| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1996;61:21-26
© 1996 The Society of Thoracic Surgeons
Cardiovascular and Pulmonary Research Center, Allegheny-Singer Research Institute, Pittsburgh, Pennsylvania
Accepted for publication July 19, 1995.
 |
Abstract |
|---|
 TopFootnotes Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Methods. We examined data for all types of coronary artery bypass graft–only operations (n = 124,793) from more than 1,200 surgeons working in more than 600 hospitals for the years 1991 through 1993. All in-hospital and 30-day out-of-hospital mortality, both observed and expected as predicted by The Society of Thoracic Surgeons risk stratification method, was plotted against annualized group practice volume. Both patient-based and practice-based sampling techniques were used.
Results. The data show that observed mortality ranged from 2.0% to 3.6% for practices of more than 100 cases through practices with more than 900 cases per year. Those practices with less than 100 cases (n = 18) had a mean mortality of 5%. Expected mortalities ranged from 2.4% to 3.9% and did not vary as a function of volume. No practice volume category had an observed/expected ratio of less than 0.8 and none had a ratio greater than 1.2, if annual volume was more than 100. Practices of less than 100 cases/year had an observed/expected ratio of 1.6% to 1.7%. There was great variation in observed and expected mortalities in the lower volume categories and less variation when volume was greater (more than 600 cases/year).
Conclusions. Although the data are practice-group–specific only, there was no clinically relevant correlation of volume to outcome except at extremely low annual volume (less than 100 cases per year). Variability of outcome was significant in lower volume practices (less than 600 cases/year) and varied little at more than 600 cases per year. There were no differences in expected mortality regardless of the size of the practice.
|
Introduction |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
Material and Methods |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
This analysis was used to look for a threshold below the level of annualized outcome that might indicate an influence on mortality. Two additional analyses were done by compartmentalizing group practices by annual volume of 100 or less, 101 to 150, 151 to 200, 201 to 300, and so on up to more than 900 cases, and by a second analysis using the number of patients in each group category. Mean, standard deviation, minimum, maximum, and lower and upper 95% confidence limits were calculated. These analyses were used to look for break points below and above categoric annualized group volume.
Statistical Analysis System software (Version 6.09 for Microsoft Windows NT; SAS Institute, Carey, NC) was used for all of the analyses. These included: (1) correlation tests between annualized ratios and outcome variables; (2) independent t-test analyses for comparison of groups in the caseload threshold categories; (3) linear regression analyses for predictive strength of any relationship between volume and outcome variables using r2 values, analysis of variance, and ß coefficients; (4) X-Y scatter plots of annualized rates versus outcome variable; and (5) Logistic regression analyses using operative mortality as the dependent variable.
Analyses for average surgeon caseload were performed in identical fashion to those used for the practice group data. The average annualized surgeon caseload was determined by dividing the annual practice volume by the number of surgeons known to be practicing within each group for each year (1991 through 1993). Actual surgeon-specific volume could not be determined because the STS database does not capture surgeon, hospital, or patient identifiers.
|
Results |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
shows the number of patients and their cumulative percentage, the percent of observed and expected mortalities, and the O/E ratios as a function of increasing caseload. Except for the lowest and highest volumes-per-year groups (<100, >900), the observed mortalities are in a range of 3.0% to 3.5%, whereas only the lowest volume groups had an O/E ratio greater than 1.2 (range, 0.9 to 1.2). Seventy-five percent of the patients were operated on by groups performing 800 or fewer CABG cases per year, and nearly two-thirds (63%) of the total caseload was done by groups performing 600 or fewer cases per year. There are no breakpoints or thresholds of statistical significance except for the lowest volume groups (n = 18), which had the highest observed and expected mortalities and the highest O/E ratio. The eight groups performing the highest volume work (>900/year) had the lowest observed and expected mortalities and O/E ratio (0.9).
|
–3). These demonstrate that for a group practice of 600 cases per year and greater (n = 22), operative mortality ranged from 1.4% to 3.8% with a mean of 2.7% and a standard deviation of 0.5%. Caseloads less than 300 per year per group had a mean mortality rate of 3.3% (range, 0.3% to 11.0%). If an acceptable standard of practice band width or a range of observed mortalities of ±0.5% is assumed about the mean of 2.9% for the 120,377 patients in this study, an acceptable range would then be 2.4% to 3.4%. Only 18 (all <100 cases/year) practices (10%) had a mortality that exceeded the upper limit, whereas 23 groups (13%) had mortalities less than the lower limit and 77% were within the range. At greater than 600 cases per year, 2 of the 22 groups (44,426 patients) had a mortality greater than 3.1%, and the lowest was 1.4% for 1,270 cases.
