- Email: [email protected]
The Society of Thoracic Surgeons Adult Cardiac Surgery Database: 2017 Update on Research
The Society of Thoracic Surgeons Adult Cardiac Surgery Database: 2017 Update on Research Vinod H. Thourani, MD, Vinay Badhwar, MD, David M. Shahian, MD, Fred H. Edwards, MD, Sean O’Brien, PhD, Robert H. Habib, PhD, John J. Kelly, BA, J. Scott Rankin, MD, Richard Prager, MD, and Jeffrey P. Jacobs, MD Division of Cardiothoracic Surgery, Emory University, Atlanta, Georgia; Division of Cardiothoracic Surgery, West Virginia University, Morgantown, West Virginia; Department of Surgery and Center for Quality and Safety, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts; Division of Cardiothoracic Surgery, University of Florida, Jacksonville, Florida; The Society of Thoracic Surgeons Research Center, Chicago, Illinois; Department of Cardiac Surgery, University of Michigan, Ann Arbor, Michigan; and Division of Cardiovascular Surgery, Department of Surgery, Johns Hopkins All Children’s Heart Institute, Johns Hopkins All Children’s Hospital and Florida Hospital for Children, Saint Petersburg, Tampa, and Orlando, Florida, and Division of Cardiac Surgery, Department of Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland
Containing more than 6 million cumulative operative records and accounting for 90% to 95% of adult cardiac surgery performed in the United States, The Society of Thoracic Surgeons Adult Cardiac Surgery Database is an invaluable resource for performance assessment, quality improvement, and clinical research. This article reviews the seven major research efforts published in 2016 that utilized the Adult Cardiac Surgery Database. Two studies evaluated national trends in clinical
practice, three assessed the effect of several risk factors on postoperative morbidity and mortality, and two developed new models to evaluate quality of care. The findings of these studies have enhanced clinical practice and delineated areas for future quality improvement research.
I
predicted risk of mortality (PROM) for major surgical procedures, including coronary artery bypass graft surgery (CABG), and aortic and mitral valve procedures [2, 6–11]. These models are maintained on the STS website for use by surgeons, cardiologists, and patients, and are regularly updated to reflect contemporary data. More recently, the STS QMTF has developed composite performance measures of risk-adjusted mortality and morbidity for major cardiac surgical procedures as a more comprehensive metric of quality of care [3, 6–11]. In addition to the site-specific quality measurement and improvement efforts of the ACSD, its comprehensiveness and accuracy also make the registry an invaluable tool for clinical research affecting changes in practice at the national level. All STS members can submit protocols for clinical research utilizing the ACSD. Collected data include demographics, baseline comorbidities, procedural details, and inhospital to 30day outcomes. Recently, the ACSD has been linked to the US Centers for Medicare and Medicaid Services claims data, making it possible to perform longer-term outcome analyses for patients 65 years of age or older [12]. This article reviews the major research efforts utilizing the ACSD that were published in 2016. Two studies analyzed patient outcomes in the ACSD to evaluate national trends in clinical practice [13, 14], three studies assessed the effect of several risk factors on
n 1989, The Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Database (ACSD) was developed to provide a more reliable performance assessment by evaluating postoperative outcomes in the context of patient demographics and risk factors [1, 2]. Over the ensuing 3 decades, the ACSD has expanded into the most comprehensive cardiac surgery registry in the world [3]. As of 2017, the database includes more than 6 million cumulative operations performed by 3,017 surgeons from 1,115 practice groups. The ACSD is highly representative of contemporary national trends in cardiac surgery, accounting for 90% to 95% of adult cardiac surgery performed in the United States [4]. The Duke Clinical Research Institute maintains the ACSD, and rigorous internal and external auditing is regularly performed to assure reliability and completeness [2, 3, 5]. Biannual reports are provided by the institute to participating sites that compare individual performance to national benchmarks as determined by review of the ACSD [3, 4]. In collaboration with the Duke Clinical Research Institute, the STS Quality Measurement Task Force (QMTF) utilized the ACSD to develop calculators for
Address correspondence to Dr Thourani, Emory Midtown Hospital, 550 Peachtree St, Medical Office Tower, 6th Flr, Atlanta, GA 30308; email: [email protected].
Ó 2017 by The Society of Thoracic Surgeons Published by Elsevier Inc.
(Ann Thorac Surg 2017;-:-–-) Ó 2017 by The Society of Thoracic Surgeons
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2017.05.013
2
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
Abbreviations and Acronyms ACSD = Adult Cardiac Surgery Database BMVR = bioprosthetic mitral valve replacement CABG = coronary artery bypass graft surgery CI = confidence interval MCS = mechanical circulatory support OR = odds ratio PROM = predicted risk of mortality QMTF = Quality Measurement Task Force RA = robotic-assisted STS = The Society of Thoracic Surgeons
postoperative morbidity and mortality [15–17], and two studies developed new models to evaluate quality of care [18, 19].
Studies Analyzing National Trends in Clinical Practice One recent study by Whellan and colleagues [13] utilized the ACSD to investigate national trends in the use and outcomes of robotic-assisted (RA) CABG [13]. Roboticassisted CABG was defined as using robotic grafting or left internal mammary artery harvesting, and included totally endoscopic CABG, minimally invasive direct CABG, and RA-CABG as a component of hybrid coronary revascularization. The investigators identified all nonemergent, elective, isolated CABG procedures utilizing a left internal mammary artery from 2006 to 2012 at 719 sites in the ACSD. Baseline characteristics and outcomes of RA-CABG procedures (n ¼ 9,862) were compared with those not using RA-CABG (n ¼ 956,349). Additionally, the use of RA-CABG within each site during the study period was analyzed. The number of sites performing RA-CABG was 148 in 2006 and 151 in 2012, and most performed between one and five procedures per year (Fig 1). Only 0.59% of all CABG operations utilized robotics in 2006, with a Fig 1. Robot-assisted coronary artery bypass graft surgery (RACABG) volume by site from 2006 to 2012. (Red bars ¼ 1 to 5 cases; green bars ¼ 6 to 10 cases; dark blue bars ¼ 11 to 20 cases; turquoise bars ¼ 21 to 40 cases; orange bars ¼ 41 to 80 cases; light blue bars ¼ more than 80 cases.)
