- Email: [email protected]
Influence of Specialty Training and Trainee Involvement on Perioperative Outcomes of Esophagectomy
Influence of Specialty Training and Trainee Involvement on Perioperative Outcomes of Esophagectomy Zeyad Khoushhal, MBBS, MPH, Joseph Canner, MHS, Eric Schneider, PhD, Miloslawa Stem, MS, Elliott Haut, MD, PhD, Benedetto Mungo, MD, Anne Lidor, MD, MPH, and Daniela Molena, MD Epidemiology and Biostatistics, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland; Department of Surgery, Taibah University School of Medicine, Madinah, Saudi Arabia; Johns Hopkins Surgery Center for Outcomes Research, Department of Surgery, Johns Hopkins University School of Medicine, Baltimore; Department of Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland; Department of Surgery, University of Wisconsin, Madison, Wisconsin; Division of Thoracic Surgery, Memorial Sloan Kettering Cancer Center, New York, New York
Background. Hospitals’ and surgeons’ volume-outcome relationship have been reported in several esophagectomy studies with an inverse association of mortality and volume. The purpose of our study was to evaluate the outcomes of esophagectomy in the United States relative to the surgeon’s specialty. Methods. This was a retrospective analysis using the American College of Surgeons National Surgical Quality Improvement Program database (2006 to 2013). All patients (18 years of age and older) who underwent esophagectomy were divided into 2 groups according to whether the operation was performed by a general surgeon (GS) or a cardiothoracic surgeon (CTS). A comparison of intraoperative and postoperative outcomes between the groups was conducted. The primary outcome was 30-day mortality. Secondary outcomes included overall and serious morbidity, discharge destination, and length of hospital stay. Results. Of the 5,142 esophagectomies identified, 70.3% were performed by GS and 29.7% by CTS. Overall, CTS patients had significantly higher comorbidities and
cancer rates (61% versus 53%). Both specialties preferred the transthoracic approach (59.41% for CTS versus 44.90% for GS). Trainee involvement was higher for CTS. There was no significant difference in mortality or overall morbidity. Patients operated on by GS had higher rates of wound infection, sepsis, shock, prolonged or unplanned intubation, and a longer hospital stay, whereas patients operated on by CTS had higher chance for bleeding and return to the operating room. Trainees’ involvement in esophagectomy was not associated with worse outcome. Conclusions. Our study showed that a large number of esophagectomies in the United States are performed by GS, with the transthoracic approach being the most popular among both specialties. Trainees’ involvement in esophagectomy did not significantly affect patients’ outcomes. However CTS specialty was associated with lower incidence of infection and a shorter hospital stay.
D
surgical oncologists, and gastrointestinal surgeons), all of which have different levels of training and expertise with the procedure. An inverse correlation between surgeons as well as hospitals volume and perioperative outcome has already been demonstrated, confirming the high skill required to perform the procedure and to care for the patient postoperatively. The purpose of our study was to evaluate the outcomes of esophagectomy in the United States relatively to the involvement of a GS or cardiothoracic surgeon (CTS) performing the procedure. Using different data sets, 4 other studies attempted to answer a similar question [10–13], and heterogeneity in
oes specialty training affect patients’ outcomes? A question few studies have tried to address [1–3]. Although we are witnessing increased trends in specialization, there is still a crossover of some operations performed by general surgeon (GS) and specialized surgeons. Several studies have been previously conducted to evaluate whether the result of an operation is influenced by the surgeon’s training [4–9]. Although esophagectomy is considered one of the most complex surgical procedure, this is performed by a variety of different surgeons (ie, GS, minimally invasive surgeons,
(Ann Thorac Surg 2016;-:-–-) Ó 2016 by The Society of Thoracic Surgeons
Accepted for publication June 13, 2016. Presented at the Fifty-second Annual Meeting of The Society of Thoracic Surgeons, Phoenix, AZ, Jan 23–27, 2016. Address correspondence to Dr Daniela Molena, Memorial Sloan Kettering Cancer Center, Thoracic Surgery Service, Department of Surgery, 1275 York Ave, New York, NY 10065; email: [email protected].
Ó 2016 by The Society of Thoracic Surgeons Published by Elsevier
Dr Haut discloses a financial relationship with Lippincott, Williams, Wilkins, and the Illinois Surgical Quality Improvement Collaborative.
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2016.06.025
2
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
these articles encouraged us to examine the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) records to analyze the 30-day outcomes of esophagectomy according to surgeon specialty and supplement the existing literature.
Material and Methods Data Source We used data from the ACS-NSQIP database according to a retrospective cohort study design between 2006 and 2013. ACS-NSQIP is a nationwide, risk adjusted, outcome-based database with a focus on quality improvement. The program collects data on over 150 variables, including preoperative risk factors, intraoperative variables, and 30-day postoperative mortality and morbidity for patients undergoing major surgical procedures in both the inpatient and outpatient setting [14]. Data gathering is done by a surgical clinical reviewer who is experienced in data extraction by completing the mandatory ACS-NSQIP training program and participating in ongoing training and conferences. Surgical specialty is a variable defined as the surgical specialty area that best characterize the principal operative procedure. This is either based on the surgeon’s self-declared specialty or the service line or specialty most closely related to the principal operative procedure if the surgeon is privileged to perform cases within multiple specialties or is Board Certified in multiple specialties. We combined cardiac surgery and thoracic surgery specialties in our analysis. The presence of a resident and the highest level of postgraduate year (PGY) is also a variable collected by NSQIP. The Institutional Review Board of the Johns Hopkins University School of Medicine approved our study.
Inclusion Criteria This study included all adult patients (18 years of age and older) who underwent esophagectomy during the period of time between 2006 and 2013. We used the following Current Procedural Terminology codes for esophagectomy: 43107, 43108, 43112, 43113, 43116, 43117, 43118, 43121, 43122, 43123, and 43124. Esophageal cancer was defined by the International Classification of DiseasesNinth Revision codes: 150, 150.1, 150.2, 150.3, 150.4, 150.5, 150.8, 150.9, 151, and 151.0. Mortality rate, clinical postoperative complications, and comorbidities are provided by NSQIP.
Baseline Characteristics of the Patients The patients were divided in 2 groups according to the operation being performed by a GS or a CTS. Clinical characteristics consisted of age, sex, and ethnicity of patients; American Society of Anesthesiology classification of patient physical condition; esophagectomy type; body mass index; emergency cases; trainee involvement; and preoperative comorbidities such as diabetes mellitus (with oral agents or insulin), current history of smoking (within 1 year before the operation), dyspnea, hypertension requiring medication, weight loss (<10% of body
Ann Thorac Surg 2016;-:-–-
weight in last 6 months), steroid use for chronic condition, history of chronic obstructive pulmonary disease, and congestive heart failure.
Outcomes Comparison of intraoperative and postoperative outcomes of esophagectomy performed by GS and CTS was conducted. Primary outcome for this study was 30-days mortality. Secondary outcomes included overall morbidity, readmission rate, hospital length of stay, and discharge destination. Overall morbidity was defined by presence of at least 1 of the following NSQIP complications: wound infection, pneumonia, urinary tract infection, return to operating room, venous thromboembolic events, cardiac complication, shock or sepsis, unplanned intubation, bleeding requiring transfusion, renal complication, ventilator dependency >48 hours, and organ or space surgical site infection.
Statistical Analysis Baseline characteristics and outcomes were compared between GS and CTS using Pearson’s Chi-square test for categorical variables and 2-sample t tests for continuous variables. Multivariable logistic regression model was implemented to predict the 30-day mortality, overall and serious morbidity, readmission, and discharge destination. Negative binomial regression model was used to estimate the hospital length of stay, comparing GS versus CTS. Adjustments of the models for clinically relevant and statistically significant patients’ characteristics were done. Some variables were adjusted for clinical relevance only, without being statistically significant. Statistical analyses were performed using Stata/MP version 13 (StataCorp LP, College Station, TX). Statistical significance was indicated by p < 0.05.
Results Patient Characteristics Of the total 5,142 esophagectomies recorded in NSQIP between 2006 and 2013, 70.3% were performed by GS while 29.7% were done by CTS (Table 1). The mean age of the patients was similar among the 2 groups. Patients treated by CTS had higher American Society of Anesthesiology scores, body mass index, and history of chronic obstructive pulmonary disease. Both GS and CTS most frequently used the transthoracic approach. The second preferable approach was transhiatal for GS and McKeown for CTS. The majority of patients underwent esophagectomy for cancer, however a larger number of patients with benign disease was treated by GS. CTS were also noted to have a significant higher percentage of trainees’ involvement when compared to GS (although this variable was collected only until 2011).
