- Email: [email protected]
Surgical trainee participation during infrainguinal bypass grafting procedures is associated with increased early postoperative graft failure
Surgical trainee participation during infrainguinal bypass grafting procedures is associated with increased early postoperative graft failure John E. Scarborough, MD, Theodore N. Pappas, MD, Mitchell W. Cox, MD, Kyla M. Bennett, MD, and Cynthia K. Shortell, MD, Durham, NC Objective: This study was conducted to determine the potential effect of surgical trainee participation during infrainguinal bypass procedures on postoperative graft patency rates. Methods: Data from the National Surgical Quality Improvement Program (NSQIP) Participant User Files from 2005 through 2009 were retrospectively reviewed, using propensity score matching, to identify all patients undergoing infrainguinal bypass grafting procedures, excluding those who had prior operation <30 days of the index procedure. A separate analysis was performed on a subset of procedures from the entire NSQIP sample that was matched on propensity for intraoperative surgical trainee participation. The primary predictor variable was intraoperative surgical trainee participation. The main outcome measure was the 30-day postoperative graft failure rate. Results: For the entire sample of 14,723 NSQIP patients undergoing infrainguinal bypass grafting, 30-day graft failure rates were significantly higher when a surgical trainee participated (5.8%) vs without participation (3.9%; P < .0001). For the cohort of 9234 patients matched on their propensity for intraoperative trainee participation, this difference in graft failure rate remained significant (5.0% with participation vs 4.0% without participation; P ⴝ .02). Conclusions: Surgical trainee participation is an independent risk factor for technical failure after infrainguinal bypass grafting. Prospective evaluation is needed to determine the cause of this increase in graft failure rates for procedures that involve surgical trainees. ( J Vasc Surg 2012;55:715-20.)
The effect of intraoperative surgical trainee participation on technical outcomes after infrainguinal bypass grafting is currently unknown. The National Surgical Quality Improvement Program (NSQIP) database offers a unique opportunity to explore this effect due to its inclusion of prospectively verified information about intraoperative trainee participation and also 30-day postoperative graft outcomes. These characteristics of NSQIP provide it with distinct advantages compared with other large multicenter data sources that rely on hospital teaching status as a marker of intraoperative trainee involvement and generally have low sensitivity for the detection of procedure-specific complications.1 Most of the previously published studies that have examined the potential effect of surgical education on nonvascular surgical quality rely on these flawed data sources or include procedures from only one or a small number of hospitals.2-14 The objective of our analysis was therefore to use data from the NSQIP Participant User File to determine whether surgical trainee involvement during lower extremity revascularization procedures affects early postoperative graft outcome. From the Department of Surgery, Duke University Medical Center. Competition of interest: none. Reprint requests: John E. Scarborough, MD, Assistant Professor of Surgery, DUMC 2837, Durham, NC 27710 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a competition of interest. 0741-5214/$36.00 Copyright © 2012 by the Society for Vascular Surgery. doi:10.1016/j.jvs.2011.06.122
METHODS The NSQIP Participant User Files for 2005 through 2009 were used for this retrospective analysis, which included all patients with a primary Current Procedural Terminology code for any of the following lower extremity infrainguinal bypass procedures: 35556 (femoral-popliteal bypass with vein), 35558 (femoral-femoral bypass with vein), 35566 (femoral-distal bypass with vein), 35570 (distal-distal bypass with vein), 35571 (popliteal-distal bypass with vein), 35583 (femoral-popliteal bypass with in situ vein), 35585 (femoral-distal bypass with in situ vein), 35587 (popliteal-distal bypass with in situ vein), 35656 (femoral-popliteal bypass with prosthesis), 35661 (femoral-femoral bypass with prosthesis), 35666 (femoraldistal bypass with prosthesis), and 35671 (popliteal-distal bypass with prosthesis). The analysis excluded patients who underwent a previous operative procedure ⱕ30 days of the index procedure. The primary outcome measure for our analysis was early postoperative graft failure, which the NSQIP defines as the mechanical failure of a graft or prosthesis requiring reoperation, an interventional radiology procedure, or balloon angioplasty ⱕ30 days of the operation. Secondary outcome measures included overall postoperative morbidity, mortality, and postoperative hospital length of stay. Complications included in postoperative morbidity were any or all of the following: superficial surgical site infection, deep surgical site infection, organ/space surgical site infection, wound dehiscence, pneumonia, unplanned intubation, ventilator dependency for ⬎48 hours, pulmonary embolism, progressive renal insufficiency, acute renal 715
JOURNAL OF VASCULAR SURGERY March 2012
716 Scarborough et al
failure, stroke, coma ⬎24 hours, peripheral nerve injury, cardiac arrest requiring cardiopulmonary resuscitation, urinary tract infection, bleeding requiring transfusion of packed red blood cells, deep venous thrombosis, sepsis, or septic shock. The primary predictor variable for our analysis was presence or absence of surgical trainee participation during the primary operative procedure. For each procedure, NSQIP contains information on the highest postgraduate year (PGY) of the trainee who scrubs for the procedure. For the purpose of our analysis, we used this information to create a dichotomous categoric variable, with PGY ⱖ1 indicating the presence of a surgical trainee and PGY ⫽ 0 indicating the presence of only an attending surgeon during the procedure. Other primary predictor variables included prior revascularization or amputation procedure for peripheral vascular disease, rest pain or gangrene, patient age, sex, white race, body mass index, American Society of Anesthesiologists classification of ⱖ3, diabetes mellitus, tobacco use within the past year, ⬎2 drinks of ethanol daily ⱕ2 weeks before operation, dyspnea at rest or on exertion, partially or totally dependent functional status preoperatively, chronic obstructive pulmonary disease, congestive heart failure, hypertension requiring medical therapy, neurologic disorder (including any or all of impaired sensorium, prior coma, hemiplegia, history of transient ischemic attacks, stroke with or without neurologic deficit, tumor involving central nervous system, paraplegia, or quadriplegia), coronary artery disease (including any or all of myocardial infarction ⱕ6 months before the operation, prior percutaneous coronary intervention, prior cardiac surgery, or angina ⱕ1 month before the operation), renal disease (including acute renal failure present ⱕ24 hours before the procedure or need for dialysis ⱕ2 weeks before the operation, or both), open wound or infected wound preoperatively, use of steroids ⱕ30 days of operation for chronic medical condition, ⬎10% loss body weight in 6 months before operation, bleeding disorder (including chronic anticoagulation therapy other than aspirin that was not discontinued preoperatively), preoperative sepsis, and emergency posting of the index operation. Intraoperative predictor variables included the wound classification of the index operative incision, nature of bypass conduit used (prosthetic vs venous), tibial or peroneal target for distal anastomosis, need for intraoperative packed red blood cell transfusion, and operative time. Preoperative and intraoperative characteristics among the infrainguinal bypass grafting procedures included in NSQIP with and without trainee participation were compared using Mann-Whitney rank sum test for continuous variables and Pearson 2 tests for categoric variables. A nonparsimonious logistic regression analysis model was created to estimate the likelihood that a surgical trainee participated in the index infrainguinal bypass grafting procedure.15 Preoperative patient characteristics and intraoperative procedural characteristics, including operative time and intraoperative requirement, were included as predictor
