OBJECTIVES To evaluate the effect of a prior authorization (PA) requirement on medication persistence in the elderly following implementation of a step therapy program for proton-pump inhibitors (PPI) in the Military Health System. METHODS Our study population comprised 33,226 TRICARE beneficiaries ≥ 65 years old with a PPI prescription between October 24, 2007 and January 31, 2008. The PA required patients without a PPI prescription in the last 180 days to receive a trial of the preferred product (omeprazole or esomeprazole) before receiving another PPI. The study cohort (n = 15,294) included patients affected by the PA requirement; the reference cohort (n = 17,932) was not affected because patients received an initial fill for the preferred product. Medication persistence was measured as 1) proportion of days covered (PDC) with persistence defined as ≥ 80% of days covered, and (2) time to first gap > 30 days following initiation. Differences in medication persistence between cohorts were detected using logistic regression and a proportional hazards model, adjusted for age, gender, and type of pharmacy (community vs. mail order). RESULTS In the 6 months following initiation of PPI treatment, 46% of study subjects were not persistent (PDC < 80%), while 77% experienced a gap in treatment > 30 days. Exposure to the PA requirement decreased the odds of being persistent with treatment at 6 months by 39% (OR 0.61; 95% CI 0.58¨C0.64) and increased the risk of experiencing a 30-day gap in treatment by 17% (HR 1.17; 95% CI 1.14¨C1.20). CONCLUSIONS The prior authorization program was associated with lower rates of medication persistence. Further study should evaluate whether the reduction in persistence had an effect on patient clinical outcomes.