[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.167.142.229. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Download PDF
Table 1. 
Derivation of the Final Study Populationa
Derivation of the Final Study Populationa
Table 2. 
Estimated Odds Ratios for All Variables Included in the Logistic Regression Model Estimating In-Hospital Mortality for FY2003 and FY2004
Estimated Odds Ratios for All Variables Included in the Logistic Regression Model Estimating In-Hospital Mortality for FY2003 and FY2004
Table 3. 
Hospital Performances by Tier for FY2003 and FY2004
Hospital Performances by Tier for FY2003 and FY2004
Table 4. 
Average Hospital Risk-Adjusted Mortality Rate by Sex for All Hospitals by Performance Tiersa
Average Hospital Risk-Adjusted Mortality Rate by Sex for All Hospitals by Performance Tiersa
Table 5. 
Potential Male and Female Deaths Avoided If Bottom Tier 4 Hospitals Improved Their Sex-Specific Risk-Adjusted Mortality Rate to the Average Rate of Top Tier 1 Hospitalsa
Potential Male and Female Deaths Avoided If Bottom Tier 4 Hospitals Improved Their Sex-Specific Risk-Adjusted Mortality Rate to the Average Rate of Top Tier 1 Hospitalsa
1.
Steinbrook  R Public report cards: cardiac surgery and beyond. N Engl J Med 2006;355 (18) 1847- 1849
PubMedArticle
2.
Hannan  ELSarrazin  MSDoran  DRRosenthal  GE Provider profiling and quality improvement efforts in coronary artery bypass graft surgery: the effect on short-term mortality among Medicare beneficiaries. Med Care 2003;41 (10) 1164- 1172
PubMedArticle
3.
Hartz  AJKuhn  EM Comparing hospitals that perform coronary artery bypass surgery: the effect of outcome measures and data sources. Am J Public Health 1994;84 (10) 1609- 1614
PubMedArticle
4.
Shahian  DM Improving cardiac surgery quality: volume, outcomes, process? JAMA 2004;291 (2) 246- 248
PubMedArticle
5.
Krumholz  HM Mathematical models and the assessment of performance in cardiology. Circulation 1999;99 (16) 2067- 2069
PubMedArticle
6.
New York State Department of Health,Coronary Artery Bypass Graft Surgery in New York State 1989-1991. Albany, NY New York State Dept of Health1992;
7.
Pennsylvania Health Care Cost Containment Council, A Consumer Guide to Coronary Artery Bypass Graft Surgery. Vol 4. Harrisburg Pennsylvania Health Care Cost Containment Council1995;
8.
Clifton  CRCardiac Surgery in 2000 in New Jersey: A Consumer Report. Trenton New Jersey Dept of Health and Senior Services2003;
9.
Parker  JPLi  ZDamberg  CL  et al. The California Report on Coronary Artery Bypass Graft Surgery: 2000-2002 Hospital Data. San Francisco, CA California Office of Statewide Health Planning and Development and the Pacific Business Group on Health2005;
10.
Naylor  CDRothwell  DMTu  JV  et al. Cardiac Care Network Steering Committee, Outcomes of coronary artery bypass graft surgery in Ontario. Naylor  CDSlaughter  PMCardiovascular Health Services in Ontario: An ICES Atlas. Toronto, ON Institute for Clinical Evaluative Sciences1999;189- 197
11.
Shahian  DMTorchiana  DFNormand  SL Implementation of a cardiac surgery report card: lessons from the Massachusetts experience. Ann Thorac Surg 2005;80 (3) 1146- 1150
PubMedArticle
12.
Marshall  MNRomano  PS Impact of reporting hospital performance. Qual Saf Health Care 2005;14 (2) 77- 78
PubMedArticle
13.
