Background: Patients treated in a usual-care setting1026 often do not comply with prescribed treatment regimens as closely as those treated in a clinical trial, for a variety of reasons. Objectives: We sought to describe compliance over 1026 3 years with a statin treatment in an observational study in a usual-care setting and to investigate potential risk factors for poor compliance. Methods: Enrollees in a Massachusetts health maintenance1026 organization (HMO) who presented with baseline low-density lipoprotein cholesterol (LDL-C) ≥130 mg/dL and who started statin treatment between January 1, 1994, and July 1, 1999, were followed until death, termination of membership in the HMO, or July 1, 1999, whichever came first, as documented by the HMO's drug-dispensing electronic records. The records included information about the demographics, service dates, laboratory data, and filled prescriptions. The outcome measures were time to treatment discontinuation, time to treatment resumption, and adherence to treatment. Results: A total of 4776 enrollees were included in 1026 the analysis. Most patients were aged 50 to 69 years (57%), men (52%), had LDL-C ≥160 mg/dL (77%), and received their prescription for antihyperlipidemia therapy from an internist or family physician (87%). The hazard of discontinuation was high during the first 6 months of therapy, with 20% of the initial population discontinuing treatment during this time period, then decreased, so that the fraction of continuing users leveled off: 74%, 65%, and 61% of statin initiators were still on treatment 1, 2, and 3 years after entry, respectively. The probability of resuming statin treatment was 51% within 24 months after last prescription filled. The proportion of adherent users stayed stable, between 53% and 55%, after the first 6 months. Risk factors of treatment discontinuation, treatment resumption, and adherence were remarkably similar and were stable over time. The significant predictors of treatment discontinuation were age <50 years (hazard ratio [HR], 1.45 [95% Cl, 1.26–1.68]), female sex (HR, 1.18 [95% CI, 1.06–1.30], and previous antihyperlipidemia treatment (HR, 0.71 [95% CI, 0.63-0.79]). Conclusions: In this HMO setting, women and individuals 1026 aged <50 years were at risk for poor compliance with statin therapy. Our analysis suggests that the association between compliance and age or gender may be quite stable over the first 3 years of treatment.