|
|
: predicted mean risk versus caseload. Those with high volumes (>600 cases) had a narrower band width of expected risk (1.7% to 3.4%) in aggregate, except for one outlier (4.9%), than did those with fewer than 600 cases (1.2% to 5.8%). This scattergram is important because it demonstrates the variability of risk of patients presenting for CABG across the country. Further, it demonstrates that high-volume practices do not necessarily attract an undue proportion of either high- or low-risk patients.
|
shows risk and outcome by the O/E mortality ratio as a function of annual caseload. A ratio of 1 indicates the observed mortality was equal to the expected mortality; 14 groups had this ratio. If a 20% variance band width of 0.8 to 1.2 is assumed as a normal range, 52 of the 158 groups (33%) with fewer than 600 cases were within the range, and 15 of the 22 groups (68%) with more than 600 per year were within this range. Fifty-one of the practices with less than 600 cases per year (32%) were below the lower limit of 0.8, as were 5 (23%) of the high-volume groups. Fifty-five or 35% of the groups with less than 600 cases per year were above the upper limit of 1.2, whereas 2 groups (9%) of those with more than 600 cases per year were above the upper limit. Linear regression and logistic regression analyses demonstrated weak to very weak inverse correlations of volume to operative mortality, with r2 values of 0.0492, a logistic coefficient of -0.0003, and an odds ratio of 1.000.
Table 2 demonstrates data specific to group practice volume, the distribution of which is shown in Figure 4 by volume categories. Again, except for the lowest (<100 cases/year) and highest volume groups (>900 cases/year), the observed and expected mortalities and the O/E ratios are well within a small range, with the exception of 15 groups performing 401 to 500 cases, whose observed mortality was 2.0%. Importantly, these data resolve the issue of case mix. The expected mortality by volume category varies little (2.5% to 3.9%), demonstrating that concerns that all the low- or high-volume practice groups get all the good- or poor-risk patients are not warranted. If that were true, the expected mortalities, as determined by the STS risk-stratification system, would vary significantly as a function of volume.
|
|
shows the distribution by the number of patients in the practice volume categories. These data are compared with the practice-based sampling in Figures 5, 6, and 7, showing a high degree of consistency across a broad spectrum of practice volume. Important to note is that there were only two group practices, with only 3,253 patients, in the 701 to 800 cases per year category. The expected mortality is high (3.6% to 3.9%) in comparison with all other categories, making these data soft.
|
|
|
|
. There are two peaks at 201 to 300 and more than 900 cases per year, demonstrating the mismatch in the distribution of the number of groups versus the number of patients. The lowest volume groups (<100 cases/year) constituted 10% of the total number of groups while operating on only 2.2% of the patients. At the other end of the spectrum at more than 900 cases per year, 8 groups (4% of the total) operated on 20,015 or 17.5% of the patients. This discordance, however, reinforces the importance of the relative commonality of the outcome data, which show that, large or small, the expected, observed, and O/E ratios are remarkably similar among all sizes of practices.
|
|
Comment |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
These data also strongly argue for a detailed ``look-and-see'' approach by quality assurance groups, managed care entities, and governmental agencies when reviewing patient outcomes. Obviously, one shoe does not fit all. It is clearly in keeping with the democratic ethos that each surgeon and each practice group be accountable for its own performance. This requires that a common risk stratification method be employed to determine outcome fairly and that each group's performance be evaluated on its merits and not on an artificial threshold set by an agency or a corporation without data to support its mandate.
Two aspects are clearly lacking from this analysis: the influence of the hospital and the influence of the surgeon. No data were obtained by the National Database for hospital, surgeon, or patient. It is clear, however, that the role of ``the process of care'' can be independent of the surgeon, ie, it is a co-variant with the surgeon. That is, when a surgeon with a low mortality begins to work in a hospital with a historically increased CABG mortality, the surgeon's mortality rate also increases. Similarly, those surgeons with increased mortality rates experience a decline in their individual mortality rates when placed in an institution with low mortality. Although one can argue that this represents a change in case mix, higher volume, peer pressure, and many other factors, the available statistical data from both community hospitals and the Veterans Affairs system do not substantiate these other influences. The process of care for patients with coronary artery disease is complex and involves multiple groups of talented, trained individuals. It is clear, however, that the surgeon/hospital combination must be considered in any outcome analysis.
