Article Text

Download PDFPDF

The Society of Thoracic Surgeons National Database
  1. David M Shahian1,
  2. Jeffrey P Jacobs2,
  3. Fred H Edwards3,
  4. J Matthew Brennan4,
  5. Rachel S Dokholyan4,
  6. Richard L Prager5,
  7. Cameron D Wright6,
  8. Eric D Peterson4,
  9. Donna E McDonald7,
  10. Frederick L Grover8
  1. 1Department of Surgery and Center for Quality and Safety, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
  2. 2Johns Hopkins Children's Heart Surgery, All Children's Hospital and Florida Hospital for Children, Saint Petersburg, Tampa, and Orlando, Florida, USA
  3. 3Shands Jacksonville, University of Florida College of Medicine—Jacksonville, Jacksonville, Florida, USA
  4. 4Duke Clinical Research Institute (DCRI), Duke University Medical Center, Durham, North Carolina, USA
  5. 5Department of Cardiac Surgery, University of Michigan Health System, Ann Arbor, Michigan, USA
  6. 6Division of Thoracic Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA
  7. 7Society of Thoracic Surgeons, Chicago, Illinois, USA
  8. 8Department of Surgery, University of Colorado Denver, School of Medicine, Aurora, Colorado, USA
  1. Correspondence to Dr David M Shahian, Department of Surgery and Center for Quality and Safety, Massachusetts General Hospital, 55 Fruit St., Boston, MA 02114, USA; dshahian{at}


Aims The Society of Thoracic Surgeons (STS) National Database collects detailed clinical information on patients undergoing adult cardiac, paediatric and congenital cardiac, and general thoracic surgical operations. These data are used to support risk-adjusted, nationally benchmarked performance assessment and feedback; voluntary public reporting; quality improvement initiatives; guideline development; appropriateness determination; shared decision making; research using cross-sectional and longitudinal registry linkages; comparative effectiveness studies; government collaborations including postmarket surveillance; regulatory compliance and reimbursement strategies.

Interventions All database participants receive feedback reports which they may voluntarily share with their hospitals or payers, or publicly report. STS analyses are regularly used as the basis for local, regional and national quality improvement efforts.

Population More than 90% of adult cardiac programmes in the USA participate, as do the majority of paediatric cardiac programmes, and general thoracic participation continues to increase. Since the inception of the Database in 1989, more than 5 million patient records have been submitted.

Baseline data Each of the three subspecialty databases includes several hundred variables that characterise patient demographics, diagnosis, medical history, clinical risk factors and urgency of presentation, operative details and postoperative course including adverse outcomes.

Data capture Data are entered by trained data abstractors and by the care team, using detailed data specifications for each element.

Data quality Quality and consistency checks assure accurate and complete data, missing data are rare, and audits are performed annually of selected participant sites.

Endpoints All major outcomes are reported including complications, status at discharge and mortality.

Data access Applications for STS Database participants to use aggregate national data for research are available at

View Full Text

Statistics from


In March 1986, the US Health Care Financing Administration (HCFA, the predecessor of CMS, the Centers for Medicare & Medicaid Services) issued the first comprehensive report of individually identifiable hospital death rates ever published in the USA.1 These death rates were based on administrative claims data and were widely criticised for their inadequate risk adjustment.2 ,3 HCFA continued to publish death rates until 1993, when the programme was discontinued. However, this initiative was a watershed event in US healthcare, as it stimulated interest in improving the data sources and risk-adjustment methodologies used for healthcare provider profiling.

Cardiac surgery was at the forefront of these activities. In October 1986, the Society of Thoracic Surgeons (STS) issued a statement of concern regarding the flawed methodology used to generate the HCFA mortality reports, and an STS ad hoc committee began to develop clinical data elements for a cardiac surgery registry to facilitate more accurate assessment of provider performance.4 The STS Database was made available to STS participants in 1989,5 and after adequate data had been collected the first risk-adjustment models were published several years later.6

Organisational structure

The STS National Database is supported by nearly 100 volunteer cardiothoracic surgeons with special interest in clinical registries, quality measurement and performance improvement. The Database is organised into both specialty-specific and functional task forces. The former encompass the three major subdivisions of cardiothoracic surgery—Adult Cardiac, Pediatric and Congenital Cardiac, and General Thoracic. The functional task forces are cross-cutting, and include Quality Measurement, Quality Initiatives, Public Reporting, International Relations, and Appropriateness. Two other Task Forces—Access and Publications (A&P) and Linked Registry Proposals—reside within the STS Research Center, which itself is closely integrated with the Database.

