OBJECTIVES: Previous work on primary nonadherence has used a wide range of follow-up periods from 30 days up to 18 months, making results difficult to compare. We sought to evaluate primary nonadherence by measuring time until filling in a cohort of elderly patients. METHODS: Data comes from a linked database of electronic health records (EHR) and claims for patients aged ≥65 years enrolled in Medicare Parts A/B/D. From these data, we identified patients receiving a new prescription for a chronic disease medication. We then followed patients for a fill of the index prescription for one year. Cox models were used to assess differences in time until filling across patient and therapeutic characteristics. RESULTS: In 28,770 new medication orders, the majority (58%) were filled within 1 day, 81% were filled within 30 days, and 93% were filled by the end of one year. The rate and timing of filling was similar across therapeutic areas except for insulin, where filling rates were much lower. Within therapeutic areas, time until filling was generally similar across medication classes. Prescriptions with supply > 30 days had a 13% (11%-15%) lower rate of filling than prescriptions with supply < 30 days. Patients with increasing numbers of current medications were much more likely to fill their new prescription. Patients with 3-4 medications had a 71% (65%-77%) higher rate of filling than patients with 2 or fewer medications. Patients with 5-7 medications had twice the rate of filling (HR: 2.06 [1.99-2.14]), and patients with more than 7 current medications had 2.4 (2.3-2.5) times the rate of filling. CONCLUSIONS: With the exception of insulin, rates of primary nonadherence were low, and the majority of prescriptions were filled quickly. Focusing interventions on patients with few current medications or those initiating specific medications, such as insulin, may be a promising approach to improve medication adherence.