Surgeon-specific data are lacking in this report. We attempted to obtain these data by verification of the number of surgeons performing CABG operations in each group for each year. The group number was then divided by the number of surgeons, yielding an equal number of CABG cases per surgeon for that group. Not unexpectedly, the data mirrored those of the group practice variability and provided no further information. Large variability in mortality occurred at fewer than 60 cases per year, and significantly less variability occurred at 130 or more cases per year. Importantly, 78% of surgeons performed 100 or fewer CABG operations per year. Observed and expected mortality rates were remarkably similar between the two extremes of less than 25 and more than 200 cases per year. Exclusion from participating in health care plans on the basis of volume alone is without basis. Variation in annual mortality may occur because of the substantial impact of a few extremely ill, high-risk emergency cases on annual mortality in smaller practices. A mortality range of 2.5% to 3.5% or 3.0% to 4.0% is currently the norm, and even 1 year with a higher or lower than expected mortality is a normal variant in less than very large practices. Thus, reporting of yearly rather than averaged 2- to 3-year data will show significant variability in most practices.
Time as a co-variant has been considered in the analysis because overall mortality as published by the STS National Database decreased from 3.7% to 3.4% from 1991 through 1993. This small decrease, although highly statistically significant because of the large numbers of patients, has little clinical relevancy, except to demonstrate a trend that continued through 1994, when the mortality rate was 3.3%. Importantly, these data demonstrate that the operative and hospital care are good and may be slightly improving in the face of a constant to very slightly increasing mean predicted risk.
In conclusion, these data, although not surgeon or hospital specific, demonstrate a weak statistical correlation of volume to mortality after CABG, which is not clinically relevant. The vast majority (88%) of surgeons practice in groups that perform fewer than 600 cases/year. There are no meaningful differences in terms of outcome except at the low extremes (100 cases per group per year) and the high extremes of volume (>600 cases per group per year). Consequently, the Committee on Clinical Privileges finds no validity to the heretofore promulgated edict that the volume of CABG operations per year is strongly related to mortality.
|
Acknowledgments |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
Footnotes |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
* The Ad Hoc Committee on Cardiac Surgery Credentialing of The Society of Thoracic Surgeons had the following members in addition to Dr Clark: Fred A. Crawford, Jr, MD, Richard P. Anderson, MD, Frederick L. Grover, MD, Nicholas T. Kouchoukos, MD, and John A. Waldhausen, MD. 
|
References |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. L. Higgins Metrics That Matter: Can Transparency Fix the Health Care System? Seminars in Cardiothoracic and Vascular Anesthesia, September 1, 2008; 12(3): 137 - 139. [PDF]
|
|
|
|
 |
|
 L. G. Glance, A. W. Dick, T. M. Osler, and D. B. Mukamel The Relation Between Surgeon Volume and Outcome Following Off-Pump vs On-Pump Coronary Artery Bypass Graft Surgery Chest, August 1, 2005; 128(2): 829 - 837. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. K. Nallamothu, S. Saint, T. P. Hofer, S. Vijan, K. A. Eagle, and S. J. Bernstein Impact of Patient Risk on the Hospital Volume-Outcome Relationship in Coronary Artery Bypass Grafting Arch Intern Med, February 14, 2005; 165(3): 333 - 337. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Shahian, E. H. Blackstone, F. H. Edwards, F. L. Grover, G. L. Grunkemeier, D. C. Naftel, S. A.M. Nashef, W. C. Nugent, and E. D. Peterson Cardiac Surgery Risk Models: A Position Article Ann. Thorac. Surg., November 1, 2004; 78(5): 1868 - 1877. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Narayan, M. Caputo, C.A. Rogers, H. Alwair, B. Mahesh, G.D. Angelini, and A.J. Bryan Early and mid-term outcomes of surgery of the ascending aorta/arch: is there a relationship with caseload? Eur. J. Cardiothorac. Surg., May 1, 2004; 25(5): 676 - 682. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. D. Peterson, L. P. Coombs, E. R. DeLong, C. K. Haan, and T. B. Ferguson Procedural Volume as a Marker of Quality for CABG Surgery JAMA, January 14, 2004; 291(2): 195 - 201. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Shahian Improving Cardiac Surgery Quality--Volume, Outcome, Process? JAMA, January 14, 2004; 291(2): 246 - 248. [Full Text] [PDF]