Ann Thorac Surg 2017;-:-–-
slight increase to 0.97% in 2012. Patients undergoing robotic procedures were slightly younger (64 versus 65 years, p < 0.0001) and had lower rates of baseline comorbidities. Cardiopulmonary bypass was used less frequently with robotic procedures (22.4% versus 80.4%, p < 0.0001), and fewer grafts were used (median 1.0 versus 3.0, p < 0.0001). After RA-CABG, major complications occurred less frequently (10.2% versus 13.5%, p < 0.0001) and, as expected, the length of postoperative admission was shorter (4 versus 5 days, p < 0.0001) compared with nonrobotic procedures. Operative mortality (defined as death within the postoperative hospitalization or after discharge within 30 days of CABG) was not significantly different after adjustment using the STS risk model (odds ratio [OR] 1.10; 95% confidence interval [CI]: 0.93 to 1.30, p ¼ 0.29). Moreover, the mortality of RA-CABG was comparable between sites with high volume (more than 20 RA-CABG per year) and low volume (20 or fewer RA-CABG per year), 1.1% versus 1.3% (p ¼ 0.51). That RA-CABG was consistently performed at low volumes and at a limited number of sites during the study period despite similar mortality and decreased morbidity compared with nonrobotic procedures merits further study. Moreover, given the improvements in robotic technology since 2012 and the accompanied resurgence in robotic cardiac operations that have occurred after the study period of this analysis, it would be most interesting to determine how this new technology can be used in contemporary practice to optimize outcomes and resource utilization. A National Heart, Lung, and Blood Institute-sponsored trial comparing hybrid CABG revascularization with multivessel stenting is soon to begin and may shed light on this novel procedure. Another study, by Schwann and colleagues [14], evaluated the patient characteristics and national variability of surgeon and hospital practice associated with the prescription of warfarin after bioprosthetic mitral valve replacement (BMVR). This analysis included 7,637 patients in the ACSD undergoing isolated, primary, nonemergent BMVR from 2008 to 2011. Patients with
Ann Thorac Surg 2017;-:-–-
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
3
preoperative warfarin or atrial fibrillation, contraindications to warfarin, and those undergoing mechanical mitral valve replacement were excluded. The median age of the cohort was 66 years and 58.7% were female. Only 58.0% were prescribed warfarin after BMVR. The patients receiving warfarin were slightly older (67 versus 65 years, p < 0.0001) and had significantly lower rates of preoperative comorbidities, including prior stroke, history of smoking, congestive heart failure, and dialysis. Patients with postoperative gastrointestinal events were less likely to receive warfarin (44.4% versus 55.6%, p < 0.001), whereas warfarin was prescribed more frequently for patients who had a postoperative myocardial infarction (75.8% versus 24.2%, p < 0.001), new atrial fibrillation after surgery (68% versus 32%, p < 0.001), intraoperative blood transfusions (55.7% versus 44.3%, p < 0.001), and postoperative blood transfusions (57% versus 43%, p < 0.03). In addition, patients prescribed warfarin had a longer postoperative hospital stay (median 8.0 versus 7.0 days, p < 0.01). No difference in the rates of warfarin prescription was observed among patients with postoperative stroke (53.6% versus 46.4%, p ¼ 0.30) or among those requiring reoperation for bleeding (54.9% versus 45.1%, p ¼ 0.20). Interestingly, the investigators reported significant regional variation among hospitals and surgeons in warfarin prescription after BMVR, with rates highest in the Northeastern United States (68.9%, p ¼ 0.12) and lowest in the South (49.7%; p ¼ 0.003) when compared with the West. Furthermore, analysis by individual surgeon and hospital after adjustment for patient characteristics revealed a bimodal distribution of warfarin prescription, in which either very few or most patients received warfarin after BMVR (Figs 2 and 3). Such variability reflects the lack of clear evidence that warfarin reduces thromboembolic events after isolated BMVR, and suggests the need for prospective, randomized clinical investigation to inform the appropriate anticoagulation strategy for this patient population.
Studies Assessing Risk Factors Acharya and colleagues [15] utilized the ACSD to investigate outcomes of CABG in patients with a recent acute myocardial infarction complicated by cardiogenic shock [15]. All patients undergoing nonelective CABG or CABG with ventricular assist device implantation within 7 days of acute myocardial infarction complicated by cardiogenic shock were included. A detailed analysis of baseline characteristics, operative data, and clinical outcomes in this high-risk cohort was performed from 2011 to 2013. A total of 5,496 patients was studied; the median age was 66 years and 71.7% were male. Surgery was classified as urgent in 31.6% of cases, emergent in 60.8%, and salvage in 7.7%. The status of mechanical circulatory support (MCS), defined as the use of devices other than an intraaortic balloon pump, was used to categorize patients as preoperative MCS (n ¼ 129; 2.3%), intraoperative or postoperative MCS (n ¼ 279; 5.1%), or no MCS (n ¼ 5,088; 92.6%).
Fig 2. Distribution of discharge warfarin use across hospitals: (A) all hospitals (n ¼ 886); (B) hospitals with more than 10 bioprosthetic mitral valve replacement cases annually (n ¼ 245). (No. ¼ number.)
Operative mortality was 18.7% for the entire cohort, decreasing from 19.3% in 2005 to 18.1% in 2013 (p < 0.001; Fig 4). Urgent, emergent, and salvage procedures had operative mortality rates of 10.3%, 18.6%, and 53.3%, respectively. Patients who underwent salvage operations also had a high incidence of postoperative respiratory failure requiring prolonged mechanical ventilation (59.7%), reoperation (21.8%), renal failure (18.5%), and multisystem organ failure (14.7%). From 2011 to 2013, the use of MCS increased from 5.8% to 8.8% (p ¼ 0.008). Patients without MCS, patients requiring preoperative MCS, and patients requiring intraoperative or postoperative MCS had operative mortality of 16.0%, 37.2%, and 58.4%, respectively. The researchers concluded that most patients undergoing CABG in the setting of acute myocardial infarction
4
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
Ann Thorac Surg 2017;-:-–-
Fig 4. Temporal trends in operative mortality for isolated coronary artery bypass graft surgery, with or without ventricular assist device.