Unadjusted and Adjusted Outcomes Unadjusted outcomes for esophagectomies showed that both GS and CTS had similar mortality and overall morbidity. GS had higher rates of wound infections,
Ann Thorac Surg 2016;-:-–-
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
3
Table 1. Baseline Demographic and Clinical Characteristics Characteristic Age Male Race White Black Hispanic Other ASA classification No disturb/mild disturb Serious disturb Life threat/moribund Approach Transthoracic (Ivor Lewis) Transhiatal McKeown Intestinal conduit Other Body mass index Diabetes Current smoker (within 1 year) Dyspnea History of COPD History of CHF Hypertension Weight loss Steroid use Esophageal cancer Emergency case Trainee involvement
Total (N ¼ 5,142)
General Surgery (n ¼ 3,617 [70.34%])
Cardiothoracic Surgery (n ¼ 1,525, [29.66%])
63.15 11.05 4,056 (78.96)
63.25 11.14 2,833 (78.43)
62.90 10.84 1,223 (80.20)
4,215 187 404 103
2,923 150 253 93
1,292 37 151 10
(85.86) (3.81) (8.23) (2.10)
(85.49) (4.39) (7.40) (2.72)
p Value 0.308 0.157
(86.71) (2.48) (10.13) (0.67)
<0.001 <0.001 <0.001 <0.001
1,050 (20.44) 3,601 (70.10) 486 (9.46)
794 (21.98) 2,492 (68.97) 327 (9.05)
256 (16.80) 1,109 (72.77) 159 (10.43)
<0.001 <0.001 <0.001
2,530 1,398 844 266 104 27.87 842 1,321 575 386 11 2,641 970 159 2,854 97 2,180
1,624 1,153 591 159 54 27.71 600 914 404 251 7 1,851 690 109 1,924 70 1,774
906 245 253 71 50 28.26 242 407 171 135 4 790 280 50 930 24 406
<0.001 <0.001 <0.001 <0.001 <0.001 0.006 0.524 0.287 0.949 0.017 0.626 0.681 0.549 0.616 <0.001 0.692 0.049
(49.20) (27.19) (16.41) (4.66) (2.02) 6.47 (16.37) (25.69) (11.18) (7.51) (0.21) (51.36) (18.86) (3.09) (55.50) (1.89) (82.17)
(44.90) (31.88) (16.34) (5.39) (1.49) 6.26 (16.59) (25.27) (11.17) (6.94) (0.19) (51.18) (19.08) (3.01) (53.19) (1.94) (81.49)
(59.41) (16.07) (16.59) (4.66) (3.28) 6.94 (15.87) (26.69) (11.21) (8.85) (0.26) (51.80) (18.36) (3.28) (60.98) (1.77) (85.29)
Values are mean SD or n (%). ASA ¼ American Society of Anesthesiology; deviation.
CHF ¼ congestive heart failure;
sepsis, unplanned and prolonged intubation and longer hospital stay (Table 2). CTS had a higher number of patients requiring transfusions and return to the operating room (Table 2). Adjusted multivariable analysis revealed that GS had significantly increased odds of wound infection, urinary tract infection, sepsis, shock, unplanned and prolonged intubation, and longer hospital stay. CTS had significantly higher odds of bleeding and return to the operating room. Mortality, overall morbidity and readmission’s rate were comparable between GS and CTS (Table 3). Because trainees were more frequently involved in cases performed by CTS, we performed a subanalysis to assess trainees’ involvement as an independent variable influencing outcomes. We noted that trainees’ involvement was not associated with worse outcomes (Table 4).
Comment In this study we have compared perioperative and postoperative clinical outcomes of esophagectomy performed
COPD ¼ chronic obstructive pulmonary disease;
SD ¼ standard
either by GS or CTS. Mortality, overall morbidity, and readmission odds were comparable between GS and CTS. However the most common complications were different between the two groups. Although GS had higher odds of infections, shock, sepsis, prolonged intubation, and longer hospital stay, CTS had higher odds of bleeding and return to the operating room. There are 4 other published studies that performed a similar comparison between these two specialties regarding the outcomes of esophagectomy and they showed mixed results [10–13]. Dimick and colleagues [10] found that thoracic surgeons had better mortality odds but hospital and surgeon volume had a stronger influence than specialty on mortality. The authors however used administrative data for their analysis and preoperative comorbidities as well as postoperative complications are not recorded for quality improvement purposes but only for billing and therefore are not as detailed as in our study. Interestingly thoracic surgeons were noted to be higher volume surgeons, working at higher volume hospitals when compared to GS. This is
4
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
Ann Thorac Surg 2016;-:-–-
Table 2. Observed Rates of Unadjusted Intraoperative and Postoperative Outcomes Outcome
Total
30-day mortality Overall morbiditya Wound infection (superficial þ deep SSI) Pneumonia Urinary tract infection Return to operating room Venous thromboembolism Cardiac complications Shock/sepsis Unplanned intubation Bleeding transfusion (surgical start time up to and including 72 hours) Renal complications (progressive þ acute renal failure) On ventilator >48 hours Organ space SSI Hospital length of stay, days Discharge destination (home versus facility) Readmission (%)
158 2,569 526 819 158 755 218 146 831 720 765
General Surgery
(3.07) (49.96) (10.23) (15.93) (3.07) (14.68) (4.24) (2.84) (16.16) (14.0) (14.88)
121 (2.35) 777 (15.11) 365 (7.10) 15.55 17.07 2,421 (84.56) 85 (12.71)
109 1,792 414 16.09 125 481 162 109 639 530 425
(3.01) (49.54) (11.45) (16.09) (3.46) (13.30) (4.48) (3.01) (17.67) (14.65) (11.75)
82 (2.27) 585 (16.17) 241 (6.66) 15.98 18.77 1,345 (83.70) 46 (12.11)
Cardiothoracic Surgery 49 777 112 237 33 274 56 37 192 190 340
p Value
(3.21) (50.95) (7.34) (15.54) (2.16) (17.97) (3.67) (2.43) (12.59) (12.46) (22.30)
0.705 0.357 <0.001 0.623 0.014 <0.001 0.190 0.247 <0.001 0.038 <0.001
39 (2.56) 192 (12.59) 124 (8.13) 14.53 12.03 1,076 (85.67) 39 (13.49)
0.530 0.001 0.061 0.006 0.147 0.593
a Overall morbidity: wound infection, pneumonia, urinary tract infection, venous thromboembolism, bleeding transfusion, renal complications, return to operating room, cardiac complications, shock/sepsis, unplanned intubation, on ventilator >48 hours, and organ space surgical site infection (SSI).
Values are n (%) or mean SD. SD ¼ standard deviation.
an interesting finding to complement our study because NSQIP does not provide information about hospital or surgeon volume. Similar results were reported by Gopaldas and colleagues [11] using a Nationwide
Inpatient Sample. They showed reduction of mortality odds for CTS. Interestingly although complication rates were similar among the two groups, failure to rescue, defined as death from a complication [11, 15, 16], seemed
Table 3. Adjusted Multivariable Analysis of Outcomes by Surgical Specialty, Comparing General Versus Cardiothoracic Surgeons (Reference Group) Outcome 30-day mortality Overall morbidity Wound infection (superficial) Wound infection (deep) Pneumonia Urinary tract infection Return to operating room Venous thromboembolism Cardiac complications Sepsis Shock Unplanned intubation Bleeding transfusion (surgical start time up to and including 72 hours) Renal complications (progressive þ acute renal failure) On ventilator >48 hours Organ space surgical site infection Hospital length of stay, daysb Discharge destination (home versus facility) Readmission a
Adjusted ORa
p Value
95% CI
0.997 0.96 1.51 1.40 1.07 1.54 0.75 1.28 1.22 1.48 1.54 1.28 0.46 0.87 1.42 0.80 1.12 0.83 0.88
0.989 0.487 0.002 0.144 0.459 0.036 0.001 0.131 0.328 0.001 0.002 0.011 <0.001 0.494 <0.001 0.071 <0.001 0.114 0.603
0.69–1.44 0.84–1.09 1.17–1.96 0.89–2.21 0.90–1.27 1.03–2.29 0.63–0.89 0.93–1.77 0.82–1.82 1.18–1.85 1.17–2.02 1.06–1.55 0.39–0.54 0.58–1.31 1.17–1.71 0.63–1.02 1.07–1.16 0.66–1.05 0.54–1.43
Adjusted for age, sex, race, American Society of Anesthesiology classification, body mass index, diabetes, hypertension smoking status, dyspnea, history of chronic obstructive pulmonary disease, history of congestive heart failure, weight loss, diagnosis of cancer steroid therapy, emergency case, operation b Negative binomial regression model was used for this outcome and the incidence rate ratio was calculated. approach.