Werner  RMAsch  DA The unintended consequences of publicly reporting quality information. JAMA 2005;293 (10) 1239- 1244
PubMedArticle
14.
Vaccarino  VAbramson  JLVeledar  EWeintraub  WS Sex differences in hospital mortality after coronary artery bypass surgery: evidence for a higher mortality in younger women. Circulation 2002;105 (10) 1176- 1181
PubMedArticle
15.
O’Rourke  DJMalenka  DJOlmstead  EM  et al.  Improving in-hospital mortality in women undergoing coronary artery bypass grafting. Ann Thorac Surg 2001;71 (2) 507- 511
PubMedArticle
16.
Hannan  ELBernard  HRKilburn  HC  et al.  Gender differences in mortality rates for coronary artery bypass surgery. Am Heart J 1992;123 (4, pt 1) 866- 872
PubMedArticle
17.
Department of Health and Human Services, Centers for Medicare and Medicaid Services, Medicare Provider Analysis and Review (MEDPAR) file. http://www.cms.hhs.gov/IdentifiableDataFiles/05_MedicareProviderAnalysisandReviewFile.asp. Accessed May 5, 2008
18.
Agency for Healthcare Research and Quality, Patient safety indicators. http://www.qualityindicators.ahrq.gov/psi_download.htm. Accessed April 20, 2008
19.
Penckofer  SMHolm  K Women undergoing coronary bypass surgery: physiological and psychosocial perspectives. Cardiovasc Nurs 1990;26 (3) 13- 18
PubMed
20.
Richardson  JVCyrus  RJ Reduced efficacy of coronary artery bypass grafting in women. Ann Thorac Surg 1986;42 (6) ((suppl)) S16- S21
PubMedArticle
21.
Gansera  BGillrath  GLieber  MAngelis  ISchimdtler  FKemkes  BM Are men treated better than women? outcome of male versus female patients after CABG using bilateral internal thoracic arteries. Thorac Cardiovasc Surg 2004;52 (5) 261- 267
PubMedArticle
22.
Edwards  MLAlbert  NMWang  CApperson-Hanson  C 1993-2003 Gender differences in coronary artery revascularization: has anything changed? J Cardiovasc Nurs 2005;20 (6) 461- 467
PubMedArticle
23.
O’Connor  GTMorton  JRDiehl  MJ  et al.  Differences between men and women in hospital mortality associated with coronary artery bypass graft surgery. Circulation 1993;88 (5, pt 1) 2104- 2110
PubMedArticle
24.
Pine  MNorusis  MJones  BRosenthal  GE Predictions of hospital mortality rates: a comparison of data sources. Ann Intern Med 1997;126 (5) 347- 354
PubMedArticle
25.
Iezzoni  LIAsh  ASShwartz  M  et al.  Judging hospitals by severity adjusted method. Am J Public Health 1996;86 (10) 1379- 1387
PubMedArticle
26.
Austin  PCTu  JVAlter  DANaylor  CD The impact of under coding of cardiac severity and comorbid diseases on the accuracy of hospital report cards. Med Care 2005;43 (8) 801- 809
PubMedArticle
27.
Weintraub  WSBecker  ERMauldin  PDCuller  SKosinski  ASKing  SB Cost of revascularization over eight years in the randomized and eligible patients in the Emory Angioplasty versus Surgery Trial (EAST). Am J Cardiol 2000;86 (7) 747- 752
PubMedArticle
28.
Zhang  ZSpertus  JAMahoney  EM  et al.  The impact of acute coronary syndrome on clinical, economic, and cardiac-specific health status after coronary artery bypass surgery versus stent-assisted percutaneous coronary intervention: 1-year results from the Stent or Surgery (SoS) trial. Am Heart J 2005;150 (1) 175- 181
PubMedArticle
29.
Clark  RE Calculating risk and outcome: the Society of Thoracic Surgeons database. Ann Thorac Surg 1996;62 (5) ((suppl)) S2- S5
PubMedArticle
Original Investigation
November 24, 2008