Fig 3. Distribution of discharge warfarin rates across surgeons: (A) all surgeons (n ¼ 1,774); (B) surgeons with more than 10 bioprosthetic mitral valve replacement cases annually (n ¼ 172). (No. ¼ number.)
complicated by cardiogenic shock have a substantial, but not prohibitive operative risk. Although the current STS model accurately predicted risk of mortality for patients without MCS, it underestimated risk in patients with MCS, especially those requiring postcardiotomy MCS. However, it must be noted that the current risk models were never intended to predict risk in patients who required MCS during or after surgery, as they are based only on the patient’s comorbid status on entering the operating room. Therefore, with patients for whom the need for MCS may be anticipated, careful assessment of the preoperative risks and benefits of surgery is required for optimal outcomes. A study by Afilalo and associates [16] investigated whether existing risk models could be improved by adding the 5-m gait speed test as a measure of frailty in
assessing the perioperative risk of elderly patients considered for cardiac surgery. This analysis included 15,171 patients aged 60 years or older in the ACSD who underwent CABG (n ¼ 9,005; 59.4%), isolated aortic or mitral valve surgery (n ¼ 3,765; 24.8%), or CABG combined with aortic or mitral valve surgery (n ¼ 2,401; 15.8%) between 2011 and 2014. Patients were divided into three groups by gait speed: fast (more tan 1.00 m/s), middle (0.83 to 1.00 m/s), and slow (less than 0.83 m/s). The primary outcome was allcause operative mortality. The STS composite of major morbidity and mortality was used as the secondary outcome, which included operative mortality, stroke, reoperation, prolonged ventilation, sternal wound infection, and acute kidney injury. The researchers reported that compared with the fastest gait speed group, unadjusted operative mortality was increased in the middle gait speed group (OR 1.77, 95% CI: 1.34 to 2.34) and highest in the slow gait speed group (OR 3.16, 95% CI: 2.31 to 4.33; Fig 5). Interestingly, gait speed remained an independent predictor of mortality after adjustment for the STS PROM and type of surgical procedure (OR 1.11 per 0.1 m/s decrease in gait speed, 95% CI: 1.07 to 1.16). Gait speed was also predictive of the composite outcome after adjustment (OR 1.03 per 0.1-m/s decrease in gait speed, 95% CI: 1.00 to 1.05). The addition of gait speed to the STS PROM model only improved the C-index by 0.005, however, and the integrated discrimination improvement was only 0.003, which the researchers concluded was likely due to the good baseline STS PROM model. Although the C-index did not change much, the predicted risk of mortality increased substantially for slower gait speed. We currently are limited in our ability to use gait speed in the STS predictive models because of the lack of completion in this data variable from the participating sites. Overall, the 5-m gait speed is an objective, useful screening test for frailty in older adults that may provide a better indication of operative risk than age alone. This study highlights the importance of frailty assessment in the accurate prediction of operative risk in elderly patients, which is essential in helping the patient and medical team select an appropriate treatment that optimizes survival and quality of life. These data may serve as a benchmark for
Ann Thorac Surg 2017;-:-–-
Fig 5. Unadjusted association between gait speed and operative mortality. Decreasing gait speed was associated with increasing odds of operative mortality. Blue shaded area indicates 95% confidence interval; dashed line indicates reference odds ratio of 1.0. (m/s ¼ meters per second.)
future studies evaluating newer transcatheter or surgical therapies, and their use by all ACSD participants should be encouraged for these patients. Another risk factor that has been historically associated with adverse outcomes after CABG is African-American race. In a recent study, Mehta and colleagues [17] compared 11,697 African-American patients with 136,362 white patients undergoing isolated CABG between 2010 and 2011 at 663 sites in the ACSD [17]. African-American patients undergoing CABG were significantly younger, more likely to be female, had a higher body mass index, and greater prevalence of comorbidities including hypertension, diabetes mellitus, smoking, peripheral vascular disease, cerebrovascular disease, use of immunosuppressive drugs, preoperative dialysis, myocardial infarction within 7 days of CABG, congestive heart failure, moderate to severe mitral regurgitation, and lower left ventricular ejection fraction compared with white patients. In addition, African-American patients had marginally reduced utilization of the internal mammary artery during CABG (93.3% versus 92.2%, p < 0.0001), but received perioperative medications for secondary prevention more frequently. African-American patients were also more likely to receive treatment in hospitals with higher risk-adjusted mortality (Fig 6). The unadjusted operative mortality rate was 2.5% for African-American patients and 1.8% for white patients (p < 0.0001), and the unadjusted major morbidity rate was 19.4% for African-American patients and 13.6% for white patients (p < 0.0001). After adjusting for baseline patient characteristics, income, hospitals, surgeons, and medications, mortality and morbidity were still significantly higher for African-American patients than for white patients (OR 1.17, 95% CI: 1.00 to 1.36, and OR 1.26, 95% CI: 1.19 to 1.34, respectively). Subgroup and propensity matched analysis confirmed these findings. That AfricanAmerican race remained an independent predictor of mortality and morbidity after current risk adjustment suggests other risk factors that have yet to be identified. Further study of the medical, biological (including genetic), social, and economic issues facing these patients is urgently needed to address the continued racial differences in CABG outcomes.
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
5
Fig 6. Proportion of black patients undergoing coronary artery bypass graft surgery at institutions ranked top to bottom based on increasing risk-adjusted mortality. Centers in quartile 1 (0% to 25%) had the lowest risk-adjusted operative mortality, whereas centers in quartile 4 (75% to 100%) had the highest risk-adjusted operative mortality. (vs ¼ versus.)
Studies Evaluating Quality of Care The STS ACSD has also been utilized to assess quality of care. Previously, the STS QMTF developed composite performance measures and risk models for isolated CABG, isolated aortic valve replacement, CABG with concurrent aortic valve replacement, and isolated mitral valve repair or replacement [6–11]. The fifth in this ongoing series was the development of a composite performance measure for mitral valve repair or replacement with concurrent CABG by reviewing outcomes of 24,740 patients undergoing mitral valve replacement or repair with concurrent CABG from 2011 to 2014 at 703 sites in the ACSD [18]. Sites with fewer than 10 cases over 3 years were excluded, whereas patients with concomitant atrial septal defect or patent foramen ovale closure, surgical ablation for atrial fibrillation, and tricuspid valve repair were included. The two-domain composite included riskadjusted operative mortality and the risk-adjusted occurrence of major complications including stroke, renal failure, reoperation, prolonged ventilation, and deep sternal infection. Hierarchical regression and 95% Bayesian credible intervals were used to calculate composite scores and assign star ratings for performance. The overall unadjusted operative mortality rate was 6.2%, and one or more major complications occurred in 30.8% (Fig 7). This composite measure classified 2% of sites (14 of 703) as one star (lower than expected performance), 95% (666 of 703) as two stars (as expected), and 3% (23 of 703) as three stars (higher than expected). Both mortality and morbidity declined as star rating scores increased. In recent years, failure to rescue (FTR), defined as mortality after the development of a complication, has gained increased attention as an alternative metric of quality. Whereas traditional measure of postoperative morbidity and mortality evaluate the procedure itself, FTR assesses perioperative institutional systems of care. Lower rates of FTR suggest that a facility has the staff,
6
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
Ann Thorac Surg 2017;-:-–-
Fig 8. Influence of number of complications on failure to rescue (mortality [blue bars]) and patient sample size (n [orange line]).