CI ¼ confidence interval;
OR ¼ odds ratio;
SD ¼ standard deviation.
Ann Thorac Surg 2016;-:-–-
5
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
Table 4. Adjusted Multivariable Analysis of Outcomes by Trainee Involvement, Comparing Trainee Involvement Versus No Trainee Involvement (Reference Group) Adjusted ORa (95%CI)
Outcome 30-day mortality (%) Overall morbidity (%) Hospital length of stay, daysb Discharge destination (home versus facility) Readmission (%)
0.84 1.19 0.96 1.01 1.33
(0.49–1.42) (0.96–1.47) (0.89–1.03) (0.40–2.52) (0.40–4.37)
p Value 0.511 0.120 0.215 0.983 0.643
a
Adjusted for age, sex, race, American Society of Anesthesiology classification, body mass index, diabetes, hypertension smoking status, dyspnea, history of chronic obstructive pulmonary disease, history of congestive heart failure, weight loss, steroid therapy, emergency case, approach, surgeon b specialty. Negative binomial regression model was used for this outcome and the incidence rate ratio was calculated.
CI ¼ confidence interval;
OR ¼ odds ratio.
to have been the cause of higher mortality in the GS group. Smith and colleagues [12] on the contrary concluded that specialty training did not affect outcomes of esophagectomy but only unadjusted clinical outcomes were reported in this study. Finally, Leigh and colleagues [13] compared mortality rates after esophagectomy between GS and CTS, and between high-volume and low-volume centers in England. They have reported higher mortality rates when the operation was performed by GS, especially if in low-volume centers. CTS surgeons had overall better outcomes than GS. Volume-outcome relationship of hospitals and surgeons were reported in several esophagectomy studies with a consistent inverse association between mortality and volume [17–19]. Nevertheless, a study conducted to identify outcome differences across high-volume hospitals (according to the leapfrog volume criteria) still found considerable variability in outcomes between these centers, suggesting that other factors rather than volume alone may play an important role in influencing outcomes [20]. Although there seem to be an association between specialty and volume in the fact that specialized surgeons usually work at higher volume hospital and are higher volume surgeons, nevertheless surgical specialty seems to be an independent factor influencing outcome consistently in several studies. Previous studies however have used administrative datasets to validate their hypothesis, with obvious biases since the data were collected for reasons different than measuring outcomes. The use of the ACSNSQIP offers unique advantages because this nationally validated surgical database was designed for evaluation of surgical outcomes and enables comparison of a large number of esophageal resections across multiple types of hospital in the United States. For this reason we believe our study adds additional information to what was previously published. Trainee involvement was more frequent in cases performed by CTS. We decided to consider the presence or absence of trainees as a variable rather than differentiating between the years of training. In fact, although NSQIP provides the PGY level of residency, the length of every residency program is different and it would have been impossible to assign level of expertise (ie, resident
versus fellow) just by looking at the PGY level. Interestingly, the presence of trainees was not associated with worse mortality and all the outcomes considered in our study were similar when comparing cases done with trainees’ involvement or not. These findings confirmed previously published data from the United Kingdom regarding esophageal surgery in particular and many other studies that have looked at trainee’s involvement with cardiac and noncardiac surgery [21–23]. It is comforting that we can continue to teach complex surgeries to young trainees without compromising results. Our study has several limitations. ACS-NSQIP data does not provide the center or surgeon volume variables and therefore it is not possible to adjust for volume to see if specialization remains independently associated to outcomes. Specialization is partly defined accordingly to Board Certification and does not provide additional information regarding fellowship-trained surgeons without a corresponding board (ie, surgical oncologists or gastrointestinal surgeons) and this limits the ability to differentiate among specific surgical skills. Moreover, often an esophagectomy is done by a GS together with a CTS. We believe that in these cases the CTS specialty is assigned to the case by the coder but we are not sure if this was consistently done or done at all. This database only reports short-term outcomes with 30 postoperative days follow-up, therefore complications and mortality happening later in time are not recorded. Also long-term outcomes and survival cannot be examined. Details about the complexity of the cases during the operation are not given in NSQIP (ie, stage of the cancer, previous surgery), another variable that might confound the relationship between specialty training and clinical outcomes. The small number of esophagectomies performed by CTS in this study might also be biased by the fact that thoracic surgery cases may not be routinely collected in NSQIP, especially in academic centers where cardiothoracic surgery is often an independent department from surgery. Finally, NSQIP includes cases with a sampling system and not all cases performed in a particular hospital. Systematic sampling process, interrater reliability audits and personnel training were introduced to make sure data collection is reliable and all cases done at one particular hospital have the same chance of being selected. Although there are certainly limitations, the
6
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
results of esophagectomies from a larger number of hospitals and practices that are not strictly university affiliated can be evaluated using a database that was specifically designed to gather information about surgical outcomes. It is interesting to notice that cancer patients and patients with greater comorbidities are more often treated by specialty surgeons. This is probably reflecting the understanding that fellowship training in fact provides exposure to cutting edge technologies, focused learning and large experience with complicated surgical techniques and procedures. Other studies have shown that specialization is often linked to higher volume of surgeons and hospitals and this should be kept in mind when discussing plans for regionalization of care. There is no question that esophagectomy is a complicated procedure and postoperative care also requires higher than average skills to prevent complication or rescue the patient when complications occur. Higher specialization should be considered an important factor for patients needing an esophagectomy. In conclusion our study shows that GS in the United States perform a large number of esophagectomies. Specialization may offer the surgeon specific skills and focused experience that lead to improved perioperative outcomes for complex surgeries such as esophagectomy. Trainee involvement does not impair the outcome of esophagectomy and therefore teaching residents complex procedures should be encouraged.
References 1. Sharp LK, Bashook PG, Lipsky MS, Horowitz SD, Miller SH. Specialty board certification and clinical outcomes: the missing link. Acad Med 2002;77:534–42. 2. Grosch EN. Does specialty board certification influence clinical outcomes? J Eval Clin Pract 2006;12:473–81. 3. Norcini JJ, Kimball HR, Lipner RS. Certification and specialization: do they matter in the outcome of acute myocardial infarction? Acad Med 2000;75:1193–8. 4. da Silva PS, de Aguiar VE, Waisberg J. Pediatric surgeon vs general surgeon: does subspecialty training affect the outcome of appendicitis? Pediatr Int 2014;56:248–53. 5. Etzioni DA, Young-Fadok TM, Cima RR, et al. Patient survival after surgical treatment of rectal cancer: Impact of surgeon and hospital characteristics. Cancer 2014;120: 2472–81. 6. Tieu B, Schipper P. Specialty matters in the treatment of lung cancer. Semin Thorac Cardiovasc Surg 2012;24:99–105. 7. von Meyenfeldt EM, Gooiker GA, van Gijn W, et al. The relationship between volume or surgeon specialty and
Ann Thorac Surg 2016;-:-–-
8.
9.
10. 11.
12. 13. 14. 15. 16. 17. 18.
19. 20.
21.
22.
23.
outcome in the surgical treatment of lung cancer: a systematic review and meta-analysis. J Thorac Oncol 2012;7:1170–8. Schipper PH, Diggs BS, Ungerleider RM, Welke KF. The influence of surgeon specialty on outcomes in general thoracic surgery: a national sample 1996 to 2005. Ann Thorac Surg 2009;88:1566–72, discussion 1572–63. Pearce WH, Parker MA, Feinglass J, Ujiki M, Manheim LM. The importance of surgeon volume and training in outcomes for vascular surgical procedures. J Vasc Surg 1999;29:768–76, discussion 777–8. Dimick JB, Goodney PP, Orringer MB, Birkmeyer JD. Specialty training and mortality after esophageal cancer resection. Ann Thorac Surg 2005;80:282–6. Gopaldas RR, Bhamidipati CM, Dao TK, Markley JG. Impact of surgeon demographics and technique on outcomes after esophageal resections: a nationwide study. Ann Thorac Surg 2013;95:1064–9. Smith BR, Hinojosa MW, Reavis KM, Nguyen NT. Outcomes of esophagectomy according to surgeon’s training: general vs. thoracic. J Gastrointest Surg 2008;12:1907–11. Leigh Y, Goldacre M, McCulloch P. Surgical specialty, surgical unit volume and mortality after oesophageal cancer surgery. Eur J Surg Oncol 2009;35:820–5. https://www.facs.org/quality-programs/acs-nsqip/programspecifics/participant-use. Accessed January 27, 2016. Ghaferi AA, Birkmeyer JD, Dimick JB. Hospital volume and failure to rescue with high-risk surgery. Med Care 2011;49: 1076–81. Ghaferi AA, Birkmeyer JD, Dimick JB. Complications, failure to rescue, and mortality with major inpatient surgery in medicare patients. Ann Surg 2009;250:1029–34. Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the united states. N Engl J Med 2002;346:1128–37. Markar SR, Karthikesalingam A, Thrumurthy S, Low DE. Volume-outcome relationship in surgery for esophageal malignancy: systematic review and meta-analysis 2000–2011. J Gastrointest Surg 2012;16:1055–63. Dimick JB, Wainess RM, Upchurch GR Jr, Iannettoni MD, Orringer MB. National trends in outcomes for esophageal resection. Ann Thorac Surg 2005;79:212–6, discussion 217–8. Varghese TK Jr, Wood DE, Farjah F, et al. Variation in esophagectomy outcomes in hospitals meeting leapfrog volume outcome standards. Ann Thorac Surg 2011;91:1003–9, discussion 1009–10. Handagala SD, Addae-Boateng E, Beggs D, Duffy JP, MartinUcar AE. Early outcomes of surgery for oesophageal cancer in a thoracic regional unit. Can we maintain training without compromising results? Eur J Cardiothorac Surg 2012;41:31–4, discussion 34–5. Virk SA, Bowman SR, Chan L, et al. Equivalent outcomes after coronary artery bypass graft surgery performed by consultant versus trainee surgeons: a systematic review and meta-analysis. J Thorac Cardiovasc Surg 2016;151:647–54.e641. Ejaz A, Spolverato G, Kim Y, et al. The impact of resident involvement on surgical outcomes among patients undergoing hepatic and pancreatic resections. Surgery 2015;158: 323–30.