|
|
|
|
|
|
L. G. Glance, A. W. Dick, D. B. Mukamel, and T. M. Osler Is the hospital volume-mortality relationship in coronary artery bypass surgery the same for low-risk versus high-risk patients? Ann. Thorac. Surg., October 1, 2003; 76(4): 1155 - 1162. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. L. Hannan, C. Wu, T. J. Ryan, E. Bennett, A. T. Culliford, J. P. Gold, A. Hartman, O. W. Isom, R. H. Jones, B. McNeil, et al. Do Hospitals and Surgeons With Higher Coronary Artery Bypass Graft Surgery Volumes Still Have Lower Risk-Adjusted Mortality Rates? Circulation, August 19, 2003; 108(7): 795 - 801. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Al-Ruzzeh, G Asimakopoulos, G Ambler, R Omar, R Hasan, B Fabri, A El-Gamel, A DeSouza, V Zamvar, S Griffin, et al. Validation of four different risk stratification systems in patients undergoing off-pump coronary artery bypass surgery: a UK multicentre analysis of 2223 patients Heart, April 1, 2003; 89(4): 432 - 435. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Shahian and S.-L. T. Normand The volume-outcome relationship: from Luft to Leapfrog Ann. Thorac. Surg., March 1, 2003; 75(3): 1048 - 1058. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. K. Nallamothu, S. Saint, S. D. Ramsey, T. P. Hofer, S. Vijan, and K. A. Eagle The role of hospital volume in coronary artery bypass grafting: is more always better? J. Am. Coll. Cardiol., December 1, 2001; 38(7): 1923 - 1930. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Shahian, G. J. Heatley, and G. A. Westcott Relationship of hospital size, case volume, and cost for coronary artery bypass surgery: Analysis of 12,774 patients operated on in Massachusetts during fiscal years 1995 and 1996 J. Thorac. Cardiovasc. Surg., July 1, 2001; 122(1): 53 - 64. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. B. Ferguson Jr, S. W. Dziuban Jr, F. H. Edwards, M. C. Eiken, A. L. W. Shroyer, P. C. Pairolero, R. P. Anderson, and F. L. Grover The STS National Database: current changes and challenges for the new millennium Ann. Thorac. Surg., March 1, 2000; 69(3): 680 - 691. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. L. Early and S. R. Roberts Excellence and low case volume: an example of the inapplicability of volume-based credentialing Ann. Thorac. Surg., January 1, 2000; 69(1): 146 - 150. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. A. Eagle, R. A. Guyton, R. Davidoff, G. A. Ewy, J. Fonger, T. J. Gardner, J. P. Gott, H. C. Herrmann, R. A. Marlow, W. C. Nugent, et al. ACC/AHA guidelines for coronary artery bypass graft surgery: A report of the American College of Cardiology/ American Heart Association task force on Practice Guidelines (Committee to revise the 1991 Guidelines for Coronary Artery Bypass Graft Surgery) J. Am. Coll. Cardiol., October 1, 1999; 34(4): 1262 - 1347. [Full Text] [PDF]
|
|
|
|
|
|
J. H. Burack, P. Impellizzeri, P. Homel, and J. N. Cunningham Jr Public reporting of surgical mortality: a survey of New York State cardiothoracic surgeons Ann. Thorac. Surg., October 1, 1999; 68(4): 1195 - 1200. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. B. Pett Jr, R. E. Clark, A. L. W. Shroyer, B. A. Warner, F. L. Grover, and K. E. Hammermeister Coronary Artery Bypass Grafting Volume and Mortality Ann. Thorac. Surg., September 1, 1996; 62(3): 941 - 943. [Full Text]
|
|
|
|
|
|
V. Parsonnet and A. D. Bernstein Outcome Versus Volume in Coronary-Bypass Operations Ann. Thorac. Surg., June 1, 1996; 61(6): 1879 - 1880. [Full Text]
|
|
|
|
|
|
F. L. Grover The Bright Future of Cardiothoracic Surgery in the Era of Changing Healthcare Delivery Ann. Thorac. Surg., February 1, 1996; 61(2): 499 - 510. [Abstract] [Full Text]
|
|
|
|
|
|
F. A. Crawford Jr, R. P. Anderson, R. E. Clark, F. L. Grover, N. T. Kouchoukos, J. A. Waldhausen, and B. R. Wilcox Volume Requirements for Cardiac Surgery Credentialing: A Critical Examination Ann. Thorac. Surg., January 1, 1996; 61(1): 12 - 16. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
