HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Mark B. Orringer Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dimick, J. B. Articles by Birkmeyer, J. D. Search for Related Content PubMed PubMed Citation Articles by Dimick, J. B. Articles by Birkmeyer, J. D. Related Collections Esophagus - cancer

Ann Thorac Surg 2005;80:282-286
© 2005 The Society of Thoracic Surgeons

---

Original article: General thoracic

Specialty Training and Mortality After Esophageal Cancer Resection

^a VA Outcomes Group, Department of Veterans Affairs Medical Center, White River Junction, Vermont
^b Center for the Evaluative Clinical Sciences, Dartmouth Medical School, Hanover, New Hampshire
^c Michigan Surgical Collaborative for Outcomes Research and Evaluation (M-SCORE), Department of Surgery, University of Michigan Medical Center, Ann Arbor, Michigan

Accepted for publication January 17, 2005.

^_* Address reprint requests to Dr Dimick, VA Outcomes Group 111B, VA Medical Center, 215 N Main St, White River Junction, VT05009 (Email: justin.b.dimick{at}dartmouth.edu).

	Abstract

Top Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
BACKGROUND: Surgeons with advanced training have lower mortality rates with some surgical procedures. The objective of the current study was to investigate the impact of thoracic surgery training on mortality rates of esophageal cancer resection.

METHODS: We studied esophageal cancer resection in the national Medicare population during 1998 and 1999. Operative mortality rates (in-hospital or 30-day) were compared for thoracic surgeons and other surgeons, adjusting for patient characteristics, hospital volume, and surgeon volume. Surgeons with specialty training in thoracic surgery were those certified by the American Board of Thoracic Surgery.

RESULTS: Of the 1,946 patients, 625 (32%) had their operation performed by a thoracic surgeon. After adjustment for patient characteristics, mortality rates were 37% (odds ratio, 1.37; 95% confidence interval, 1.02 to 1.82) higher for surgeons without specialty training compared with thoracic surgeons (adjusted mortality 16.5% versus 12.4%; p = 0.01). However, differences in mortality between high-volume and low-volume hospitals (24.3% versus 11.4%; p < 0.001) and surgeons (20.7% versus 10.7%; p < 0.001) were larger than those between thoracic and general surgeons. Although thoracic surgeons had lower mortality rates after adjusting for hospital volume, the effect of thoracic surgery training was no longer significant after accounting for surgeon volume (odds ratio, 1.23; 95% confidence interval, 0.92 to 1.63).

CONCLUSIONS: Specialty training in thoracic surgery has an independent association with lower mortality after esophageal resection. But specialty training appears to be less important than hospital and surgeon volume.

	Introduction

Top Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
Mortality after esophageal resection is strongly related to the experience of the hospital [1–4]. Because of this strong relationship, several groups endorse using hospital volume to identify high-quality providers for this operation. For example, the National Cancer Policy Board recommends hospital volume as a quality indicator for esophageal resection [5]. Taking an even more aggressive approach, the Leapfrog Group, a coalition of large employers and other payers, recommends selective referral of esophagectomy to high-volume hospitals in their evidence-based hospital referral initiative [6].

However, many feel that these efforts are focusing on the wrong provider-level variable. In particular, some argue that the attributes of the surgeon are more important than those of the hospital in which they work. Indeed, emerging evidence has shown that surgeon volume is strongly linked to operative mortality. In one recent study from the Medicare population, low-volume surgeons had mortality rates of more than twofold greater than high-volume surgeons. Despite this new evidence, there is another surgeon-level variable that has largely been ignored: specialty training in thoracic surgery. With other high-risk operations, surgeon specialty is a strong independent predictor of postoperative outcomes [7–10].

To determine the relationship of surgeon specialty and mortality for esophageal resection, we studied all patients who had this operation in the national Medicare population. We identified surgeons with thoracic specialty training as those who were board certified according to the American Board of Thoracic Surgery. Because we wanted to determine the independent effect of specialty training, we adjusted for hospital and surgeon volume in our analysis.

	Material and Methods

Top Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
Study Population
We studied all esophageal resections for cancer performed in the national Medicare population during a 2-year period (1998 through 1999). Hospital discharge abstracts were obtained from the Medicare Provider Analysis Review File (MEDPAR database). Operative cases were identified using the appropriate International Classification of Diseases, Ninth Revision (ICD-9) procedure codes for esophageal resection: partial esophagectomy ICD-9-CM code 42.41, total esophagectomy ICD-9-CM code 42.42, and esophagectomy, not otherwise specified ICD-9-CM code 42.40. Given the concern of confounding as a result of differences in the indication for operation, we limited our analysis to patients who underwent esophagectomy specifically for cancer, which was identified using a combination of ICD-9 diagnostic codes [3]. We sought to create a homogeneous group for comparison (all cancer patients) because little clinical detail is available for risk-adjustment in administrative data. Including a broader group of diagnoses would create more opportunity for residual differences (confounding) owing to differences in the indication among surgeons. Medicare beneficiaries younger than 65 years or older than 99 years or those enrolled in health maintenance organizations (approximately 10%) were also excluded from our analysis.