Sex Differences in Hospital Risk-Adjusted Mortality Rates for Medicare Beneficiaries Undergoing CABG Surgery

Author Affiliations

Author Affiliations: Rollins School of Public Health (Drs Culler and Rask) and Emory Center on Health Outcomes and Quality (Dr Rask), Emory University, Atlanta, Georgia; Cardiac Data Solutions Inc, Atlanta (Ms Simon and Dr Brown); Henry Ford Hospital, Detroit, Michigan (Dr Kugelmass); and Beth Israel Deaconess Medical Center, Boston, Massachusetts (Dr Reynolds).

Arch Intern Med. 2008;168(21):2317-2322. doi:10.1001/archinte.168.21.2317
Abstract

Background  The primary purpose of this study was to rank US hospitals performing coronary artery bypass graft (CABG) surgery on Medicare beneficiaries into 4 performance tiers and determine if there were overall and sex-specific differences in the risk-adjusted mortality rates across performance tiers.

Methods  A retrospective analysis was done using a Medicare Provider Analysis and Review (MEDPAR) file of all Medicare beneficiaries who underwent CABG surgery without valve repair or replacement during fiscal years 2003 and 2004. Logistic regression models controlling for demographic characteristics, comorbidities, and cardiac risk factors were used to predict the probability of in-hospital mortality. Hospitals performing at least 52 CABG surgeries during a fiscal year (at least 17 female patients) were ranked into 4 tiers. Rankings were based on the number of lives saved, calculated as the expected number of risk-adjusted deaths minus the actual number of deaths in the hospital during each fiscal year.

Results  Average risk-adjusted mortality rate was stable and declining over the 2 years: 3.68% in 2003 and 3.61% in 2004. In 2004, the average risk-adjusted mortality rate ranged from 1.39% in tier 1 hospitals to 6.40% in tier 4 hospitals. The sex-specific mortality rate was consistently higher for women in all tiers, with the differential smallest (0.68%) in tier 1 hospitals and greatest (2.67%) in tier 4 hospitals.

Conclusion  The sex differential increases from top- to bottom-tier hospitals, suggesting female beneficiaries could benefit from having CABG performed at tier 1 hospitals.

Hospital performance on the basis of coronary artery bypass graft (CABG) surgery outcomes are being published by a growing number of public and private organizations.13 While many of these reports rank hospitals according to a broad spectrum of factors including hospital structure, scope, processes, reputation, and postoperative clinical outcomes, inpatient mortality rate is the most commonly used measure of cardiovascular surgical care quality.4,5 New York, Pennsylvania, New Jersey, California, and Ontario, Canada, have all published risk-adjusted hospital rankings for cardiac surgery outcomes demonstrating differences in CABG mortality across hospitals.610 With CABG surgery increasingly becoming an elective procedure performed by an increasing number of hospitals, there is a growing interest and debate over making hospital risk-adjusted mortality rates (RAMRs) publicly available.1113 In addition, the growing awareness of sex differences in cardiovascular outcomes has raised concerns about whether the sex differential in risk-adjusted mortality rate varies across hospital performance tiers.1416

The primary purpose of this study was to rank all US hospitals performing CABG surgery on Medicare beneficiaries into 4 performance tiers based on the total number of lives saved, calculated as the difference between the risk-adjusted expected number of deaths and the actual number of deaths. We chose this ranking method because it incorporates both a quality of care and volume component into the ranking. The hospital rankings are then used to answer 3 questions. First, are there statistically significant differences in the overall and sex-specific average RAMRs across performance tiers? Second, is the sex differential (average male RAMR minus average female RAMR) stable across hospital performance tiers, or does it change between tiers? Third, what would be the potential mortality impact (percentage of deaths avoided) among Medicare beneficiaries if bottom-tier hospitals would improve their RAMR to the average rate of hospitals performing in the top tier?

METHODS
DATA SOURCE

The Medicare Provider Analysis and Review (MEDPAR) files17 for fiscal year 2003 (FY2003) (October 1, 2002, through September 30, 2003) and fiscal year 2004 (FY2004) (October 1, 2003, through September 30, 2004) were the data sources for this retrospective analysis. The MEDPAR file is an administrative database maintained by the Centers for Medicare and Medicaid Services containing all claims submitted by hospitals for services provided to Medicare beneficiaries. For each hospitalization, the MEDPAR record includes patient information on age, sex, race, date of admission, date of discharge, the principal diagnosis code, 8 secondary diagnosis codes, 6 procedure codes, discharge status, total charges, and total reimbursement.

STUDY POPULATION

The population in this study consists of all US hospitals not in the Veterans Affairs Hospital System that performed at least 52 CABG procedures without concomitant valve surgery on Medicare beneficiaries during FY2003 or FY2004. The final study sample was derived as follows: First, the respective fiscal year MEDPAR file was searched for all hospital admissions with an International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) procedure code of 36.10 through 36.19 or 36.2, indicating that the patient underwent a primary or secondary CABG procedure during that admission. A Medicare beneficiary's CABG hospitalization was excluded from the study if the beneficiary also underwent concomitant value surgery (ICD-9-CM procedure code of 35.00 through 35.04, 35.10 through 35.14, 35.20 through 35.28, or 35.31 through 35.39).

