Analysis of factors associated with statin adherence in a hierarchical model considering physician, pharmacy, patient, and prescription characteristics.

BACKGROUND Adherence with maintenance drug therapy such as HMG-CoA reductase inhibitors (statins) is typically analyzed from the perspective of patient characteristics. OBJECTIVE To determine the effects of physician and pharmacy characteristics in addition to patient characteristics on variation in adherence rates for 4 statin drugs (atorvastatin, pravastatin, rosuvastatin, and simvastatin) for patients who patronized only 1 pharmacy and 1 prescriber of a statin. METHODS A retrospective cohort study of 6,436 patients who initiated statin therapy was performed from computerized pharmacy records of 2 large national pharmacy chains. Adherence was defined as the number of 30-day refills within 12 months after initiation of statin therapy. Physician, pharmacy, prescription, and patient covariates were considered in a cross-classified hierarchical regression model. RESULTS The average number of refills dispensed was 4.75 per patient. Patients younger than 50 years had, on average, 13.6% fewer refills per year than did patients older than 70 years (P<0.001). Women were 4.4% less adherent than men (P = 0.041). Patients residing in southern states were significantly less adherent than were other patients; they had 19.4% fewer refills per year than did patients from western states (P<0.001). Each prescription dispensed for comorbid conditions increased adherence by 2.0% (P =0.002), and patients with a history of cardiovascular drug use were 14.1% more adherent than were other patients (P <0.001). Patients on a higher statin dose appeared to be 8.4% less adherent than were patients on a lower dose (P <0.001). Adherence was greater as the number of prescribed refills increased, with a rate of 2.1% per refill (P <0.001). Adherence was lower for patients with higher copayments, with a rate of 2.2% per each additional $10 of copayment (P <0.001). For patients treated by physicians in the top 2.5 percentile and bottom 2.5 percentile of statin adherence, mean refill counts per year were 6.1 and 2.9, respectively. For patients who patronized pharmacies in the top 2.5 percentile and bottom 2.5 percentile of statin adherence, mean refill counts per year were 6.6 and 2.5, respectively. Adherence increased at a rate of 28.4% per each additional 100 statin patients per patronized pharmacy (P <0.001) and decreased at a rate of about 6.5% per each additional 10 statin patients per treating physician (P <0.001). CONCLUSION Because of the variability in adherence rates across pharmacies and physicians, further assessment of pharmacy and physician characteristics in addition to patient characteristics may be of value in improving adherence.