Accuracy of clinical data in a population based vascular registry.

INTRODUCTION: Clinical databases are increasingly being employed to evaluate the quality of treatments, including patients with peripheral vascular disease. Valid data is vital to the value of these analyses.
OBJECTIVE: To assess the validity of clinical data in a population-based national vascular registry.
DESIGN: Traditional reproducibility study was supplemented by refilling of data by an independent observer, thereby creating three data sets for comparison.
MATERIALS AND METHODS: Twenty prospectively recorded electronic forms from each department were selected randomly from the Danish National Vascular Registry. Data forms were refilled by the surgeons of the department concerned, and by an independent member of the board of the Danish National Vascular Registry. Refilling was performed blinded to the original forms.
CONCLUSIONS: A high degree of accuracy of clinical data can be achieved. An independent observer makes it possible to evaluate the classification of observer dependent parameters and explain differences in the reproducibility of data.