Development of a refill pattern method to measure polypharmacy in administrative claims databases.

PURPOSE: The purposes of the study were to develop a refill pattern method to identify polypharmacy in pharmacy billing records and to compare the method with traditional days' supply overlap algorithms. METHODS: This method is characterized by the assessment of prescription refill pattern. Concomitant therapy is assumed when two drugs are dispensed repeatedly during the active days' supply of each other. We tested the refill pattern method in a simplified scenario in which two drugs (methylphenidate/dexmethylphenidate and atomoxetine) for attention deficit/hyperactivity disorder (ADHD) were considered. Children who had at least one prescription of methylphenidate/dexmethylphenidate or atomoxetine in 2008 were included for the calculation of 2-year prevalence of ADHD treatment polypharmacy. Results were compared with traditional method that requires a minimum overlap of 30, 60 or 90 days of filled prescriptions. We compared polypharmacy prevalence estimated by the two methods and explored reasons for disagreement. RESULTS: Among 131 385 children who had at least one prescription of methylphenidate/dexmethylphenidate or atomoxetine, the refill pattern method identified 4021 patients who had ADHD treatment polypharmacy (2-year prevalence = 3.1%). This prevalence estimate fell between those from a 30- to 60-day overlap method. The Cohen's kappa regarding determination of polypharmacy was 0.83, 0.92 and 0.80 considering 90-, 60- and 30-day overlap method, respectively. CONCLUSIONS: The refill pattern method can be used as another way to measure polypharmacy in administrative claims databases and can be adapted to a wide variety of research questions, diseases and study populations. Copyright © 2016 John Wiley & Sons, Ltd.