variables to adjust for patient condition and complexity of the index procedure. A propensity score for resident participation (range, 0-1) was calculated for each patient using the logit coefficients for the predictors of trainee participation. The propensity scores were then used to create two evenly matched groups according to trainee participation using a calipermatching algorithm with a caliper distance of .005, with controls being used only once in the matching. The preoperative and intraoperative characteristics of the matched cohort of patients were compared using Wilcoxon signed rank sum tests for continuous variables and McNemar 2 tests for categoric variables.16 For the entire NSQIP sample, primary and secondary outcome measures between the “no trainee” and “trainee” groups were compared using 2 tests or Wilcoxon rank sum tests, where appropriate. For the cohorts of patients matched on propensity for trainee participation, comparison of the primary and secondary outcome measures for the matched pairs was performed using Wilcoxon signed rank sum tests or McNemar 2 tests, where appropriate. All statistical analyses were performed using Stata 11.0 software (StataCorp, College Station, Tex). RESULTS The Table reports the preoperative and intraoperative characteristics for infrainguinal bypass procedures performed without or with trainee participation intraoperatively for the entire sample of NSQIP procedures and for the matched cohort. Comparison of the two groups using the entire NSQIP sample of infrainguinal bypass grafting procedures shows significant and potentially important differences between the two groups. Most notably, those patients whose procedures involved surgical trainee participation had a higher incidence of prior revascularization or amputation surgery for peripheral vascular disease and were significantly more likely to present with rest pain or gangrene, or both, than patients whose procedures were performed by attending surgeons alone. Procedures performed with trainee participation also appeared to be more complex on average than procedures performed by attendings alone. Procedures with trainee participation took significantly longer, were more likely to require intraoperative packed red blood cell transfusion, and were more likely to involve distal anastomoses beyond the level of the tibioperoneal trunk. When the two groups of patients were matched by the propensity for intraoperative trainee participation, the differences in preoperative and intraoperative characteristics were no longer statistically or clinically different (Table 1). The 30-day graft failure rates were significantly higher in procedures involving surgical trainee participation for the entire NSQIP sample of infrainguinal bypass grafting procedures and also for the cohorts matched on propensity for resident participation (Fig 1). In the latter comparison, the relative risk of graft failure was 25% greater for procedures that included surgical trainees than for procedures
JOURNAL OF VASCULAR SURGERY Volume 55, Number 3
Scarborough et al 717
Table I. Preoperative characteristics of patients undergoing infrainguinal bypass grafting with and without trainee participation All infrainguinal bypass patients in NSQIP Variablea Patients Age, years Female White Body mass index, kg/m2 ASA classification ⱖ3 Diabetes mellitus Tobacco use Heavy ethanol use Dyspnea Dependent functional status On mechanical ventilation Chronic obstructive pulmonary disease Preoperative pneumonia Ascites Esophageal varices Congestive heart failure Coronary artery disease Hypertension Renal disease Neurologic disorder Disseminated cancer Preoperative infected wound Steroids Weight loss Bleeding disorder Preoperative transfusion Chemotherapy ⱕ30 days Radiotherapy ⱕ90 days Preoperative sepsis None SIRS Sepsis Septic shock Emergency procedure Prior revascularization or amputation for PVD Rest pain or gangrene Prosthetic graft Tibial or peroneal distal anastomosis Wound classification of incision Clean Clean-contaminated Contaminated Dirty/infected Operative time, min Intraoperative PRBC transfusion
No trainee
Trainee
5562 (37.8) 68.7 ⫾ 11.5 2068 (37.2) 4110 (80.9) 27.1 ⫾ 6.3 1110 (20.0) 2202 (39.6) 2361 (42.5) 340 (6.1) 1097 (19.7) 78 (1.4) 10 (0.2) 840 (15.1) 14 (0.3) 10 (0.2) 5 (0.1) 129 (2.3) 2126 (38.2) 4585 (82.4) 348 (6.3) 1119 (20.1) 26 (0.5) 1790 (32.2) 197 (3.5) 90 (1.6) 1185 (21.3) 9 (0.2) 22 (0.4) 10 (0.2)
9161 (62.2) 67.0 ⫾ 12.1 3252 (35.5) 6697 (77.9) 27.0 ⫾ 6.5 1716 (18.7) 3792 (41.4) 3955 (43.2) 529 (5.8) 1644 (18.0) 152 (1.7) 22 (0.2) 1168 (12.8) 29 (0.3) 22 (0.2) 12 (0.1) 237 (2.6) 3476 (37.9) 7645 (83.5) 671 (7.3) 1832 (20.0) 49 (0.5) 3167 (34.6) 380 (4.2) 151 (1.7) 2166 (23.6) 15 (0.2) 39 (0.4) 25 (0.3)
5306 (95.4) 185 (3.3) 65 (1.2) 6 (0.1) 318 (5.7) 2838 (51.0) 2387 (42.9) 2203 (39.6) 1793 (32.2)
8664 (94.6) 379 (4.1) 99 (1.1) 19 (0.2) 511 (5.6) 4066 (55.3) 4399 (48.0) 3388 (37.0) 3554 (38.8)
5110 (91.9) 8282 (90.4) 144 (2.6) 175 (1.9) 164 (3.0) 401 (4.4) 144 (2.6) 303 (3.3) 192.1 ⫾ 93.3 240.0 ⫾ 105.2 598 (10.8) 1783 (19.5)
P ⬍.0001 .04 ⬍.0001 .44 .07 .03 .39 .4 .007 .22 .45 ⬍.0001 .48 .45 .48 .31 .73 .11 .01 .86 .58 .003 .07 .89 .001 .98 .78 .26 .04
.72 ⬍.0001 ⬍.0001 .001 ⬍.0001 ⬍.0001
⬍.0001 ⬍.0001
Cohorts matched on propensity for trainee participation No trainee
Trainee
4617 68.1 ⫾ 11.5 1686 (36.5) 3712 (80.4) 27.1 ⫾ 6.3 900 (19.5) 1854 (40.2) 1990 (43.1) 280 (6.1) 894 (19.4) 67 (1.5) 9 (0.2) 671 (14.5) 11 (0.2) 10 (0.2) 5 (0.1) 109 (2.4) 1760 (38.1) 3849 (83.4) 298 (6.5) 928 (20.1) 23 (0.5) 1471 (31.9) 166 (3.6) 76 (1.7) 1001 (21.7) 8 (0.2) 19 (0.4) 9 (0.2)
4617 68.5 ⫾ 11.7 1708 (37.0) 3727 (80.7) 27.1 ⫾ 6.5 917 (19.9) 1855 (40.2) 1988 (43.1) 293 (6.4) 907 (19.5) 62 (1.3) 8 (0.2) 710 (15.4) 10 (0.2) 9 (0.2) 6 (0.1) 111 (2.4) 1770 (38.3) 3811 (82.5) 305 (6.6) 932 (20.2) 24 (0.5) 1509 (32.7) 172 (3.7) 74 (1.6) 985 (21.3) 6 (0.1) 19 (0.4) 9 (0.2)
4403 (95.4) 160 (3.5) 49 (1.1) 5 (0.1) 231 (5.0) 2404 (52.1) 2.029 (44.0) 1855 (40.2) 1514 (32.8)
4410 (95.5) 153 (3.3) 48 (1.0) 6 (0.1) 232 (5.0) 2411 (52.2) 1992 (43.1) 1858 (40.2) 1510 (32.7)
4243 (91.9) 102 (2.2) 147 (3.2) 125 (2.7) 199.8 ⫾ 88.7 542 (11.7)
4250 (92.1) 107 (2.3) 142 (3.1) 118 (2.6) 199.0 ⫾ 82.0 539 (11.7)
P .07 .65 .71 .5 .67 ⬎.99 .98 .60 .75 .72 ⬎.99 .27 ⬎.99 ⬎.99 ⬎.99 .95 .85 .31 .80 .94 ⬎.99 .41 .78 .94 .71 .79 ⬎.99 ⬎.99 .77
⬎.99 .90 .44 0.97 .95 .82
.88 .94
ASA, American Society of Anesthesiologists; NSQIP, National Surgical Quality Improvement Program; PRBC, packed red blood cells; PVD, peripheral vascular disease; SIRS, systemic inflammatory response syndrome. a Categoric data are presented as number (%), and continuous data as mean ⫾ standard deviation.
