The following are key points to remember from this state-of-the-art review of the National Cardiovascular Data Registry Data Quality Program 2020:

1. The National Cardiovascular Data Registry (NCDR) is a group of hospital-based and outpatient registries maintained by the American College of Cardiology (ACC) Foundation to support quality improvement, assess effectiveness and safety of therapies, and inform research. Given the broad use of the NCDR registries, ensuring high-quality data is critical. This paper is an update to a previous report examining the data collection process and assessing data for completeness and integrity.
2. Data Collection—Registry data elements are well defined in a data dictionary and include measures that are important quality metrics and support risk-adjustment models. These data elements undergo periodic review. For those participating in the registry, online resources, regular teleconferences and calls, annual meetings, and clinical support infrastructure is available to help promote high-quality data. Data entry is completed through a combination of manual abstraction and database transfer from compatible clinical applications for hospital-based registries, and through automated data extraction from the medical record for outpatient registries.
3. Data Quality Checks and Data Quality Reports—Data submissions go through quality checks to ensure completeness and integrity. Data Quality Reports are generated and given back to hospitals to review, and an opportunity is given to complete or correct submissions. Submissions that pass these quality checks get added to the registry data warehouse and are available for use. Data that pass integrity checks but not completeness checks are still added to the data warehouse but not included for use in research and aggregate registry national reports. Submissions that do not pass integrity checks are not added to the data warehouse. Based on 2019 data, 95.8% of all submissions met all quality checks.
4. Data Outlier Report—For each registry, biannual reports are created to identify potential outliers for specific data elements and metrics, such as lower than expected event rates or fewer than expected submissions of follow-up testing. Hospitals that are found to be outliers are notified via a Data Outlier Report and sites are asked to review data for accuracy. Approximately 5% of registry sites received such a report.
5. Data Audit Program—For each registry, a random sample of 7-10% of participating sites undergo an audit to assess if selected data elements are accurate and if all eligible episodes of care are reported. Each site randomly selects a cohort of patients for review and submits the clinical records to an independently contracted organization for audit. Data elements are abstracted and compared to the original submission. Senior or supervisory auditors also performed secondary audits on a subset of cases to ensure internal consistency. Hospitals receive reports of these audits and can appeal results if desired. These audits are meant to improve future data quality and original data submissions are not changed. A smaller group of the audited hospitals also submit patient lists based on billing codes to be compared to registry submissions to assess if all eligible patients have been submitted.
6. Clinical Event Adjudication—For two registries, the STS (Society for Thoracic Surgeons)/ACC TVT Registry and LAAO (left atrial appendage occlusion) Registry, reporting of select clinical adverse events required additional information to be collected through an adjudication form. This is done to ensure consistent reporting based on registry data definitions for post-approval study requirements. As a point of reference, for the TVT registry, hospital-submitted and adjudicated event reporting matched 89% of the time.
7. Data Audit Report—When audits are complete, a summary is generated by the independent contracting organization that includes facility accuracy, completeness of records, accuracy of abstracted data elements, and interrater reliability of the audit abstractors. Regarding audits for long-term outcomes data, the current approach is limited by data only available from index hospitalizations and lack of linked administrative data to account for all care episodes.
8. Data Quality of the NCDR Registries—Data from the most recent audits of the registries demonstrates the significant effort put into improving data quality. Across the registries, interrater reliability of the data abstraction was high (range 93%-99%) and mean agreement with sites was high (range 90%-95%).
9. Improving Data Quality—The NCDR continues to expand its data quality efforts. Attention is being paid to enhancing data collection instruments, ensuring consistency across registries, shortening audit times, expanding outlier reporting, and capturing all eligible patients.
10. Final Message—The NCDR Data Quality Program has put forward a tremendous effort to ensure data quality for these important registries. Current analysis shows that the registry data are generally complete and accurate, though variability exists in some areas and represents a point for future improvement.

Clinical Topics: Cardiac Surgery, Cardiovascular Care Team, Invasive Cardiovascular Angiography and Intervention, Prevention, Cardiac Surgery and Arrhythmias

Keywords: Atrial Appendage, Cardiac Surgical Procedures, Data Accuracy, Data Management, Data Warehousing, Episode of Care, LAAO Registry, Medical Audit, Medical Records, National Cardiovascular Data Registries, Outpatients, Quality Improvement, Quality of Health Care, Registries, Secondary Prevention, STS/ACC TVT Registry, Telecommunications

< Back to Listings