Table 1 summarizes the characteristics of the final study population. The initial search identified 143 462 Medicare beneficiaries (66.7% men) who were treated in 1069 hospitals during FY2003 and 134 707 Medicare beneficiaries (67.1% men) treated in 1108 hospitals during FY2004. Hospitals (and thus the Medicare beneficiaries they treated) were excluded from the study for either or both of the following reasons: (1) they performed fewer than 52 CABG procedures (on average <1/wk) on Medicare beneficiaries during the fiscal year, and/or (2) they performed fewer than 17 CABG procedures on female Medicare beneficiaries during the fiscal year. A minimum of 17 female patients was set because it represents approximately 33% of 52 CABG procedures.

STATISTICAL ANALYSIS

Univariate differences in hospital outcomes were assessed using χ2 analysis or the Fisher exact test for discrete variables and 2-sample student t tests for continuous variables. For each fiscal year, a multivariate logistic regression model was estimated to predict the expected probability of in-hospital mortality for each patient. The identical statistical analysis was conducted in 2 consecutive fiscal years to provide insight into the stability of both the average RAMR and the size of the sex difference by hospital performance class. All analyses were performed with SAS version 9.1 statistical software (SAS Institute, Cary, North Carolina).

ESTIMATING EXPECTED IN-HOSPITAL MORTALITY

The primary purpose of the multivariate logistic regression model was to predict the probability of a Medicare beneficiary experiencing in-hospital mortality during a CABG hospitalization, controlling for demographic, comorbidity, and cardiac risk factors. The same multivariate logistic regression model was estimated in each fiscal year. The estimated logistic regression model controlled for demographic variables (age, sex, and race) and for the following comorbidities and cardiac risk factors coded from the administrative database: prior CABG procedure, prior percutaneous coronary intervention, prior myocardial infarction, unstable angina, type of primary acute myocardial infarction, congestive heart failure, chronic obstructive pulmonary disease, chronic renal failure, renal failure unspecified, chronic liver disease, hypertension, conduction disorders, endocarditis, aortic aneurysm, history of cerebral vascular disease, diabetes mellitus, obesity, and current smoking status.

RANKING HOSPITALS

This study ranks hospitals on their CABG surgery performance based on the potential number of lives saved by each hospital. The number of lives saved by each hospital is defined as the difference between the hospital's expected number of deaths and its observed number of deaths. A hospital's expected number of deaths is calculated by summing the expected probability of in-hospital mortality (estimated from the logistic regression equation) for each patient treated at that hospital during the fiscal year. Hospitals were divided into quartile performance tiers based on the total number of lives saved, with the top quartile (tier 1) hospitals saving the most lives and tier 4 the fewest.

We chose to rank hospitals based on the number of lives saved instead of risk-adjusted mortality rates for several reasons. First, for a hospital to save lives, the hospital must have a below average (better) RAMR because a hospital can only save lives by definition if the number of observed deaths is smaller than its expected number of deaths adjusted for the demographic characteristics, comorbidities, and cardiac risk factors of its CABG patients. Second, the better a hospital's actual performance relative to the average hospital, the more lives it will save with any given set of patients. Likewise, between 2 hospitals with the same number of observed deaths, the hospital with the more difficult case mix will save more lives than the hospital with the less difficult case mix because increasing disease severity increases the hospital's expected number of deaths. Alternatively, between 2 hospitals with the same RAMR, the lives saved method will rank the hospital that treated the most patients higher than the hospital that treated fewer patients.