that were performed by attending surgeons alone. Conversely, overall morbidity and 30-day postoperative mortality were not significantly associated with intraoperative trainee participation for the matched cohort (Fig 2 and Fig 3). Postoperative hospital length of stay was significantly longer for procedures involving trainee participation for the entire NSQIP sample (mean, 5.6; median, 4; interquartile range [IQR], 3-7 days for the “no trainee” group vs mean, 6.6; median, 5; IQR, 3-7 days for the “trainee” group P ⬍
.0001) as well as for the matched cohort of patients (mean, 5.5; median, 4; IQR, 3-6 days for the “no trainee” group vs mean, 5.9; median, 4; IQR, 3-7 days, P ⫽ .14). Fig 4 shows the postoperative graft failure rate for the entire NSQIP sample, stratified by the PGY of the highestlevel participating surgical trainee. In general, graft failure rates had a direct association with PGY training level of the participating trainee and were highest in those procedures that involved a trainee in PGY ⱖ6 of training.
JOURNAL OF VASCULAR SURGERY March 2012
30-Day Postoperave Mortality Rate
718 Scarborough et al
3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% No Resident (n=5,562)
Resident (n=9,161)
All NSQIP Paents
No Resident
Resident
(n=4,617)
(n=4,617)
Matched Cohort
Fig 3. The 30-day day postoperative mortality rate is shown in patients undergoing infrainguinal bypass grafting with and without trainee participation. NSQIP, National Surgical Quality Improvement Program.
30-Day Overall Postoperave Morbidity Rate
Fig 1. The 30-day early graft failure rate is shown for patients undergoing infrainguinal bypass grafting with or without trainee participation. NSQIP, National Surgical Quality Improvement Program. 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% No Resident (n=5,562)
Resident (n=9,161) All NSQIP Paents
No Resident
Resident
(n=4,617)
(n=4,617)
Fig 4. Postoperative graft failure rate is shown for entire National Surgical Quality Improvement Program (NSQIP) sample stratified by highest postgraduate year (PGY) training level of the participating trainee. CI, Confidence interval.
Matched Cohort
Fig 2. Overall morbidity is shown for patients undergoing infrainguinal bypass grafting with and without trainee participation. NSQIP, National Surgical Quality Improvement Program.
DISCUSSION Our analysis of NSQIP data from 2005 through 2009 demonstrates that infrainguinal bypass grafting procedures performed with surgical trainees have a 52% higher unadjusted risk of 30-day postoperative graft failure than procedures performed by attending surgeons alone. When comparing a subset of patients that are well matched for patient comorbidities and procedure complexity, the association between trainee participation and increased technical complications rates remains significant. To our knowledge, this is the first time that intraoperative surgical trainee participation has been identified as an independent predictor of negative outcomes after revascularization procedures for peripheral vascular disease. Support for the potential effect of intraoperative trainee participation on postoperative technical outcomes comes from a prospective, randomized trial of inguinal herniorrhaphy by Wilkiemeyer et al.17 In every procedure per-
formed in that trial, a surgical trainee was the operating surgeon and an attending was the assisting surgeon. The authors found that the risk of postoperative hernia recurrence was inversely related to the training level of the participating surgical resident. Results from a study of simulated vascular anastomoses performed by surgical residents provides further evidence that operator experience influences technical outcomes.18 Interestingly, our findings suggest that early infrainguinal bypass graft failure rates increase, instead of decrease, with increasing seniority of the participating surgical trainee. A major difference between our analysis and the aforementioned studies is that the procedures included in the NSQIP database are not controlled with respect to the level of intraoperative attending surgeon supervision. Conversely, the protocol in the Wilkiemeyer study required the level of attending supervision to be the same for every procedure, regardless of the training level of the participating resident. Similarly, in the study assessing the effect of trainee seniority on technical quality of simulated vascular anastomoses, each anastomosis was performed entirely by the trainee, with no active intervention from an instructor. In the real world, however, there is considerable variation in the level of supervision that an attending provides
JOURNAL OF VASCULAR SURGERY Volume 55, Number 3
intraoperatively and in the amount of the technical portions of the procedure performed by the attending rather than the trainee. Attending surgeons are much more likely to serve as the operating surgeon when performing a procedure with a junior trainee but as the assisting surgeon when the procedure is performed by a more advanced trainee. Within this traditional framework of graded responsibility, the actual overall experience level of the operating surgeon can then be expected to decrease as the training level of the participating resident or fellow increases.19 It is not necessarily surprising, therefore, that the technical complication rate of bypass grafting procedures that involve trainees increases as the seniority of the participating trainee increases. Our analysis has several important limitations: First, we cannot determine the exact cause or causes of the increased graft failure rate for procedures that involve surgical trainees. Although early graft failure is generally attributable to the intraoperative technical conduct of a bypass grafting procedure, the procedure involves several important steps, including vein harvesting, vessel handling, anastomosis, and wound closure, that cannot be adequately assessed retrospectively. Second, a number of potentially confounding variables were not included in our propensity matching procedure because they are not tracked by the NSQIP. Specific to infrainguinal bypass grafting procedures, we do not know the Society for Vascular Surgery runoff score or other potential indicators of procedure complexity. Although we did attempt to adjust for procedure complexity by including operative time and intraoperative transfusion requirement in the propensity-matching procedure, there may possibly be systematic differences in the burden of vascular disease or other factors among patients who undergo infrainguinal bypass grafting procedures at teaching vs nonteaching hospitals that could explain the increased incidence of graft failure in the former group. More generally, hospital-level variables, such as annual procedure volume, and patient-level variables, such as socioeconomic status, may affect postoperative outcomes but have not been included as confounders in our analysis because they are not recorded in the NSQIP database.20 Finally, our analysis may have miscategorized patients with early postoperative graft failure who declined further intervention. Although the number of such patients cannot be known, their disproportionate grouping into one or the other cohort might prevent an accurate comparison of graft failure rates for the two groups of our analysis. In light of the aforementioned limitations, we believe that prospective evaluation is needed to confirm the potential effect of intraoperative trainee participation on technical outcomes after infrainguinal bypass grafting. If such evaluation confirms our findings, efforts will be needed to minimize the potential effect of trainee participation on the quality of intraoperative surgical care provided to patients with peripheral vascular disease. Prohibiting trainees from participating in surgical procedures is not an option. Such participation is a critical component of
Scarborough et al 719
