Predictors of Medication Adherence Using a Multidimensional Adherence Model in Patients with Heart Failure

Abstract

Objective: The purpose of this study is to explore important factors contributing to medication adherence in patients with heart failure (HF) based on the World Health Organization's multidimensional adherence model (MAM). Background: Medication adherence in HF is a complex and poorly understood phenomenon. It is imperative that medication adherence be studied using an appropriate model and multivariate approach to get comprehensive information about predictors of medication adherence. Methods and Results: Data from 134 patients with HF were analyzed to determine the best multivariate medication adherence model given variables derived from the MAM. Medication adherence was measured objectively using the Medication Event Monitoring System (MEMS) for three months. Four indicators of adherence were assessed by the MEMS: 1) dose-count, the percentage of prescribed doses taken; 2) dose-days, the percentage of days the correct number of doses taken; 3) dose-time, the percentage of doses taken on schedule; and 4) therapeutic coverage, the percentage of time therapeutic drug action maintained. Barriers to medication adherence, ethnicity, and perceived social support composed the best group of variables associated with dose-count (P < .001). New York Heart Association (NYHA) functional class, barriers to medication adherence, financial status, and perceived social support were the best model associated with dose-day (P < .001). Barriers to medication adherence and financial status were the best group of predictors associated with dose-time (P = .005). The best group of variables of therapeutic coverage was NYHA, barriers to medication adherence, and ethnicity (P < .001). The patient-related and healthcare system-related variables proposed by the MAM were not significant predictors. Conclusion: The most influential variable associated with medication adherence was barriers to medication adherence. Assessing patients’ barriers to taking prescribed medications should be the first step in implementing any intervention to improve medication adherence. Minorities and those who had more severe symptoms as reflected by NYHA functional class, less income, and less social support should be considered at high risk for poor medication adherence. Objective: The purpose of this study is to explore important factors contributing to medication adherence in patients with heart failure (HF) based on the World Health Organization's multidimensional adherence model (MAM). Background: Medication adherence in HF is a complex and poorly understood phenomenon. It is imperative that medication adherence be studied using an appropriate model and multivariate approach to get comprehensive information about predictors of medication adherence. Methods and Results: Data from 134 patients with HF were analyzed to determine the best multivariate medication adherence model given variables derived from the MAM. Medication adherence was measured objectively using the Medication Event Monitoring System (MEMS) for three months. Four indicators of adherence were assessed by the MEMS: 1) dose-count, the percentage of prescribed doses taken; 2) dose-days, the percentage of days the correct number of doses taken; 3) dose-time, the percentage of doses taken on schedule; and 4) therapeutic coverage, the percentage of time therapeutic drug action maintained. Barriers to medication adherence, ethnicity, and perceived social support composed the best group of variables associated with dose-count (P < .001). New York Heart Association (NYHA) functional class, barriers to medication adherence, financial status, and perceived social support were the best model associated with dose-day (P < .001). Barriers to medication adherence and financial status were the best group of predictors associated with dose-time (P = .005). The best group of variables of therapeutic coverage was NYHA, barriers to medication adherence, and ethnicity (P < .001). The patient-related and healthcare system-related variables proposed by the MAM were not significant predictors. Conclusion: The most influential variable associated with medication adherence was barriers to medication adherence. Assessing patients’ barriers to taking prescribed medications should be the first step in implementing any intervention to improve medication adherence. Minorities and those who had more severe symptoms as reflected by NYHA functional class, less income, and less social support should be considered at high risk for poor medication adherence.