Classifying Thoracic Surgeons
The American Board of Thoracic Surgery is the main certifying body for thoracic surgeons in the United States. We first determined the Medicare Unique Physician Identity Number (UPIN) for the operating surgeon of each patient who had an esophageal resection during the 2-year study period. Using a list of all surgeons who were board certified by the American Board of Thoracic Surgery up to 2002, we assigned each surgeon to one of two categories: general surgeon or thoracic surgeon. Thoracic surgeons were those who were board certified and general surgeons were those who were not.

Hospital and Surgeon Volume
Using unique hospital and surgeon identifiers, we determined the number of cases performed by each hospital and each surgeon during the 2-year study period. For both hospital and surgeon volume, we created three approximately equal size groups on the basis of terciles of volume: low volume, medium volume, and high volume. Although we excluded patients without a cancer diagnosis from the analysis of outcomes, the calculation of hospital and surgeon volume included all cases performed regardless of the indication. To extrapolate Medicare volumes to total hospital volumes, we used data from the Nationwide Inpatient Sample to determine the proportion of the total number of cases (all payers) performed in Medicare patients. We then estimated the total volume by multiplying each hospital’s observed Medicare volume by the overall total to Medicare ratio for hospitals and surgeons. The volume cut points for total hospital volume were as follows: low volume, fewer than 5 cases per year; medium volume, 5 to 12 cases per year; and high volume, more than 12 cases per year. For surgeon volume, the cut points were as follows: low volume, fewer than 2 cases per year; medium volume, 2 to 5 cases per year; and high volume, more than 5 cases per year.

Statistical Analysis
Our primary goal was to assess the relationship of operative mortality, which was defined as death before discharge or within 30 days of the procedure, and board certification in thoracic surgery. Beyond this primary analysis, we were also interested in determining the effect of thoracic surgery training on mortality after adjusting for hospital and surgeon volume. We used multiple logistic regression to compare the risk of operative mortality with thoracic and general surgeons, adjusting for provider volume (either surgeon or hospital volume). For these analyses, we entered hospital and surgeon volume into the regression as continuous variables. However, we present them stratified by terciles for ease of presentation. Because of the high degree of correlation between hospital and surgeon volume, separate logistic regression models were created for each volume variable. In these analyses, we also adjusted for patient demographics (age, sex, race), coexisting diseases, and admission acuity to account for case-mix differences among surgeons. Only those independent variables with p less than 0.10 were included in the final model. Clustering within surgeons and hospitals was accounted for using the generalized estimating equation in each of the stratified analyses. All statistical analyses were conducted using STATA 8.0 (Statacorp, College Station, TX).

	Results

Top Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
Patient Characteristics
During the 2-year study period, 1,946 Medicare patients had esophageal cancer resection with surgeon-level information available. Of these, 625 (32%) had their operation performed by a thoracic surgeon and 1,321 (68%) by a general surgeon. Patients having operations by each type of surgeon were similar with respect to sex, race, admission acuity, and comorbid diseases (Table 1). However, patients having their operation performed by a thoracic surgeon were less likely to be older than 75 years (27.8% versus 32.3%) and more likely to be in the lowest income group (26.6% versus 24.3%).

View larger version (42K): [in this window] [in a new window]   Fig 1. Operative mortality rates for esophageal resection performed by thoracic versus general surgeons, high-volume versus low-volume surgeons, and at high-volume versus low-volume hospitals.

View larger version (37K): [in this window] [in a new window]   Fig 2. Risk-adjusted mortality rates for general and thoracic surgeons stratified by hospital volume (A) and surgeon volume (B).

	Comment

Top Abstract Introduction Material and Methods Results Comment Acknowledgments References

We should consider these findings in the context of certain weaknesses in our study design. First, our study is based on the Medicare inpatient database and is subject to certain limitations. One chief concern is the adequacy of risk adjustment with administrative data. This shortcoming would be problematic if patients operated on by thoracic surgeons were more severely ill than the patients of general surgeons. However, we have no data to indicate whether this is the case. Another potential limitation of our study is the accuracy of our definition of specialty training in thoracic surgery. We believe our measure was accurate inasmuch as we used actual board certification by the American Board of Thoracic Surgery to classify surgeon specialty. However, there may have been some surgeons in our sample of general surgeons who had completed a fellowship but have not yet taken or passed the board examination for thoracic surgery. If this were the case, some surgeons with advanced thoracic training would be labeled as general surgeons. Misclassification would likely bias our results toward the null hypothesis, ie, no difference between thoracic and general surgeons. Finally, many surgeons who are board certified in thoracic surgery focus their practice on cardiac surgery. In our analysis, we did not consider what types of cases each thoracic surgeon performed. Our results may have been different if we focused on cardiac versus noncardiac thoracic surgeons. But our goal was to demonstrate the effect of thoracic surgery training on outcomes—not the effect of further clinical specialization within a surgeon’s practice after training.