resident and fellow education and is essential for maintaining the ongoing quality and competency of the vascular surgery workforce.21 Efforts should focus on better preparing surgical trainees for patient operations without limiting their exposure to such operations. Foremost among the potential strategies for ensuring that trainees are well prepared are the earlier incorporation of skills laboratories into surgical training programs and the adaptation of technical proficiency in the skills laboratory as a prerequisite for trainee participation in the operating room.22-24 Although potentially costly, greater use of such adjuncts may help to ensure that the dual goals of patient safety and surgical training do not conflict. CONCLUSIONS The results of our analysis suggest that surgical trainee participation represents an independent risk factor for graft failure after infrainguinal bypass procedures. This finding should serve as sufficient impetus for a more detailed and prospective evaluation of the potential effect of trainee participation on patient outcomes. Although active participation in operative procedures will always remain a critical element of proper surgical training, earlier use of simulation laboratories in such training and the development of proficiency-based training criteria may help to minimize the potential effect of intraoperative trainee participation on postoperative technical outcomes. AUTHOR CONTRIBUTIONS Conception and design: JS, TP, CS Analysis and interpretation: JS, KB, MC, CS Data collection: JS Writing the article: JS, CS, MC, TP Critical revision of the article: JS, TP, CS, KB Final approval of the article: JS, TP, MC, KB, CS Statistical analysis: JS, KB Obtained funding: Not applicable Overall responsibility: JS REFERENCES 1. Best WR, Khuri SF, Phelan M, Hur K, Henderson WG, Demakis JG, et al. Identifying patient preoperative risk factors and postoperative adverse events in administrative databases: results from the Department of Veterans Affairs National Surgical Quality Improvement Program. J Am Coll Surg 2002;194:257-66. 2. Yaghoubian A, de Virgilio C, Lee SL. Appendicitis outcomes are better at resident teaching institutions: a multi-institutional analysis. Am J Surg 2010;200:810-3. 3. Lee SL, Shekherdimian S, Chiu VY. Comparison of pediatric appendicitis outcomes between teaching and nonteaching hospitals. J Pediatr Surg 2010;45:894-7. 4. Babineau TJ, Becker J, Gibbons G, Sentovich S, Hess D, Robertson S, et al. The “cost” of operative training for surgical residents. Arch Surg 2004;139:366-9. 5. Silber JH, Rosenbaum PR, Romano PS, Rosen AK, Wang Y, Teng Y, et al. Hospital teaching intensity, patient race, and surgical outcomes. Arch Surg 2009;144:113-20. 6. Itani KM, DePalma RG, Schifftner T, Sanders KM, Chang BK, Henderson WG, et al. Surgical resident supervision in the operating room and outcomes of care in Veterans Affairs hospitals. Am J Surg 2005;190: 725-31.
720 Lee
7. Hutter MM, Glasgow RE, Mulvihill SJ. Does the participation of a surgical trainee adversely impact patient outcomes? A study of major pancreatic resections in California. Surgery 2000;128:286-92. 8. Dimick JB, Cowan JA, Colletti LM, Upchurch GR. Hospital teaching status and outcomes of complex surgical procedures in the United States. Arch Surg 2009;139:137-41. 9. Duggirala AV, Chen FM, Gergen PJ. Postoperative adverse events in teaching and nonteaching hospitals. Fam Med 2009;36:508-13. 10. Vartak S, Ward MM, Vaughn TE. Do postoperative complications vary by hospital teaching status? Med Care 2008;46:25-32. 11. Khuri SF, Najjar SF, Daley J, Krasnicka B, Hossain M, Henderson WG, et al. Comparison of surgical outcomes between teaching and nonteaching hospitals in the Department of Veterans Affairs. Ann Surg 2001;234:370-82. 12. Wong K, Duncan T, Pearson A. Unsupervised laparoscopic appendicectomy by surgical trainees is safe and time-effective. Asian J Surg 2007; 30:161-6. 13. Bencini L, Bernini M, Martini F, Rossi M, Tommasi C, Miranda E, et al. Laparoscopic appendectomy performed by residents and experienced surgeons. JSLS 2009;13:391-7. 14. Hwang CS, Pagano CR, Wichterman KA, Dunnington GL, Alfrey EJ. Resident versus no resident: a single institutional study on operative complications, mortality, and cost. Surgery 2008;144:339-44. 15. Hemmila MR, Birkmeyer NJ, Arbabi S, Osborne NH, Wahl WL, Dimick JB. Introduction to propensity scores: a case study on the comparative effectiveness of laparoscopic versus open appendectomy. Arch Surg 2010;145:939-45.
JOURNAL OF VASCULAR SURGERY March 2012
16. Austin PC. A critical appraisal of propensity-score matching in the medical literature between 1996 and 2003. Statist Med 2008;27: 2037-49. 17. Wilkiemeyer M, Pappas TN, Giobbie-Hurder A, Itani KM, Jonasson O, Neumayer LA. Does resident post graduate year influence the outcomes of inguinal hernia repair? Ann Surg 2005;241:879-84. 18. Wilasrusmee C, Phromsopha N, Lertsitichai P, Kittur DS. A new vascular anastomosis model: relation between outcome and experience. Eur J Vasc Endovasc Surg 2007;33:208-13. 19. O’Shea JS. Individual and social concerns in American surgical education: paying patients, prepaid health insurance, Medicare and Medicaid. Acad Med 2010;85:854-62. 20. Bennett KM, Scarborough JE, Pappas TN, Kepler TB. Patient socioeconomic status is an independent predictor of operative mortality. Ann Surg 2010;252:552-7; Discussion:5578-558. 21. Welling RE. The importance of vascular surgery training in a general surgery practice. Curr Surg 2000;57:381-383. 22. Wilasrusmee C, Lertsithichai P, Kittur DS. Vascular anastomosis model: relation between competency in a laboratory-based model and surgical competency. Eur J Vasc Endovasc Surg 2007;34:405-10. 23. Chaer RA, Derubertis BG, Lin SC, Bush HL, Karwowski JK, Birk D, et al. Simulation improves resident performance in catheter-based intervention: results of a randomized, controlled study. Ann Surg 2006;244: 343-52. 24. Reed AB, Crafton C, Giglia JS, Hutto JD. Back to basics: use of fresh cadavers in vascular surgery training. Surgery 2009;146:757-62. Submitted Apr 6, 2011; accepted Jun 18, 2011.
INVITED COMMENTARY
Jason T. Lee, MD, Stanford, Calif We probably all remember the satisfying sound of the multiphasic Doppler signal immediately after our first successful lower extremity bypass as a trainee, particularly as the foot pinked up when the clamps were removed. Lower extremity revascularization and limb salvage remain cornerstones of vascular surgery education and require intense attention to technical skill, delicacy, and some amount of swiftness. Very few vascular operations or interventions have outcomes that are so black or white when they fail, and the tolerance of success is measured in millimeters and often related to the technical details of the exercise. Volume– outcome relationships are well established for leg bypass, also indicating the precision and technical efficiency required for optimal limb salvage. Dr Scarborough and colleagues from Duke University Medical Center report in this provocative article the negative effect that surgical trainee participation has on early postoperative graft failure when analyzing the National Surgical Quality Improvement Program (NSQIP) database. The finding that having surgical residents as part of the operative team essentially has a 29% higher chance of early failure brings to light several issues plaguing surgical educators and residency program directors. Fortunately, morbidity and mortality are not adversely affected in the propensity-matched data by trainee involvement. In this new era of duty hour restrictions, many attendings already feel the trainees are not operating enough. More and more surgical residents are relying on fellowships or junior faculty posi-
tions to be “finishing schools” to successfully transition into practice. Additional pressures on faculty to operate more, faster, and better also affect the level of involvement and autonomy the trainee has during all types of vascular procedures. Findings in this NSQIP analysis may therefore force us to more prospectively and accurately determine what the trainee-related causes are that worsen surgical outcomes. Although I obviously agree with the authors that the solution to this dilemma is not to prohibit trainees from participating in surgery, much more emphasis needs to be placed on skills acquisition and competency assessments as methods to prepare surgical trainees for ultimate success when they are in the operating room. Much like the putting green before a round of golf, the simulation center/skills laboratory needs to be a warm-up zone for surgical trainees. National resources need to be provided to allow program directors to safely and efficiently train the future vascular surgery workforce. Formalized and mentored skills milestones of various vascular technical skills need to be identified, validated, and tested to create a skills curriculum for all future vascular surgeons. Although one-on-one faculty-to-trainee teaching in the operating room can never be completely replaced, those valuable interactions can be made much more meaningful when basic technical skills and operative plans have already been deliberately practiced outside of the operating room environment.