Fig 7. Distributions of risk-adjusted (A) mortality rate and (B) morbidity rate (n ¼ 703 participants). (IQR ¼ interquartile range.)
resources, and systems in place to adequately care for patients who have complications. A recent study by Edwards and colleagues [19] analyzed the outcomes of 604,154 patients in the ACSD undergoing isolated CABG from 2010 to 2014 at 1,105 centers to calculate rates of FTR and develop a model predictive of FTR after CABG. The investigators defined FTR as death after stroke, renal failure, reoperation, and prolonged ventilation. The FTR was calculated for each of these complications, as well as for a composite of the four complications. Individual rates of FTR for renal failure, stroke, reoperation, prolonged ventilation, and the composite were 22.3%, 16.4%, 12.4%, 12.1%, and 10.5%, respectively. The FTR increased in patients with more than one and certain combinations of complications (Fig 8). Multivariable logistic regression was used to develop a model to predict FTR, which had a C-index of 0.792 and was well calibrated with a 1.0% difference between observed to expected rates of FTR. When centers were grouped into tertiles by mortality, the complication rate increased only modestly from lowest to highest group (11.4% to 15.7%), whereas mortality varied directly with FTR, doubling from the lowest to highest group (6.8% to 13.9%). Interestingly, centers with the highest complication rates had the lowest FTR, whereas sites with the lowest complication rates had the highest FTR. Using the model developed in this study, the FTR observed to expected ratio was 1.14 in the lowest complication tertile and 0.91 in the highest complication tertile. The researchers postulated that perhaps tertiary
centers who are referred very complex and high-risk patients have complication rates that are correspondingly higher. However, these centers may also have the expertise, staffing, and advanced care technologies to salvage these patients when they do have complications. These findings will be further investigated by the STS QMTF [3]. The quality of care data reported by these two studies have provided national benchmarks that can be used to counsel patients, aid in clinical decision making, and encourage centers to reevaluate and restructure their practices to improve outcomes.
Conclusion Cardiac surgeons have remained on the forefront of the arduous task of self-reflection by continuing to reassess and refine their surgical techniques. One of the most important tools to serve this endeavor is the STS ACSD. With more than 6 million cumulative operative records, it remains the most comprehensive health care registry in the world. In addition to performance assessment and quality improvement, the ACSD is an invaluable resource for clinical research. The findings of studies published in 2016 that utilized this database have enhanced clinical practice and delineated areas for future quality improvement research.
References 1. Clark RE. It is time for a national cardiothoracic surgical data base. Ann Thorac Surg 1989;48:755–6. 2. Winkley Shroyer AL, Bakaeen F, Shahian DM, et al. The Society of Thoracic Surgeons Adult Cardiac Surgery Database: the driving force for improvement in cardiac surgery. Semin Thorac Cardiovasc Surg 2015;27:144–51. 3. D’Agostino RS, Jacobs JP, Badhwar V, et al. The Society of Thoracic Surgeons Adult Cardiac Surgery Database: 2017 update on outcomes and quality. Ann Thorac Surg 2017;103: 18–24. 4. Jacobs JP, Shahian DM, Prager RL, et al. The Society of Thoracic Surgeons National Database 2016 annual report. Ann Thorac Surg 2016;102:1790–7. 5. Shahian DM. The Society of Thoracic Surgeons National Database: “What’s past is prologue.” Ann Thorac Surg 2016;101:841–5.
Ann Thorac Surg 2017;-:-–-
6. Shahian DM, O’Brien SM, Filardo G, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 1— coronary bypass grafting surgery. Ann Thorac Surg 2009;88(Suppl):2–22. 7. O’Brien SM, Shahian DM, Filardo G, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 2— isolated valve surgery. Ann Thorac Surg 2009;88(Suppl):23–42. 8. Shahian DM, O’Brien SM, Filardo G, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 3— valve plus coronary artery bypass grafting surgery. Ann Thorac Surg 2009;88(Suppl):43–62. 9. Shahian DM, He X, Jacobs JP, et al. The Society of Thoracic Surgeons isolated aortic valve replacement (AVR) composite score: a report of the STS Quality Measurement Task Force. Ann Thorac Surg 2012;94:2166–71. 10. Shahian DM, He X, Jacobs JP, et al. The STS AVRþCABG composite score: a report of the STS Quality Measurement Task Force. Ann Thorac Surg 2014;97:1604–9. 11. Badhwar V, Rankin JS, He X, et al. The Society of Thoracic Surgeons mitral repair/replacement composite score: a report of The Society of Thoracic Surgeons Quality Measurement Task Force. Ann Thorac Surg 2016;101:2265–71. 12. Jacobs JP, Shahian DM, He X, et al. Penetration, completeness, and representativeness of The Society of Thoracic Surgeons Adult Cardiac Surgery Database. Ann Thorac Surg 2016;101:33–41. 13. Whellan DJ, McCarey MM, Taylor BS, et al. Trends in robotic-assisted coronary artery bypass grafts: a study of The
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
14.
15.
16. 17.
18.
19.
7
Society of Thoracic Surgeons Adult Cardiac Surgery Database, 2006 to 2012. Ann Thorac Surg 2016;102:140–6. Schwann TA, Habib RH, Suri RM, et al. Variation in warfarin use at hospital discharge after isolated bioprosthetic mitral valve replacement: an analysis of The Society of Thoracic Surgeons Adult Cardiac Surgery Database. Chest 2016;150: 597–605. Acharya D, Gulack BC, Loyaga-Rendon RY, et al. Clinical characteristics and outcomes of patients with myocardial infarction and cardiogenic shock undergoing coronary artery bypass surgery: data from The Society of Thoracic Surgeons National Database. Ann Thorac Surg 2016;101:558–66. Afilalo J, Kim S, O’Brien S, et al. Gait speed and operative mortality in older adults following cardiac surgery. JAMA Cardiol 2016;1:314–21. Mehta RH, Shahian DM, Sheng S, et al. Association of hospital and physician characteristics and care processes with racial disparities in procedural outcomes among contemporary patients undergoing coronary artery bypass grafting surgery. Circulation 2016;133:124–30. Rankin JS, Badhwar V, He X, et al. The Society of Thoracic Surgeons mitral valve repair/replacement plus coronary artery bypass grafting composite score: a report of The Society of Thoracic Surgeons Quality Measurement Task Force. Ann Thorac Surg 2017;103:1475–81. Edwards FH, Ferraris VA, Kurlansky PA, et al. Failure to rescue rates after coronary artery bypass grafting: an analysis from The Society of Thoracic Surgeons Adult Cardiac Surgery Database. Ann Thorac Surg 2016;102:458–64.
