DOI: 10.1161/CIRCULATIONAHA.107.756379 2008


Published on

1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

DOI: 10.1161/CIRCULATIONAHA.107.756379 2008

  1. 1. Long-Term Results of Heart Operations Performed by Surgeons-in-TrainingSerban C. Stoica, Dimitri Kalavrouziotis, Billie-Jean Martin, Karen J. Buth, Gregory M. Hirsch, John A. Sullivan and Roger J.F. Baskett Circulation 2008;118;S1-S6 DOI: 10.1161/CIRCULATIONAHA.107.756379 Circulation is published by the American Heart Association. 7272 Greenville Avenue, Dallas, TX 72514Copyright © 2008 American Heart Association. All rights reserved. Print ISSN: 0009-7322. Online ISSN: 1524-4539 The online version of this article, along with updated information and services, is located on the World Wide Web at: Subscriptions: Information about subscribing to Circulation is online at Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Fax: 410-528-8550. E-mail: Reprints: Information about reprints can be found online at from at Vanderbilt University Eskind Biomedical Library on October 1, 2008
  2. 2. Long-Term Results of Heart Operations Performed by Surgeons-in-Training Serban C. Stoica, MD; Dimitri Kalavrouziotis, MD; Billie-Jean Martin, MD; Karen J. Buth, MSc; Gregory M. Hirsch, MD; John A. Sullivan, MD; Roger J.F. Baskett, MDBackground—We investigated the association between trainees performing supervised operations and late outcomes of patients undergoing cardiac surgery.Methods and Results—Data were prospectively collected on patients who underwent coronary artery bypass graft surgery, aortic valve replacement, or a combination of these between 1998 and 2005 at the Maritime Heart Center, Halifax, Canada. In-hospital mortality and a composite outcome of in-hospital mortality, stroke, bleeding, intra-aortic balloon pump insertion, renal failure, and sternal infection was compared between teaching (n 1054) and nonteaching cases (n 5877). Late survival and cardiovascular hospital readmissions were also examined. To adjust for baseline risk disparities, we used logistic regression for dichotomous in-hospital outcomes and Cox proportional hazards regression for survival data. Resident cases were significantly more likely to have high-risk features such as depressed ventricular function, redo operation, and urgent or emergent procedure. Resident as primary operator was not independently associated with in-hospital mortality (OR, 1.09; 95% CI, 0.75 to 1.58; P 0.66) or with the composite outcome (OR, 1.01; 95%, CI 0.82 to 1.26; P 0.90). The Kaplan-Meier event-free survival of the 2 groups was equivalent at 1, 3, and 5 years (log-rank P 0.06). By Cox regression, resident cases were not associated with late death or cardiovascular rehospitalization (hazard ratio, 1.05; 95% CI, 0.94 to 1.17; P 0.42).Conclusions—Cases performed by senior-level cardiac surgery residents were more likely to have greater acuity and complexity than staff surgeon-performed cases. However, clinical outcomes were similar in the short- and long-term. Allowing residents to perform cardiac surgery is not associated with adverse patient outcomes. (Circulation. 2008; 118[suppl 1]:S1–S6.) Key Words: morbidity mortality risk factors surgeryT eaching young surgeons how to perform heart operations can be a balancing act. The literature is sparse, but itseems that in our profession, this apprenticeship has taken responsible for good early results, it is conceivable that a few less-than-perfect stitches in a CABG surgery can reduce graft patency over the long term, which may in turn result inplace over entire decades within tight quality control. In suboptimal clinical results. We therefore sought to examinerecent years, a number of factors conspire in potentially the outcomes of resident-performed cases compared withdiluting the overall experience available to trainees. Patients staff cases in terms of event-free long-term survival.coming to cardiac surgery have a higher risk profile andmany programs throughout the world have to deal with Patients and Methodslegislation implementing a reduction in working hours.1,2 Consecutive cases performed by residents and staff surgeons at theAt the same time, institutions report that training surgeons Maritime Heart Center between January 1, 1998, and December 31,in this field remains possible and safe, whether they 2005, were reviewed retrospectively. All patient data were collected prospectively in the institutional database, and audits reveal