Specialty-specific databases

Adult Cardiac Surgery Database

The largest and oldest of the three specialty databases is the STS Adult Cardiac Surgery Database (STS-ACSD), which has participants from 49 US states and several international sites (figure 1). The 1073 ACSD participant groups include 3031 surgeons practicing in 1069 hospitals, representing over 90% of cardiac surgery programmes in the USA (table 1). They have contributed over 4.8 million patient records encompassing the entire spectrum of adult cardiac procedures (eg, coronary artery bypass grafting (CABG), valve replacement, thoracic aortic surgery, arrhythmia ablation, ventricular assist, etc).

Table 1

STS National Database participation (as of 28 November 2012)

Figure 1

US national penetration of the Society of Thoracic Surgeons Adult Cardiac Surgery Database.

Collection of granular clinical data and feedback of risk-adjusted, nationally benchmarked results to participants was the initial motivation for development of the STS ACSD, and these functions remain its primary purpose. Over the 23 years since its inception, the ACSD has gone through multiple revisions of its data elements, consistent with evolving clinical science. Increasingly sophisticated statistical risk models have been developed to more fully capture differences in patient severity, and both individual and composite performance measures for most major cardiac procedures have been created and nationally endorsed. Consistent with the STS's belief that patients have an ethical right to know their provider's performance, STS initiated a voluntary public reporting system for ACSD participants beginning in 2010. The ACSD has served as the basis for over 150 peer-reviewed publications and numerous externally funded grants, and external linkage initiatives have substantially expanded both its cross-sectional and longitudinal utility. New outcomes measures are being explored, including patient-reported outcomes, cost effectiveness and demonstration of appropriateness, as described subsequently in this report. The Quality Initiatives (QI) Taskforce, described in a subsequent section, has used evidence-based best practices to enhance the care of cardiothoracic patients. Finally, the STS ACSD has established numerous relationships with the federal government, including direct submission of Physician Quality Reporting System (PQRS) data on behalf of STS participant surgeons, and the recent development of a 30-day CABG readmission measure based on risk-adjusted clinical data.

General Thoracic Database

Modelled after the STS-ACSD, the STS General Thoracic Database (STS-GTDB) was initially conceived more than two decades ago and formally implemented in 2002. Its goal is to provide feedback of quality assessment data to participants and to serve as the basis for clinical practice improvement in general thoracic surgery. The GTDB currently has 231 participant groups and continues to grow, with expansion planned for Canada and other countries. Seven hundred eighty-six cardiothoracic surgeons and 18 general surgeons from 271 hospitals across 40 states have contributed more than 312 000 general thoracic surgical records (figure 2; table 1).

Figure 2

US national penetration of the Society of Thoracic Surgeons General Thoracic Database.

The GTDB is a procedure-based database which currently collects in-hospital and 30 day mortality. However, because many general thoracic operations are performed for cancer, plans for longer follow-up are being developed, some of which will involve an expanded network of linkages to other clinical and claims registries.

Data reports are provided to GTDB participants semiannually, listing all their cases for the reporting period in a rolling 3-year window. Benchmark results are based on an aggregate of all the STS-GTDB participants. Mortality and morbidity are reported for each procedure using currently available general thoracic risk models: morbidity, mortality and composite endpoints for lung cancer resections; prolonged length of stay after lobectomy for lung cancer; and morbidity and mortality associated with esophagectomy performed for cancer.7–9

The STS GTDB is a valuable source of clinical data for research projects, all of which are initially screened through the STS A&P Task Force using the STS website. Examples of this research include the lobectomy and esophagectomy risk models; studies of the effect of smoking cessation on perioperative morbidity; and a comparison of lobectomy by open or video-assisted thoracoscopic surgery (VATS) techniques.7 ,9–11