Containing more than 6 million cumulative operative records and accounting for 90% to 95% of adult cardiac surgery performed in the United States, The Society of Thoracic Surgeons Adult Cardiac Surgery Database is an invaluable resource for performance assessment, quality improvement, and clinical research. This article reviews the seven major research efforts published in 2016 that utilized the Adult Cardiac Surgery Database. Two studies evaluated national trends in clinical
practice, three assessed the effect of several risk factors on postoperative morbidity and mortality, and two developed new models to evaluate quality of care. The findings of these studies have enhanced clinical practice and delineated areas for future quality improvement research.
I
predicted risk of mortality (PROM) for major surgical procedures, including coronary artery bypass graft surgery (CABG), and aortic and mitral valve procedures [2, 6–11]. These models are maintained on the STS website for use by surgeons, cardiologists, and patients, and are regularly updated to reflect contemporary data. More recently, the STS QMTF has developed composite performance measures of risk-adjusted mortality and morbidity for major cardiac surgical procedures as a more comprehensive metric of quality of care [3, 6–11]. In addition to the site-specific quality measurement and improvement efforts of the ACSD, its comprehensiveness and accuracy also make the registry an invaluable tool for clinical research affecting changes in practice at the national level. All STS members can submit protocols for clinical research utilizing the ACSD. Collected data include demographics, baseline comorbidities, procedural details, and inhospital to 30day outcomes. Recently, the ACSD has been linked to the US Centers for Medicare and Medicaid Services claims data, making it possible to perform longer-term outcome analyses for patients 65 years of age or older [12]. This article reviews the major research efforts utilizing the ACSD that were published in 2016. Two studies analyzed patient outcomes in the ACSD to evaluate national trends in clinical practice [13, 14], three studies assessed the effect of several risk factors on
n 1989, The Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Database (ACSD) was developed to provide a more reliable performance assessment by evaluating postoperative outcomes in the context of patient demographics and risk factors [1, 2]. Over the ensuing 3 decades, the ACSD has expanded into the most comprehensive cardiac surgery registry in the world [3]. As of 2017, the database includes more than 6 million cumulative operations performed by 3,017 surgeons from 1,115 practice groups. The ACSD is highly representative of contemporary national trends in cardiac surgery, accounting for 90% to 95% of adult cardiac surgery performed in the United States [4]. The Duke Clinical Research Institute maintains the ACSD, and rigorous internal and external auditing is regularly performed to assure reliability and completeness [2, 3, 5]. Biannual reports are provided by the institute to participating sites that compare individual performance to national benchmarks as determined by review of the ACSD [3, 4]. In collaboration with the Duke Clinical Research Institute, the STS Quality Measurement Task Force (QMTF) utilized the ACSD to develop calculators for
Address correspondence to Dr Thourani, Emory Midtown Hospital, 550 Peachtree St, Medical Office Tower, 6th Flr, Atlanta, GA 30308; email: [email protected].
Ó 2017 by The Society of Thoracic Surgeons Published by Elsevier Inc.
(Ann Thorac Surg 2017;-:-–-) Ó 2017 by The Society of Thoracic Surgeons
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2017.05.013
2
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
Abbreviations and Acronyms ACSD = Adult Cardiac Surgery Database BMVR = bioprosthetic mitral valve replacement CABG = coronary artery bypass graft surgery CI = confidence interval MCS = mechanical circulatory support OR = odds ratio PROM = predicted risk of mortality QMTF = Quality Measurement Task Force RA = robotic-assisted STS = The Society of Thoracic Surgeons
postoperative morbidity and mortality [15–17], and two studies developed new models to evaluate quality of care [18, 19].
Studies Analyzing National Trends in Clinical Practice One recent study by Whellan and colleagues [13] utilized the ACSD to investigate national trends in the use and outcomes of robotic-assisted (RA) CABG [13]. Roboticassisted CABG was defined as using robotic grafting or left internal mammary artery harvesting, and included totally endoscopic CABG, minimally invasive direct CABG, and RA-CABG as a component of hybrid coronary revascularization. The investigators identified all nonemergent, elective, isolated CABG procedures utilizing a left internal mammary artery from 2006 to 2012 at 719 sites in the ACSD. Baseline characteristics and outcomes of RA-CABG procedures (n ¼ 9,862) were compared with those not using RA-CABG (n ¼ 956,349). Additionally, the use of RA-CABG within each site during the study period was analyzed. The number of sites performing RA-CABG was 148 in 2006 and 151 in 2012, and most performed between one and five procedures per year (Fig 1). Only 0.59% of all CABG operations utilized robotics in 2006, with a Fig 1. Robot-assisted coronary artery bypass graft surgery (RACABG) volume by site from 2006 to 2012. (Red bars ¼ 1 to 5 cases; green bars ¼ 6 to 10 cases; dark blue bars ¼ 11 to 20 cases; turquoise bars ¼ 21 to 40 cases; orange bars ¼ 41 to 80 cases; light blue bars ¼ more than 80 cases.)
