Assessment of potential drug–drug interactions with a prescription claims database

PURPOSE: The prevalence of 25 clinically important potential drug-drug interactions (DDIs) in a population represented by the drug claims database of a pharmacy benefit management company (PBM) was studied. METHODS: A retrospective cross-sectional analysis of pharmaceutical claims for almost 46 million participants in a PBM was conducted to determine the frequency of 25 DDIs previously identified as clinically important. A DDI was counted when drugs in potentially interacting combinations were dispensed within 30 days of each other during a 25-month period between April 2000 and June 2002. RESULTS: The number of DDIs ranged from 37 for pimozide and an azole antifungal to 127,684 for warfarin and a nonsteroidal antiinflammatory drug (NSAID). The highest prevalence (278.56 per 100,000 persons) and highest case-exposure rate (242.7 per 1,000 warfarin recipients) occurred with the warfarin-NSAID combination. The combination with the lowest overall prevalence (cyclosporine and a rifamycin, 0.10/100,000) differed from the combination with the lowest case-exposure rate (pimozide and an azole antifungal, 0.028 per 1,000 azole antifungal recipients). Number of cases, prevalence, and case-exposure rates for both sexes generally increased with age. An estimated 374,000 plan participants were exposed to a clinically important DDI during a 25-month period. Between 20% and 46% of prescription drug claims were reversed (canceled) for a medication with a drug interaction when a warning about the interaction was sent to the pharmacy. CONCLUSION: Analysis of prescription claims data from a major PBM found that 374,000 of 46 million plan participants had been exposed to a potential DDI of clinical importance.