GENERAL THORACIC SURGERY: The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
Video-Assisted Thoracoscopic Lobectomy Is Less Costly and Morbid Than Open Lobectomy: A Retrospective Multiinstitutional Database Analysis Scott J. Swanson, MD, Bryan F. Meyers, MD, Candace L. Gunnarsson, EdD, Matthew Moore, MHA, John A. Howington, MD, Michael A. Maddaus, MD, Robert J. McKenna, MD, and Daniel L. Miller, MD Division of Thoracic Surgery, Brigham and Women’s Hospital and the Dana Farber Cancer Institute, Boston, Massachusetts; Division of Cardiothoracic Surgery, Washington University, St. Louis, Missouri; S2 Statistical Solutions, Inc, Cincinnati, Ohio; Ethicon Endo-Surgery, Inc, Cincinnati, Ohio; Department of Surgery, Evanston Hospital, Evanston, Illinois; Division of Thoracic Surgery, University of Minnesota, Duluth, Minnesota; Division of Thoracic Surgery, Cedars Sinai Medical Center, Los Angeles, California; and Division of Cardiothoracic Surgery, Emory University Clinic, Atlanta, Georgia
Background. The Premier Perspective Database (Premier Inc, Charlotte, NC) was used to compare hospital costs and perioperative outcomes for video-assisted thoracoscopic surgery (VATS) and open lobectomy procedures in the United States. Methods. Eligible patients underwent a lobectomy for cancer by a thoracic surgeon, by VATS or open thoracotomy and were captured in the database between third quarter of 2007 and through 2008. Multivariable logistic regression analyses were performed for binary outcomes. Ordinary least-squares regressions were used to estimate continuous outcomes. All models were adjusted for patient and hospital characteristics. Results. A total of 3,961 patients underwent a lobectomy by a thoracic surgeon by open (n ⴝ 2,907) or VATS (n ⴝ 1,054) approach. Hospital costs were higher for open versus VATS; $21,016 versus $20,316 (p ⴝ 0.027). Adjustment for surgeon experience with VATS over the 6
months prior to each operation showed a significant association between surgeon experience and cost. Average costs ranged from $22,050 for low volume surgeons to $18,133 for high volume surgeons. For open lobectomies, cost differences by surgeon experience were not significant and both levels were estimated at $21,000. Length of stay was 7.83 versus 6.15 days, for open versus VATS (p ⴝ 0.000). Surgery duration was shorter for open procedures at 3.75 versus 4.09 for VATS (p ⴝ 0.000). The risk of adverse events was significantly lower in the VATS group, odds ratio of 1.22 (p ⴝ 0.019). Conclusions. Lobectomy performed by the VATS approach as compared with an open technique results in shorter length of stay, fewer adverse events, and less cost to the hospital. Economic impact is magnified as the surgeon’s experience increases. (Ann Thorac Surg 2012;93:1027–32) © 2012 by The Society of Thoracic Surgeons
T
profile since its description in 1992, but a multiinstitutional prospective national study demonstrated the feasibility and provided a uniform definition for VATS lobectomy for lung cancer [4]. Despite this definition, VATS, a minimally-invasive technique, has had slow adoption due to concerns about oncologic principles, possible increased costs, complications, and lack of specialized surgeon training. The literature lists the potential benefits of VATS for lung procedures as smaller incisions, less pain, less blood loss, less respiratory compromise, shortened
oday, thoracic surgeons have many choices in the approach to the diagnosis and treatment of lung cancer [1, 2]. Thoracoscopic procedures are typically used for diagnosis of lung disease, excision of indeterminate pulmonary nodules, and pleurodesis for palliation of recurrent pleural effusions; these procedures are now typically performed with the use of video assistance (ie, video-assisted thoracoscopic surgery [VATS]) [3]. Video-assisted thoracoscopic surgery lobectomy has suffered from an imprecise definition and unclear safety
Accepted for publication June 1, 2011. Presented at the Poster Session of the Forty-seventh Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31–Feb 2, 2011. Address correspondence to Dr Swanson, Minimally Invasive Thoracic Surgery, Brigham and Women’s Hospital, Dana-Farber Cancer Institute, 75 Francis St, Boston, MA 02115; e-mail: sjswanson@partners.org.