Paediatric and congenital cardiac surgery

The STS Congenital Heart Surgery Database (STS-CHSD) is the largest database in the world that tracks the outcomes of patients with paediatric and congenital cardiac disease.12–14 The STS-CHSD has 106 participant groups from 109 hospitals (figure 3) in the USA and Canada (86% of hospitals in the USA), and 339 surgeons who have contributed more than 200 000 records15 ,16 (table 1). Further collaborative efforts are ongoing between congenital and paediatric cardiac surgeons and other subspecialties, including paediatric cardiac anaesthesiologists, paediatric cardiac intensivists, and paediatric cardiologists. A congenital cardiac anaesthesiology component was added to the CHSD in 2010, including 222 anaesthesiologists from 31 CHSD programmes. The STS-CHSD has enhanced its cross-sectional registry capabilities by linking to other databases through probabilistic matching using indirect identifiers. It plans to similarly augment its longitudinal follow-up capabilities by linking with other sources of data through both probabilistic matching and deterministic matching with direct identifiers.17

Figure 3

US national penetration of the Society of Thoracic Surgeons Congenital Heart Surgery Database.

Over the past 15 years, important advances have been made by STS-CHSD in seven domains: (1) use of a common language and nomenclature, (2) development of a uniform core dataset, (3) incorporation of a mechanism to evaluate case complexity, (4) robust audit to verify data completeness and accuracy, (5) collaboration between medical and surgical subspecialties, (6) development of strategies for life-long follow-up and (7) implementation of standardised tools for quality assessment and improvement. In 2000, The European Association for Cardio-Thoracic Surgery (EACTS) and the STS jointly adopted a common nomenclature and core minimal dataset for congenital heart surgery,12 which were subsequently cross-mapped with the European Paediatric Cardiac Code to create the International Paediatric and Congenital Cardiac Code (IPCCC). This common nomenclature, the IPCCC, and the common minimum database data set created by the International Congenital Heart Surgery Nomenclature and Database Project, are now utilised by STS, EACTS, and The Japan Congenital Cardiovascular Surgery Database (JCCVSD). The IPCCC is also being used as the platform for creation of the congenital cardiac portion of the 11th edition of the International Classification of Diseases (ICD-11).

Three major multi-institutional efforts have been used to measure and adjust for the complexity of congenital cardiac surgery, including The STS—EACTS Congenital Heart Surgery Mortality Categories (STAT Mortality Categories), Aristotle Basic Complexity Levels, and RACHS-1 Categories.18–26 In comparison with the other two methodologies which were largely based on expert opinion, STAT Mortality Categories are based on objective data from STS-CHSD and EACTS-CHSD. STS and the Congenital Heart Surgeons’ Society recently published a series of quality measures for congenital and paediatric cardiac surgery,27 three of which have been endorsed by the National Quality Forum (NQF).

STS data elements

Although the three subspecialty databases each have many unique data elements pertaining to their specific patients and procedures, the general structure of these databases is similar. Each contains demographic fields that allow for analysis of patients and outcomes geographically and by race, ethnicity, age and gender. Detailed preoperative data elements characterise patient history, comorbidities and urgency of operation, and these, in turn, are used for risk adjustment of outcomes, analysis of patient selection and appropriateness of care. Operative fields capture surgical approach, techniques and intraoperative management. Implant data on heart valves and ventricular assist devices are catalogued and can be used for postmarket surveillance. Analysis of resource utilisation is facilitated by variables that track intensive care unit (ICU) and ventilator hours, hospital length of stay and blood usage. Postoperative events, complications, readmission and mortality fields complete the record.

Specification upgrades

To maintain relevance, the STS National Database evolves with surgical practice. As a part of this evolution, data elements undergo periodic updates, with the goal of clarifying existing variables, removing irrelevant variables, and adding new variables of interest. Database versions are reviewed and revised on a 3-year cycle, with one of the three specialty databases addressed each year. A team of surgeons, data managers and technical advisors assess existing and proposed fields and definitions based on feedback from database participants, with particular attention to variables that require clarification. Content experts from various disciplines, such as cardiologists, anaesthesiologists, nephrologists, pulmonologists and perfusionists are consulted to optimise data definitions. Every effort is made to harmonise definitions with national and international data standards and definitions, as well as other related databases, such as those maintained by the American College of Cardiology (ACC), and future versions will incorporate clinical data standards compatible with electronic health records. Each official version of STS data resulting from these specification upgrades is numbered and is fully described by the data collection form and data specifications, which serve as the central tool with which STS National Database stakeholders (Duke Clinical Research Institute (DCRI), STS, software vendors, data managers, etc) communicate regarding specific data elements. The data specifications document contains all the parameters (such as variable name, valid values set, definition, etc) for each data element.28


In addition to these major version upgrades, variables that are of interest to a narrower spectrum of participants are made available for data collection through DCRI-supported online web modules. These web modules are useful tools for piloting new or project-specific variables (eg, atrial fibrillation ablation surgery).

