Editor's Choice - International Validation of the Vascular Registry of Switzerland, Swissvasc: A VASCUNET Report.

Abstract

Data validity and completeness are essential for registry data, and a meaningful interpretation of differences between countries can only be made if the data are robust within each of the countries. Therefore, the VASCUNET collaboration has performed validation of several vascular registries and so far published the validation results of national vascular registries of Sweden, Denmark, and Malta. The findings of the validation of the Swissvasc Registry, the vascular registry of Switzerland, are presented here. Five of 34 hospitals participating in Swissvasc were visited by two international validators. Independent evaluation of the procedures of carotid endarterectomy, infrarenal abdominal aortic aneurysm (AAA) repair, and bypass surgery for lower extremity arterial disease (LEAD) was performed. Local administrative data and Swissvasc Registry data were compared for external validation, as well as registry data of 15 randomly picked cases within each group with data from individual patient records for internal validation. Hospital administrative data identified 265 carotid, 399 AAA, and 399 LEAD procedures, whereas Swissvasc identified 265 carotid, 398 AAA, and 395 LEAD procedures. Four LEAD procedures were found in hospital administrative data but not in Swissvasc. External validity was 100% for carotid, 99.8% for AAA, and 99% for LEAD. In internal validation, 5.2% of the variable data were missing in Swissvasc, and 2.4% were different from data in patient records, resulting in external validity of 92.8% for carotid, 94.4% for AAA, and 90.1% for LEAD. The VASCUNET template was used for international validation of registry data and demonstrated that Swissvasc is a reliable system of quality data collection for the participating vascular centres, with excellent external validity and good internal validity. Quality of the Swissvasc registry data was clearly superior to local administrative data.