THE SOCIETY OF Thoracic Surgeons (STS) database was created in 1989 as a patient safety initiative for adults undergoing cardiothoracic surgery. In 2005, pediatric and congenital cardiac patient data began being collected by a registry operated by the Congenital Cardiac Anesthesia Society. Collaboration between STS and Congenital Cardiac Anesthesia Society fostered eventual development of the Congenital Heart Surgery Database (CHSD), a specialized section within the STS database. The STS CHSD collects data to provide participants with risk-adjusted outcomes information to allow for improvement in individual surgical or program components.
The STS predated meaningful use of electronic health records (EHRs), an initiative encouraging health organizations to move from anecdotal-based evidence to systems-based practice in decision making. The meaningful use mandate accompanied widespread adoption of the EHR, an administrative database allowing capture of clinical information to facilitate quality improvement.
1
In parallel, sophisticated databases and patient registries such as the STS CHSD, which are housed and operated outside a given health care system, provide an opportunity to explore associations, answer clinical questions, and ensure quality care.Such patient information repositories also create a nearly infinite amount of data. This “big data” is conceptualized in terms of the following 5 Vs: volume, variety, velocity of entry into the EHR, value, and veracity.
1
The volume of patient data requires specialized software to manage and navigate; it often must be housed in a central location outside a given health care institution. The variety of data in the database is unstructured and exists in raw form, making it difficult to query or integrate into other databases. New data are constantly created, and the integrity of any EHR or database depends on the velocity of entry of into the system. The value of the patient data must be considered as well, particularly through the lens of volume and variety; all captured data are not necessary or helpful. Finally, data are only useful if they are accurate and complete; ensuring veracity of information in an EHR or database is mandatory.In the March 2019 edition of the Annals of Thoracic Surgery, Jacobs et al. published findings from the most recent report from the STS CHSD, describing morbidity and mortality data from January 2014 to December 2017.
2
Their report contains national aggregate outcomes and participant mortality data for 10 specific types of surgery. This report is the fourth update from the STS CHSD, but it is the first since the STS began public outcomes reporting in 2015; public reporting increases transparency and recognizes variation among centers regarding treatment strategies and outcomes.3
The risk-adjusted mortality is reported as an observed-to-expected (O:E) ratio for each participating institution; this O:E ratio is the basis for the star quality rating.Even though the star quality ratings may be controversial, the STS CHSD continues to improve the current risk models used to generate those scores. The current star ratings are based on surgical mortality and are tied to specific procedures; however, because surgery for congenital heart disease has become safer, with very low mortality rates, the STS CHSD is shifting from a focus on procedures to a focus on cardiac diagnoses. Composite scores will evolve to include more measures of morbidity and cost.
3
It is in this shift that the role of anesthesia will become scrutinized. This is appropriate because advances in anesthesia care, access, medical management, and expertise have contributed to improved morality in congenital heart surgery, so much so that anesthesia in general should not contribute to mortality in any significant manner.As congenital cardiac anesthesiologists, we must harness the “big data” in EHRs and databases to address our contribution to morbidity and mortality.
4
Our responsibility is to proactively engage with databases to ensure that meaningful information is collected to address morbidity. For example, recent changes were implemented in the STS CHSD that improve granularity and value in data surrounding transfusion and blood management. Anesthesiologists now enter actual doses of factor concentrates administered, milliliters of blood rather than total unit exposures, and information regarding laboratory evaluation during cardiopulmonary bypass. These changes are due to the active involvement of anesthesiologists in the process.Our expertise should be used to modify other parts of the STS CHSD. For example, we enter medications used on bypass and arrival to the intensive care unit, but we do not enter total doses or infusion rates. This information, particularly as it pertains to sedative-hypnotic medications, may have implications for anesthetic neurotoxicity in the congenital cardiac population. Furthermore, the specific analgesic medication(s) used should be entered, instead of the generic “narcotic.” Even though this addition would add to the volume of information in the database, the additional detail to medication parameters may allow for the assessment of best practices regarding early extubation, inotropic management, and sedation protocols.
We also must recognize the limits of database research. First, veracity—the information in any database is only useful if it is complete and accurate. Different institutions have varied resources available for data entry to contribute to the STS CHSD; administrative staff may be available to input data, or surgeons, perfusionists, and anesthesiologists may be responsible for entering their respective data for a given patient. Accuracy and completeness are a function of individual ownership and may not be detected by quality control mechanisms. Moreover, differing interpretations of the definitions of variables, for example “preoperative ventilation,” allow for the potential for data manipulation; variables can be manipulated in such a way as to favor an improved O:E ratio.
Related to the idea of database accuracy is the fact that the STS CHSD is far from the only database addressing quality in outcomes in children with congenital heart disease (Table 1).
Joined the Society for Thoracic Surgeons database in January 2018.