© 2012 by The Society of Thoracic Surgeons Published by Elsevier Inc
Drs Swanson, Gunnarsson, Moore, and Miller disclose that they have financial relationships with Ethicon Endo-Surgery.
0003-4975/$36.00 doi:10.1016/j.athoracsur.2011.06.007
GENERAL THORACIC
ORIGINAL ARTICLES: GENERAL THORACIC
GENERAL THORACIC
1028
SWANSON ET AL VATS LOBECTOMY VERSUS OPEN LOBECTOMY
hospital lengths of stay, and superior survival rates [5, 6]. The drawbacks to VATS include higher equipment costs, longer operating room times, steeper learning curves for surgeons and operating room personnel, and uncertain oncologic control when performed for a cancer diagnosis [7]. In order to assess the current use of VATS in lobectomies for lung cancer, we analyzed real-world data from a large, nationally representative database of hospital claims data. The primary objectives of our analysis were to assess unbiased estimates of outcomes to compare the safety, utilization, and cost profiles of VATS versus open thoracotomy for lobectomy in lung cancer.
Material and Methods A protocol describing the analysis objectives, criteria for patient selection, data elements of interest, and statistical methods was submitted to the New England Institutional Review Board. Exemption was obtained.
Data Source This study utilizes the Premier Perspective Database (Premier Inc, Charlotte, NC) the largest hospital clinical and economic database in the US developed for quality and utilization benchmarking [8]. The database contains clinical and utilization information on patients receiving care in over 600 US hospitals and ambulatory surgery centers across the nation. Complete patient billing, hospital cost, and coding histories from more than 25 million inpatient discharges and 175 million hospital outpatient visits are contained in the database. Approximately 65 million service records are added monthly, and annually, more than five million hospital discharges are processed and recorded. The data undergo quality checks and cost information is reconciled with the hospitals’ financial statements. Upon receiving data from participating hospitals, Premier undertakes an extensive 7-part data validation and correction process that includes more than 95 quality assurance checks. Once all validations are complete, it is moved to the Perspective data warehouse to populate and maintain the databases for health services research. Patient diagnosis and procedures in Premier’s database are coded using the International Classification of Diseases, 9th Revision Clinical Modification (ICD-9CM) classification system. Video-assisted thoracoscopic surgery is a new technology, so the analyzable dataset from the database was restricted to procedures occurring from third quarter of 2007 through 2008. Only data that were anonymized with regard to patient identifiers were used.
Patients and Procedures Eligible patients were those of any age undergoing lobectomy for cancer by a thoracic surgeon, using either a VATS approach or an open thoracotomy. The inclusion of only lobectomy procedures by a thoracic surgeon for cancer was intended to reduce variation in our sample. The ICD-9 diagnosis codes and procedure codes for identifying the lobectomy procedures, cancer diagnoses, comor-
Ann Thorac Surg 2012;93:1027–32
bid conditions, and adverse events are available from the corresponding author at sjswanson@partners.org.
Statistical Analyses Initial counts, percentages, means, and standard deviations for patient demographics, comorbid conditions, hospital characteristics, safety, utilization, and cost outcomes were summarized for open lobectomy and VATS groups using descriptive statistics. The safety outcomes of interest were pertinent adverse events occurring during or up to 60 days post surgery; utilization outcomes were surgery duration (hours) and hospital length of stay (days). Cost outcomes included total hospital costs per patient, both fixed and variable, but did not include costs for initial acquisition or upkeep of the VATS equipment. Also, cost outcomes did not include charges or reimbursement, but rather, actual costs to the hospital, specifically. Univariate analyses were performed to compare utilization and adverse event frequencies between the two comparison groups, open lobectomy versus VATS lobectomy. Multivariable logistic regression analysis was utilized to adjust binary outcomes. Ordinary least-squares regression was used to adjust continuous outcomes, such as hospital costs, surgery time, and length of stay. For all models, the following explanatory variables were included: open lobectomy versus VATS, age, gender, race, marital status, insurance type, diagnosis (metastasis vs primary cancer), comorbid conditions (eg, diabetes), all patient refined-diagnosis related groups (APR-DRG; 3M, Salt Lake City, UT) severity index (an index of comorbidity unique to the Premier database that reflects preoperative severity level), census region of hospital, rural versus urban hospitals, teaching versus non-teaching hospitals, and number of hospital beds. Using these explanatory variables, multivariable models were estimated to isolate the effects of open lobectomy versus VATS on hospital costs, surgery time, and length of stay. Because these three continuous dependent variables were skewed, they were converted to natural logarithms to help normalize their distributions. Although estimation was by ordinary least squares, due to the logarithmic nature of the dependent variables, a common practice of smearing estimates was used to obtain predicted values for hospital costs, surgery time, and length of stay. Cases with missing data or values of zero were not included in the ordinary least-squares regression models. Weights provided in the Premier database were used to transform the results in a manner that permitted generalizability to the US population. All analyses were performed using SAS Version 9.1 (SAS Institute, Cary, NC).