that lessperform coronary artery bypass graft surgery (CABG), than 5% of data points are missing.5 For the purpose of this study, weaortic valve replacement (AVR),3,4 or operations that are excluded a minority of patients who were out of province referralstechnically more challenging such as mitral valve surgery5 and did not have subsequent links to the provincial administrationor off-pump revascularization.6,7 databases. The training system and how the surgical teaching takes One component that is missing in previously published place at this institution have been described previously.3,5 Briefly, in the Canadian system, cardiac training is a 6-year, full-fledged standreports is an analysis of long-term outcomes. If residents’ alone program to which medical students can match directly afteroperating does not seem to affect hospital morbidity and medical school without prior general surgical training. In our study,mortality, does this also apply to the long-term success of the resident operators were senior-level cardiac surgery trainees inoperation? Despite the best supervision, which may largely be postgraduate years 3 to 6. Because of the presence of a high From the Maritime Heart Center, Dalhousie University, Halifax, Nova Scotia, Canada. Presented at the American Heart Association Scientific Sessions, November 4 –7, 2007, Orlando, Fla. Correspondence to Roger J.F. Baskett, MD, Queen Elizabeth II Health Sciences Centre, 1796 Summer Street, Room 2269, Halifax, Nova Scotia,Canada B3H 3A7. E-mail © 2008 American Heart Association, Inc. Circulation is available at DOI: 10.1161/CIRCULATIONAHA.107.756379 S1 Downloaded from at Vanderbilt University Eskind Biomedical Library on October 1, 2008
  3. 3. S2 Circulation September 30, 2008trainers-to-trainees ratio, a senior-level resident has the privilege of Table 2. A Comparison of the Operative Variables and Crudeselecting the cases with which he or she is to be involved. A resident In-Hospital Outcomes of Resident- versus Staffcase is defined as a procedure performed from “skin to skin” by the Surgeon-Performed Casestrainee with the staff surgeon acting as the first assistant orsupervising while another person directly assists the resident. This Variables (% unless Resident Cases Staff Casesmeans the resident performed all technical aspects of the operation otherwise indicated) (n 1054) (n 5877) P Value(cannulation for bypass, vascular anastomoses, valve implantation, Operative characteristics Acuity of operation 0.01Table 1. A Comparison of the Baseline Characteristics of Elective 49.7 51.5Resident- versus Staff Surgeon-Performed Cases Waiting in-hospital 33.8 36.3Preoperative Variables Resident Cases Staff Cases Urgent ( 24 hours) or 16.5 12.2(% unless otherwise indicated) (n 1054) (n 5877) P Value emergent Type of procedure 0.0001Demographics Isolated CABG 79.2 87.0 Age at operation, years 0.18 Isolated AVR 11.2 6.6 60 30.0 31.3 CABG and AVR 9.6 6.4 60–69 31.7 31.1 Among patients undergoing 0.18 70–79 28.8 30.2 CABG 80 9.5 7.4 No. of distal anastomoses, 3.02 1.06 2.97 0.99 Mean age, SD 65.2 11.2 65.0 10.8 0.32 mean SD Female 28.1 25.5 0.08 Two or more arterial 40.9 30.8 0.0001Cardiovascular risk distalsfactors/diseases Off-pump 9.6 5.3 0.0001 Cigarette smoking 0.67 Operative durations, minutes, Never 31.6 31.9 mean SD Former 52.9 53.3 Isolated CABG Active smoker within 30 15.5 14.8 Aortic clamp time 79.4 29.2 70.8 26.9 0.0001 days of surgery CPB time 117.5 36.6 107.9 35.8 0.0001 Body mass index* 0.13 Isolated AVR 18.5 0.8 0.6 Aortic clamp time 88.6 24.1 87.0 26.9 0.41 18.5–24.9 21.4 21.0 CPB time 120.8 32.2 123.4 37.1 0.43 25–29.9 44.0 41.4 CABG and AVR 30 33.8 36.9 Aortic clamp time 132.5 41.9 128.7 36.8 0.65 Diabetes mellitus 35.6 35.1 0.78 CPB time 179.2 53.9 183.3 48.5 0.17 Hyperlipidemia 81.0 84.2 0.01 Postoperative (in-hospital) Hypertension 63.1 65.6 0.12 data, unadjusted Peripheral vascular disease 18.3 16.7 0.21 Composite* 12.71 11.32 0.19 Cerebrovascular disease 13.6 15.0 0.23 Mortality 3.98 2.91 0.06 Left ventricular ejection 0.003 Stroke 2.09 1.92 0.72 fraction Reoperation for bleeding 3.04 2.57 0.38 0.50 70.3 74.3 Intra-aortic balloon pump 1.33 1.60 0.51 0.30–0.50 23.2 20.7 inserted intra- or 0.30 6.6 5.0 postoperatively Clinical congestive heart 20.6 16.5 0.001 New renal failure 5.69 5.55 0.85 failure Deep sternal wound 1.04 0.73 0.29 Previous myocardial infarction 52.8 54.1 0.43 infection Preoperative atrial fibrillation 11.5 8.6 