Abstract
Background: Research on medication adherence is primarily conducted using insurance or pharmacy claims data, focusing on patients who have already obtained their first prescription for a defined condition. Claims data alone cannot be used to assess first-fill rates.
Methods: We conducted a retrospective cohort study that examined first-fill rates for patients prescribed medications to treat diabetes by linking prescribing information from electronic health records (EHR) to pharmacy claims data of one insurer. EHR data were obtained from the Geisinger Clinic, a large group practice serving a 31-county area of central and northeastern Pennsylvania. The Clinic includes 41 community practice sites with primary and multi-specialty care, all of which have used an EHR since 2001. Claims data were obtained from Geisinger Health Plan, a payer that accounts for 30% of Geisinger Clinic patient volume. Adherence was defined as a claim generated for the first-fill prescription within 30 days of the index date.
Results: Of 1132 patients written a new, first-time prescription for any diabetes medication, 962 (85%) generated a corresponding claim within 30 days. Therapeutic class, refills, co-pay, baseline A1c values, and change in A1c values were associated with first-fill rates. Gender, race, age, comorbidity score, number of office visits, and number of medications prescribed had no association (P<0.05).
Conclusions: Adherence relates directly to higher A1c levels and is associated with therapeutic drug class, higher number of refills, and lower co-pay. As expected, patients with higher adherence exhibited greater change in A1c levels. Patients with relatively lower baseline A1c levels, while at lower risk for diabetes complications, are at higher risk of not adhering to diabetes medications.
- Received September 11, 2008.