Ann Thorac Surg 2017;-:-–-
slight increase to 0.97% in 2012. Patients undergoing robotic procedures were slightly younger (64 versus 65 years, p < 0.0001) and had lower rates of baseline comorbidities. Cardiopulmonary bypass was used less frequently with robotic procedures (22.4% versus 80.4%, p < 0.0001), and fewer grafts were used (median 1.0 versus 3.0, p < 0.0001). After RA-CABG, major complications occurred less frequently (10.2% versus 13.5%, p < 0.0001) and, as expected, the length of postoperative admission was shorter (4 versus 5 days, p < 0.0001) compared with nonrobotic procedures. Operative mortality (defined as death within the postoperative hospitalization or after discharge within 30 days of CABG) was not significantly different after adjustment using the STS risk model (odds ratio [OR] 1.10; 95% confidence interval [CI]: 0.93 to 1.30, p ¼ 0.29). Moreover, the mortality of RA-CABG was comparable between sites with high volume (more than 20 RA-CABG per year) and low volume (20 or fewer RA-CABG per year), 1.1% versus 1.3% (p ¼ 0.51). That RA-CABG was consistently performed at low volumes and at a limited number of sites during the study period despite similar mortality and decreased morbidity compared with nonrobotic procedures merits further study. Moreover, given the improvements in robotic technology since 2012 and the accompanied resurgence in robotic cardiac operations that have occurred after the study period of this analysis, it would be most interesting to determine how this new technology can be used in contemporary practice to optimize outcomes and resource utilization. A National Heart, Lung, and Blood Institute-sponsored trial comparing hybrid CABG revascularization with multivessel stenting is soon to begin and may shed light on this novel procedure. Another study, by Schwann and colleagues [14], evaluated the patient characteristics and national variability of surgeon and hospital practice associated with the prescription of warfarin after bioprosthetic mitral valve replacement (BMVR). This analysis included 7,637 patients in the ACSD undergoing isolated, primary, nonemergent BMVR from 2008 to 2011. Patients with
Ann Thorac Surg 2017;-:-–-
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
3
preoperative warfarin or atrial fibrillation, contraindications to warfarin, and those undergoing mechanical mitral valve replacement were excluded. The median age of the cohort was 66 years and 58.7% were female. Only 58.0% were prescribed warfarin after BMVR. The patients receiving warfarin were slightly older (67 versus 65 years, p < 0.0001) and had significantly lower rates of preoperative comorbidities, including prior stroke, history of smoking, congestive heart failure, and dialysis. Patients with postoperative gastrointestinal events were less likely to receive warfarin (44.4% versus 55.6%, p < 0.001), whereas warfarin was prescribed more frequently for patients who had a postoperative myocardial infarction (75.8% versus 24.2%, p < 0.001), new atrial fibrillation after surgery (68% versus 32%, p < 0.001), intraoperative blood transfusions (55.7% versus 44.3%, p < 0.001), and postoperative blood transfusions (57% versus 43%, p < 0.03). In addition, patients prescribed warfarin had a longer postoperative hospital stay (median 8.0 versus 7.0 days, p < 0.01). No difference in the rates of warfarin prescription was observed among patients with postoperative stroke (53.6% versus 46.4%, p ¼ 0.30) or among those requiring reoperation for bleeding (54.9% versus 45.1%, p ¼ 0.20). Interestingly, the investigators reported significant regional variation among hospitals and surgeons in warfarin prescription after BMVR, with rates highest in the Northeastern United States (68.9%, p ¼ 0.12) and lowest in the South (49.7%; p ¼ 0.003) when compared with the West. Furthermore, analysis by individual surgeon and hospital after adjustment for patient characteristics revealed a bimodal distribution of warfarin prescription, in which either very few or most patients received warfarin after BMVR (Figs 2 and 3). Such variability reflects the lack of clear evidence that warfarin reduces thromboembolic events after isolated BMVR, and suggests the need for prospective, randomized clinical investigation to inform the appropriate anticoagulation strategy for this patient population.
Studies Assessing Risk Factors Acharya and colleagues [15] utilized the ACSD to investigate outcomes of CABG in patients with a recent acute myocardial infarction complicated by cardiogenic shock [15]. All patients undergoing nonelective CABG or CABG with ventricular assist device implantation within 7 days of acute myocardial infarction complicated by cardiogenic shock were included. A detailed analysis of baseline characteristics, operative data, and clinical outcomes in this high-risk cohort was performed from 2011 to 2013. A total of 5,496 patients was studied; the median age was 66 years and 71.7% were male. Surgery was classified as urgent in 31.6% of cases, emergent in 60.8%, and salvage in 7.7%. The status of mechanical circulatory support (MCS), defined as the use of devices other than an intraaortic balloon pump, was used to categorize patients as preoperative MCS (n ¼ 129; 2.3%), intraoperative or postoperative MCS (n ¼ 279; 5.1%), or no MCS (n ¼ 5,088; 92.6%).
Fig 2. Distribution of discharge warfarin use across hospitals: (A) all hospitals (n ¼ 886); (B) hospitals with more than 10 bioprosthetic mitral valve replacement cases annually (n ¼ 245). (No. ¼ number.)
Operative mortality was 18.7% for the entire cohort, decreasing from 19.3% in 2005 to 18.1% in 2013 (p < 0.001; Fig 4). Urgent, emergent, and salvage procedures had operative mortality rates of 10.3%, 18.6%, and 53.3%, respectively. Patients who underwent salvage operations also had a high incidence of postoperative respiratory failure requiring prolonged mechanical ventilation (59.7%), reoperation (21.8%), renal failure (18.5%), and multisystem organ failure (14.7%). From 2011 to 2013, the use of MCS increased from 5.8% to 8.8% (p ¼ 0.008). Patients without MCS, patients requiring preoperative MCS, and patients requiring intraoperative or postoperative MCS had operative mortality of 16.0%, 37.2%, and 58.4%, respectively. The researchers concluded that most patients undergoing CABG in the setting of acute myocardial infarction
4
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
Ann Thorac Surg 2017;-:-–-
Fig 4. Temporal trends in operative mortality for isolated coronary artery bypass graft surgery, with or without ventricular assist device.
Fig 3. Distribution of discharge warfarin rates across surgeons: (A) all surgeons (n ¼ 1,774); (B) surgeons with more than 10 bioprosthetic mitral valve replacement cases annually (n ¼ 172). (No. ¼ number.)