The effect of intraoperative surgical trainee participation on technical outcomes after infrainguinal bypass grafting is currently unknown. The National Surgical Quality Improvement Program (NSQIP) database offers a unique opportunity to explore this effect due to its inclusion of prospectively verified information about intraoperative trainee participation and also 30-day postoperative graft outcomes. These characteristics of NSQIP provide it with distinct advantages compared with other large multicenter data sources that rely on hospital teaching status as a marker of intraoperative trainee involvement and generally have low sensitivity for the detection of procedure-specific complications.1 Most of the previously published studies that have examined the potential effect of surgical education on nonvascular surgical quality rely on these flawed data sources or include procedures from only one or a small number of hospitals.2-14 The objective of our analysis was therefore to use data from the NSQIP Participant User File to determine whether surgical trainee involvement during lower extremity revascularization procedures affects early postoperative graft outcome. From the Department of Surgery, Duke University Medical Center. Competition of interest: none. Reprint requests: John E. Scarborough, MD, Assistant Professor of Surgery, DUMC 2837, Durham, NC 27710 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a competition of interest. 0741-5214/$36.00 Copyright © 2012 by the Society for Vascular Surgery. doi:10.1016/j.jvs.2011.06.122
METHODS The NSQIP Participant User Files for 2005 through 2009 were used for this retrospective analysis, which included all patients with a primary Current Procedural Terminology code for any of the following lower extremity infrainguinal bypass procedures: 35556 (femoral-popliteal bypass with vein), 35558 (femoral-femoral bypass with vein), 35566 (femoral-distal bypass with vein), 35570 (distal-distal bypass with vein), 35571 (popliteal-distal bypass with vein), 35583 (femoral-popliteal bypass with in situ vein), 35585 (femoral-distal bypass with in situ vein), 35587 (popliteal-distal bypass with in situ vein), 35656 (femoral-popliteal bypass with prosthesis), 35661 (femoral-femoral bypass with prosthesis), 35666 (femoraldistal bypass with prosthesis), and 35671 (popliteal-distal bypass with prosthesis). The analysis excluded patients who underwent a previous operative procedure ⱕ30 days of the index procedure. The primary outcome measure for our analysis was early postoperative graft failure, which the NSQIP defines as the mechanical failure of a graft or prosthesis requiring reoperation, an interventional radiology procedure, or balloon angioplasty ⱕ30 days of the operation. Secondary outcome measures included overall postoperative morbidity, mortality, and postoperative hospital length of stay. Complications included in postoperative morbidity were any or all of the following: superficial surgical site infection, deep surgical site infection, organ/space surgical site infection, wound dehiscence, pneumonia, unplanned intubation, ventilator dependency for ⬎48 hours, pulmonary embolism, progressive renal insufficiency, acute renal 715
JOURNAL OF VASCULAR SURGERY March 2012
716 Scarborough et al
failure, stroke, coma ⬎24 hours, peripheral nerve injury, cardiac arrest requiring cardiopulmonary resuscitation, urinary tract infection, bleeding requiring transfusion of packed red blood cells, deep venous thrombosis, sepsis, or septic shock. The primary predictor variable for our analysis was presence or absence of surgical trainee participation during the primary operative procedure. For each procedure, NSQIP contains information on the highest postgraduate year (PGY) of the trainee who scrubs for the procedure. For the purpose of our analysis, we used this information to create a dichotomous categoric variable, with PGY ⱖ1 indicating the presence of a surgical trainee and PGY ⫽ 0 indicating the presence of only an attending surgeon during the procedure. Other primary predictor variables included prior revascularization or amputation procedure for peripheral vascular disease, rest pain or gangrene, patient age, sex, white race, body mass index, American Society of Anesthesiologists classification of ⱖ3, diabetes mellitus, tobacco use within the past year, ⬎2 drinks of ethanol daily ⱕ2 weeks before operation, dyspnea at rest or on exertion, partially or totally dependent functional status preoperatively, chronic obstructive pulmonary disease, congestive heart failure, hypertension requiring medical therapy, neurologic disorder (including any or all of impaired sensorium, prior coma, hemiplegia, history of transient ischemic attacks, stroke with or without neurologic deficit, tumor involving central nervous system, paraplegia, or quadriplegia), coronary artery disease (including any or all of myocardial infarction ⱕ6 months before the operation, prior percutaneous coronary intervention, prior cardiac surgery, or angina ⱕ1 month before the operation), renal disease (including acute renal failure present ⱕ24 hours before the procedure or need for dialysis ⱕ2 weeks before the operation, or both), open wound or infected wound preoperatively, use of steroids ⱕ30 days of operation for chronic medical condition, ⬎10% loss body weight in 6 months before operation, bleeding disorder (including chronic anticoagulation therapy other than aspirin that was not discontinued preoperatively), preoperative sepsis, and emergency posting of the index operation. Intraoperative predictor variables included the wound classification of the index operative incision, nature of bypass conduit used (prosthetic vs venous), tibial or peroneal target for distal anastomosis, need for intraoperative packed red blood cell transfusion, and operative time. Preoperative and intraoperative characteristics among the infrainguinal bypass grafting procedures included in NSQIP with and without trainee participation were compared using Mann-Whitney rank sum test for continuous variables and Pearson 2 tests for categoric variables. A nonparsimonious logistic regression analysis model was created to estimate the likelihood that a surgical trainee participated in the index infrainguinal bypass grafting procedure.15 Preoperative patient characteristics and intraoperative procedural characteristics, including operative time and intraoperative requirement, were included as predictor