Results Of 8,228 patients in the database with elective, inpatient resections for lung cancer, 3,961 patients underwent lobectomy by a thoracic surgeon using either an open technique (n ⫽ 2,907) or a VATS approach (n ⫽ 1,054). A patient attrition diagram is shown in Figure 1. Characteristics of eligible patients are summarized in Table 1.
SWANSON ET AL VATS LOBECTOMY VERSUS OPEN LOBECTOMY
1029
Fig 1. Attrition diagram thoracotomy: Open versus video-assisted thoracoscopic surgery– lobectomy. (Premier DB ⫽ Premier database.)
There were more females than males in both groups, and most patients were over 60 years of age and covered by Medicare. Most patients were Caucasian, with primary (as opposed to metastatic) cancer of the lung, and only minimal to moderate comorbid illness severity, as measured by the APR-DRG severity index. The distribution of specific comorbidities is shown in Table 2. The most frequent comorbidities reported were chronic obstructive pulmonary disease, diabetes mellitus, and heart disease. The distribution of these conditions is very well balanced between groups except for a modest excess of patients with chronic obstructive pulmonary disease in the open group (60%) versus 52% compared with the VATS group. Of the 201 hospitals contributing data on lobectomies, almost all hospitals (96%) performed open lobectomies (n ⫽ 194) and 57% performed VATS resections (n ⫽ 114) (see Table 3). Of the hospitals performing open lobectomies, there were more non-teaching (64%) than teaching hospitals (36%). The same is true for hospitals perform-
ing VATS procedures, although the split between nonteaching and teaching hospitals is not as great (57% and 43%, respectively). There are proportionately more patients in both groups (open 31% and VATS 38%) in the largest hospitals (⬎600 beds), even though these hospitals comprise proportionately smaller percentages (open 14% and VATS 21%) of the total hospitals. Lastly, it is noteworthy that more VATS patients were managed in teaching hospitals (58% vs 42% non-teaching) compared with the open group patients (47% teaching vs 53% non-teaching). In summary, it appears that lobectomy patients are more likely to have surgery in larger hospitals and VATS operations are more likely to occur in teaching hospitals. Clinically relevant adverse events including pneumonia, persistent air leaks, and red blood cell transfusions for each procedure were examined. Pneumonia occurred more frequently in the open group (9.91%) versus VATS (8.91%) although this was not statistically significant. There were significantly more arrhythmias, other cardiac
GENERAL THORACIC
Ann Thorac Surg 2012;93:1027–32
GENERAL THORACIC
1030
SWANSON ET AL VATS LOBECTOMY VERSUS OPEN LOBECTOMY
Ann Thorac Surg 2012;93:1027–32
Table 1. Patient Demographics Lobectomy a
Procedure
Open
VATS
p Value
Total N 2907 1054 (% of Total N ⫽ 3,961) (73.39%) (26.61%) Age average (SD) 67.40 (10.12) 66.97 (10.73) ⬍40 0.01 0.01 0.8185 41-50 0.05 0.06 51-60 0.17 0.18 61-70 0.36 0.35 71-80 0.32 0.31 ⬎80 0.08 0.09 Race Caucasian 0.81 0.84 0.0002 African American 0.06 0.06 Hispanic 0.02 0.03 Other 0.11 0.07 Gender Female 0.51 0.54 0.1025 Male 0.49 0.46 Marital status Married 0.55 0.6 0.0023 Unmarried 0.45 0.4 Insurance type Commercial 0.07 0.06 ⬍0.0001 Medicare 0.64 0.61 Medicaid 0.03 0.02 Managed Care 0.22 0.28 Other 0.04 0.02 Malignancy indicationb Primary neoplasm of 0.96 0.93 0.0001 the lung Metastases from other 0.04 0.07 primary malignancy Illness severity level APR-DRG Severity 0.72 0.79 ⬍0.0001 Level (1, 2) APR-DRG Severity 0.28 0.21 Level (3, 4) a
All procedures are inpatient. CPT and ICD codes for resections in b ICD codes for lung cancer in Appendix B. Appendix A. These Appendices are available from the corresponding author at sjswanson@partners.org. APR-DRG ⫽ all patient refined-diagnosis related groups; CPT ⫽ current procedural terminology; ICD ⫽ International Classification of Diseases; VATS ⫽ video-assisted thoracoscopic surgery.
