Completeness and validity of the Danish fracture database.

Abstract

To calculate completeness of the Danish Fracture Database (DFDB) overall and stratified by hospital volume and to calculate the validity of independently assessed variables in the DFDB. In this completeness and validation study, cases registered in the DFDB with fracture-related surgery in 2016 were retrospectively reviewed. All cases had undergone fracture-related surgery at a Danish hospital reporting to the DFDB in 2016. The Danish health care system is fully tax-funded providing equal and free access to all residents. Completeness was calculated as sensitivity and validity was calculated as positive predictive values (PPVs). OVERALL COMPLETENESS WAS 55.4% (95% CI: : 54.7-56.0). For small-volume hospitals it was 60% (95% CI: 58.9-61.1), and for large-volume hospitals, it was 52.9% (95% CI: 52.0-53.7). The PPV for variables of interest ranged from 81% to 100%. The PPV of key variables was 98% (95% CI: 95-98) for operated side, 98% (95% CI: 96-98) for date of surgery, and 98% (95% CI: 98-100) for surgery type. We found low completeness of data reported to the DFDB in 2016; however, in the same period, the validity of data in the DFDB was high.