External validation of the Swedvasc registry: a first-time individual cross-matching with the unique personal identity number.

Abstract

To study external validity of the Swedvasc registry concerning numbers of procedures and mortality. Vascular registry data for carotid, infrainguinal bypass and aortic aneurysm (AAA) procedures were compared to the Swedish Hospital Discharge Register (SHDR) data, and the National Population Registry (for mortality) by matching every individual patient using the unique personal identity numbers (PINs). The time-period studied was 2000-2004 (5 years) for carotid and infrainguinal procedures. A separate analysis was performed for AAA-surgery in 2006. The external validity for carotid, infrainguinal bypass and AAA repair was 93.4%, 93.0% and 93.1%, respectively. The 30-day mortality was 0.86% after carotid and 2.9% after infrainguinal bypass procedures. Mortality was 2.6% after planned and 25.9% after unplanned AAA repair. Although there was a general trend towards inferior outcomes after procedures not registered in the Swedvasc, those procedures were so few that in none of the analyses did the inclusion of non-registered procedures affect general outcomes significantly. Combining data from both registries, the incidence for carotid, infrainguinal bypass and AAA procedures was 7.8, 15.2 and 13.6 per 100,000 person-years, respectively. In the hospital-specific analysis for 2006 it was shown that the non-registered procedures for AAA were localized to one non-compliant county hospital, and small district hospitals not performing elective AAA-surgery but only rare emergency operations. The external and internal validity of the Swedvasc registry allows to confidently assess volumes of, and mortality after, vascular surgery in Sweden.