OVERALL AND SEX-SPECIFIC HOSPITAL RAMR

The primary outcome of interest in this study is the overall and sex-specific hospital RAMR among Medicare beneficiaries. We adopted the linear method of calculating a hospital's RAMR used in the Patient Safety Indicator software written by the Agency for Healthcare Research and Quality.18 A hospital's overall RAMR was calculated by subtracting that hospital's expected mortality rate from its observed rate and adding the difference to the national average mortality rate for all beneficiaries undergoing CABG surgery. We report the sex differential in RAMR as the male rate minus the female rate. As a result, a negative sex differential implies that the RAMR for women is higher than for men. We report RAMR by performance tier as the unweighted average of each hospital's overall or sex-specific RAMR in a given performance tier. Nearly identical results were obtained by calculating a patient-weighted average RAMR (data not shown).

RESULTS

The final study population consisted of 134 407 Medicare beneficiaries and 802 hospitals in FY2003 and 122 231 Medicare beneficiaries and 768 hospitals in FY2004 (Table 1). While excluded hospitals represent 25% and 31% of all hospitals performing CABG surgery on beneficiaries in the respective fiscal years, the excluded hospitals averaged fewer than 37 CABG patients per hospital. As a result, the final study population included approximately 93% and 91% of all Medicare beneficiaries undergoing CABG surgery in FY2003 and FY2004, respectively.

Table 2 lists the multivariate estimated odds ratios and 95% confidence intervals for each of the variables included in the logistic regression model. Overall, there were very few differences in the estimated odds ratios for individual variables between FY2003 and FY2004; the goodness of fit, as measured by the C statistic, was approximately 0.78 in both years.

The number of estimated lives saved in study hospitals ranged from a +15.91 to a −21.86 during FY2003 and from +15.06 to −15.82 during FY2004 (Table 3). As expected, the estimated number of lives saved decreased from tier 1 to tier 4; however, the average CABG volume per hospital per year was higher in tier 1 and tier 4 than in the middle 2 tiers.

Overall, there was a slight decline in the average hospital observed mortality rate (3.72% vs 3.63%) and RAMR (3.68% vs 3.61%) among Medicare beneficiaries undergoing CABG surgery from FY2003 to FY2004. In addition, compared with female beneficiaries, male beneficiaries had lower observed mortality rates (3.7% vs 4.8% in FY2003 and 3.1% vs 4.7% in FY2004) and RAMRs (3.2% vs 4.7% in FY2003 and 3.1% vs 4.7% in FY2004). As a result, the sex differential (risk-adjusted mortality advantage to male patients) for all hospitals changed little (from 1.5% to 1.6%).

Table 4 lists mean hospital RAMRs overall and by sex for each performance tier. The following observations can be made: First, the mean hospital RAMR increases consistently, both overall and for male and female beneficiaries, from the top performance tier to the bottom in both fiscal years. While not listed in Table 4, the mean overall and sex-specific RAMR achieved by tier 1 hospitals was significantly lower than the comparable RAMR achieved by hospitals in any of the other 3 tiers in both fiscal years (P < .001).

Second, the sex differential in the average RAMR for all hospitals increases from the top to the bottom performance tier. For example, in FY2003, among top-tier hospitals, the average male RAMR was 1.24% compared with the average female RAMR of 1.96% (−0.72 percentage point differential), while among bottom-tier hospitals, the average male RAMR was 5.68% compared with the average female rate of 8.40% (−2.72 percentage point differential). The increase in the average hospital RAMR sex differential was statistically significant between tier 1 hospitals and hospitals in each of the other 3 performance tiers (P < .05).

Finally, while the risk-adjusted mortality rates for both sexes vary slightly in any performance tier between the 2 fiscal years, the average sex differential in RAMR in each of the performance tiers is very stable.

POTENTIAL MORTALITY IMPACT

Table 5 details the potential mortality impact of improving outcomes in bottom-tier hospitals. The simulation estimates the number of deaths avoided if bottom-tier hospitals could improve their sex-specific RAMR to the average sex-specific rate achieved in top-tier hospitals. Overall, this simulation indicates that between 74% and 76% of all observed deaths in bottom-tier hospitals could potentially be avoided—a total of roughly 1500 deaths per year. Finally, while women have a higher RAMR in tier 4 hospitals than men, both sexes benefit nearly equally from the improvement in hospital outcomes with very similar rates of potentially avoidable mortality, in part because a sex differential remains, even in top-tier hospitals.