complicated by cardiogenic shock have a substantial, but not prohibitive operative risk. Although the current STS model accurately predicted risk of mortality for patients without MCS, it underestimated risk in patients with MCS, especially those requiring postcardiotomy MCS. However, it must be noted that the current risk models were never intended to predict risk in patients who required MCS during or after surgery, as they are based only on the patient’s comorbid status on entering the operating room. Therefore, with patients for whom the need for MCS may be anticipated, careful assessment of the preoperative risks and benefits of surgery is required for optimal outcomes. A study by Afilalo and associates [16] investigated whether existing risk models could be improved by adding the 5-m gait speed test as a measure of frailty in
assessing the perioperative risk of elderly patients considered for cardiac surgery. This analysis included 15,171 patients aged 60 years or older in the ACSD who underwent CABG (n ¼ 9,005; 59.4%), isolated aortic or mitral valve surgery (n ¼ 3,765; 24.8%), or CABG combined with aortic or mitral valve surgery (n ¼ 2,401; 15.8%) between 2011 and 2014. Patients were divided into three groups by gait speed: fast (more tan 1.00 m/s), middle (0.83 to 1.00 m/s), and slow (less than 0.83 m/s). The primary outcome was allcause operative mortality. The STS composite of major morbidity and mortality was used as the secondary outcome, which included operative mortality, stroke, reoperation, prolonged ventilation, sternal wound infection, and acute kidney injury. The researchers reported that compared with the fastest gait speed group, unadjusted operative mortality was increased in the middle gait speed group (OR 1.77, 95% CI: 1.34 to 2.34) and highest in the slow gait speed group (OR 3.16, 95% CI: 2.31 to 4.33; Fig 5). Interestingly, gait speed remained an independent predictor of mortality after adjustment for the STS PROM and type of surgical procedure (OR 1.11 per 0.1 m/s decrease in gait speed, 95% CI: 1.07 to 1.16). Gait speed was also predictive of the composite outcome after adjustment (OR 1.03 per 0.1-m/s decrease in gait speed, 95% CI: 1.00 to 1.05). The addition of gait speed to the STS PROM model only improved the C-index by 0.005, however, and the integrated discrimination improvement was only 0.003, which the researchers concluded was likely due to the good baseline STS PROM model. Although the C-index did not change much, the predicted risk of mortality increased substantially for slower gait speed. We currently are limited in our ability to use gait speed in the STS predictive models because of the lack of completion in this data variable from the participating sites. Overall, the 5-m gait speed is an objective, useful screening test for frailty in older adults that may provide a better indication of operative risk than age alone. This study highlights the importance of frailty assessment in the accurate prediction of operative risk in elderly patients, which is essential in helping the patient and medical team select an appropriate treatment that optimizes survival and quality of life. These data may serve as a benchmark for
Ann Thorac Surg 2017;-:-–-
Fig 5. Unadjusted association between gait speed and operative mortality. Decreasing gait speed was associated with increasing odds of operative mortality. Blue shaded area indicates 95% confidence interval; dashed line indicates reference odds ratio of 1.0. (m/s ¼ meters per second.)
future studies evaluating newer transcatheter or surgical therapies, and their use by all ACSD participants should be encouraged for these patients. Another risk factor that has been historically associated with adverse outcomes after CABG is African-American race. In a recent study, Mehta and colleagues [17] compared 11,697 African-American patients with 136,362 white patients undergoing isolated CABG between 2010 and 2011 at 663 sites in the ACSD [17]. African-American patients undergoing CABG were significantly younger, more likely to be female, had a higher body mass index, and greater prevalence of comorbidities including hypertension, diabetes mellitus, smoking, peripheral vascular disease, cerebrovascular disease, use of immunosuppressive drugs, preoperative dialysis, myocardial infarction within 7 days of CABG, congestive heart failure, moderate to severe mitral regurgitation, and lower left ventricular ejection fraction compared with white patients. In addition, African-American patients had marginally reduced utilization of the internal mammary artery during CABG (93.3% versus 92.2%, p < 0.0001), but received perioperative medications for secondary prevention more frequently. African-American patients were also more likely to receive treatment in hospitals with higher risk-adjusted mortality (Fig 6). The unadjusted operative mortality rate was 2.5% for African-American patients and 1.8% for white patients (p < 0.0001), and the unadjusted major morbidity rate was 19.4% for African-American patients and 13.6% for white patients (p < 0.0001). After adjusting for baseline patient characteristics, income, hospitals, surgeons, and medications, mortality and morbidity were still significantly higher for African-American patients than for white patients (OR 1.17, 95% CI: 1.00 to 1.36, and OR 1.26, 95% CI: 1.19 to 1.34, respectively). Subgroup and propensity matched analysis confirmed these findings. That AfricanAmerican race remained an independent predictor of mortality and morbidity after current risk adjustment suggests other risk factors that have yet to be identified. Further study of the medical, biological (including genetic), social, and economic issues facing these patients is urgently needed to address the continued racial differences in CABG outcomes.
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
5
Fig 6. Proportion of black patients undergoing coronary artery bypass graft surgery at institutions ranked top to bottom based on increasing risk-adjusted mortality. Centers in quartile 1 (0% to 25%) had the lowest risk-adjusted operative mortality, whereas centers in quartile 4 (75% to 100%) had the highest risk-adjusted operative mortality. (vs ¼ versus.)
Studies Evaluating Quality of Care The STS ACSD has also been utilized to assess quality of care. Previously, the STS QMTF developed composite performance measures and risk models for isolated CABG, isolated aortic valve replacement, CABG with concurrent aortic valve replacement, and isolated mitral valve repair or replacement [6–11]. The fifth in this ongoing series was the development of a composite performance measure for mitral valve repair or replacement with concurrent CABG by reviewing outcomes of 24,740 patients undergoing mitral valve replacement or repair with concurrent CABG from 2011 to 2014 at 703 sites in the ACSD [18]. Sites with fewer than 10 cases over 3 years were excluded, whereas patients with concomitant atrial septal defect or patent foramen ovale closure, surgical ablation for atrial fibrillation, and tricuspid valve repair were included. The two-domain composite included riskadjusted operative mortality and the risk-adjusted occurrence of major complications including stroke, renal failure, reoperation, prolonged ventilation, and deep sternal infection. Hierarchical regression and 95% Bayesian credible intervals were used to calculate composite scores and assign star ratings for performance. The overall unadjusted operative mortality rate was 6.2%, and one or more major complications occurred in 30.8% (Fig 7). This composite measure classified 2% of sites (14 of 703) as one star (lower than expected performance), 95% (666 of 703) as two stars (as expected), and 3% (23 of 703) as three stars (higher than expected). Both mortality and morbidity declined as star rating scores increased. In recent years, failure to rescue (FTR), defined as mortality after the development of a complication, has gained increased attention as an alternative metric of quality. Whereas traditional measure of postoperative morbidity and mortality evaluate the procedure itself, FTR assesses perioperative institutional systems of care. Lower rates of FTR suggest that a facility has the staff,
6
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
Ann Thorac Surg 2017;-:-–-
Fig 8. Influence of number of complications on failure to rescue (mortality [blue bars]) and patient sample size (n [orange line]).
