VALIDATION OF COMPUTERIZED SWEDISH DOG AND CAT INSURANCE DATA AGAINSTVETERINARY PRACTICE RECORDS

Large computerized medical databases offer great potential for epidemi ological research. However, data-quality issues must be addressed. Thi s study evaluated the agreement between veterinary practice records an d computerized insurance data in a large Swedish claims database. For the year 1995, the company insured over 320000 dogs and cats. A total of 470 hard-copy records were sampled from claims for health care (n=2 36) and life insurance (n=234). Computerized insurance data for these claims were accessed and records from the attending veterinary practic es were collected. For health and life claims, respectively, 79.2 and 72.8% of practice records were retrieved. Variables compared between t he computer and practice records were breed, sex, date of birth and di agnosis for the claim. The degree of agreement was categorized as agre ement, minor disagreement, major disagreement or data missing. Multiva riable logistic regression was used to examine factors associated with errors. The observed agreement for breed and sex was excellent. There was 28.9 and 33.8% minor disagreement for the date of birth for healt h and life claims, respectively. This was mainly because, prior to 199 3, the date of birth was coded as the first of January when more compl ete information was unavailable. Major disagreement (different year of birth) was low for both health and life claims. For health and life c laims, the observed agreement for diagnosis was 84.0 and 84.9%, minor disagreement was 6.4 and 4.3%, and major disagreement 5.9 and 9.4%, re spectively. Although there was no difference based on size of the vete rinary clinic involved, there was a tendency for better agreement betw een the practice record and the computerized insurance data for claims from clinics viith computerized practice records (n=40) than for clin ics with manual practice records (n=286). Rates of discrepancy were af fected by the clerk who processed the claims. Records processed by two of the 21 clerks had significantly more disagreements. Given the natu re of the data, it was only possible to calculate a measure of observe d agreement. In general, the agreement between data in the insurance-c ompany database and from the practice records was excellent for demogr aphic data such as breed and sex and fair for diagnostic information a nd date of birth. In general, the data are adequate to support ongoing research with due considerations of certain limitations. (C) 1998 Pub lished by Elsevier Science B.V.