COMMENT

This study examined 3 questions concerning hospital performance rankings for CABG surgery. First, this study found that when hospitals were ranked according to the number of lives saved, Medicare beneficiaries of either sex undergoing CABG surgery in the typical top-tier hospital were significantly less likely to experience in-hospital mortality than beneficiaries being treated by the typical hospital in the other 3 tiers. The difference in RAMR between the average top- and bottom-tier hospital is quite striking. The relative risk of mortality for undergoing CABG surgery in a bottom-tier hospital was 4.4 times that of a top-tier hospital.

Second, this study found that women had a higher mortality rate than men by 1.5 percentage points. However, the sex differential increases from 0.7 percentage points in the average top-tier hospital to 2.7 percentage points in the average bottom-tier hospital.

Finally, potentially 75% of the observed deaths of either male or female Medicare beneficiaries treated in tier 4 hospitals could be avoided if tier 4 hospitals could improve their performance to the average performance of top-tier hospitals. Clearly, to achieve these simulated reductions in mortality, the causes of interhospital variation in mortality must be identified.

These findings raise a number of questions. First, is the statistically significant difference in average RAMR between hospital tiers clinically meaningful? We believe that a 4-fold difference in RAMR is a clinically meaningful difference between hospitals in the top and bottom tiers. Approximately 1500 deaths that occurred in tier 4 hospitals could potentially have been avoided in each fiscal year if hospitals in the bottom tier achieved outcomes similar to the average hospital in tier 1. While there will always be bottom-tier hospitals in any performance ranking, this study demonstrates the potential mortality impact of improving the quality of care. Better understanding of what the nearly 200 top-performing hospitals are doing to achieve their relatively low average RAMR may provide insight into how to improve the average performance in bottom-tier hospitals.

Second, this study also finds that even in the best performing hospitals there is a sex differential in mortality rates. This finding may support those who argue for clinical reasons (eg, artery size) why mortality equality between men and women may be difficult to achieve among patients undergoing CABG surgery.1923 However, after risk adjusting, there was a nearly 4-fold increase in the sex differential between the top- and bottom-tier hospitals. While we are aware of the limitations associated with RAMR models for predicting subsequent performance, the size of the increase in the sex differential between top- and bottom-tier hospitals clearly demonstrates the importance of examining hospital CABG outcomes, especially for various subpopulations.

Third, these analyses suggest that both male and female Medicare beneficiaries undergoing elective CABG surgery could improve their outcomes by carefully selecting their CABG hospital. The proportional consistency of the sex-based mortality differences across tiers implies that the quality problems of poor performers apply equally to male and female patients, especially since the difference in average RAMR between top performing and lower performing hospitals is greater than the sex difference in RAMR in the 2 bottom tiers. However, for beneficiaries to be able to exercise this choice, the following conditions must exist: (1) hospitals’ rankings must be relatively stable over time and (2) beneficiaries must be able to access surgeons operating in top-tier hospitals and have financial and geographical access to top-tier hospitals. A brief examination of geographical access indicates that 37 states had at least 1 tier 1 hospital in FY2003 and that 40 states had at least 1 tier 1 hospital in FY2004. Furthermore, 20 states had at least 1 hospital ranked in the top tier and 1 in the bottom tier in both fiscal years.

Several limitations to this study must be acknowledged. First, the analyses and risk-adjustment models used to calculate RAMRs by sex are based on information contained in an administrative database.24,25 As a result, the risk-adjustment models we estimated might not adequately control for variations in patients' severity of illness because of a lack of clinical detail: the MEDPAR file is limited to 9 diagnosis codes per patient. Nevertheless, previous studies have shown that the predictive power of administrative databases, as measured by the C statistic, is nearly identical to that of clinical data sets for predicting the probability of death in a group of patients.26 Missing clinical data would not bias the results unless the missing factors were disproportionately distributed across Medicare beneficiaries and sex. The results of our risk-adjusted mortality models compare favorably to the results reported in the literature.610

A second limitation is that this study uses in-hospital mortality rates as opposed to 30-day or 6-month mortality rates. For CABG surgery outcomes, we believe that in-hospital mortality is appropriate because studies examining long-term survival rates have shown that very few CABG patients who survive to discharge die in the first year after surgery.2729

A third limitation is that low-volume hospitals may not treat enough patients of each sex (especially women) to provide statistically meaningful and stable rankings. However, the lives saved method of ranking hospitals makes it difficult for low-volume hospitals to be ranked in either the top or bottom tier. And finally, this study population included only Medicare beneficiaries undergoing CABG without heart valve surgery and may not represent hospital performance for all CABG patients treated in US hospitals.