events, and bleeding issues in the open group than in the VATS group. Vascular, neurologic, and wound complications were quite infrequent in both groups and did not differ significantly between groups. Average hospital costs, surgery time, and length of hospital stay for each group were examined prior to multivariable modeling. The data suggest that, on average, open lobectomy cost hospitals more than VATS lobectomy and is associated with longer lengths of hospital stay. Furthermore, the frequency of patients with
prolonged lengths of stay (ie, ⱖ 14 days) was higher in the open group than in the VATS group. Conversely, operative time for open lobectomy was slightly shorter than for VATS.
Multivariable Findings Given the possibility of confounders in these group comparisons of outcomes, we performed multivariable regression analyses, adjusting for a number of potential confounders including patient demographics, metastatic versus primary cancer, comorbid conditions, APR-DRG severity index, and hospital characteristics. The results of these adjusted analyses of costs, surgery time, and length of stay, as well as selected surgical complications, are shown in Table 4. Even after adjusting for confounding variables, the differences in utilization outcomes noted in the unadjusted analyses persisted, and all were statistically significant. Hospital costs remained significantly higher for open lobectomy than for VATS lobectomy; $21,016 versus $20,316 (p ⫽ 0.027), an average difference of $700. Average length of stay was significantly longer for open lobectomy than for VATS lobectomy, at 7.83 days versus 6.15 days (p ⬍ 0.000). The risk of patients needing prolonged hospital stays 14 days or greater was also significantly greater in the open lobectomy group than in the VATS lobectomy group (odds ratio [OR], 1.53; 95% confidence interval [CI], 1.13 to 2.09). Conversely, surgery time remained significantly shorter for open lobectomy at 3.75 hours versus 4.09 hours for VATS lobectomy (p ⬍ 0.000). After adjusting for potential confounders, the open group had a significantly higher risk of adverse events overall (OR, 1.22; 95% CI, 1.03 to 1.44) and of transfusions (OR,
Table 2. Comorbid Conditions (Existing for Patient any Time During or Before Procedure Stay in Premier Data) Lobectomy a,b
Comorbid Conditions
Total n (3,961) Myocardial infarction, acute or old Congestive heart failure Other chronic or unspecified heart failure Peripheral vascular disease Dementia Chronic pulmonary disease Connective tissue disease Liver disease Chronic viral hepatitis Renal insufficiency, chronic Diabetes mellitus
Open
VATS
p Value
2,907 0.12
1,054 0.11
0.4546 0.9205
0.07 0.01
0.07 0.02
0.1029 0.0339
0.12 0.03 0.60 0.03 0.04 0.01 0.06 0.22
0.10 0.03 0.52 0.03 0.05 0.01 0.05 0.18
0.9287 ⬍0.0001 0.7786 0.4734 0.5406 0.7831 0.0188
a
b Premier data available begins in 2000. ICD codes for these variables are found in Appendix D (this Appendix is available from the corresponding author at sjswanson@partners.org).
ICD ⫽ International Classification of Diseases; sisted thoracoscopic surgery.