Although no previous investigation has specifically addressed esophageal resection, the broader issue of specialty training and mortality rates has been studied for several other high-risk cancer operations. One recent study from New York State compared mortality rates for surgeons with and without subspecialty interest for two other gastrointestinal operations [8]. For colon resection, mortality rates were much lower (2.4% versus 4.8%) for surgeons who were members of the Society of Colorectal Surgery. Similarly, for gastric resection, mortality rates were lower (6.5% versus 8.7%) for members of the Society of Surgical Oncology. It is important to note that for this study in New York State, membership in the relevant professional society was used as a proxy for subspecialty training. This method stands in contrast to our study, which used actual board-certification status. In another recent study, Goodney and colleagues [7] found that surgeons who are board certified in thoracic surgery have lower mortality rates (7.6% versus 5.8%) for resection of lung cancer in the national Medicare population.

Most of these previous analyses also adjust for surgeon and hospital volume and, similar to our findings, often demonstrate an independent effect of specialty training on mortality. For instance, both studies mentioned above showed a large, independent effect of surgeon specialty on short-term outcomes after adjusting for hospital and surgeon volume. The procedures included in these previous studies and our study—colon resection, gastric resection, lung resection, and esophageal resection—are all relatively high-risk cancer operations. For such operations, it is not surprising that both surgeon-level and hospital-level variables exert an independent effect. Individual surgeon experience and training are necessary but not sufficient for optimal outcomes. Many hospital-wide resources are needed to ensure good outcomes, including the availability of preoperative evaluation, sufficient nurse-to-patient ratios, and well-staffed intensive care units.

For some other operations, however, the effect of surgeon specialty completely disappears after adjusting for provider volume. For example, Cowan and colleagues [9] demonstrated that mortality rates after carotid endarterectomy varied across surgeon specialty, hospital volume, and surgeon volume. When considering the three provider-level variables together, however, they found that only surgeon volume was an independent predictor of mortality. In contrast to high-risk cancer operations, it makes sense that for carotid endarterectomy, few resources beyond the skill of the individual surgeon are needed.

The findings of our study can be applied in two ways to improve the quality of care for patients undergoing esophageal resection. The first option is to selectively refer patients to surgeons or hospitals that meet certain criteria. Given the strong volume-outcome effect for this operation, selective referral is an attractive option for this procedure. The Leapfrog Group, a coalition of large public and private health-care purchasers, currently suggests referral based only on hospital volume. Should these standards be changed to focus on selective referral to thoracic surgeons? The results of our study would argue against this approach. Both hospital and surgeon volume are stronger predictors of operative mortality than thoracic specialty training. Because many thoracic surgeons are high-volume surgeons or work at high-volume hospitals, however, it is likely that selective referral in any form would result in more operations done by board-certified thoracic surgeons.

The second option to improve quality for this operation is to identify what high-volume or board-certified thoracic surgeons do differently to achieve better outcomes. Identifying the underlying details that contribute to these differences in outcomes is a major priority for surgeons and health services researchers. Indeed, great success has been found in linking processes to outcomes and using this information to guide quality improvement for some cardiovascular operations. For example, Hannan and colleagues [10] used a large prospective data set in New York State to isolate several discrete processes of care that account for the better outcomes of vascular surgeons with carotid endarterectomy. They found that the lower mortality rates achieved by vascular surgeons were directly attributable to several specific processes of care—the eversion technique, use of protamine, and use of shunts. Increasing the use of these techniques by nonvascular surgeons could improve the outcomes for all providers.

Although process improvement may work for more common procedures, certain challenges exist for less common operations, such as esophagectomy. For example, most individual hospitals perform only a few esophageal resections per year, which is a number much too small to support the measurement of precise mortality or morbidity rates. Any efforts to link processes to outcomes for this operation will need to undertake multicenter data collection. Alternatively, hospitals could consider only those processes of care that are common across multiple operations (eg, antibiotic prophylaxis to prevent wound infection). Beyond this problem with small sample size, the processes of care that contribute to superior outcomes have yet to be elucidated for esophageal resection.