variables to adjust for patient condition and complexity of the index procedure. A propensity score for resident participation (range, 0-1) was calculated for each patient using the logit coefficients for the predictors of trainee participation. The propensity scores were then used to create two evenly matched groups according to trainee participation using a calipermatching algorithm with a caliper distance of .005, with controls being used only once in the matching. The preoperative and intraoperative characteristics of the matched cohort of patients were compared using Wilcoxon signed rank sum tests for continuous variables and McNemar 2 tests for categoric variables.16 For the entire NSQIP sample, primary and secondary outcome measures between the “no trainee” and “trainee” groups were compared using 2 tests or Wilcoxon rank sum tests, where appropriate. For the cohorts of patients matched on propensity for trainee participation, comparison of the primary and secondary outcome measures for the matched pairs was performed using Wilcoxon signed rank sum tests or McNemar 2 tests, where appropriate. All statistical analyses were performed using Stata 11.0 software (StataCorp, College Station, Tex). RESULTS The Table reports the preoperative and intraoperative characteristics for infrainguinal bypass procedures performed without or with trainee participation intraoperatively for the entire sample of NSQIP procedures and for the matched cohort. Comparison of the two groups using the entire NSQIP sample of infrainguinal bypass grafting procedures shows significant and potentially important differences between the two groups. Most notably, those patients whose procedures involved surgical trainee participation had a higher incidence of prior revascularization or amputation surgery for peripheral vascular disease and were significantly more likely to present with rest pain or gangrene, or both, than patients whose procedures were performed by attending surgeons alone. Procedures performed with trainee participation also appeared to be more complex on average than procedures performed by attendings alone. Procedures with trainee participation took significantly longer, were more likely to require intraoperative packed red blood cell transfusion, and were more likely to involve distal anastomoses beyond the level of the tibioperoneal trunk. When the two groups of patients were matched by the propensity for intraoperative trainee participation, the differences in preoperative and intraoperative characteristics were no longer statistically or clinically different (Table 1). The 30-day graft failure rates were significantly higher in procedures involving surgical trainee participation for the entire NSQIP sample of infrainguinal bypass grafting procedures and also for the cohorts matched on propensity for resident participation (Fig 1). In the latter comparison, the relative risk of graft failure was 25% greater for procedures that included surgical trainees than for procedures
JOURNAL OF VASCULAR SURGERY Volume 55, Number 3
Scarborough et al 717
Table I. Preoperative characteristics of patients undergoing infrainguinal bypass grafting with and without trainee participation All infrainguinal bypass patients in NSQIP Variablea Patients Age, years Female White Body mass index, kg/m2 ASA classification ⱖ3 Diabetes mellitus Tobacco use Heavy ethanol use Dyspnea Dependent functional status On mechanical ventilation Chronic obstructive pulmonary disease Preoperative pneumonia Ascites Esophageal varices Congestive heart failure Coronary artery disease Hypertension Renal disease Neurologic disorder Disseminated cancer Preoperative infected wound Steroids Weight loss Bleeding disorder Preoperative transfusion Chemotherapy ⱕ30 days Radiotherapy ⱕ90 days Preoperative sepsis None SIRS Sepsis Septic shock Emergency procedure Prior revascularization or amputation for PVD Rest pain or gangrene Prosthetic graft Tibial or peroneal distal anastomosis Wound classification of incision Clean Clean-contaminated Contaminated Dirty/infected Operative time, min Intraoperative PRBC transfusion
No trainee
Trainee
5562 (37.8) 68.7 ⫾ 11.5 2068 (37.2) 4110 (80.9) 27.1 ⫾ 6.3 1110 (20.0) 2202 (39.6) 2361 (42.5) 340 (6.1) 1097 (19.7) 78 (1.4) 10 (0.2) 840 (15.1) 14 (0.3) 10 (0.2) 5 (0.1) 129 (2.3) 2126 (38.2) 4585 (82.4) 348 (6.3) 1119 (20.1) 26 (0.5) 1790 (32.2) 197 (3.5) 90 (1.6) 1185 (21.3) 9 (0.2) 22 (0.4) 10 (0.2)
9161 (62.2) 67.0 ⫾ 12.1 3252 (35.5) 6697 (77.9) 27.0 ⫾ 6.5 1716 (18.7) 3792 (41.4) 3955 (43.2) 529 (5.8) 1644 (18.0) 152 (1.7) 22 (0.2) 1168 (12.8) 29 (0.3) 22 (0.2) 12 (0.1) 237 (2.6) 3476 (37.9) 7645 (83.5) 671 (7.3) 1832 (20.0) 49 (0.5) 3167 (34.6) 380 (4.2) 151 (1.7) 2166 (23.6) 15 (0.2) 39 (0.4) 25 (0.3)
5306 (95.4) 185 (3.3) 65 (1.2) 6 (0.1) 318 (5.7) 2838 (51.0) 2387 (42.9) 2203 (39.6) 1793 (32.2)
8664 (94.6) 379 (4.1) 99 (1.1) 19 (0.2) 511 (5.6) 4066 (55.3) 4399 (48.0) 3388 (37.0) 3554 (38.8)
5110 (91.9) 8282 (90.4) 144 (2.6) 175 (1.9) 164 (3.0) 401 (4.4) 144 (2.6) 303 (3.3) 192.1 ⫾ 93.3 240.0 ⫾ 105.2 598 (10.8) 1783 (19.5)
P ⬍.0001 .04 ⬍.0001 .44 .07 .03 .39 .4 .007 .22 .45 ⬍.0001 .48 .45 .48 .31 .73 .11 .01 .86 .58 .003 .07 .89 .001 .98 .78 .26 .04
.72 ⬍.0001 ⬍.0001 .001 ⬍.0001 ⬍.0001
⬍.0001 ⬍.0001
Cohorts matched on propensity for trainee participation No trainee
Trainee
4617 68.1 ⫾ 11.5 1686 (36.5) 3712 (80.4) 27.1 ⫾ 6.3 900 (19.5) 1854 (40.2) 1990 (43.1) 280 (6.1) 894 (19.4) 67 (1.5) 9 (0.2) 671 (14.5) 11 (0.2) 10 (0.2) 5 (0.1) 109 (2.4) 1760 (38.1) 3849 (83.4) 298 (6.5) 928 (20.1) 23 (0.5) 1471 (31.9) 166 (3.6) 76 (1.7) 1001 (21.7) 8 (0.2) 19 (0.4) 9 (0.2)
4617 68.5 ⫾ 11.7 1708 (37.0) 3727 (80.7) 27.1 ⫾ 6.5 917 (19.9) 1855 (40.2) 1988 (43.1) 293 (6.4) 907 (19.5) 62 (1.3) 8 (0.2) 710 (15.4) 10 (0.2) 9 (0.2) 6 (0.1) 111 (2.4) 1770 (38.3) 3811 (82.5) 305 (6.6) 932 (20.2) 24 (0.5) 1509 (32.7) 172 (3.7) 74 (1.6) 985 (21.3) 6 (0.1) 19 (0.4) 9 (0.2)
4403 (95.4) 160 (3.5) 49 (1.1) 5 (0.1) 231 (5.0) 2404 (52.1) 2.029 (44.0) 1855 (40.2) 1514 (32.8)
4410 (95.5) 153 (3.3) 48 (1.0) 6 (0.1) 232 (5.0) 2411 (52.2) 1992 (43.1) 1858 (40.2) 1510 (32.7)
4243 (91.9) 102 (2.2) 147 (3.2) 125 (2.7) 199.8 ⫾ 88.7 542 (11.7)
4250 (92.1) 107 (2.3) 142 (3.1) 118 (2.6) 199.0 ⫾ 82.0 539 (11.7)
P .07 .65 .71 .5 .67 ⬎.99 .98 .60 .75 .72 ⬎.99 .27 ⬎.99 ⬎.99 ⬎.99 .95 .85 .31 .80 .94 ⬎.99 .41 .78 .94 .71 .79 ⬎.99 ⬎.99 .77
⬎.99 .90 .44 0.97 .95 .82
.88 .94
ASA, American Society of Anesthesiologists; NSQIP, National Surgical Quality Improvement Program; PRBC, packed red blood cells; PVD, peripheral vascular disease; SIRS, systemic inflammatory response syndrome. a Categoric data are presented as number (%), and continuous data as mean ⫾ standard deviation.
that were performed by attending surgeons alone. Conversely, overall morbidity and 30-day postoperative mortality were not significantly associated with intraoperative trainee participation for the matched cohort (Fig 2 and Fig 3). Postoperative hospital length of stay was significantly longer for procedures involving trainee participation for the entire NSQIP sample (mean, 5.6; median, 4; interquartile range [IQR], 3-7 days for the “no trainee” group vs mean, 6.6; median, 5; IQR, 3-7 days for the “trainee” group P ⬍
.0001) as well as for the matched cohort of patients (mean, 5.5; median, 4; IQR, 3-6 days for the “no trainee” group vs mean, 5.9; median, 4; IQR, 3-7 days, P ⫽ .14). Fig 4 shows the postoperative graft failure rate for the entire NSQIP sample, stratified by the PGY of the highestlevel participating surgical trainee. In general, graft failure rates had a direct association with PGY training level of the participating trainee and were highest in those procedures that involved a trainee in PGY ⱖ6 of training.