DISCUSSION DR PAUL SCHIPPER (Portland, OR): I have two comments and then a question. My comment. Depending on your practice situation, it is actually entirely possible for general surgeons and thoracic surgeons to work together, and as long as you are willing to subjugate your ego to outcomes, and pay attention to how you are doing, you can have outcomes that are maybe better than what you would each achieve individually.
My second comment. I think the solid message out of this paper is the resident involvement, and that is what I want to ask questions about. You are showing there was no difference in outcomes whether residents were involved or not. Do you know the nature of the residents who were involved, meaning were they general surgery residents or thoracic residents? Two-thirds of these faculty were general surgeons and one-third were cardiothoracic surgeons. Do you know if there was any
Ann Thorac Surg 2016;-:-–-
cross-training that was happening, general surgery residents working with cardiothoracic teaching faculty or cardiothoracic residents working with general surgery teaching faculty? DR KHOUSHHAL: Those are actually very good questions. Unfortunately, the NSQIP database does not specify the status of the residents, whether they are general surgery residents or cardiothoracic residents. They only give you the variable with the level of the resident and whether the operation was involving residents or not. That is it. DR ANTOON E. M. R. LERUT (Leuven, Belgium): I have a comment. I think we should stop comparing cardiothoracic surgeons versus general surgeons. I think what we really need to compare is thoracic surgeons versus upper GI surgeons because that is really where the experts are and where the question comes up if groups are working together, eg, in the Scandinavian countries the abdominal part is done by the upper GI surgeons or the esophagus, cardiac, and stomach surgeons while the thoracic part is done by the thoracic surgeons. Another comment is that 30-day mortality should no longer be used as a standard. It is at least 90-day, and for me, preferably 6month mortality rates. DR KHOUSHHAL: Thanks a lot. DR ROSS M. BREMNER (Phoenix, AZ): Dr Lerut, do you have any sense of the proportion of esophagectomies that are done in Europe by foregut general surgeons versus thoracic surgeons, or is the training such that halfway through, you start moving into that organ system as a surgeon? DR LERUT: In Europe, as a whole, I think the vast majority of esophagectomies are done by visceral surgeons, where we see a split in to upper GI, hepatobiliary, and colorectal surgery. There are some countries, like in the Scandinavian countries, where there is a collaboration between upper GI surgeons and thoracic surgeons. There are a few (eg, in France, the UK) and like us, of course, who are typically in line with the North American model where thoracic surgeons are doing the entire spectrum of the esophagus. DR SHANDA H. BLACKMON (Rochester, MN): How did you classify those cases where they were collaborating? I know at UAB all of the esophagectomies are performed half by a general surgeon who does the abdominal part, and then the thoracic part is performed by the thoracic surgeon. I think it is also that way at Emory and some other cases. How did you code those cases? And then just a note. If you have PGY-6 or above, they are likely a thoracic resident and below they are likely a general surgery resident. So that would be the way that you could delineate those two categories. DR KHOUSHHAL: What I understand from your question, I think the way that NSQIP classifies them, they just designate the surgical specialty for each surgeon doing the operation, and if there was a crossover, like if the surgeon had a board certification in general surgery and in cardiothoracic surgery, the data collector would ask the surgeon, “How do you want to designate this specific operation?” DR DANIELA MOLENA (New York, NY): And the NSQIP data is a sample, so the experts who sample the data will look at who did the operation. Of course, it is hard to know for sure because we have no control over the coder, but if there was a thoracic surgeon involved, then the designation would be thoracic
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
7
surgeon. You cannot really tell in NSQIP if there is a collaboration between two surgeons doing the operation and their specialties, but if there is a thoracic surgeon in the operation, then you would be coded as a thoracic surgeon. As regards to the PGY level, it is a little bit simplistic to say if they are 6 or above they are cardiothoracic residents, because some people do 8 or 10 years of training, especially if they go into a PhD or master program during their research time. We decided that since we really do not know the level of training, it would have been more appropriate to just look at overall involvement of residents or not. DR LUIS G. CAMERO (Detroit, MI): I want to echo the comment on ego. I do my surgeries with a general surgeon who does at least 10 laparoscopic cases a day versus my occasional laparoscopic case. This guy flies through it. In an abdominal case, the patient comes the same day of surgery, and by 10:00, the gastric mobilization, the feeding jejunostomy, and the myotomy are done, with an excellent lymph node dissection, no problem. Then I do the thoracic surgery. I do an open thoracotomy in order to obtain the best paraesophageal lymph node dissection, with an average of 20 to 30 lymph nodes, with outstanding results, no leaks. We cover the gastrostomy part with the pleura. At the dissection of the esophagus, I create a flap with the pleura. By the end of the anastomosis, I flip the gastric component of the suture line and bury it with the pleura, so there is no suture line exposed within the chest. We have zero leaks. I manage the patients postop. I am not admitting them anymore to the ICU because they are walking in the afternoon of their case. Success does not come easy, and I think that involving the general surgeon in all of the cases not only gives you a selectivity that you are involved, but they are involved. Second, you have the best postoperative care for all the patients, with the best outcomes, and more goes around in town and people keep coming to you because the results are outstanding. DR KHOUSHHAL: Thanks. DR GAIL E. DARLING (Toronto, Canada): I would echo some of the comments from the previous speakers. I think what we should be really interested in is the quality outcomes and not necessarily who is doing the surgery. Although I am a thoracic surgeon and I think we should be doing the esophagectomies, I cannot really argue that I am doing a better job than a foregut surgeon who is fully trained and does lots of esophagectomies. I really think in future studies we should be focusing on the quality of the surgery that is done and we should be developing quality indicators, whether it is the number of lymph node nodes or whatever, and we should be looking at the volume of surgery. If you are a cardiothoracic surgeon and you do 2 esophagectomies a year, I do not think that is as good as a foregut surgeon who is doing 30 esophagectomies a year. I think the volume of surgery of the individual surgeon and the type of training received is more important than the name of the training program. DR KHOUSHHAL: A couple of papers actually reported that the volume is the most important factor in this comparison. DR STEVEN R. DEMEESTER (Los Angeles, CA): I am going philosophical again. I am old school. I am board certified in general surgery. I know the abdomen as well as a general surgeon. I am board certified in thoracic surgery. I can do the thoracic component. I think that one surgeon who does the entire operation is always going to be better than a team approach. That is just my philosophy.
8
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
Now, here is the question for you. In an era where we are moving toward shorter residencies, I-6, people are not going to be board certified in general surgery. What are the limitations of that system moving forward? Are we going to be forced to have dual surgeons trying to take on esophagectomies?
Ann Thorac Surg 2016;-:-–-
with a more specialized approach, a thoracic surgeon will be able to do both. I think that it is important to be there for the abdominal portion of the procedure because the conduit is the most important part of the reconstruction, and you want to know that the conduit is a safe conduit when you reconstruct the esophagus. I think this is a very important point.
