Aims: To study whether there is agreement between cardiovascular drugs presented at patient interview and pharmacy records in an elderly population.
Setting: The Rotterdam Elderly Study, a prospective population based cohort study of people older than 55 years of age.
Methods: We compared cardiovascular drug use as presented during patient interview with the computerized pharmacy medication history. Concordance of cardiovascular drug use in patient interview and pharmacy data was measured by merging the two data sets and assessed in two episodes: in a period of 6 months before patient interview and within the legend duration of each cardiovascular drug.
Results: We found 2706 concordant pairs in a total of 3365 prescriptions (80.4%) in the merged data. There were 195 prescriptions presented at patient interview which had not been filled at the pharmacy, and 464 which had been filled but were not presented by the patient. The percentage of concordant pairs slightly increased to 80.6% (2275 of a total of 2824) for prescriptions of which the legend duration included the date of patient interview. The highest agreement was noted with beta-adrenoceptor blocking agents with Kappa statistics of 0.97 for atenolol and metoprolol. The lowest Kappa statistic was found for organo-heparinoid, mainly as ointments against haemorrhoids (0.26).
Conclusion: The agreement between patient interview and pharmacy records is good for prescription only drugs, and pharmacy records are a useful resource for pharmacoepidemiological studies on cardiovascular agents.