Introduction of a population-based vascular registry: Validity of data and limitations of registration

A study was performed to assess the validity of the data in a nationwide vascular registry, Finnvasc, initiated in 1991. Data recorded during the first year were of 3507 procedures (immediate reoperations were excluded); 2436 (69 per cent) were surgical and 1071 (31 per cent) endovascular in nature. The Finnvasc data were compared with 2325 hospital records and re-registration of a random sample of 233 patients treated during October 1991 in nine centres was also carried out. In all, 21 per cent of patients from hospitals participating in Finnvasc were not registered and so it was calculated that about 4500 vascular procedures, 3125 surgical and 1375 endovascular, were performed in Finland in 1991. This represents an average of 900,625 and 275 procedures per million population per year respectively. The non-registration rate in October was 17 per cent. Apart from operation classification, and function and patency at 1 month, all simple binary and code data at re-registration were in agreement in at least 92 per cent of cases; checklist and numerical data did not show such good conformity. During the first year of the registry many data were either missing or inadequately recorded, emphasizing the need for cross-validation and re-registration of a random sample of cases. Furthermore, the non-uniformity of coding between hospital records and the Finnvasc registry illustrated deficiencies in the current coding system.