Fig 7. Distributions of risk-adjusted (A) mortality rate and (B) morbidity rate (n ¼ 703 participants). (IQR ¼ interquartile range.)
resources, and systems in place to adequately care for patients who have complications. A recent study by Edwards and colleagues [19] analyzed the outcomes of 604,154 patients in the ACSD undergoing isolated CABG from 2010 to 2014 at 1,105 centers to calculate rates of FTR and develop a model predictive of FTR after CABG. The investigators defined FTR as death after stroke, renal failure, reoperation, and prolonged ventilation. The FTR was calculated for each of these complications, as well as for a composite of the four complications. Individual rates of FTR for renal failure, stroke, reoperation, prolonged ventilation, and the composite were 22.3%, 16.4%, 12.4%, 12.1%, and 10.5%, respectively. The FTR increased in patients with more than one and certain combinations of complications (Fig 8). Multivariable logistic regression was used to develop a model to predict FTR, which had a C-index of 0.792 and was well calibrated with a 1.0% difference between observed to expected rates of FTR. When centers were grouped into tertiles by mortality, the complication rate increased only modestly from lowest to highest group (11.4% to 15.7%), whereas mortality varied directly with FTR, doubling from the lowest to highest group (6.8% to 13.9%). Interestingly, centers with the highest complication rates had the lowest FTR, whereas sites with the lowest complication rates had the highest FTR. Using the model developed in this study, the FTR observed to expected ratio was 1.14 in the lowest complication tertile and 0.91 in the highest complication tertile. The researchers postulated that perhaps tertiary
centers who are referred very complex and high-risk patients have complication rates that are correspondingly higher. However, these centers may also have the expertise, staffing, and advanced care technologies to salvage these patients when they do have complications. These findings will be further investigated by the STS QMTF [3]. The quality of care data reported by these two studies have provided national benchmarks that can be used to counsel patients, aid in clinical decision making, and encourage centers to reevaluate and restructure their practices to improve outcomes.
Conclusion Cardiac surgeons have remained on the forefront of the arduous task of self-reflection by continuing to reassess and refine their surgical techniques. One of the most important tools to serve this endeavor is the STS ACSD. With more than 6 million cumulative operative records, it remains the most comprehensive health care registry in the world. In addition to performance assessment and quality improvement, the ACSD is an invaluable resource for clinical research. The findings of studies published in 2016 that utilized this database have enhanced clinical practice and delineated areas for future quality improvement research.
References 1. Clark RE. It is time for a national cardiothoracic surgical data base. Ann Thorac Surg 1989;48:755–6. 2. Winkley Shroyer AL, Bakaeen F, Shahian DM, et al. The Society of Thoracic Surgeons Adult Cardiac Surgery Database: the driving force for improvement in cardiac surgery. Semin Thorac Cardiovasc Surg 2015;27:144–51. 3. D’Agostino RS, Jacobs JP, Badhwar V, et al. The Society of Thoracic Surgeons Adult Cardiac Surgery Database: 2017 update on outcomes and quality. Ann Thorac Surg 2017;103: 18–24. 4. Jacobs JP, Shahian DM, Prager RL, et al. The Society of Thoracic Surgeons National Database 2016 annual report. Ann Thorac Surg 2016;102:1790–7. 5. Shahian DM. The Society of Thoracic Surgeons National Database: “What’s past is prologue.” Ann Thorac Surg 2016;101:841–5.
Ann Thorac Surg 2017;-:-–-
6. Shahian DM, O’Brien SM, Filardo G, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 1— coronary bypass grafting surgery. Ann Thorac Surg 2009;88(Suppl):2–22. 7. O’Brien SM, Shahian DM, Filardo G, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 2— isolated valve surgery. Ann Thorac Surg 2009;88(Suppl):23–42. 8. Shahian DM, O’Brien SM, Filardo G, et al. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 3— valve plus coronary artery bypass grafting surgery. Ann Thorac Surg 2009;88(Suppl):43–62. 9. Shahian DM, He X, Jacobs JP, et al. The Society of Thoracic Surgeons isolated aortic valve replacement (AVR) composite score: a report of the STS Quality Measurement Task Force. Ann Thorac Surg 2012;94:2166–71. 10. Shahian DM, He X, Jacobs JP, et al. The STS AVRþCABG composite score: a report of the STS Quality Measurement Task Force. Ann Thorac Surg 2014;97:1604–9. 11. Badhwar V, Rankin JS, He X, et al. The Society of Thoracic Surgeons mitral repair/replacement composite score: a report of The Society of Thoracic Surgeons Quality Measurement Task Force. Ann Thorac Surg 2016;101:2265–71. 12. Jacobs JP, Shahian DM, He X, et al. Penetration, completeness, and representativeness of The Society of Thoracic Surgeons Adult Cardiac Surgery Database. Ann Thorac Surg 2016;101:33–41. 13. Whellan DJ, McCarey MM, Taylor BS, et al. Trends in robotic-assisted coronary artery bypass grafts: a study of The
QUALITY REPORT THOURANI ET AL STS ACSD 2017 RESEARCH UPDATE
14.
15.
16. 17.
18.
19.
7
Society of Thoracic Surgeons Adult Cardiac Surgery Database, 2006 to 2012. Ann Thorac Surg 2016;102:140–6. Schwann TA, Habib RH, Suri RM, et al. Variation in warfarin use at hospital discharge after isolated bioprosthetic mitral valve replacement: an analysis of The Society of Thoracic Surgeons Adult Cardiac Surgery Database. Chest 2016;150: 597–605. Acharya D, Gulack BC, Loyaga-Rendon RY, et al. Clinical characteristics and outcomes of patients with myocardial infarction and cardiogenic shock undergoing coronary artery bypass surgery: data from The Society of Thoracic Surgeons National Database. Ann Thorac Surg 2016;101:558–66. Afilalo J, Kim S, O’Brien S, et al. Gait speed and operative mortality in older adults following cardiac surgery. JAMA Cardiol 2016;1:314–21. Mehta RH, Shahian DM, Sheng S, et al. Association of hospital and physician characteristics and care processes with racial disparities in procedural outcomes among contemporary patients undergoing coronary artery bypass grafting surgery. Circulation 2016;133:124–30. Rankin JS, Badhwar V, He X, et al. The Society of Thoracic Surgeons mitral valve repair/replacement plus coronary artery bypass grafting composite score: a report of The Society of Thoracic Surgeons Quality Measurement Task Force. Ann Thorac Surg 2017;103:1475–81. Edwards FH, Ferraris VA, Kurlansky PA, et al. Failure to rescue rates after coronary artery bypass grafting: an analysis from The Society of Thoracic Surgeons Adult Cardiac Surgery Database. Ann Thorac Surg 2016;102:458–64.