Data Warehouse and Analytic Center

Since 1999, the DCRI has served as the Data Warehouse and Analytic Center for the STS National Database, collaborating with STS staff and volunteer leaders in the management of database operations. The DCRI team contributes general expertise in the design and implementation of clinical registries and observational clinical research, as well as specific expertise in project management, data management, site management, and statistical programming and analysis.

Data entry

STS National Database software is licensed by contracted independent software vendors who translate the updated database elements for online data submission. These vendors must comply with STS data and software specifications and undergo validation for every STS National Database software version they distribute.

One of the most important elements contributing to the success of the STS Database is its dedicated data managers, most of whom are employed by participant sites and manage both the collection and submission of patient data. Many participants use a team approach to achieve point-of-care data capture. STS hosts an annual 3-day educational conference that is typically attended by 500 or more data managers. In addition, online resources are available to answer data abstraction questions, and a list is available for data manager networking. STS data manager core groups from all three databases participate in frequent phone conferences during which they provide valuable feedback regarding data specifications and difficult data abstraction issues. These suggestions and clarifications are made available in STS online reference and teaching materials and they are incorporated into subsequent data specification upgrades. Finally, regional groups provide education, support and promote QI for Database participants.

Data harvests

Participants submit their data to the DCRI during what is referred to as ‘data harvest’ (figure 4). Data harvests take place four times annually for the ACSD, and twice annually for both the Congenital Heart Surgery and General Thoracic Surgery databases. Participants export their data as an American Standard Code for Information Interchange text file from their STS software and submit it to the DCRI via secure web-based upload.

Figure 4

Society of Thoracic Surgeons National Database Harvest Cycle.

Data quality checks

Mandatory consistency checks, valid ranges and discrete data elements are designed to optimise data quality. Data files that arrive at the DCRI are automatically reviewed by data processing programmes that perform a preliminary analysis to ensure that it can be accepted (the surgeon is on file with STS, data vendor is certified, etc). If the file passes this initial stage it undergoes a series of checks at the level of individual record and data element. The results of this check are output into a Data Quality Report (DQR) that is e-mailed to the participant. The DQR contains summary information about the submitted data file, as well as detailed and record-specific information about missing data elements and data inconsistencies. Participants generally receive their DQR within hours of submitting their data file, sometimes within minutes. Participants can make changes to their data and resubmit a file to the DCRI at any time, and as many times as necessary, during the 3-week harvest window.

Participant feedback reporting

Upon completion of each data harvest, database participant feedback reports are created. The feedback reports contain detailed participant-level information on patient demographics, preoperative medications, operative and process of care information, outcomes and discharge status. Major mortality and morbidity outcomes are risk-adjusted using both hierarchical and non-hierarchical modelling approaches. Twice a year, the Adult Cardiac Surgery Report also contains results of the STS Quality Composite Rating.29 ,30 In addition to the participant feedback reports provided to each organisation that submits data, the DCRI also creates custom reports for the STS upon request, usually for regional groups of multiple participants such as a regional quality improvement (QI) initiative or statewide quality assurance (QA) programme. Content for all reports is approved by the STS, and each participant provides permission before their data are included in any report provided to a third party.

Privacy protection

To enhance the utility of the Database for QA and quality improvement initiatives, the STS collects direct patient identifiers without individual patient consent, having obtained a waiver from the Duke University Health System institutional review board (IRB). However, access to directly identified patient data is limited, and separate IRB review is obtained for all human subjects research that uses data that can directly identify individual patients.31

Data audit

The STS National Database has implemented both an internal validation programme and an external audit. Submitted data are validated during each harvest process, as described previously, at the time of submission to the data warehouse. STS contracts with an independent external auditing firm to perform audits. Currently, 5% of ACSD participants are randomly selected for an audit, and this proportion will increase to 8% in 2013. The most recent year of clinical data submitted to STS's data warehouse are validated against the corresponding medical records by trained reviewers, and selected charts are evaluated for accuracy on nearly 80 data fields. Hospital surgical logs are compared with submitted cases to be certain all cases are included. The latter results have shown nearly 100% completeness of case ascertainment, and audits of individual data elements are typically recorded with agreement rates in excess of 95%.