0.003 Prolonged mechanical 14.52 10.33 0.0001 ventilation 24 hours Preoperative intra-aortic 9.9 6.5 0.0001 balloon pump Length of hospital stay in 6 (5–9) 6 (5–8) 0.006 days (median, IQR)† Previous cardiac surgery 7.2 5.0 0.003 *Any of: death, stroke, reoperation for bleeding, intra- or postoperativeMedical comorbidities insertion of intra-aortic balloon pump, postoperative new renal failure, or deep Hemoglobin 12 g/dL 21.7 18.7 0.02 sternal wound infection. Renal failure 5.4 5.3 0.90 †Among survivors to hospital discharge. CPB indicates cardiopulmonary bypass; IQR, interquartile range. Chronic obstructive 17.7 15.6 0.08 pulmonary disease *Body mass index weight in kg/(height in m)2. Downloaded from at Vanderbilt University Eskind Biomedical Library on October 1, 2008
  4. 4. Stoica et al Heart Operations by Surgeons-in-Training S3and so on) as well as, and importantly, being involved in the constant Meier technique and multivariate analysis was performed using Coxintraoperative dialogue that takes place with the anesthetist and the proportional hazards regression to adjust for baselines disparities inperfusionist to determine the conduct of the operation (myocardial risk.10 Variables that were significantly associated with the primaryprotection, weaning from bypass, and so on). If a senior-level outcome on univariate analysis at a probability value 0.10 or wereresident was scrubbed in a case but did not perform it as primary clinically relevant were included in the model. Effect estimates wereoperator for whatever reason and, instead, assisted a staff surgeon, described using hazard ratios and their associated 95% CIs. Thethis case was labeled a staff surgeon case. proportionality assumption of the Cox regression model was as- Resident cases were identified using case logs compiled by the sessed graphically and with the use of Schoenfeld residuals.11,12residents themselves. In this period, there were 6 residents; for the Risk-adjusted event-free survival curves were generated using thepurposes of this study, they were treated as a single surgeon. During corrected group prognosis method.13 With respect to the longitudinalthe study period, the resident trainees were involved in a variety of outcomes analysis, the study had a statistical power of 82% (betaoperations. However, to minimize potential heterogeneity among error 0.18) to detect a 30% increase in the relative risk of thecases and to allow for more meaningful comparisons, we included in composite outcome of late death or cardiovascular rehospitalizationthe study only those adult patients undergoing isolated CABG, for those patients who had a resident versus staff surgeon as theisolated AVR, or combined CABG and AVR either as a primary or primary operator.redo procedure. The incidence of an in-hospital composite outcome, Statistical analysis was performed using Statistical Analysis Sys-including death, stroke, reoperation for bleeding, intra-aortic balloon tem software version 8.2 (SAS Institute, Cary, NC). Continuouspump inserted during the operation or postoperatively, new postop- variables were compared using unpaired t tests or nonparametricerative renal failure, and deep sternal wound infection was compared equivalents where appropriate. Proportions were compared using 2between cases that were performed by residents and those performed tests or Fisher exact test where appropriate and Mantel-Haenszel 2by staff surgeons. The rates of the individual components of the statistics with one degree of freedom were used to test for trends.composite were also examined among resident- and staff-performed All probability values were 2-sided. The study was in compliancecases. Separate logistic regression models of both in-hospital mor- with our local institutional research ethics board and received fulltality and the composite outcome as the dependent variables were fit which all variables with a univariate association significant at aprobability value 0.10 or that were felt to be clinically relevant Statement of Responsibilitywere included in the model. A variable selection algorithm was not The authors had full access to and take full responsibility for theused. Effect estimates were described using ORs and their associated integrity of the data. All authors have read and agree to the95% CIs. Each model’s discriminatory capacity and calibration were manuscript as written.assessed by computing the area under the receiver operating char-acteristic