Predictors and consequences of nonadherence to antihypertensive medication

Context: Nonadherence to antihypertensive therapy attributes to uncontrolled blood pressure. This may worsen the severity of hypertension and ultimately increase health care costs. This underlines the importance of identifying predictors and consequences of nonadherence to antihypertensive therapy. Objectives: (I) To empirically determine the length of gap between antihypertensive prescription refills that predicts long term prescription discontinuation. This gap can be used to define antihypertensive medication nonadherence. (II) To examine predictors of nonadherence and (III) To evaluate the role of nonadherence on rates of hospitalization and emergency room visits. Design, settings and subjects: For Objectives I and II, retrospective cohort designs were employed on 51,615 subjects enrolled in a large United States pharmacy benefit manager during Jan 1st 2003 and May 31st 2006. Subjects were included if they had [greater than or equal to] 2 prescriptions of antihypertensive medication, were new users and were [greater than or equal to] 30 years. For Objective III, a cross-sectional design was employed on 9,945 subjects aged 30-64 years continuously enrolled in New Jersey Medicaid between Jan 1999 and Dec 2001. For Objective I, Receiver Operating Characteristics (ROC) analysis was performed using maximum gap in anti hypertensive therapy as a predictor of long term treatment discontinuation. For Objective II, time to nonadherence was analyzed with the use of Cox Proportional Hazard Regression model. For Objective III, Log-Linear Regression analysis was utilized to estimate the risk of health care utilization associated with nonadherence Results: For Objective I, ROC analysis generated a C-statistic of 0.87. The cut-off value for maximum gap between refills that optimized sensitivity (0.81) and specificity (0.79) was 75 days. For Objective II, region of the country in which the subjects resided and subjects living in a census block with high percentage of African American population and low levels of income were found to be significant predictors of nonadherence. Subjects who were treated by cardiologists and younger physicians also had improved adherence. Objective III showed that subjects who were nonadherent to antihypertensive therapy had significantly higher rates of hospitalizations as well as emergency visits. Conclusion: Timely corrective interventions to improve adherence will have significant impact on the cost-effectiveness in the treatment of hypertension.Ph.D.Includes bibliographical references