VATS ⫽ video-as-
SWANSON ET AL VATS LOBECTOMY VERSUS OPEN LOBECTOMY
1031
Table 3. Hospital and Patient Demographics (Total: Hospitals, n ⫽ 201a, Patients n ⫽ 3,961) Lobectomy Open Demographics Number (% of total number) Census region Northeast Midwest South West Location Urban Not urban Type Teaching Non-teaching Bed count ⬍ 200 201–400 401–600 ⬎600
VATS
Hospitals
Patients
Hospitals
Patients
194 (96.52%)
2,907 (73.39%)
114 (56.72%)
1,054 (22.61%)
0.12 0.27 0.40 0.20
0.09 0.22 0.49 0.20
0.16 0.20 0.46 0.18
0.21 0.11 0.51 0.17
0.90 0.10
0.91 0.09
0.92 0.08
0.96 0.04
0.36 0.64
0.47 0.53
0.43 0.57
0.58 0.42
0.10 0.45 0.31 0.14
0.04 0.38 0.28 0.31
0.05 0.40 0.33 0.21
0.03 0.23 0.35 0.38
a One hundred ninety-four hospitals have open procedures and 114 hospitals have video-assisted thoracic surgery procedures, but the total hospital number is only 201. This is because the majority of hospitals do both procedures.
VATS ⫽ video-assisted thoracoscopic surgery.
1.65; 95% CI ⫽ 1.20 to 2.27). There were no group differences in pulmonary infections and air leaks. Given that there is both a reduction in adverse events and a 1.68 day reduction in length of stay with VATS, we Table 4. Multivariable Model Resultsa Lobectomy Procedure Dependent Variable Hospital costs (dollars) Open VATS Surgery time (hours) Open VATS Length of stay (days) Open VATS
Adverse event Pulmonary infections Air leak Transfusion Length of stay 14 days or more
Adjusted Outcome
Standard Deviation
$21,016.04 $20,316.19
$5,645.14 $5,457.15
0.027
3.75 4.09
0.47 0.52
0.000
7.83 6.15
2.05 1.61
0.000
p Value
Odds Ratio
Confidence Interval
p Value
1.22 0.99 1.17 1.65 1.53
1.03 to 1.44 .81 to 1.20 .98 to 1.41 1.20 to 2.27 1.13 to 2.09
0.019 0.914 0.083 0.002 0.006
VATS ⫽ video-assisted thoracoscopic surgery.
expected the difference in cost between open and VATS to be greater than $700 or 3% of the overall cost of the procedure. Thus, an additional post hoc analysis was conducted to determine if surgeon experience had an effect on overall hospital cost. The additional analysis was performed on a subset of the lobectomies (2,390) for which the surgeon’s experience could be considered. The post hoc analysis calculated each surgeon’s volume 6 months prior to the date of each patient’s surgery and then adjusted for experience in two separate multivariable cost models (VATS and open). When considering the cost of VATS lobectomies, the difference between cases by surgeons with low volume versus surgeons with high volume (16 surgeries or more in a 6-month time period) was significant, with average adjusted costs ranging from $22,050 for cases by surgeons with little experience to an average of $18,133 for cases by high volume surgeons. In the open lobectomy group, there were no differences in costs according to surgeon experience and the cost for both cohorts was estimated at around $21,000.
Comment In this retrospective analysis of a large, nationally representative database of more than 600 hospitals, VATS lobectomies for lung cancer were performed in approximately one-half of all hospitals and one-quarter of all patients undergoing lobectomy for cancer. Patients with VATS lobectomy had significantly shorter lengths of stay and lower hospital costs than patients undergoing open
GENERAL THORACIC
Ann Thorac Surg 2012;93:1027–32
GENERAL THORACIC
1032
SWANSON ET AL VATS LOBECTOMY VERSUS OPEN LOBECTOMY
Ann Thorac Surg 2012;93:1027–32
lobectomy for lung cancer. These differences persisted even after adjusting for potentially important differences in patient and hospital characteristics. Furthermore, the VATS group had a significantly lower frequency of overall adverse events than the open group. Long-term survival for patients who have undergone a VATS lobectomy appear at least equal to a thoracotomy approach in various comparative studies [9] and in two meta-analyses [10, 11]. The National Comprehensive Cancer Network guidelines for treatment of lung cancer recognize a VATS approach as a reasonable method for the treatment of lung cancer. Thus, it is highly recommended that surgeons learn this technique. Courses are available along with preceptorships for those surgeons who are already out in practice. The technology continues to improve, making the surgical procedure safer and easier. As lung cancers get detected at an increasingly smaller size and more peripheral location, minimally invasive surgical techniques become more important, particularly because other nonsurgical techniques such as stereotactic radiation become more appealing.