5
Currently, data regarding care of children with congenital heart disease are scattered across registries and databases with overlapping target populations and variables but inability to interact. For example, the IMproving Pediatric and Adult Congenital Treatments Registry collects data on cardiac catheterization but is not set up for information sharing with the STS CHSD.6
Institutional involvement in more than 1 registry duplicates data entry effort and financial commitments. However, the variety in data fields across institutional EHRs and/or databases hinders linkage between systems. The STS CHSD is working toward a global patient identifier that works across databases, but this is not yet a reality.7
Table 1Databases and Registries Involving Care of Children With Congenital Heart Disease in North America
| Abbreviation | Database and/or Registry Name | website |
|---|---|---|
| C3PO | Congenital Cardiac Catheterization Project on Outcomes | https://c3po-r3.chboston.org/#/home |
| CCISC | Congenital Cardiovascular Interventional Study Consortium | https://ccisc.med.wayne.edu |
| CHSS | Congenital Heart Surgeons’ Society Data Center | http://data-center.chss.org |
| ECHSA | European Congenital Heart Surgeons Association Database | http://www.echsa.org/congenital-database-history |
| ELSO | Extracorporeal Life Support Organization Registry | https://www.elso.org/Registry.aspx |
| IFCIR | International Fetal Cardiac Intervention Registry | http://www.ifcir.org |
| iPHFR | International Pediatric Heart Failure Registry | https://ishlt.org/registries/international-pediatric-heart-failure-registry |
| ISHLT | International Society for Heart and Lung Transplantation | https://ishlt.org/registries/overview |
| PCMR | Pediatric Cardiomyopathy Registry | https://dev.childrenscardiomyopathy.org/Pediatric-Cardiomyopathy-Registry-71-315 |
| STS CHSD | Society for Thoracic Surgeons Congenital Heart Surgery Database | https://www.sts.org/registries-research-center/sts-national-database/sts-congenital-heart-surgery-database |
| TOPP-2 | Tracking Outcomes and Practice in Pediatric Pulmonary Hypertension | https://www.peph-association.org/index.php/topp-2.html |
| PAC3 | Pediatric Acute Care Cardiology Collaborative | https://pac3quality.org |
| PC4 | Pediatric Cardiac Critical Care Consortium | http://pc4quality.org |
| CNOC | Cardiac Neurodevelopmental Outcome Collaborative | https://www.cardiacneuro.org/year-1-accomplishments |
| IMPACT | Improving Pediatric and Adult Congenital Treatments | https://cvquality.acc.org/NCDR-Home/Registries/Hospital-Registries |
| NPC-QIC | National Pediatric Cardiology Quality Improvement Collaborative | https://npcqic.org |
† Part of International Society for Heart and Lung Transplantation.
Second, quality improvement is a dynamic process that outpaces the ability of a large database to collect, analyze, and report information. The velocity of entry of new data into the EHR or given database is so rapid that statisticians cannot synthesize and use data in a real-time fashion. For example, Jacobs et al. reported outcomes from 2014–2017 in the winter of 2019. During this 4-year period at our institution, we adopted a hemostasis management system to monitor anticoagulation in children undergoing cardiac surgery with cardiopulmonary bypass, eliminated fresh frozen plasma in pump prime of children more than 4 kg, reduced transfusion of platelets and cryoprecipitate through an initiative to check laboratory values on bypass, and instituted transfusion algorithms to standardize blood management. In parallel, our perfusionists changed equipment and practices, and we had a change in surgeon staffing. Our dynamic environment outpaces the ability to collect, analyze, and report meaningful data. The fluidity of progress is lost if we must wait years to determine the effect of a given practice change.
Third, databases exist to allow for continuous quality improvement, but analyzing information to assess quality is not easy. The STS CHSD is housed at the Duke Clinical Research Institute, and access to data is restricted due to the cost of data harvest and statistical analysis. An anesthesiologist with a clinical question related to quality must submit a data request.
7
Due to the complexities of big data, even a minor data request may take months. A major data request, which addresses a significant clinical question with anticipated publication, may take a year or more.However, individual institutions do have full access to all of their own data, which can be internally analyzed without Duke Clinical Research Institute support. This is an important tool that participants may use to conduct ongoing quality assurance. Moreover, recent advances in the willingness of institutional review boards to allow multi-institutional studies and de-identified data sharing enhance the ability of participating institutions to partner and share information to improve practice. In parallel, clinical researchers should actively work with local institutional review board personnel to address data ownership and change consent paradigms to allow more research projects using big data.
8
Anesthesia should never be the cause of major morbidity or mortality in children undergoing surgery for congenital heart disease, but the shifting focus on morbidity and public outcomes reporting requires that we address our role forthright. We must demonstrate value by engaging with our cardiology, cardiac surgery, intensive care, and perfusion colleagues. We are well-poised to lead initiatives in blood management, medication safety, and neurotoxicity awareness and must increase involvement in database tasks to have our voices and expertise included.
Data are not knowledge. Databases facilitate access to all that we need to improve care and outcomes for our patients, but they are a victim of their own success. It is the analysis of data and collaboration among providers and health systems that allow transformation of data to knowledge and knowledge into improved outcomes, and this must be our goal as anesthesiologists.
Conflict of interest
There is no conflicts of interest.
References
- Big data and paediatric cardiovascular disease in the era of transparency in healthcare.Cardiol Young. 2016; 26: 1597-1602
- The Society of Thoracic Surgeons Congenital Heart Surgery Database: 2019 update on outcomes and quality.Ann Thorac Surg. 2019; 107: 691-704
- The Society of Thoracic Surgeons Congenital Heart Surgery Database public reporting initiative.Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2017; 20: 43-48
- How big data shape paediatric anaesthesia.Br J Anaesth. 2017; 119: 448-451
- Clinical databases and registries in congenital and pediatric cardiac surgery, cardiology, critical care, and anesthesiology worldwide.World J Pediatr Congenit Heart Surg. 2017; 8: 77-87
- IMPACT Registry and National Pediatric Cardiology Quality Improvement Collaborative: Contributions to quality in congenital heart disease.World J Pediatr Congenit Heart Surg. 2019; 10: 72-80
- Anesthesia and databases: Pediatric cardiac disease as a role model.Anesth Analg. 2017; 124: 572-581
- Dilemmas in medical ethics in the age of big data.Mult Scler. 2017; 23: 1306-1308
Article info
Publication history
Published online: May 08, 2019
Identification
Copyright
© 2019 Elsevier Inc. All rights reserved.