Performance measurement

The primary motivation for development of the STS Database was a desire to provide highly credible performance assessment based on the best available clinical data, with appropriate risk adjustment to account for differences in patient severity across surgeons and institutions. Although the Database has subsequently evolved into a multifunctional resource serving the cardiothoracic community, performance assessment remains its core purpose.

A major limitation of the original release of mortality data by HCFA was their reliance on administrative claims data. Such data are collected primarily for billing purposes and generally lack clinical specificity. Studies by Mack and colleagues32 and Shahian and associates33 showed substantial differences in profiling results when claims data were used compared with clinical registry data. By using more granular, accurate data, registry-based performance profiling accounts more fully for the inherent differences among providers, thus protecting surgeons and hospitals that care for more seriously ill patients. Risk adjustment has become an integral feature of every credible performance profiling and public reporting initiative in cardiac surgery,34–36 and statistical risk modelling has been a fundamental component of the STS performance measurement approach since its inception.6 In 2008, the STS constructed 27 new risk models for isolated CABG, isolated valve and combined valve+CABG surgery.37–39 These include separate models for each of nine major outcomes (death, stroke, reoperation, renal failure, sternal infection, prolonged ventilation, short length of stay, prolonged length of stay, composite adverse event) for each of the three types of procedures. The conceptual and methodological considerations relevant to these models were presented in detail in these publications, including model intercepts and coefficients. In 2012, a new risk model was developed for multiple valve operations, and details of its development will be published in 2013.

Historically, performance reports in cardiac surgery have used risk-adjusted mortality as the primary endpoint,40 but this assesses only one performance domain and has limited ability to discriminate among providers when death rates are generally quite low. In view of these limitations, the STS developed a CABG composite measure29–30consisting of four domains: risk-adjusted mortality; risk-adjusted occurrence of any of the five major complications of CABG; use of an internal mammary artery (IMA); and use of all four NQF-endorsed perioperative medications. Using a composite score derived from these 11 individual measures, a more comprehensive evaluation of provider quality is possible, and many more providers can be differentiated as low- or high-performing than could be done using death rates alone. The numerical scores of this CABG composite are then also converted to a star rating system to facilitate consumer interpretation.41 ,42 This STS CABG Composite Score is NQF endorsed and is provided to cardiac surgical programmes twice a year. A similar composite performance measure for isolated aortic valve replacement (AVR) has recently been developed.

Public reporting

STS believes that public reporting of cardiothoracic surgical outcomes is a professional society's ethical responsibility and one of several effective ways to stimulate performance improvement;43–45 consistent with these beliefs, the Society has been a national leader in such initiatives. In collaboration with Consumer Reports, STS began voluntary, consumer-friendly public reporting of outcomes results for isolated coronary artery bypass (CABG) surgery in September 2010, using the NQF-endorsed STS CABG Composite Score. The STS CABG Composite Score is publicly reported on both the Consumer Reports website and the STS website, and STS AVR Composite Scores will be publicly reported on both of these websites beginning in 2013. One additional new procedure will be added to the public reporting portfolio each year. STS also plans to expand this public reporting initiative to the STS CHSD and the STS GTDB.

Quality improvement initiatives

The STS Task Force on QI has several major functions, the underlying goals of which are to maintain the STS’ position as the quality leader in adult and congenital cardiothoracic surgery. The first responsibility of the QI Task Force relates to performance measures developed by the Quality Measurement Task Force. QI Task Force surgeons and STS staff have developed special expertise in adapting these measures to the format required by external endorsement organisations, such as the NQF and CMS. They prepare and submit these measures, respond to public comments, and make necessary modifications when required. The QI Task Force is also responsible for periodic measure maintenance required by these organisations, including justification for continuing the measures.

There are currently 31 STS-sponsored, NQF-endorsed measures based on the three STS databases.

A second major function of the QI Task Force is to use STS data, as well as best-practice evidence from other sources, to improve patient outcomes. There are multiple examples of statewide, regional and national collaborations utilising the STS data to energise broad quality improvement, including the Virginia Cardiac Surgical Quality Initiative,46–48 the Michigan Society of Thoracic and Cardiovascular Surgeon's Quality Initiative,49 and national efforts to increase use of β-blockers and internal mammary artery grafts.50 The QI Task Force has created a programme of quality webinars on numerous topics of broad interest, such as blood conservation and glycemic control, and a web-based library of best-practice articles that will link with the topics of future webinars. Thirty-minute e-learning modules will be created from each webinar and will be available for continuing education credit. STS data will increasingly be used to identify best practices in areas such as ventilator management and blood conservation, and to base national quality initiative efforts on these best practices.