curve (C statistic) and by using the Hosmer-Lemeshowgoodness-of-fit test, respectively.8,9 Results The association between resident trainee versus staff as operator Between January 1998 and December 2005, 9510 patientsand longitudinal clinical outcomes postcardiac surgery was also were operated on at our institution, of which 1648 wereexamined and constituted the main focus of this study. Patients that resident cases (17.3%). There were 7 staff surgeons for thewere discharged alive after CABG and/or AVR were linked to entire time period and 6 residents over the 7-year period. Thegovernment administrative hospitalization and mortality databases.The primary clinical end point of interest in this study was the breakdown of training cases was (percent of total cases stafffreedom from all-cause mortality and a first hospital readmission for and resident combined): CABG only: 947 (14.1%);cardiovascular ischemic events (unstable angina, myocardial infarc- AVR CABG: 248 (21.9%); mitral valve repair/tion, transient ischemic attack, or stroke), prosthetic valve-related replacement CABG: 142 (27.8%); aortic root/arch/descend-complications (recurrent stenosis/insufficiency or endocarditis), dys- ing or thoracoabdominal pathology: 91 (31.1%); transplant:rhythmias (atrial or ventricular fibrillation or flutter, ventriculartachycardia, or cardiac arrest), congestive heart failure, and cardiac 17 (25.4%); and other: 203 (25.0%).procedures (valve repair/replacement, percutaneous or surgical re- For this analysis, we included only CABG and/or AVR ofvascularization). Survival times were analyzed using the Kaplan- patients residing in the province of Nova Scotia (to allow for Table 3. The Association Between Resident as Operator and (1) In-Hospital Mortality; and (2) the Composite Outcome in Fully Adjusted Multivariate Logistic Regression Models Stratified by Procedure* Model Accuracy Adjusted OR 95% CI P Value C Statistic GOF P Value† In-hospital mortality All procedures: CABG, AVR, and combined CABG/AVR 1.09 0.75–1.58 0.66 0.82 0.52 Isolated CABG 1.14 0.75–1.74 0.55 0.84 0.50 AVR CABG 1.01 0.41–2.50 0.98 0.83 0.42 Composite outcome‡ All procedures: CABG, AVR, and combined CABG/AVR 1.01 0.82–1.26 0.90 0.74 0.29 Isolated CABG 1.00 0.79–1.28 0.97 0.75 0.31 AVR CABG 0.84 0.53–1.34 0.46 0.73 0.12 *The statistically significant covariates that emerged as independent predictors of outcome have been omitted for clarity. †A P value 0.05 suggests that the model is well calibrated with good agreement between the values observed in the data set and those expected by the model. ‡Any of: death, stroke, reoperation for bleeding, intra- or postoperative insertion of intra-aortic balloon pump, postoperative new renal failure, or deep sternal wound infection. GOF indicates Hosmer and Lemeshow goodness-of-fit test. Downloaded from at Vanderbilt University Eskind Biomedical Library on October 1, 2008
  5. 5. S4 Circulation September 30, 2008linkage to administrative data). Therefore, the study popula- Table 4. Multivariate Predictors of Late Death or Hospitaltion included 6931 patients, of which residents performed Readmission for Cardiovascular Events/Procedures Among1054 and staff surgeons performed 5877 operations. Patient Patients Discharged Alive After CABG, AVR, and Combineddemographics at baseline were similar between the 2 groups Procedures (Cox proportional hazards model)with respect to many prognostically important variables such Variable Adjusted HR 95% CI P Valueas age, diabetes, and extracardiac vasculopathy. However, Resident case 1.05 0.94–1.17 0.42residents operated on patients with a significantly higher Age, yearsproportion of left ventricular dysfunction, clinical congestive 60 1.0 (referent)heart failure, preoperative atrial fibrillation, preoperative 60–69 1.05 0.94–1.17 0.36intra-aortic balloon pump, previous cardiac surgery, andanemia (all P 0.05; Table 1). In addition, resident cases were 70–79 1.30 1.16–1.46 0.0001significantly more likely to be urgent or emergent operations 80 1.40 1.19–1.65 0.0001and combined procedures (Table 2). The durations of cardio- Cerebrovascular disease 1.20 1.08–1.34 0.0006pulmonary bypass and aortic clamping were on average 10 Peripheral vascular disease 1.25 1.13–1.38 0.0001minutes longer for resident-performed isolated CABG cases, CHF 1.18 1.06–1.32 0.003but operative durations were not significantly different be- LVEFtween the 2 groups for isolated AVR and CABG/AVR cases 