In conclusion, this study provides evidence that Medicare beneficiaries should pay attention to hospital performance rankings for CABG surgery based on lives saved. We also suggest that improving the quality of care in poorer-performing hospitals may be an important opportunity for improving CABG surgery outcomes. The difference in average RAMR across hospital tiers is substantially larger than the overall reduction in CABG mortality that has been achieved over the last decade with clinical advances in performing the procedure. The sex differential is smallest at high-performing hospitals meaning that both male and female beneficiaries benefit from having CABG performed at tier 1 hospitals. Future research is needed to focus on the processes and structures that differentiate the top performers and how these lessons can be translated to other hospitals.

Back to top
Article Information

Correspondence: Steven D. Culler, PhD, Rollins School of Public Health, Emory University, 1518 Clifton Rd NE, Atlanta, GA 30322 (sculler@sph.emory.edu).

Accepted for Publication: May 19, 2008.

Author Contributions:Study concept and design: Culler, Simon, Brown, and Reynolds. Acquisition of data: Culler and Simon. Analysis and interpretation of data: Culler, Kugelmass, and Rask. Drafting of the manuscript: Culler, Simon, and Brown. Critical revision of the manuscript for important intellectual content: Culler, Kugelmass, Reynolds, and Rask. Statistical analysis: Culler and Reynolds. Obtained funding: Simon. Administrative, technical, and material support: Culler, Brown, Kugelmass, and Rask. Study supervision: Simon, Kugelmass, and Reynolds.

Financial Disclosure: None reported.

Funding/Support: Ms Simon, in her capacity as president of Cardiac Data Solutions Inc, Atlanta, Georgia, purchased the MEDPAR files analyzed in this study.