The original 1964 STS constitution stated that the Society was to ‘encourage clinical … research’ and thereby ‘improve the quality and practice of thoracic and cardiovascular surgery.’ STS believes that optimal patient management must be based on objective scientific evidence, and this guiding principle has been the foundation for STS research over the last half century.

Early research reports from STS members, especially those reported in Annals of Thoracic Surgery, usually originated from academic institutions. However, with the establishment of the STS National Database, STS research was forever changed. Although the Database was originally designed as the basis for national benchmarking and quality improvement activities, its research potential quickly became apparent. Initial research efforts from the Database were STS reports generated at the direction of STS leadership.6 ,51–53 With continued database growth, STS members increasingly recognised that the Database contained a wealth of objective data for research studies to address practical clinical issues. In response, STS leadership established a formal mechanism by which members could access the data for research as a benefit of database participation. This was initially carried out through the Risk Assessment Subcommittee, and later through the A&P Task Force.

The A&P Task Force process has facilitated access to the database for a wide range of research projects, and has made meaningful research activity possible even for CT surgeons not practicing at major academic centres. Database participants submit research requests to the A&P Task Force by completing a simple form that is evaluated by Task Force leadership, which consists of surgeons and statisticians. Once formally accepted, the requestor is provided with a research team that includes seasoned STS researchers and DCRI biostatisticians. This process has produced well over 150 reports in the peer-reviewed literature, and a large number of presentations at regional and national meetings.

In more recent years, the STS Database has become widely recognised as a valuable national resource for clinical information. Particularly after the American Recovery and Reinvestment Act of 2009, numerous funding opportunities for STS research became available. With these new opportunities, STS found it necessary to create an organisational structure to acquire and manage funded research, and the STS Research Center was formed in 2011 to facilitate these tasks. In its first year of operation, the Research Center has streamlined the process of grant operations and has worked closely with ACC, Food and Drug Administration (FDA), and CMS to develop the STS/ACC Transcatheter Valve Therapy Registry for postapproval monitoring, a new paradigm for future device surveillance registries.

Cross-registry linkage activities

Through linkages with other clinical registries and administrative data sources, the STS Database has been enriched with valuable, previously unavailable information which has substantially enhanced its value for research and many other activities. These linkages have been both cross-sectional (eg, data from other cardiovascular specialties, such as interventional cardiology) and longitudinal (eg, linkages with claims data bases such as CMS MEDPAR and the Social Security Death Master File (SSDMF)).54–57 These innovative linkages permit clinical registries to expand beyond their current condition or procedure-specific scope and their focus on short-term clinical outcomes. The resulting combined data sources are well suited to healthcare reform priorities that emphasise care coordination and reimbursement over the continuum of a disease process. Ethical and regulatory issues relevant to linkages have been studied extensively by STS, and every effort is made to assure full compliance with all relevant federal regulations.31

Matching with indirect identifiers has been used to link the STS Database to administrative databases, such as the CMS Medicare Database and the Pediatric Health Information System (PHIS) Database. The resulting linked datasets are valuable sources of information about long-term morbidity and mortality, rates of rehospitalisation and reintervention, medication usage and resource utilisation.56 ,57 Deterministic matching with shared unique identifiers has been used to link the STS Databases to the SSDMF to verify life status over time.55

Numerous analyses have now been performed using the STS Database linked datasets.55–61 The National Heart, Lung, and Blood Institute has funded an STS-ACC collaborative project, the ASCERT study, to investigate the comparative effectiveness of percutaneous coronary intervention (PCI) and CABG for the treatment of stable coronary artery disease.62 ASCERT compares long-term outcomes following stent placement versus CABG surgery for two- and three-vessel disease, using existing clinical databases from the ACC and STS, linked with CMS MEDPAR administrative data.63 ASCERT investigators also used the STS-CMS link to develop statistical models for long-term survival following isolated CABG and PCI, adjusting for patient risk.64 ,65 Several Agency for Healthcare Research and Quality (AHRQ)-sponsored DEcIDE (Developing Evidence to Inform Decisions about Effectiveness) Network investigations have used linked STS-MEDPAR data to study aortic valve surgery outcomes,61 ,66 and similar linkages have also been used to study long-term outcomes after mitral valve surgery.58 Efforts are now underway to use the STS-CMS link to study longitudinal outcomes after surgery for lung cancer and oesophageal cancer. In paediatric cardiac surgery, the STS-PHIS link has been used to perform several clinical and comparative effectiveness analyses.17 ,56 ,59 ,60

