OBJECTIVES: While managed care data can be used to assess medication compliance and persistence, these databases generally do not permit observation of long-term outcomes. We therefore applied modeling techniques to explore the impacts of compliance and persistence with hypertension therapy on long-term cardiovascular outcomes among managed care hypertension patients. METHODS: We analyzed 12-month blood pressure (BP) and compliance/persistence using data for hypertension patients receiving valsartan therapy from the Geisinger Clinic, a U.S. regional healthcare network. Long-term cardiovascular outcomes (myocardial infarction, stroke, coronary artery bypass, angioplasty, and cardiovascular death) were based on algorithms from the Physicians' Health Study (13 years follow-up) and Women's Health Study (7 years) as presented by Glynn et al. (2002). No significant differences were observed between the compliant/persistent and non-compliant/non-persistent groups for other cardiovascular risk factors (age, diabetes, smoking status, cholesterol, and alcohol). RESULTS: Compliant patients (those with medication-possession ratios > 80%) had 12-month average systolic and diastolic BP values 3.5 mmHg (p<0.0005) and 0.5 mmHg (non-significant) lower, respectively, than those of non-compliant patients. Assuming these differences remain over time, the Glynn algorithms predict a 4.6% decrease (95% confidence interval 2.4%, 6.9%) in the relative risk of adverse cardiovascular outcomes for men and a 16.3% decrease (95% CI 12.2%, 20.6%) for women. Persistent patients (those remaining on valsartan therapy for 12 months) had final systolic and diastolic BP values 4.1 mmHg and 2.5 mmHg (both p<0.05) lower than those for non-persistent patients, projecting an 8.4% decrease (95% CI 3.1%, 13.9%) in relative risk of adverse cardiovascular outcomes for men and a 26.6% decrease (95% CI 19.8%, 33.9%) for women. CONCLUSIONS: Compliance with and persistence to hypertension therapy predicted substantial decreases in adverse cardiovascular outcomes. These impacts may also lead to improved patient-reported outcomes and decreased costs, and can be used to assess long-term treatment cost-effectiveness.