References
1.
Steinbrook  R Public report cards: cardiac surgery and beyond. N Engl J Med 2006;355 (18) 1847- 1849
PubMedArticle
2.
Hannan  ELSarrazin  MSDoran  DRRosenthal  GE Provider profiling and quality improvement efforts in coronary artery bypass graft surgery: the effect on short-term mortality among Medicare beneficiaries. Med Care 2003;41 (10) 1164- 1172
PubMedArticle
3.
Hartz  AJKuhn  EM Comparing hospitals that perform coronary artery bypass surgery: the effect of outcome measures and data sources. Am J Public Health 1994;84 (10) 1609- 1614
PubMedArticle
4.
Shahian  DM Improving cardiac surgery quality: volume, outcomes, process? JAMA 2004;291 (2) 246- 248
PubMedArticle
5.
Krumholz  HM Mathematical models and the assessment of performance in cardiology. Circulation 1999;99 (16) 2067- 2069
PubMedArticle
6.
New York State Department of Health,Coronary Artery Bypass Graft Surgery in New York State 1989-1991. Albany, NY New York State Dept of Health1992;
7.
Pennsylvania Health Care Cost Containment Council, A Consumer Guide to Coronary Artery Bypass Graft Surgery. Vol 4. Harrisburg Pennsylvania Health Care Cost Containment Council1995;
8.
Clifton  CRCardiac Surgery in 2000 in New Jersey: A Consumer Report. Trenton New Jersey Dept of Health and Senior Services2003;
9.
Parker  JPLi  ZDamberg  CL  et al. The California Report on Coronary Artery Bypass Graft Surgery: 2000-2002 Hospital Data. San Francisco, CA California Office of Statewide Health Planning and Development and the Pacific Business Group on Health2005;
10.
Naylor  CDRothwell  DMTu  JV  et al. Cardiac Care Network Steering Committee, Outcomes of coronary artery bypass graft surgery in Ontario. Naylor  CDSlaughter  PMCardiovascular Health Services in Ontario: An ICES Atlas. Toronto, ON Institute for Clinical Evaluative Sciences1999;189- 197
11.
Shahian  DMTorchiana  DFNormand  SL Implementation of a cardiac surgery report card: lessons from the Massachusetts experience. Ann Thorac Surg 2005;80 (3) 1146- 1150
PubMedArticle
12.
Marshall  MNRomano  PS Impact of reporting hospital performance. Qual Saf Health Care 2005;14 (2) 77- 78
PubMedArticle
13.
Werner  RMAsch  DA The unintended consequences of publicly reporting quality information. JAMA 2005;293 (10) 1239- 1244
PubMedArticle
14.
Vaccarino  VAbramson  JLVeledar  EWeintraub  WS Sex differences in hospital mortality after coronary artery bypass surgery: evidence for a higher mortality in younger women. Circulation 2002;105 (10) 1176- 1181
PubMedArticle
15.
O’Rourke  DJMalenka  DJOlmstead  EM  et al.  Improving in-hospital mortality in women undergoing coronary artery bypass grafting. Ann Thorac Surg 2001;71 (2) 507- 511
PubMedArticle
16.
Hannan  ELBernard  HRKilburn  HC  et al.  Gender differences in mortality rates for coronary artery bypass surgery. Am Heart J 1992;123 (4, pt 1) 866- 872
PubMedArticle
17.
Department of Health and Human Services, Centers for Medicare and Medicaid Services, Medicare Provider Analysis and Review (MEDPAR) file. http://www.cms.hhs.gov/IdentifiableDataFiles/05_MedicareProviderAnalysisandReviewFile.asp. Accessed May 5, 2008
18.
Agency for Healthcare Research and Quality, Patient safety indicators. http://www.qualityindicators.ahrq.gov/psi_download.htm. Accessed April 20, 2008
19.
Penckofer  SMHolm  K Women undergoing coronary bypass surgery: physiological and psychosocial perspectives. Cardiovasc Nurs 1990;26 (3) 13- 18
PubMed
20.
Richardson  JVCyrus  RJ Reduced efficacy of coronary artery bypass grafting in women. Ann Thorac Surg 1986;42 (6) ((suppl)) S16- S21
PubMedArticle
21.
Gansera  BGillrath  GLieber  MAngelis  ISchimdtler  FKemkes  BM Are men treated better than women? outcome of male versus female patients after CABG using bilateral internal thoracic arteries. Thorac Cardiovasc Surg 2004;52 (5) 261- 267
PubMedArticle
22.
Edwards  MLAlbert  NMWang  CApperson-Hanson  C 1993-2003 Gender differences in coronary artery revascularization: has anything changed? J Cardiovasc Nurs 2005;20 (6) 461- 467
PubMedArticle
23.
O’Connor  GTMorton  JRDiehl  MJ  et al.  Differences between men and women in hospital mortality associated with coronary artery bypass graft surgery. Circulation 1993;88 (5, pt 1) 2104- 2110
PubMedArticle
24.
Pine  MNorusis  MJones  BRosenthal  GE Predictions of hospital mortality rates: a comparison of data sources. Ann Intern Med 1997;126 (5) 347- 354
PubMedArticle
25.
Iezzoni  LIAsh  ASShwartz  M  et al.  Judging hospitals by severity adjusted method. Am J Public Health 1996;86 (10) 1379- 1387
PubMedArticle
26.
Austin  PCTu  JVAlter  DANaylor  CD The impact of under coding of cardiac severity and comorbid diseases on the accuracy of hospital report cards. Med Care 2005;43 (8) 801- 809
PubMedArticle
27.
Weintraub  WSBecker  ERMauldin  PDCuller  SKosinski  ASKing  SB Cost of revascularization over eight years in the randomized and eligible patients in the Emory Angioplasty versus Surgery Trial (EAST). Am J Cardiol 2000;86 (7) 747- 752
PubMedArticle
28.
Zhang  ZSpertus  JAMahoney  EM  et al.  The impact of acute coronary syndrome on clinical, economic, and cardiac-specific health status after coronary artery bypass surgery versus stent-assisted percutaneous coronary intervention: 1-year results from the Stent or Surgery (SoS) trial. Am Heart J 2005;150 (1) 175- 181
PubMedArticle
29.
Clark  RE Calculating risk and outcome: the Society of Thoracic Surgeons database. Ann Thorac Surg 1996;62 (5) ((suppl)) S2- S5
PubMedArticle
×