgery in north west England.
Methods
A total of 18 913 patients underwent surgery in the region during the study period, 5678 of whom were included in the learning curve analysis. Table 1 shows details of the new consultant appointments and number of patients operated on. The number of patients pooled into the first, second, third, and fourth years were 1967, 1534, 1283, and 894 respectively. The average number of patients per surgeon was 379. The learning curve dataset included 30% of the total number of patients undergoing surgery.
Table 1 Number of cases dealt with by new consultants according to year of appointment
Overall, 374 patients (2.0%) undergoing isolated coronary artery surgery during the study period died. Mortality in patients operated on by surgeons in their first four years after appointment and that of established consultants was similar (109/5678 (1.9%, 95% confidence interval 1.6% to 2.3%) and 265/13 235 (2.0%, 1.8% to 2.2%), respectively, P = 0.71). The mean predicted mortality for patients operated on by consultants in the first four years after appointment (EuroSCORE mean 3.1, median 3, 25th and 75th centiles 1 and 4) was similar to that of those operated on by more established surgeons (mean 3, median 3, 25th and 75th centiles 1 and 4) (P = 0.79).
The ratio of observed to expected mortality is the best risk adjusted summary of performance. The ratios for the patients operated on by surgeons in the first four years of practice and more established surgeons were 0.61 and 0.66, respectively.
There was a progressive decrease in observed mortality during the first four years after appointment, from 2.2% (1.6% to 3.0%) to 1.2% (0.5% to 2.1%) in the first and fourth years, respectively (P = 0.05, table 2). The predicted mortality progressively increased over the four years from a mean EuroSCORE of 3.0 to 3.3 (P = 0.03). The ratio reduced by a factor of 50% from the first to the fourth year after appointment (P < 0.001). Exclusion of locum appointments from the dataset had no effect on the results. After we accounted for the effect of time and case mix by using the logistic model, the adjusted mortality decreased from 2.3% in the first year to 1.0% in the fourth year (P = 0.019, table 2).
Table 2 Observed mortality, predicted mortality, observed to expected mortality ratios, and adjusted mortality in patients operated on by newly appointed surgeons according to year after appointment
Table 3 shows changes in mortality in the low and high risk groups. There was a significant reduction in mortality in the low risk group with increasing length of time since appointment that was not seen in high risk patients. Table 4 shows that for established consultants there was a significant reduction in mortality over the time of the study (P = 0.01).
Table 3 Changes in observed mortality in low and high risk patients for newly appointed surgeons by year after appointment. Figures are percentage mortality
Table 4 Overall mortality for patients undergoing surgery by established consultants for each year of study period
Discussion
Dyer O. Heart surgeons are to be rated according to bypass surgery success. BMJ 2003;326: 1053.
Bridgewater B, Grayson AD, Jackson M, Brooks N, Grotte GJ, Keenan DJ, et al. Surgeon specific mortality in adult cardiac surgery: comparison between crude and risk stratified data. BMJ 2003;327: 13-7.
North west quality improvement programme in cardiac interventions. www.nwheartaudit.nhs.uk (accessed 12 Jan 2004).
Roques F, Nashef SA, Michel P, Guaducheau E, de Vincentiis C, Baudet E, et al. Risk factors and outcome in European cardiac surgery; analysis of the EuroSCORE multinational database of 19 030 patients. Eur J Cardiothorac Surg 1999;15: 816-23.
Hosmer D, Lemeshow S. Applied logistic regression. New York, NY: John Wiley, 1989.
Hannan EL, O'Donnell JF, Kilburn H Jr, Bernard HR, Yazici A. Investigation of the relationship between volume and mortality for surgical procedures performed in New York state hospitals. JAMA 1989;262: 503-10.
Keogh B, Kinsman R. National adult cardiac surgical database report 2000-2001. London: Society of Cardiothoracic Surgeon of Great Britain and Ireland, 2002.
Roques F, Michel P, Goldstone AR, Nashef SAM. The logistic EuroSCORE. Eur Heart J 2003;24: 1-2.
Michel P, Roques F, Nashef SA; EuroSCORE Project Group. Logistic or additive EuroSCORE for high-risk patients? Eur J Cardiothorac Surg 2003;23: 684-7.
Hassan A, Pozzi M, Hamilton L. New surgical procedures: can we minimise the learning curve? BMJ 2000;320: 171-3.
Bull C, Yates R, Sarkar D, Deanfield J, de Leval M. Scientific, ethical, and logistical considerations in introducing a new operation: a retrospective cohort study from paediatric cardiac surgery. BMJ 2000;320: 1168-73.
Orr RK, Hoehn JL, Col NF. The learning curve for sentinel node biopsy in breast cancer: practical considerations. Arch Surg 1999;134: 764-9.
Gates EA. New surgical procedures: can our patients benefit while we learn? Am J Obstet Gynecol 1997;176: 1293-9.
Dincler S, Koller MT, Steurer J, Bachmann LM, Christen D, Buchmann P. Multidimensional analysis of learning curves in laparoscopic sigmoid resection: eight-year results. Dis Colon Rectum 2003;46: 1371.
Soot SJ, Eshraghi N, Farahmand M, Sheppard BC, Deveney CW. Transition from open to laparascopic fundoplication: the learning curve. Arch Surg 1999;134: 278-81.
Ford WD, Crameri JA, Holland AJ. The learning curve for laparoscopic pyloromyotomy. J Paediatr Surg 1997;32: 552-4.
Ikhena SE, Oni M, Naftalin NJ, Konje JC. The effect of the learning curve on the duration and peri-operative complications of laparoscopically assisted vaginal hysterectomy. Acta Obstet Gynecol Scand 1999;78: 632-5.
Waldman DJ, Steven A, Ypurstine SA, Smith HL. Learning curves in health care. Health Care Manage Rev 2003;28: 41-54.(Ben Bridgewater, consulta)