Federal government collaborations

Because of the granularity and clinical relevance of the STS ACSD, CMS has contracted with STS to develop a 30-day all-cause readmission risk model for CABG that might potentially be used as a quality metric. This is the first initiative of its kind and reflects both the high perceived quality of STS data, its acceptance by a broad spectrum of stakeholders, and the high national penetration of the STS Database.

STS is an approved registry that directly submits PQRS data to CMS on behalf of providers, thereby providing substantial additional value to STS participants. As previously described, the FDA has worked collaboratively with the STS and ACC to develop a new paradigm for postapproval monitoring of new devices, the first application of which is the ACC-STS Transcatheter Valve Therapy registry.67

International expansion

International activities of the STS Database span more than a decade and include collaborations with the European Association for Cardiothoracic Surgery, the Japan Adult Cardiovascular Surgery Database, and the Asian Society for Cardiovascular and Thoracic Surgery. There have also been highly successful international collaborations involving the STS Congenital Database Task Force, the European Association for Cardiothoracic Surgery, and the JCCVSD.

International expansion of the STS Database has generally focused on regions where no CT registries currently exist, and it is the intention of the STS Database to collaborate with rather than compete with registries where they are currently in widespread use. The STS International Database Relations Task Force is currently in contact with over 20 international surgical centres. These programmes have expressed an interest in participating in the STS Database, and receiving individual feedback reports benchmarked against the results from US sites; some have also requested summaries of programme performance for their region or country. Additional international centres have been contacted with regard to joining the CHSD, and similar plans have been envisioned for the General Thoracic Surgery Database.

Introduction of the STS Database involves additional personnel to collect the data, and familiarity of these individuals with the nuances of STS Database definitions. Startup costs are higher over the first few years, but ongoing costs are then expected to more closely approach those of established US programmes. Accurate and complete data acquisition is of paramount importance, and audit can be performed onsite or, more cost efficiently, via remote audit of electronic records.

STS leadership believes that making the STS Database available for our international colleagues will greatly enhance their ability to assess and improve quality and conduct meaningful research, thus enhancing the care that cardiothoracic surgeons deliver world-wide.


As part of current health reform efforts, there is increasing pressure to assure that high cost procedures and technologies are utilised only when indicated. Although underuse may be an equally important healthcare challenge, overuse is the major current focus, as its financial impact is more immediate. For certain procedures, such as elective PCI and CABG, demonstration of appropriateness will likely be required in the future, and clinical registries provide an ideal mechanism by which to make evidence-based assessments of appropriateness.68–70

In 2012, the STS National Database established a Task Force on Appropriateness, the goal of which is to map the clinical characteristics of every STS CABG patient to the 2011 ACC Foundation/American Heart Association Task Force CABG guidelines71 as well as the 2012 Appropriateness Criteria for Coronary Revascularisation.70 This mapping process will allow real-time confirmation that evidence-based guidelines and criteria support an intervention. Work to date suggests that most elements required for full assessment of guideline or appropriate use criteria adherence are found in the Database. However, a number of data elements are currently not available and will be added in the 2013 specification upgrade (eg, stress test results, maximum goal-directed medical therapy).


From its inception 23 years ago, the STS National Database has collected granular clinical data for all patients undergoing cardiothoracic surgery. Although these data are still used primarily to assess risk-adjusted provider performance, robust STS data are now used to facilitate a much broader range of functions including public reporting; determination of appropriateness; development of guidelines and performance measures; quality improvement initiatives; clinical and comparative effectiveness research; shared decision making; government collaborations; regulatory compliance; reimbursement; and centre of excellence designation.


View Abstract


  • Funding Internal.

  • Competing interests All authors are paid or volunteer staff of the Society of Thoracic Surgeons or the Duke Clinical Research Institute (the data warehouse and analytical centre for the STS National Database).

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.