JOURNAL OF VASCULAR SURGERY March 2012
30-Day Postoperave Mortality Rate
718 Scarborough et al
3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% No Resident (n=5,562)
Resident (n=9,161)
All NSQIP Paents
No Resident
Resident
(n=4,617)
(n=4,617)
Matched Cohort
Fig 3. The 30-day day postoperative mortality rate is shown in patients undergoing infrainguinal bypass grafting with and without trainee participation. NSQIP, National Surgical Quality Improvement Program.
30-Day Overall Postoperave Morbidity Rate
Fig 1. The 30-day early graft failure rate is shown for patients undergoing infrainguinal bypass grafting with or without trainee participation. NSQIP, National Surgical Quality Improvement Program. 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% No Resident (n=5,562)
Resident (n=9,161) All NSQIP Paents
No Resident
Resident
(n=4,617)
(n=4,617)
Fig 4. Postoperative graft failure rate is shown for entire National Surgical Quality Improvement Program (NSQIP) sample stratified by highest postgraduate year (PGY) training level of the participating trainee. CI, Confidence interval.
Matched Cohort
Fig 2. Overall morbidity is shown for patients undergoing infrainguinal bypass grafting with and without trainee participation. NSQIP, National Surgical Quality Improvement Program.
DISCUSSION Our analysis of NSQIP data from 2005 through 2009 demonstrates that infrainguinal bypass grafting procedures performed with surgical trainees have a 52% higher unadjusted risk of 30-day postoperative graft failure than procedures performed by attending surgeons alone. When comparing a subset of patients that are well matched for patient comorbidities and procedure complexity, the association between trainee participation and increased technical complications rates remains significant. To our knowledge, this is the first time that intraoperative surgical trainee participation has been identified as an independent predictor of negative outcomes after revascularization procedures for peripheral vascular disease. Support for the potential effect of intraoperative trainee participation on postoperative technical outcomes comes from a prospective, randomized trial of inguinal herniorrhaphy by Wilkiemeyer et al.17 In every procedure per-
formed in that trial, a surgical trainee was the operating surgeon and an attending was the assisting surgeon. The authors found that the risk of postoperative hernia recurrence was inversely related to the training level of the participating surgical resident. Results from a study of simulated vascular anastomoses performed by surgical residents provides further evidence that operator experience influences technical outcomes.18 Interestingly, our findings suggest that early infrainguinal bypass graft failure rates increase, instead of decrease, with increasing seniority of the participating surgical trainee. A major difference between our analysis and the aforementioned studies is that the procedures included in the NSQIP database are not controlled with respect to the level of intraoperative attending surgeon supervision. Conversely, the protocol in the Wilkiemeyer study required the level of attending supervision to be the same for every procedure, regardless of the training level of the participating resident. Similarly, in the study assessing the effect of trainee seniority on technical quality of simulated vascular anastomoses, each anastomosis was performed entirely by the trainee, with no active intervention from an instructor. In the real world, however, there is considerable variation in the level of supervision that an attending provides
JOURNAL OF VASCULAR SURGERY Volume 55, Number 3
intraoperatively and in the amount of the technical portions of the procedure performed by the attending rather than the trainee. Attending surgeons are much more likely to serve as the operating surgeon when performing a procedure with a junior trainee but as the assisting surgeon when the procedure is performed by a more advanced trainee. Within this traditional framework of graded responsibility, the actual overall experience level of the operating surgeon can then be expected to decrease as the training level of the participating resident or fellow increases.19 It is not necessarily surprising, therefore, that the technical complication rate of bypass grafting procedures that involve trainees increases as the seniority of the participating trainee increases. Our analysis has several important limitations: First, we cannot determine the exact cause or causes of the increased graft failure rate for procedures that involve surgical trainees. Although early graft failure is generally attributable to the intraoperative technical conduct of a bypass grafting procedure, the procedure involves several important steps, including vein harvesting, vessel handling, anastomosis, and wound closure, that cannot be adequately assessed retrospectively. Second, a number of potentially confounding variables were not included in our propensity matching procedure because they are not tracked by the NSQIP. Specific to infrainguinal bypass grafting procedures, we do not know the Society for Vascular Surgery runoff score or other potential indicators of procedure complexity. Although we did attempt to adjust for procedure complexity by including operative time and intraoperative transfusion requirement in the propensity-matching procedure, there may possibly be systematic differences in the burden of vascular disease or other factors among patients who undergo infrainguinal bypass grafting procedures at teaching vs nonteaching hospitals that could explain the increased incidence of graft failure in the former group. More generally, hospital-level variables, such as annual procedure volume, and patient-level variables, such as socioeconomic status, may affect postoperative outcomes but have not been included as confounders in our analysis because they are not recorded in the NSQIP database.20 Finally, our analysis may have miscategorized patients with early postoperative graft failure who declined further intervention. Although the number of such patients cannot be known, their disproportionate grouping into one or the other cohort might prevent an accurate comparison of graft failure rates for the two groups of our analysis. In light of the aforementioned limitations, we believe that prospective evaluation is needed to confirm the potential effect of intraoperative trainee participation on technical outcomes after infrainguinal bypass grafting. If such evaluation confirms our findings, efforts will be needed to minimize the potential effect of trainee participation on the quality of intraoperative surgical care provided to patients with peripheral vascular disease. Prohibiting trainees from participating in surgical procedures is not an option. Such participation is a critical component of
Scarborough et al 719
resident and fellow education and is essential for maintaining the ongoing quality and competency of the vascular surgery workforce.21 Efforts should focus on better preparing surgical trainees for patient operations without limiting their exposure to such operations. Foremost among the potential strategies for ensuring that trainees are well prepared are the earlier incorporation of skills laboratories into surgical training programs and the adaptation of technical proficiency in the skills laboratory as a prerequisite for trainee participation in the operating room.22-24 Although potentially costly, greater use of such adjuncts may help to ensure that the dual goals of patient safety and surgical training do not conflict. CONCLUSIONS The results of our analysis suggest that surgical trainee participation represents an independent risk factor for graft failure after infrainguinal bypass procedures. This finding should serve as sufficient impetus for a more detailed and prospective evaluation of the potential effect of trainee participation on patient outcomes. Although active participation in operative procedures will always remain a critical element of proper surgical training, earlier use of simulation laboratories in such training and the development of proficiency-based training criteria may help to minimize the potential effect of intraoperative trainee participation on postoperative technical outcomes. AUTHOR CONTRIBUTIONS Conception and design: JS, TP, CS Analysis and interpretation: JS, KB, MC, CS Data collection: JS Writing the article: JS, CS, MC, TP Critical revision of the article: JS, TP, CS, KB Final approval of the article: JS, TP, MC, KB, CS Statistical analysis: JS, KB Obtained funding: Not applicable Overall responsibility: JS REFERENCES 1. Best WR, Khuri SF, Phelan M, Hur K, Henderson WG, Demakis JG, et al. Identifying patient preoperative risk factors and postoperative adverse events in administrative databases: results from the Department of Veterans Affairs National Surgical Quality Improvement Program. J Am Coll Surg 2002;194:257-66. 2. Yaghoubian A, de Virgilio C, Lee SL. Appendicitis outcomes are better at resident teaching institutions: a multi-institutional analysis. Am J Surg 2010;200:810-3. 3. Lee SL, Shekherdimian S, Chiu VY. Comparison of pediatric appendicitis outcomes between teaching and nonteaching hospitals. J Pediatr Surg 2010;45:894-7. 4. Babineau TJ, Becker J, Gibbons G, Sentovich S, Hess D, Robertson S, et al. The “cost” of operative training for surgical residents. Arch Surg 2004;139:366-9. 5. Silber JH, Rosenbaum PR, Romano PS, Rosen AK, Wang Y, Teng Y, et al. Hospital teaching intensity, patient race, and surgical outcomes. Arch Surg 2009;144:113-20. 6. Itani KM, DePalma RG, Schifftner T, Sanders KM, Chang BK, Henderson WG, et al. Surgical resident supervision in the operating room and outcomes of care in Veterans Affairs hospitals. Am J Surg 2005;190: 725-31.