We have demonstrated that specialty board certification in thoracic surgery is independently related to improved operative mortality rates after esophageal resection. However, hospital and surgeon volume are more strongly related to mortality. With small number of esophageal resections performed in each hospital and by each surgeon, measuring outcomes with enough precision for quality improvement is not possible. For related reasons, identifying processes of care that are associated with improved outcomes may prove difficult. As a result, serious attention should be paid to selective referral on the basis of proxy measures of quality. Although surgeon specialty is important, these efforts at selective referral could be based on variables that are equally actionable but more important, including hospital and surgeon volume.

	Acknowledgments

Top Abstract Introduction Material and Methods Results Comment Acknowledgments References

 
Doctor Dimick was supported by a Veterans Affairs Special Fellowship Program in Outcomes Research. This study was also supported by a grant to Dr Birkmeyer from the National Cancer Institute (1R01CA098481–01A1).

	References

Top Abstract Introduction Material and Methods Results Comment Acknowledgments References

 

1. Halm EA, Lee C, Chassin MR. Is volume related to outcome in health care? A systematic review and methodologic critique of the literature Ann Intern Med 2002;137:511-520.[Abstract/Free Full Text]
2. Dudley RA, Johansen KL, Brand R, Rennie DJ, Milstein A. Selective referral to high-volume hospitalsestimating potentially avoidable deaths. JAMA 2000;283:1159-1166.[Abstract/Free Full Text]
3. Birkmeyer JD, Siewers AE, Finlayson EV, et al. Hospital volume and surgical mortality in the United States N Engl J Med 2002;346:1128-1137.[Abstract/Free Full Text]
4. Dimick JB, Cowan Jr JA, Ailawadi G, Wainess RM, Upchurch Jr GR. National variation in operative mortality rates for esophageal resection and the need for quality improvement Arch Surg 2003;138:1305-1309.[Abstract/Free Full Text]
5. Hewitt M, Petitti D, National Cancer Policy Board Interpreting the volume-outcome relationship in the context of cancer careWashington, DC: National Academy Press; 2001.
6. Milstein A, Galvin RS, Delbanco SF, Salber P, Buck Jr CR. Improving the safety of health carethe Leapfrog initiative. Effective Clinical Practice 2000;3:313-316.
7. Goodney PP, Lucas FL, Stukel TA, Birkmeyer JD, Buck Jr CR. Surgeon specialty and operative mortality with lung resection Ann Surg 2005;241:179-184.[Medline]
8. Callahan MA, Christos PJ, Gold HT, Mushlin AI, Daly JM. Influence of surgical subspecialty training on in-hospital mortality for gastrectomy and colectomy patients Ann Surg 2003;238:629-636.[Medline]
9. Cowan Jr JA, Dimick JB, Thompson BG, Stanley JC, Upchurch Jr GR. Surgeon volume as an indicator of outcomes after carotid endarterectomyan effect independent of specialty practice and hospital volume. J Am Coll Surg 2002;195:814-821.[Medline]
10. Hannan EL, Popp AJ, Feustel P, et al. Association of surgical specialty and processes of care with patient outcomes for carotid endarterectomy Stroke 2001;32:2890-2897.[Abstract/Free Full Text]

This article has been cited by other articles:

				B. D. Kozower, G. J. Stukenborg, C. L. Lau, and D. R. Jones Measuring the Quality of Surgical Outcomes in General Thoracic Surgery: Should Surgical Volume Be Used to Direct Patient Referrals? Ann. Thorac. Surg., November 1, 2008; 86(5): 1405 - 1408. [Full Text] [PDF]

				E. S. Weiss, R. A. Meguid, N. D. Patel, S. D. Russell, A. S. Shah, W. A. Baumgartner, and J. V. Conte Increased mortality at low-volume orthotopic heart transplantation centers: should current standards change? Ann. Thorac. Surg., October 1, 2008; 86(4): 1250 - 1260. [Abstract] [Full Text] [PDF]

				J. Ra, E. C. Paulson, J. Kucharczuk, K. Armstrong, C. Wirtalla, R. Rapaport-Kelz, L. R. Kaiser, and F. R. Spitz Postoperative Mortality After Esophagectomy for Cancer: Development of a Preoperative Risk Prediction Model Ann. Surg. Oncol., June 1, 2008; 15(6): 1577 - 1584. [Abstract] [Full Text] [PDF]

				A. Pennathur and J. D. Luketich Resection for Esophageal Cancer: Strategies for Optimal Management Ann. Thorac. Surg., February 1, 2008; 85(2): S751 - S756. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Mark B. Orringer Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dimick, J. B. Articles by Birkmeyer, J. D. Search for Related Content PubMed PubMed Citation Articles by Dimick, J. B. Articles by Birkmeyer, J. D. Related Collections Esophagus - cancer