DR KHOUSHHAL: I am sorry, can you repeat the question? DR MOLENA: I want to compliment Zeyad. He is an MPH student, so he might not really know all the details. He has done a great job with this paper. (Applause.) As Dr Darling was saying, I think it depends on the expertise. As thoracic surgeons, we learn how to do the operation in the abdomen. We do not need the general surgery specialization. I think that moving forward, as the teaching programs develop
DR LERUT: I would like to add something to that. It is probably not the thoracic surgeons who have to work with the general surgeons. It is just vice versa. We should convince them that if they want to do that kind of surgery—and it depends on countries, of course, which is what I tried to illustrate—that they should come to the thoracic surgeons and work as a team with thoracic surgeons to get the best quality. Usually it is not the other way around. I am a general thoracic surgeon.
Background. Hospitals’ and surgeons’ volume-outcome relationship have been reported in several esophagectomy studies with an inverse association of mortality and volume. The purpose of our study was to evaluate the outcomes of esophagectomy in the United States relative to the surgeon’s specialty. Methods. This was a retrospective analysis using the American College of Surgeons National Surgical Quality Improvement Program database (2006 to 2013). All patients (18 years of age and older) who underwent esophagectomy were divided into 2 groups according to whether the operation was performed by a general surgeon (GS) or a cardiothoracic surgeon (CTS). A comparison of intraoperative and postoperative outcomes between the groups was conducted. The primary outcome was 30-day mortality. Secondary outcomes included overall and serious morbidity, discharge destination, and length of hospital stay. Results. Of the 5,142 esophagectomies identified, 70.3% were performed by GS and 29.7% by CTS. Overall, CTS patients had significantly higher comorbidities and
cancer rates (61% versus 53%). Both specialties preferred the transthoracic approach (59.41% for CTS versus 44.90% for GS). Trainee involvement was higher for CTS. There was no significant difference in mortality or overall morbidity. Patients operated on by GS had higher rates of wound infection, sepsis, shock, prolonged or unplanned intubation, and a longer hospital stay, whereas patients operated on by CTS had higher chance for bleeding and return to the operating room. Trainees’ involvement in esophagectomy was not associated with worse outcome. Conclusions. Our study showed that a large number of esophagectomies in the United States are performed by GS, with the transthoracic approach being the most popular among both specialties. Trainees’ involvement in esophagectomy did not significantly affect patients’ outcomes. However CTS specialty was associated with lower incidence of infection and a shorter hospital stay.
D
surgical oncologists, and gastrointestinal surgeons), all of which have different levels of training and expertise with the procedure. An inverse correlation between surgeons as well as hospitals volume and perioperative outcome has already been demonstrated, confirming the high skill required to perform the procedure and to care for the patient postoperatively. The purpose of our study was to evaluate the outcomes of esophagectomy in the United States relatively to the involvement of a GS or cardiothoracic surgeon (CTS) performing the procedure. Using different data sets, 4 other studies attempted to answer a similar question [10–13], and heterogeneity in
oes specialty training affect patients’ outcomes? A question few studies have tried to address [1–3]. Although we are witnessing increased trends in specialization, there is still a crossover of some operations performed by general surgeon (GS) and specialized surgeons. Several studies have been previously conducted to evaluate whether the result of an operation is influenced by the surgeon’s training [4–9]. Although esophagectomy is considered one of the most complex surgical procedure, this is performed by a variety of different surgeons (ie, GS, minimally invasive surgeons,
(Ann Thorac Surg 2016;-:-–-) Ó 2016 by The Society of Thoracic Surgeons
Accepted for publication June 13, 2016. Presented at the Fifty-second Annual Meeting of The Society of Thoracic Surgeons, Phoenix, AZ, Jan 23–27, 2016. Address correspondence to Dr Daniela Molena, Memorial Sloan Kettering Cancer Center, Thoracic Surgery Service, Department of Surgery, 1275 York Ave, New York, NY 10065; email: [email protected].
Ó 2016 by The Society of Thoracic Surgeons Published by Elsevier
Dr Haut discloses a financial relationship with Lippincott, Williams, Wilkins, and the Illinois Surgical Quality Improvement Collaborative.
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2016.06.025
2
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
these articles encouraged us to examine the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) records to analyze the 30-day outcomes of esophagectomy according to surgeon specialty and supplement the existing literature.
Material and Methods Data Source We used data from the ACS-NSQIP database according to a retrospective cohort study design between 2006 and 2013. ACS-NSQIP is a nationwide, risk adjusted, outcome-based database with a focus on quality improvement. The program collects data on over 150 variables, including preoperative risk factors, intraoperative variables, and 30-day postoperative mortality and morbidity for patients undergoing major surgical procedures in both the inpatient and outpatient setting [14]. Data gathering is done by a surgical clinical reviewer who is experienced in data extraction by completing the mandatory ACS-NSQIP training program and participating in ongoing training and conferences. Surgical specialty is a variable defined as the surgical specialty area that best characterize the principal operative procedure. This is either based on the surgeon’s self-declared specialty or the service line or specialty most closely related to the principal operative procedure if the surgeon is privileged to perform cases within multiple specialties or is Board Certified in multiple specialties. We combined cardiac surgery and thoracic surgery specialties in our analysis. The presence of a resident and the highest level of postgraduate year (PGY) is also a variable collected by NSQIP. The Institutional Review Board of the Johns Hopkins University School of Medicine approved our study.
Inclusion Criteria This study included all adult patients (18 years of age and older) who underwent esophagectomy during the period of time between 2006 and 2013. We used the following Current Procedural Terminology codes for esophagectomy: 43107, 43108, 43112, 43113, 43116, 43117, 43118, 43121, 43122, 43123, and 43124. Esophageal cancer was defined by the International Classification of DiseasesNinth Revision codes: 150, 150.1, 150.2, 150.3, 150.4, 150.5, 150.8, 150.9, 151, and 151.0. Mortality rate, clinical postoperative complications, and comorbidities are provided by NSQIP.
Baseline Characteristics of the Patients The patients were divided in 2 groups according to the operation being performed by a GS or a CTS. Clinical characteristics consisted of age, sex, and ethnicity of patients; American Society of Anesthesiology classification of patient physical condition; esophagectomy type; body mass index; emergency cases; trainee involvement; and preoperative comorbidities such as diabetes mellitus (with oral agents or insulin), current history of smoking (within 1 year before the operation), dyspnea, hypertension requiring medication, weight loss (<10% of body
Ann Thorac Surg 2016;-:-–-
weight in last 6 months), steroid use for chronic condition, history of chronic obstructive pulmonary disease, and congestive heart failure.
Outcomes Comparison of intraoperative and postoperative outcomes of esophagectomy performed by GS and CTS was conducted. Primary outcome for this study was 30-days mortality. Secondary outcomes included overall morbidity, readmission rate, hospital length of stay, and discharge destination. Overall morbidity was defined by presence of at least 1 of the following NSQIP complications: wound infection, pneumonia, urinary tract infection, return to operating room, venous thromboembolic events, cardiac complication, shock or sepsis, unplanned intubation, bleeding requiring transfusion, renal complication, ventilator dependency >48 hours, and organ or space surgical site infection.
Statistical Analysis Baseline characteristics and outcomes were compared between GS and CTS using Pearson’s Chi-square test for categorical variables and 2-sample t tests for continuous variables. Multivariable logistic regression model was implemented to predict the 30-day mortality, overall and serious morbidity, readmission, and discharge destination. Negative binomial regression model was used to estimate the hospital length of stay, comparing GS versus CTS. Adjustments of the models for clinically relevant and statistically significant patients’ characteristics were done. Some variables were adjusted for clinical relevance only, without being statistically significant. Statistical analyses were performed using Stata/MP version 13 (StataCorp LP, College Station, TX). Statistical significance was indicated by p < 0.05.
Results Patient Characteristics Of the total 5,142 esophagectomies recorded in NSQIP between 2006 and 2013, 70.3% were performed by GS while 29.7% were done by CTS (Table 1). The mean age of the patients was similar among the 2 groups. Patients treated by CTS had higher American Society of Anesthesiology scores, body mass index, and history of chronic obstructive pulmonary disease. Both GS and CTS most frequently used the transthoracic approach. The second preferable approach was transhiatal for GS and McKeown for CTS. The majority of patients underwent esophagectomy for cancer, however a larger number of patients with benign disease was treated by GS. CTS were also noted to have a significant higher percentage of trainees’ involvement when compared to GS (although this variable was collected only until 2011).
