Objectives: To review the data quality and validity in the nationwide Swedish Heart Surgery register for patients operated on the proximal thoracic aorta.
Design: Medical records from a random sample of 300 patients in The Swedish Heart Surgery register were reviewed with register data items systematically re-reported. Variable reporting frequency, proportion of adequately reported data, and number and correctness of diagnostic and procedural codes were analysed.
Results: After exclusions, 251 patients (84%) remained in the analysis. Reporting frequency for individual items varied from 12% to 100% (median 61%). For core variables, reporting frequency was 96%-100%. In 40 of 43 (93%) reviewed variables, registry data were at least 85% correct. A total of 485 diagnoses and 673 procedures were reported, compared to 617 diagnoses and 758 procedures identified in the review process.
Conclusions: The register data quality and validity for patients operated on the proximal thoracic aorta was satisfactory overall, but need further improvement for complications. The register coverage and completeness was very high. Register-based reports should be accompanied by review of data quality.