0.50 1.0 (referent)(Table 2). The observed unadjusted in-hospital mortality was 0.30–0.50 1.16 1.06–1.28 0.003higher among resident cases; however, this difference was not 0.30 1.22 1.03–1.45 0.02statistically significant (3.98% versus 2.91%, P 0.06). In Previous cardiac surgery 1.35 1.16–1.58 0.0002addition, the incidence of the composite end point did notsignificantly differ between the 2 groups (Table 2). Diabetes mellitus 1.28 1.17–1.39 0.0001 In multivariate analysis using logistic regression modeling, Hypertension 1.21 1.11–1.32 0.0001resident as operator was not independently associated with Hyperlipidemia 0.89 0.81–0.99 0.03in-hospital mortality or the composite outcome of death, Preoperative renal failure 1.50 1.29–1.74 0.0001stroke, reoperation for bleeding, intra- or postoperative inser- COPD 1.30 1.18–1.44 0.0001tion of an intra-aortic balloon pump, renal failure, or deep Preoperative atrial fibrillation 1.67 1.49–1.89 0.0001sternal wound infection (Table 3). Similarly, resident was not Acuity of operationassociated with in-hospital mortality or the composite out- Elective 1.0 (referent)come when strata of procedure type was examined separately. Waiting in-hospital 1.22 1.11–1.34 0.0001All models had good discriminatory capacity and were a goodfit to the data (Table 3). Urgent ( 24 hours) or 1.36 1.20–1.55 0.0001 emergent Late outcomes were also examined. Of 6931 patients Type of procedureundergoing CABG and/or AVR at our institution, 6718patients were discharged alive (96.9%). Among these, 4 Isolated CABG 1.0 (referent)patients did not have a valid health card number and we were Isolated AVR 1.01 0.85–1.20 0.91unable to link them to the healthcare utilization databases. CABG and AVR 1.25 1.08–1.45 0.004The remaining 6714 patients were successfully linked to the Preoperative hemoglobin, g/dLadministrative hospitalization and mortality data sets and 14.4 1.0 (referent)were included in the longitudinal analyses (5703 staff 13.2–14.3 1.06 0.94–1.18 0.34surgeon-performed cases and 1011 resident-performed 12.0–13.1 1.18 1.04–1.34 0.009cases). Median follow-up to event or censor was 2.7 years 12.0 1.44 1.26–1.64 0.0001(interquartile range 1.0 to 4.7 years, maximum 8.2 years). Cigarette smokingThere were 2522 late deaths or rehospitalizations for cardio-vascular events or procedures during the study period Never 1.0 (referent)(37.6%). The Kaplan-Meier event-free survival of staff- Former 1.08 0.98–1.18 0.13versus resident-performed cases was equivalent at 1, 3, and 5 Active smoker within 30 days 1.19 1.04–1.36 0.01years: 81.3% versus 79.1%, 68.2% versus 66.7%, and 58.6% of surgeryversus 55.8%, respectively (log rank P 0.06). In a fully Year of operationrisk-adjusted time-to-first event multivariate Cox propor- 1998–1999 1.0 (referent)tional hazards model, resident as operator was not associated 2000–2001 0.83 0.75–0.92 0.0004with long-term death or hospital readmission for cardiovas- 2002–2003 0.82 0.73–0.91 0.0003cular reasons (adjusted hazard ratio, 1.05; 95% CI, 0.94 to 2004–2005 0.81 0.70–0.93 0.0031.17; P 0.42). The variables that emerged as independent CHF indicates congestive heart failure; COPD, chronic obstructive pulmonarypredictors of late death and rehospitalization are shown in disease; HR, hazard ratio; LVEF, left ventricular ejection fraction.Table 4. Crude and adjusted survival curves are depicted inthe Figure. Assessment of the Cox model using tests ofproportionality and Schoenfeld residuals revealed that the Downloaded from at Vanderbilt University Eskind Biomedical Library on October 1, 2008
  6. 6. Stoica et al Heart Operations by Surgeons-in-Training S5 100 90 Staff cases, crude Staff cases, adjusted 80 Resident cases, crude Event-free survival (%) Resident cases, adjusted 70 60 50 40 0 1 2 3 4 5 6 7 8 Years post-discharge Patients remaining Staff cases 5,703 4,322 3,439 2,527 1,887 1,308 865 449 76 Resident cases 1,011 722 606 487 347 192 98 20 3 Figure. Crude Kaplan-Meier and risk-adjusted event-free survival curves.effects of the predictor variables were the same at all values influencing medium-term outcome in this report, all seem toof time (constant hazard ratios over time) suggesting that the have a small hazard ratio, suggesting small individual con-proportionality assumption of the model was not violated. tributions. The usual predictors of poorer long-term outcome A separate Cox model was fit for patients undergoing AVR are confirmed in this analysis (age, comorbidities and acuity(with or without concomitant CABG). A similar association of procedure; Table 4).between resident case and outcomes was observed (hazard Training and quality assurance go hand in hand and byratio, 1.10; 95% CI, 0.86 to 1.40; P 0.46). examining the few existing reports, it can be seen how crude and adjusted results, as well as the tools used to evaluateComment them, are evolving. In reference to the Bristol report onPreviously published reports have shown that allowing resi- off-pump training, Blackstone describes the sophisticateddents to perform cardiac surgical operations under staff analyses (and their caveats) to which these cases are amena-supervision does not adversely affect in-hospital patient ble.7,14 Lim and Tsui showed how adopting a team approachoutcomes.2–5,7 Our data support these findings using a much to training ensures a continued good level of exposure despitelarger series of patients. Furthermore, we have found that an absolute reduction in hours worked.2 In an interesting butthese results are maintained over a median follow-up of 3 not unexpected finding in their analysis, the main factorsyears with a maximum follow-up of 8 years. These results influencing whether an operation became a teaching casesuggest that resident training does not adversely affect mor- were the consultant in charge and the complexity of thebidity and mortality in the intermediate term after cardiac procedure. Although few in number, these publications showsurgery. collectively that the subtle chemistry of training can be The absolute differences in the pre- and intraoperative broken down into smaller components and examined incharacteristics of the 2 groups studied here are small. Overall, analytic fashion. Of utmost importance, patient safety isresident cases were sicker and more complex patients. This preserved in all these positive reports. Intuitively, this mustwas expected because the residents choose which cases they be the case in the overwhelming majority of programs.want to scrub for and tend to pick those that are more Extrapolating from the level of an individual institution to thecomplex. This, we believe, is beneficial in terms of residency wider world is conceptually difficult but not impossible. Haantraining and exposure to difficult decision-making. It appears and colleagues examined the STS database and drew athat the clustering of risk factors in the 2 groups has such an comparison between residency and nonresidency programs.effect that short- and medium-term outcomes remain compa- Without teasing out further which cases were done byrable. For example, prolonged ventilation in the resident residents, training programs had longer crossclamp and per-group in unadjusted analysis may be associated with a higher fusion times but short-term patient outcomes did not appearrate of preoperative intra-aortic balloon pump use. A more to be affected.4detailed analysis of observed versus expected outcomes was What cannot be described in simple, retrospective reportsdone previously3 and is not duplicated here. such as ours is how the team, containing trainees in this The main focus was what happens to patients after dis- instance, adapts to change for an optimal outcome. In thecharge from our service. Of the many independent risk factors conduct of an operation there is a permanent interaction Downloaded from at Vanderbilt University Eskind Biomedical Library on October 1, 2008
  7. 7. S6 Circulation September 30, 2008between patient and system factors or in a temporal sequence Disclosuresbetween pre- and intraoperative factors.15 In a landmark study None.from Great Ormond St Hospital in London, de Leval andcolleagues used human factors researchers to observe teambehavior and adjustment in performing complex pediatriccardiac surgery.16 Recovery from major as well as serial References 1. DeLaRosa J, Thourani VH, Wheatley GH III, McMullan DM, Karaman-minor adverse events leads to avoidance of death and/or oukian RL, Greene MG, Morales DL. Impact of resident duty hournear-miss. This sophisticated type of analysis may be appli- standards on cardiothoracic residents and program directors. Ann Thoraccable in the future to other fields of cardiac surgery, including Surg. 2005;80:1569 – 2. Lim E, Tsui S, on behalf of the registrars and consultant cardiac surgeons Several limitations should be noted. The single-center of Papworth Hospital 2003–2005. Impact of the European Working Time