An important strength of the Premier database is that it provides very large numbers of patients, surgeons, and procedures on a nationwide scale. Obtaining this extremely large sample size from a real world setting allows researchers to better understand the effectiveness and cost effectiveness of a surgical technology-procedure in every day practice; hence, providing hospitals, patients, and surgeons with a realistic view of the clinical and economic impact of VATS versus open lobectomy during the study period. In conclusion, our analysis of a large, nationally representative claims database provides strong evidence showing that VATS lobectomy for lung cancer has both clinical and economic advantages over traditional open thoracotomy for lobectomy.
Economic Implications
1. DeCamp MM Jr, Jaklitsch MT, Mentzer SJ, Harpole DH Jr, Sugarbaker DJ. The safety and versatility of videothoracoscopy: a prospective analysis of 895 consecutive cases. J Am Coll Surg 1995;181:113–20. 2. Winer-Muram HT. The solitary pulmonary nodule. Radiology 2006;239:34 – 49. 3. Da Silva MC, Swanson SJ. Video-assisted thoracic surgery. In: Ashley SW, ed chair. ACS Surgery: Principles & Practice. Ontario, Canada: Decker Publishing; 2008:Sect 4/Chapter 10. Available at http://www.acssurgery.com/acssurgery/ institutional/payPerAdd.action?chapterId⫽part04_ch10. Accessed January 25, 2011. 4. Swanson SJ, Herndon II JE, D’Amico TA, et al. Video-assisted thoracic surgery lobectomy: report of CALGB 39802 – A prospective, multi-institutional feasibility study. J Clin Oncol 2007; 25:4993–7. 5. Mahtabifard A, Fuller CB, McKenna RJ Jr. Video-assisted thoracic surgery sleeve lobectomy: a case series. Ann Thorac Surg 2008;85:S729 –32. 6. Nicastri DG, Wisnivesky JP, Litle VR, et al. Thoracoscopic lobectomy: Report on safety, discharge independence, pain and chemotherapy tolerance. J Thorac Cardiovasc Surg 2008;135:642–7. 7. Jones RO, Casali G, Walker WS. Does failed video-assisted lobectomy for lung cancer prejudice immediate and longterm outcomes? Ann Thorac Surg 2008;86:235–39. 8. Premier Research Services. Premier, Inc. Charlotte, NC. Available at: http://www.premierinc.com/prs/. Accessed January 25, 2011. 9. Flores RM, Alam N. Video-assisted thoracic surgery lobectomy (VATS), open thoracotomy, and the robot for lung cancer. Ann Thorac Surg 2008;85:S710 –5. 10. Whitson BA, Groth SS, Duval SJ, Swanson SJ, Maddaus MA. Surgery for early-stage non-small cell lung cancer: a systematic review of video-assisted thoracoscopic surgery versus thoracotomy approaches to lobectomy. Ann Thorac Surg 2008:86:2008 –18. 11. Yan TD, Black D, Bannon PG, McCaughan BC. Systematic review and meta-analysis of randomized and nonrandomized trials on safety and efficacy of video-assisted thoracic surgery lobectomy for early-stage non-small-cell lung cancer. J Clin Oncol 2009;27:2553– 62.
The initial cost of capital equipment for hospitals to purchase and implement VATS (eg, video and endoscopic equipment) has not been considered in this analysis. The cost of disposable instrumentation, however, has been included in the cost to the hospital, and therefore has been included in this analysis. Similar to other minimally invasive approaches, this analysis indicates that VATS compared with an open technique offers both clinical and economic advantages. Where there may be concern over the cost of the thoracoscopic equipment required for VATS, the significant hospital savings combined with better outcomes, particularly when an experienced surgeon performs the procedure, are achievable with VATS. Hospitals typically have videoscopic equipment, which is also used in laparoscopic procedures, and all types of thoracoscopic procedures are increasingly routine.
Study Limitations This is a retrospective analysis from a transactional database rather than a prospective analysis where randomization could have reduced bias as well as provided more detailed information about patients and procedures. For instance, it would have been of interest to examine the influence of additional patient characteristics, such as weight or body mass index, and more procedurerelated details. Other data such as tumor size and nodal stations sampled would have helped to be sure the cohorts are similar. These data are not available in the Premier database. It may be possible in the future using more clinically rich datasets, namely those resulting from the use of electronic medical records in hospital settings, to extend our analyses in these directions.
This study was funded by Ethicon Endo-Surgery, Inc, Cincinnati, Ohio.
References