720 Lee
7. Hutter MM, Glasgow RE, Mulvihill SJ. Does the participation of a surgical trainee adversely impact patient outcomes? A study of major pancreatic resections in California. Surgery 2000;128:286-92. 8. Dimick JB, Cowan JA, Colletti LM, Upchurch GR. Hospital teaching status and outcomes of complex surgical procedures in the United States. Arch Surg 2009;139:137-41. 9. Duggirala AV, Chen FM, Gergen PJ. Postoperative adverse events in teaching and nonteaching hospitals. Fam Med 2009;36:508-13. 10. Vartak S, Ward MM, Vaughn TE. Do postoperative complications vary by hospital teaching status? Med Care 2008;46:25-32. 11. Khuri SF, Najjar SF, Daley J, Krasnicka B, Hossain M, Henderson WG, et al. Comparison of surgical outcomes between teaching and nonteaching hospitals in the Department of Veterans Affairs. Ann Surg 2001;234:370-82. 12. Wong K, Duncan T, Pearson A. Unsupervised laparoscopic appendicectomy by surgical trainees is safe and time-effective. Asian J Surg 2007; 30:161-6. 13. Bencini L, Bernini M, Martini F, Rossi M, Tommasi C, Miranda E, et al. Laparoscopic appendectomy performed by residents and experienced surgeons. JSLS 2009;13:391-7. 14. Hwang CS, Pagano CR, Wichterman KA, Dunnington GL, Alfrey EJ. Resident versus no resident: a single institutional study on operative complications, mortality, and cost. Surgery 2008;144:339-44. 15. Hemmila MR, Birkmeyer NJ, Arbabi S, Osborne NH, Wahl WL, Dimick JB. Introduction to propensity scores: a case study on the comparative effectiveness of laparoscopic versus open appendectomy. Arch Surg 2010;145:939-45.
JOURNAL OF VASCULAR SURGERY March 2012
16. Austin PC. A critical appraisal of propensity-score matching in the medical literature between 1996 and 2003. Statist Med 2008;27: 2037-49. 17. Wilkiemeyer M, Pappas TN, Giobbie-Hurder A, Itani KM, Jonasson O, Neumayer LA. Does resident post graduate year influence the outcomes of inguinal hernia repair? Ann Surg 2005;241:879-84. 18. Wilasrusmee C, Phromsopha N, Lertsitichai P, Kittur DS. A new vascular anastomosis model: relation between outcome and experience. Eur J Vasc Endovasc Surg 2007;33:208-13. 19. O’Shea JS. Individual and social concerns in American surgical education: paying patients, prepaid health insurance, Medicare and Medicaid. Acad Med 2010;85:854-62. 20. Bennett KM, Scarborough JE, Pappas TN, Kepler TB. Patient socioeconomic status is an independent predictor of operative mortality. Ann Surg 2010;252:552-7; Discussion:5578-558. 21. Welling RE. The importance of vascular surgery training in a general surgery practice. Curr Surg 2000;57:381-383. 22. Wilasrusmee C, Lertsithichai P, Kittur DS. Vascular anastomosis model: relation between competency in a laboratory-based model and surgical competency. Eur J Vasc Endovasc Surg 2007;34:405-10. 23. Chaer RA, Derubertis BG, Lin SC, Bush HL, Karwowski JK, Birk D, et al. Simulation improves resident performance in catheter-based intervention: results of a randomized, controlled study. Ann Surg 2006;244: 343-52. 24. Reed AB, Crafton C, Giglia JS, Hutto JD. Back to basics: use of fresh cadavers in vascular surgery training. Surgery 2009;146:757-62. Submitted Apr 6, 2011; accepted Jun 18, 2011.
INVITED COMMENTARY
Jason T. Lee, MD, Stanford, Calif We probably all remember the satisfying sound of the multiphasic Doppler signal immediately after our first successful lower extremity bypass as a trainee, particularly as the foot pinked up when the clamps were removed. Lower extremity revascularization and limb salvage remain cornerstones of vascular surgery education and require intense attention to technical skill, delicacy, and some amount of swiftness. Very few vascular operations or interventions have outcomes that are so black or white when they fail, and the tolerance of success is measured in millimeters and often related to the technical details of the exercise. Volume– outcome relationships are well established for leg bypass, also indicating the precision and technical efficiency required for optimal limb salvage. Dr Scarborough and colleagues from Duke University Medical Center report in this provocative article the negative effect that surgical trainee participation has on early postoperative graft failure when analyzing the National Surgical Quality Improvement Program (NSQIP) database. The finding that having surgical residents as part of the operative team essentially has a 29% higher chance of early failure brings to light several issues plaguing surgical educators and residency program directors. Fortunately, morbidity and mortality are not adversely affected in the propensity-matched data by trainee involvement. In this new era of duty hour restrictions, many attendings already feel the trainees are not operating enough. More and more surgical residents are relying on fellowships or junior faculty posi-
tions to be “finishing schools” to successfully transition into practice. Additional pressures on faculty to operate more, faster, and better also affect the level of involvement and autonomy the trainee has during all types of vascular procedures. Findings in this NSQIP analysis may therefore force us to more prospectively and accurately determine what the trainee-related causes are that worsen surgical outcomes. Although I obviously agree with the authors that the solution to this dilemma is not to prohibit trainees from participating in surgery, much more emphasis needs to be placed on skills acquisition and competency assessments as methods to prepare surgical trainees for ultimate success when they are in the operating room. Much like the putting green before a round of golf, the simulation center/skills laboratory needs to be a warm-up zone for surgical trainees. National resources need to be provided to allow program directors to safely and efficiently train the future vascular surgery workforce. Formalized and mentored skills milestones of various vascular technical skills need to be identified, validated, and tested to create a skills curriculum for all future vascular surgeons. Although one-on-one faculty-to-trainee teaching in the operating room can never be completely replaced, those valuable interactions can be made much more meaningful when basic technical skills and operative plans have already been deliberately practiced outside of the operating room environment.