Directive on exposure to operative cardiac surgical training. Eur Jnature of the study limits its generalization. However, it is Cardiothorac Surg. 2006;30:574 –577.reasonable to believe that similar results could be observed in 3. Baskett RJ, Buth KJ, Legare JF, Hassan A, Hancock Friesen C, Hirschother jurisdictions and programs. Despite risk adjustment GM, Ross DB, Sullivan JA. Is it safe to train residents to perform cardiacusing multivariable analysis that controls for differences in surgery? Ann Thorac Surg. 2002;74:1043–1049.patient baseline characteristics, intra- and postoperative vari- 4. Haan CK, Milford-Beland S, O’Brien S, Mark D, Dullum M, Ferguson TB, Peterson ED. Impact of residency status on perfusion times andables, the study’s observational design, and lack of random outcomes for coronary artery bypass graft surgery. Ann Thorac Surg.treatment allocation cannot completely exclude residual con- 2007;83:2103–2110.founding from unrecognized or incompletely measured vari- 5. Baskett RJF, Kalavrouziotis D, Buth KJ, Hirsch GM, Sullivan JAP.ables. The potential for selection bias is all the more impor- Training residents in mitral valve surgery. Ann Thorac Surg. 2004;78:tant because residents are free to choose which cases they 1236 –1240. 6. Asimakopoulos G, Karagounis AP, Valencia O, Rose D, Niranjan G,perform. Furthermore, there may be some intangible factor at Chandrasekaran V. How safe is it to train residents to perform off-pumpplay that attracts residents to certain cases and confers a coronary artery bypass surgery? Ann Thorac Surg. 2006;81:568 –572.protective effect, which is falsely attributed to resident as 7. Caputo M, Reeves BC, Rogers CA, Ascione R, Angelini GD. Monitoringprimary operator. Such a covariate may obviate the possible the performance of residents during training in off-pump coronaryexistence of a negative, deleterious association between surgery. J Thorac Cardiovasc Surg. 2004;128:907–915. 8. Cook NR. Use and misuse of the receiver operating characteristic curveresident as primary operator and clinical outcome after in risk prediction. Circulation. 2007;115:928 –935.cardiac surgery. This would seem unlikely given that by most 9. Lemeshow S, Hosmer DW. A review of goodness of fit statistics for useimportant well-established risk factors the adverse outcomes in the development of logistic regression models. Am J Epidemiol. 1982;the resident cases were higher risk, yet the outcomes were not 115:92–106.significantly different. Therefore, the influence of a residual 10. Cox DR. Regression analysis and life tables. J R Stat Soc. 1972;34: 187–220.confounding variable, if present, is likely to be minimal. 11. Barlow WE, Prentice RL. Residuals for relative risk regression.Another limitation is the use of clinical outcomes instead of Biometrika. 1988;75:65–74.more sensitive measures of technical success such as angio- 12. Fisher LD, Lin DY. Time-dependent covariates in the Cox proportional-graphic patency. However, the objective of this study was to hazards regression model. Annu Rev Public Health. 1999;20:145–157.examine “hard” clinical end points such as death and ische- 13. Ghali WA, Quan H, Brant R, van Melle G, Norris CM, Faris PD, Galbraith PD, Knudtson ML, for the APPROACH (Alberta Provincialmic events requiring rehospitalization, the opportunity to Project for Outcome Assessment in Coronary Heart Disease) Investi-acquire communitywide outcomes being a strength of this gators. Comparison of 2 methods for calculating adjusted survivalwork. Despite these limitations, our results strongly support curves from proportional hazards models. JAMA. 2001;286:the view that cardiac operations performed by surgeons in 1494 – are not associated with adverse outcomes after 14. Blackstone EH. Monitoring surgical performance. J Thorac Cardiovasc Surg. 2004;128:807– discharge. 15. Stoica SC, Sharples LD, Ahmed I, Roques F, Large SR, Nashef SA. Preoperative risk prediction and intraoperative events in cardiac surgery. Sources of Funding Eur J Cardiothorac Surg. 2002;21:41– 46.D.K. was supported by a Canadian Institutes of Health Research 16. De Leval MR, Carthey J, Wright DJ, Farewell VT, Reason JT. HumanTeam Grant in Cardiovascular Outcomes Research to the Canadian factors and cardiac surgery: a multicenter study. J Thorac CardiovascCardiovascular Outcomes Research Team. Surg. 2000;119:661– 672. Downloaded from at Vanderbilt University Eskind Biomedical Library on October 1, 2008