Unadjusted and Adjusted Outcomes Unadjusted outcomes for esophagectomies showed that both GS and CTS had similar mortality and overall morbidity. GS had higher rates of wound infections,
Ann Thorac Surg 2016;-:-–-
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
3
Table 1. Baseline Demographic and Clinical Characteristics Characteristic Age Male Race White Black Hispanic Other ASA classification No disturb/mild disturb Serious disturb Life threat/moribund Approach Transthoracic (Ivor Lewis) Transhiatal McKeown Intestinal conduit Other Body mass index Diabetes Current smoker (within 1 year) Dyspnea History of COPD History of CHF Hypertension Weight loss Steroid use Esophageal cancer Emergency case Trainee involvement
Total (N ¼ 5,142)
General Surgery (n ¼ 3,617 [70.34%])
Cardiothoracic Surgery (n ¼ 1,525, [29.66%])
63.15 11.05 4,056 (78.96)
63.25 11.14 2,833 (78.43)
62.90 10.84 1,223 (80.20)
4,215 187 404 103
2,923 150 253 93
1,292 37 151 10
(85.86) (3.81) (8.23) (2.10)
(85.49) (4.39) (7.40) (2.72)
p Value 0.308 0.157
(86.71) (2.48) (10.13) (0.67)
<0.001 <0.001 <0.001 <0.001
1,050 (20.44) 3,601 (70.10) 486 (9.46)
794 (21.98) 2,492 (68.97) 327 (9.05)
256 (16.80) 1,109 (72.77) 159 (10.43)
<0.001 <0.001 <0.001
2,530 1,398 844 266 104 27.87 842 1,321 575 386 11 2,641 970 159 2,854 97 2,180
1,624 1,153 591 159 54 27.71 600 914 404 251 7 1,851 690 109 1,924 70 1,774
906 245 253 71 50 28.26 242 407 171 135 4 790 280 50 930 24 406
<0.001 <0.001 <0.001 <0.001 <0.001 0.006 0.524 0.287 0.949 0.017 0.626 0.681 0.549 0.616 <0.001 0.692 0.049
(49.20) (27.19) (16.41) (4.66) (2.02) 6.47 (16.37) (25.69) (11.18) (7.51) (0.21) (51.36) (18.86) (3.09) (55.50) (1.89) (82.17)
(44.90) (31.88) (16.34) (5.39) (1.49) 6.26 (16.59) (25.27) (11.17) (6.94) (0.19) (51.18) (19.08) (3.01) (53.19) (1.94) (81.49)
(59.41) (16.07) (16.59) (4.66) (3.28) 6.94 (15.87) (26.69) (11.21) (8.85) (0.26) (51.80) (18.36) (3.28) (60.98) (1.77) (85.29)
Values are mean SD or n (%). ASA ¼ American Society of Anesthesiology; deviation.
CHF ¼ congestive heart failure;
sepsis, unplanned and prolonged intubation and longer hospital stay (Table 2). CTS had a higher number of patients requiring transfusions and return to the operating room (Table 2). Adjusted multivariable analysis revealed that GS had significantly increased odds of wound infection, urinary tract infection, sepsis, shock, unplanned and prolonged intubation, and longer hospital stay. CTS had significantly higher odds of bleeding and return to the operating room. Mortality, overall morbidity and readmission’s rate were comparable between GS and CTS (Table 3). Because trainees were more frequently involved in cases performed by CTS, we performed a subanalysis to assess trainees’ involvement as an independent variable influencing outcomes. We noted that trainees’ involvement was not associated with worse outcomes (Table 4).
Comment In this study we have compared perioperative and postoperative clinical outcomes of esophagectomy performed
COPD ¼ chronic obstructive pulmonary disease;
SD ¼ standard
either by GS or CTS. Mortality, overall morbidity, and readmission odds were comparable between GS and CTS. However the most common complications were different between the two groups. Although GS had higher odds of infections, shock, sepsis, prolonged intubation, and longer hospital stay, CTS had higher odds of bleeding and return to the operating room. There are 4 other published studies that performed a similar comparison between these two specialties regarding the outcomes of esophagectomy and they showed mixed results [10–13]. Dimick and colleagues [10] found that thoracic surgeons had better mortality odds but hospital and surgeon volume had a stronger influence than specialty on mortality. The authors however used administrative data for their analysis and preoperative comorbidities as well as postoperative complications are not recorded for quality improvement purposes but only for billing and therefore are not as detailed as in our study. Interestingly thoracic surgeons were noted to be higher volume surgeons, working at higher volume hospitals when compared to GS. This is
4
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
Ann Thorac Surg 2016;-:-–-
Table 2. Observed Rates of Unadjusted Intraoperative and Postoperative Outcomes Outcome
Total
30-day mortality Overall morbiditya Wound infection (superficial þ deep SSI) Pneumonia Urinary tract infection Return to operating room Venous thromboembolism Cardiac complications Shock/sepsis Unplanned intubation Bleeding transfusion (surgical start time up to and including 72 hours) Renal complications (progressive þ acute renal failure) On ventilator >48 hours Organ space SSI Hospital length of stay, days Discharge destination (home versus facility) Readmission (%)
158 2,569 526 819 158 755 218 146 831 720 765
General Surgery
(3.07) (49.96) (10.23) (15.93) (3.07) (14.68) (4.24) (2.84) (16.16) (14.0) (14.88)
121 (2.35) 777 (15.11) 365 (7.10) 15.55 17.07 2,421 (84.56) 85 (12.71)
109 1,792 414 16.09 125 481 162 109 639 530 425
(3.01) (49.54) (11.45) (16.09) (3.46) (13.30) (4.48) (3.01) (17.67) (14.65) (11.75)
82 (2.27) 585 (16.17) 241 (6.66) 15.98 18.77 1,345 (83.70) 46 (12.11)
Cardiothoracic Surgery 49 777 112 237 33 274 56 37 192 190 340
p Value
(3.21) (50.95) (7.34) (15.54) (2.16) (17.97) (3.67) (2.43) (12.59) (12.46) (22.30)
0.705 0.357 <0.001 0.623 0.014 <0.001 0.190 0.247 <0.001 0.038 <0.001
39 (2.56) 192 (12.59) 124 (8.13) 14.53 12.03 1,076 (85.67) 39 (13.49)
0.530 0.001 0.061 0.006 0.147 0.593
a Overall morbidity: wound infection, pneumonia, urinary tract infection, venous thromboembolism, bleeding transfusion, renal complications, return to operating room, cardiac complications, shock/sepsis, unplanned intubation, on ventilator >48 hours, and organ space surgical site infection (SSI).
Values are n (%) or mean SD. SD ¼ standard deviation.
an interesting finding to complement our study because NSQIP does not provide information about hospital or surgeon volume. Similar results were reported by Gopaldas and colleagues [11] using a Nationwide
Inpatient Sample. They showed reduction of mortality odds for CTS. Interestingly although complication rates were similar among the two groups, failure to rescue, defined as death from a complication [11, 15, 16], seemed
Table 3. Adjusted Multivariable Analysis of Outcomes by Surgical Specialty, Comparing General Versus Cardiothoracic Surgeons (Reference Group) Outcome 30-day mortality Overall morbidity Wound infection (superficial) Wound infection (deep) Pneumonia Urinary tract infection Return to operating room Venous thromboembolism Cardiac complications Sepsis Shock Unplanned intubation Bleeding transfusion (surgical start time up to and including 72 hours) Renal complications (progressive þ acute renal failure) On ventilator >48 hours Organ space surgical site infection Hospital length of stay, daysb Discharge destination (home versus facility) Readmission a
Adjusted ORa
p Value
95% CI
0.997 0.96 1.51 1.40 1.07 1.54 0.75 1.28 1.22 1.48 1.54 1.28 0.46 0.87 1.42 0.80 1.12 0.83 0.88
0.989 0.487 0.002 0.144 0.459 0.036 0.001 0.131 0.328 0.001 0.002 0.011 <0.001 0.494 <0.001 0.071 <0.001 0.114 0.603
0.69–1.44 0.84–1.09 1.17–1.96 0.89–2.21 0.90–1.27 1.03–2.29 0.63–0.89 0.93–1.77 0.82–1.82 1.18–1.85 1.17–2.02 1.06–1.55 0.39–0.54 0.58–1.31 1.17–1.71 0.63–1.02 1.07–1.16 0.66–1.05 0.54–1.43
Adjusted for age, sex, race, American Society of Anesthesiology classification, body mass index, diabetes, hypertension smoking status, dyspnea, history of chronic obstructive pulmonary disease, history of congestive heart failure, weight loss, diagnosis of cancer steroid therapy, emergency case, operation b Negative binomial regression model was used for this outcome and the incidence rate ratio was calculated. approach.
CI ¼ confidence interval;
OR ¼ odds ratio;
SD ¼ standard deviation.
