Does uncollected medication reduce the validity of pharmacy dispensing data?

Purpose When using pharmacy data as collected in the InterAction database (IADB) for pharmacoepidemiology studies, we tend to ignore the fact that filled prescriptions are not always collected by the patient. This study investigated whether unclaimed prescriptions pose a validity threat for pharmacy data, by estimating the percentage of filled prescriptions that are not collected, the percentage of patients who do not collect their filled prescriptions and describing the items remaining unclaimed. Methods Prospective study in three independent pharmacies in the region that is covered by the IADB. All prescriptions that entered these pharmacies on 3 days in 1 week in October 2002 were monitored for a month with respect to whether and when they were claimed. Results A total of 3946 prescriptions concerning 3082 patients were filled. The majority of prescriptions were collected the day they were filled. In total 18 prescription items (0.46%) were not collected within 1 month; excluding health products and homeopathic drugs 13 remained. These 13 covered a variety of drug groups; 0.45% of the patients did not claim their medication within 1 month. Conclusions Primary non‐compliance due to not claiming medication has little impact on the validity of pharmacy dispensing data in the region under study. Copyright © 2004 John Wiley & Sons, Ltd.