Adherence to Prescription Drugs and Adverse Health Events for the Chronically Ill

Yoon, Jean

Abstract

Previous research has consistently found a strong, positive association of adherence to prescription drugs with health outcomes for chronic conditions. However, few clinical drug trials nor observational studies have been able to interpret these effects without concerns about bias due to omitted variables (e.g., the possibility that patients who are unobservably more severely ill are more likely to adhere but have worse outcomes, or that those with a higher unobservable propensity to engage in risk prevention behaviors are more likely to adhere and more likely to have a good outcome). Many of these studies have also focused exclusively on the elderly. The proposed study will analyze the impact of adherence to prescription drugs on health outcomes in a large, national sample of working-age adults and early retirees covered by employer-sponsored health insurance coverage. This study will focus on hypertension and depression as these conditions are among the most prevalent chronic conditions among adults in the U.S. This study will 1) predict adherence to prescription drugs with both personal and health insurance characteristics, 2) test a semiparametric model of drug adherence to see if personal and health insurance characteristics influence adherence more at higher or lower levels of adherence, and 3) determine the effect of adherence on health outcomes by using simultaneous equations methods to account for biased estimates arising from omitted variables in cross-sectional analyses. By estimating adherence and health outcomes simultaneously, the model allows for the error terms of the two equations to be correlated; that is, the same unobservable characteristics influencing adherence are also allowed to influence health outcomes, thereby eliminating any bias due to omitted variables. This study will use medical and pharmacy claims for a short, two-year panel design. Adherence will be measured in study year one to look at the impact on health outcomes in study year two. These results can help inform clinicians, employers, and policymakers about how much effort to spend on increasing adherence to prescription drugs and how mutable adherence is with specific drug benefit designs and targeted clinical interventions. ? ? ? ?