Ann Thorac Surg 2016;-:-–-
5
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
Table 4. Adjusted Multivariable Analysis of Outcomes by Trainee Involvement, Comparing Trainee Involvement Versus No Trainee Involvement (Reference Group) Adjusted ORa (95%CI)
Outcome 30-day mortality (%) Overall morbidity (%) Hospital length of stay, daysb Discharge destination (home versus facility) Readmission (%)
0.84 1.19 0.96 1.01 1.33
(0.49–1.42) (0.96–1.47) (0.89–1.03) (0.40–2.52) (0.40–4.37)
p Value 0.511 0.120 0.215 0.983 0.643
a
Adjusted for age, sex, race, American Society of Anesthesiology classification, body mass index, diabetes, hypertension smoking status, dyspnea, history of chronic obstructive pulmonary disease, history of congestive heart failure, weight loss, steroid therapy, emergency case, approach, surgeon b specialty. Negative binomial regression model was used for this outcome and the incidence rate ratio was calculated.
CI ¼ confidence interval;
OR ¼ odds ratio.
to have been the cause of higher mortality in the GS group. Smith and colleagues [12] on the contrary concluded that specialty training did not affect outcomes of esophagectomy but only unadjusted clinical outcomes were reported in this study. Finally, Leigh and colleagues [13] compared mortality rates after esophagectomy between GS and CTS, and between high-volume and low-volume centers in England. They have reported higher mortality rates when the operation was performed by GS, especially if in low-volume centers. CTS surgeons had overall better outcomes than GS. Volume-outcome relationship of hospitals and surgeons were reported in several esophagectomy studies with a consistent inverse association between mortality and volume [17–19]. Nevertheless, a study conducted to identify outcome differences across high-volume hospitals (according to the leapfrog volume criteria) still found considerable variability in outcomes between these centers, suggesting that other factors rather than volume alone may play an important role in influencing outcomes [20]. Although there seem to be an association between specialty and volume in the fact that specialized surgeons usually work at higher volume hospital and are higher volume surgeons, nevertheless surgical specialty seems to be an independent factor influencing outcome consistently in several studies. Previous studies however have used administrative datasets to validate their hypothesis, with obvious biases since the data were collected for reasons different than measuring outcomes. The use of the ACSNSQIP offers unique advantages because this nationally validated surgical database was designed for evaluation of surgical outcomes and enables comparison of a large number of esophageal resections across multiple types of hospital in the United States. For this reason we believe our study adds additional information to what was previously published. Trainee involvement was more frequent in cases performed by CTS. We decided to consider the presence or absence of trainees as a variable rather than differentiating between the years of training. In fact, although NSQIP provides the PGY level of residency, the length of every residency program is different and it would have been impossible to assign level of expertise (ie, resident
versus fellow) just by looking at the PGY level. Interestingly, the presence of trainees was not associated with worse mortality and all the outcomes considered in our study were similar when comparing cases done with trainees’ involvement or not. These findings confirmed previously published data from the United Kingdom regarding esophageal surgery in particular and many other studies that have looked at trainee’s involvement with cardiac and noncardiac surgery [21–23]. It is comforting that we can continue to teach complex surgeries to young trainees without compromising results. Our study has several limitations. ACS-NSQIP data does not provide the center or surgeon volume variables and therefore it is not possible to adjust for volume to see if specialization remains independently associated to outcomes. Specialization is partly defined accordingly to Board Certification and does not provide additional information regarding fellowship-trained surgeons without a corresponding board (ie, surgical oncologists or gastrointestinal surgeons) and this limits the ability to differentiate among specific surgical skills. Moreover, often an esophagectomy is done by a GS together with a CTS. We believe that in these cases the CTS specialty is assigned to the case by the coder but we are not sure if this was consistently done or done at all. This database only reports short-term outcomes with 30 postoperative days follow-up, therefore complications and mortality happening later in time are not recorded. Also long-term outcomes and survival cannot be examined. Details about the complexity of the cases during the operation are not given in NSQIP (ie, stage of the cancer, previous surgery), another variable that might confound the relationship between specialty training and clinical outcomes. The small number of esophagectomies performed by CTS in this study might also be biased by the fact that thoracic surgery cases may not be routinely collected in NSQIP, especially in academic centers where cardiothoracic surgery is often an independent department from surgery. Finally, NSQIP includes cases with a sampling system and not all cases performed in a particular hospital. Systematic sampling process, interrater reliability audits and personnel training were introduced to make sure data collection is reliable and all cases done at one particular hospital have the same chance of being selected. Although there are certainly limitations, the
6
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
results of esophagectomies from a larger number of hospitals and practices that are not strictly university affiliated can be evaluated using a database that was specifically designed to gather information about surgical outcomes. It is interesting to notice that cancer patients and patients with greater comorbidities are more often treated by specialty surgeons. This is probably reflecting the understanding that fellowship training in fact provides exposure to cutting edge technologies, focused learning and large experience with complicated surgical techniques and procedures. Other studies have shown that specialization is often linked to higher volume of surgeons and hospitals and this should be kept in mind when discussing plans for regionalization of care. There is no question that esophagectomy is a complicated procedure and postoperative care also requires higher than average skills to prevent complication or rescue the patient when complications occur. Higher specialization should be considered an important factor for patients needing an esophagectomy. In conclusion our study shows that GS in the United States perform a large number of esophagectomies. Specialization may offer the surgeon specific skills and focused experience that lead to improved perioperative outcomes for complex surgeries such as esophagectomy. Trainee involvement does not impair the outcome of esophagectomy and therefore teaching residents complex procedures should be encouraged.
References 1. Sharp LK, Bashook PG, Lipsky MS, Horowitz SD, Miller SH. Specialty board certification and clinical outcomes: the missing link. Acad Med 2002;77:534–42. 2. Grosch EN. Does specialty board certification influence clinical outcomes? J Eval Clin Pract 2006;12:473–81. 3. Norcini JJ, Kimball HR, Lipner RS. Certification and specialization: do they matter in the outcome of acute myocardial infarction? Acad Med 2000;75:1193–8. 4. da Silva PS, de Aguiar VE, Waisberg J. Pediatric surgeon vs general surgeon: does subspecialty training affect the outcome of appendicitis? Pediatr Int 2014;56:248–53. 5. Etzioni DA, Young-Fadok TM, Cima RR, et al. Patient survival after surgical treatment of rectal cancer: Impact of surgeon and hospital characteristics. Cancer 2014;120: 2472–81. 6. Tieu B, Schipper P. Specialty matters in the treatment of lung cancer. Semin Thorac Cardiovasc Surg 2012;24:99–105. 7. von Meyenfeldt EM, Gooiker GA, van Gijn W, et al. The relationship between volume or surgeon specialty and
Ann Thorac Surg 2016;-:-–-
8.
9.
10. 11.
12. 13. 14. 15. 16. 17. 18.
19. 20.
21.
22.
23.
outcome in the surgical treatment of lung cancer: a systematic review and meta-analysis. J Thorac Oncol 2012;7:1170–8. Schipper PH, Diggs BS, Ungerleider RM, Welke KF. The influence of surgeon specialty on outcomes in general thoracic surgery: a national sample 1996 to 2005. Ann Thorac Surg 2009;88:1566–72, discussion 1572–63. Pearce WH, Parker MA, Feinglass J, Ujiki M, Manheim LM. The importance of surgeon volume and training in outcomes for vascular surgical procedures. J Vasc Surg 1999;29:768–76, discussion 777–8. Dimick JB, Goodney PP, Orringer MB, Birkmeyer JD. Specialty training and mortality after esophageal cancer resection. Ann Thorac Surg 2005;80:282–6. Gopaldas RR, Bhamidipati CM, Dao TK, Markley JG. Impact of surgeon demographics and technique on outcomes after esophageal resections: a nationwide study. Ann Thorac Surg 2013;95:1064–9. Smith BR, Hinojosa MW, Reavis KM, Nguyen NT. Outcomes of esophagectomy according to surgeon’s training: general vs. thoracic. J Gastrointest Surg 2008;12:1907–11. Leigh Y, Goldacre M, McCulloch P. Surgical specialty, surgical unit volume and mortality after oesophageal cancer surgery. Eur J Surg Oncol 2009;35:820–5. https://www.facs.org/quality-programs/acs-nsqip/programspecifics/participant-use. Accessed January 27, 2016. Ghaferi AA, Birkmeyer JD, Dimick JB. Hospital volume and failure to rescue with high-risk surgery. Med Care 2011;49: 1076–81. Ghaferi AA, Birkmeyer JD, Dimick JB. Complications, failure to rescue, and mortality with major inpatient surgery in medicare patients. Ann Surg 2009;250:1029–34. Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the united states. N Engl J Med 2002;346:1128–37. Markar SR, Karthikesalingam A, Thrumurthy S, Low DE. Volume-outcome relationship in surgery for esophageal malignancy: systematic review and meta-analysis 2000–2011. J Gastrointest Surg 2012;16:1055–63. Dimick JB, Wainess RM, Upchurch GR Jr, Iannettoni MD, Orringer MB. National trends in outcomes for esophageal resection. Ann Thorac Surg 2005;79:212–6, discussion 217–8. Varghese TK Jr, Wood DE, Farjah F, et al. Variation in esophagectomy outcomes in hospitals meeting leapfrog volume outcome standards. Ann Thorac Surg 2011;91:1003–9, discussion 1009–10. Handagala SD, Addae-Boateng E, Beggs D, Duffy JP, MartinUcar AE. Early outcomes of surgery for oesophageal cancer in a thoracic regional unit. Can we maintain training without compromising results? Eur J Cardiothorac Surg 2012;41:31–4, discussion 34–5. Virk SA, Bowman SR, Chan L, et al. Equivalent outcomes after coronary artery bypass graft surgery performed by consultant versus trainee surgeons: a systematic review and meta-analysis. J Thorac Cardiovasc Surg 2016;151:647–54.e641. Ejaz A, Spolverato G, Kim Y, et al. The impact of resident involvement on surgical outcomes among patients undergoing hepatic and pancreatic resections. Surgery 2015;158: 323–30.
