Quality control activities associated with registries in interventional cardiology and surgery.

OBJECTIVE: To describe and outline audit and quality control activities of the multicentre interventional and cardiac surgery registry in Victoria as a potential model for a national registry. DESIGN, SETTING, AND PATIENTS: The Melbourne Interventional Group (MIG) database is a prospective multicentre registry recording consecutive percutaneous coronary interventional (PCI) procedures across eight Victorian hospitals. Similarly, the Australasian Society of Cardiac and Thoracic Surgeons (ASCTS) database captures cardiac surgical activity across six Victorian hospitals. Auditing of each registry involved systematic selection of baseline, clinical and procedural variables from 5% of procedures to examine for data integrity and mismatches. MAIN OUTCOME MEASURES: Performance trend and data accuracy of each registry was assessed by the number of mismatches detected during the auditing process for different demographic, clinical and procedural variables and across different (de-identified) sites.
RESULTS: Over two auditing phases from 2004-2006 and 2007, 10 (4.3%) of variables from 3% of all PCI procedures and 15 (6.4%) variables from 5% of PCI procedures were analysed. There was 96.5% agreement during the first auditing phase of the MIG registry with an average of 0.35 mismatches per audit (CI 0.28-0.42), whereas during the second audit phase, agreement was up to 97% with 0.32 mismatches per 10 fields per audit (CI 0.25-0.40). The ASCTS database audit selected 39 (14.8%) variables from 5% of annual surgical cases across six cardiac surgical centres with an overall 96.7% agreement.
CONCLUSION: The current auditing process of these two databases is rigorous, robust and reflects a high degree of accuracy of data collected by participating hospitals.