Comparison of measures of medication persistency using a prescription drug database

Purpose Studies of medication persistency using drug databases use different definitions of persistency, making it difficult to compare the results from separate studies. We undertook a study of persistency to statins using various definitions to compare the results obtained with the different definitions. Methods All patients with an acute myocardial infarction in the province of Quebec between April 1999 and March 2004 who filled a prescription for a statin within 30 days of discharge were identified. The main outcomes were the 1-year rates of persistency defined as (1) the proportion of individuals with a medication possession ratio ≥80%, (2) the proportion of individuals having filled a prescription in the last 60 days of the year, and (3) the proportion of individuals with continuous exposure after the initial prescription, using a grace period of 7 days or 25% of the duration of the previous prescription between successive prescriptions. Kaplan-Meier analysis was also performed to assess continuous persistency over time using a 7-day grace period. Results Of the 20 239 patients identified, 1 year persistency to statins ranged from 41% to 90%, depending on the definition used: 85% had a medication possession ratio for statins ≥80%, 90% filled a prescription in the last 60 days of the year, 41% to 44% had continuous persistency, and, in survival analysis, the probability of continuous persistency was 41%. Conclusions Measures of medication persistency yield different results, depending on the definition used to define persistency. Results of studies of persistency should thus be interpreted with caution