DISCUSSION DR PAUL SCHIPPER (Portland, OR): I have two comments and then a question. My comment. Depending on your practice situation, it is actually entirely possible for general surgeons and thoracic surgeons to work together, and as long as you are willing to subjugate your ego to outcomes, and pay attention to how you are doing, you can have outcomes that are maybe better than what you would each achieve individually.
My second comment. I think the solid message out of this paper is the resident involvement, and that is what I want to ask questions about. You are showing there was no difference in outcomes whether residents were involved or not. Do you know the nature of the residents who were involved, meaning were they general surgery residents or thoracic residents? Two-thirds of these faculty were general surgeons and one-third were cardiothoracic surgeons. Do you know if there was any
Ann Thorac Surg 2016;-:-–-
cross-training that was happening, general surgery residents working with cardiothoracic teaching faculty or cardiothoracic residents working with general surgery teaching faculty? DR KHOUSHHAL: Those are actually very good questions. Unfortunately, the NSQIP database does not specify the status of the residents, whether they are general surgery residents or cardiothoracic residents. They only give you the variable with the level of the resident and whether the operation was involving residents or not. That is it. DR ANTOON E. M. R. LERUT (Leuven, Belgium): I have a comment. I think we should stop comparing cardiothoracic surgeons versus general surgeons. I think what we really need to compare is thoracic surgeons versus upper GI surgeons because that is really where the experts are and where the question comes up if groups are working together, eg, in the Scandinavian countries the abdominal part is done by the upper GI surgeons or the esophagus, cardiac, and stomach surgeons while the thoracic part is done by the thoracic surgeons. Another comment is that 30-day mortality should no longer be used as a standard. It is at least 90-day, and for me, preferably 6month mortality rates. DR KHOUSHHAL: Thanks a lot. DR ROSS M. BREMNER (Phoenix, AZ): Dr Lerut, do you have any sense of the proportion of esophagectomies that are done in Europe by foregut general surgeons versus thoracic surgeons, or is the training such that halfway through, you start moving into that organ system as a surgeon? DR LERUT: In Europe, as a whole, I think the vast majority of esophagectomies are done by visceral surgeons, where we see a split in to upper GI, hepatobiliary, and colorectal surgery. There are some countries, like in the Scandinavian countries, where there is a collaboration between upper GI surgeons and thoracic surgeons. There are a few (eg, in France, the UK) and like us, of course, who are typically in line with the North American model where thoracic surgeons are doing the entire spectrum of the esophagus. DR SHANDA H. BLACKMON (Rochester, MN): How did you classify those cases where they were collaborating? I know at UAB all of the esophagectomies are performed half by a general surgeon who does the abdominal part, and then the thoracic part is performed by the thoracic surgeon. I think it is also that way at Emory and some other cases. How did you code those cases? And then just a note. If you have PGY-6 or above, they are likely a thoracic resident and below they are likely a general surgery resident. So that would be the way that you could delineate those two categories. DR KHOUSHHAL: What I understand from your question, I think the way that NSQIP classifies them, they just designate the surgical specialty for each surgeon doing the operation, and if there was a crossover, like if the surgeon had a board certification in general surgery and in cardiothoracic surgery, the data collector would ask the surgeon, “How do you want to designate this specific operation?” DR DANIELA MOLENA (New York, NY): And the NSQIP data is a sample, so the experts who sample the data will look at who did the operation. Of course, it is hard to know for sure because we have no control over the coder, but if there was a thoracic surgeon involved, then the designation would be thoracic
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
7
surgeon. You cannot really tell in NSQIP if there is a collaboration between two surgeons doing the operation and their specialties, but if there is a thoracic surgeon in the operation, then you would be coded as a thoracic surgeon. As regards to the PGY level, it is a little bit simplistic to say if they are 6 or above they are cardiothoracic residents, because some people do 8 or 10 years of training, especially if they go into a PhD or master program during their research time. We decided that since we really do not know the level of training, it would have been more appropriate to just look at overall involvement of residents or not. DR LUIS G. CAMERO (Detroit, MI): I want to echo the comment on ego. I do my surgeries with a general surgeon who does at least 10 laparoscopic cases a day versus my occasional laparoscopic case. This guy flies through it. In an abdominal case, the patient comes the same day of surgery, and by 10:00, the gastric mobilization, the feeding jejunostomy, and the myotomy are done, with an excellent lymph node dissection, no problem. Then I do the thoracic surgery. I do an open thoracotomy in order to obtain the best paraesophageal lymph node dissection, with an average of 20 to 30 lymph nodes, with outstanding results, no leaks. We cover the gastrostomy part with the pleura. At the dissection of the esophagus, I create a flap with the pleura. By the end of the anastomosis, I flip the gastric component of the suture line and bury it with the pleura, so there is no suture line exposed within the chest. We have zero leaks. I manage the patients postop. I am not admitting them anymore to the ICU because they are walking in the afternoon of their case. Success does not come easy, and I think that involving the general surgeon in all of the cases not only gives you a selectivity that you are involved, but they are involved. Second, you have the best postoperative care for all the patients, with the best outcomes, and more goes around in town and people keep coming to you because the results are outstanding. DR KHOUSHHAL: Thanks. DR GAIL E. DARLING (Toronto, Canada): I would echo some of the comments from the previous speakers. I think what we should be really interested in is the quality outcomes and not necessarily who is doing the surgery. Although I am a thoracic surgeon and I think we should be doing the esophagectomies, I cannot really argue that I am doing a better job than a foregut surgeon who is fully trained and does lots of esophagectomies. I really think in future studies we should be focusing on the quality of the surgery that is done and we should be developing quality indicators, whether it is the number of lymph node nodes or whatever, and we should be looking at the volume of surgery. If you are a cardiothoracic surgeon and you do 2 esophagectomies a year, I do not think that is as good as a foregut surgeon who is doing 30 esophagectomies a year. I think the volume of surgery of the individual surgeon and the type of training received is more important than the name of the training program. DR KHOUSHHAL: A couple of papers actually reported that the volume is the most important factor in this comparison. DR STEVEN R. DEMEESTER (Los Angeles, CA): I am going philosophical again. I am old school. I am board certified in general surgery. I know the abdomen as well as a general surgeon. I am board certified in thoracic surgery. I can do the thoracic component. I think that one surgeon who does the entire operation is always going to be better than a team approach. That is just my philosophy.
8
KHOUSHHAL ET AL ESOPHAGECTOMY AND SPECIALTY TRAINING
Now, here is the question for you. In an era where we are moving toward shorter residencies, I-6, people are not going to be board certified in general surgery. What are the limitations of that system moving forward? Are we going to be forced to have dual surgeons trying to take on esophagectomies?
Ann Thorac Surg 2016;-:-–-
with a more specialized approach, a thoracic surgeon will be able to do both. I think that it is important to be there for the abdominal portion of the procedure because the conduit is the most important part of the reconstruction, and you want to know that the conduit is a safe conduit when you reconstruct the esophagus. I think this is a very important point.
DR KHOUSHHAL: I am sorry, can you repeat the question? DR MOLENA: I want to compliment Zeyad. He is an MPH student, so he might not really know all the details. He has done a great job with this paper. (Applause.) As Dr Darling was saying, I think it depends on the expertise. As thoracic surgeons, we learn how to do the operation in the abdomen. We do not need the general surgery specialization. I think that moving forward, as the teaching programs develop
DR LERUT: I would like to add something to that. It is probably not the thoracic surgeons who have to work with the general surgeons. It is just vice versa. We should convince them that if they want to do that kind of surgery—and it depends on countries, of course, which is what I tried to illustrate—that they should come to the thoracic surgeons and work as a team with thoracic surgeons to get the best quality. Usually it is not the other way around. I am a general thoracic surgeon.