An Approach to Combining Adherence-to-Therapy and Patient Preference Models for Evaluation of Therapies in Patient-Centered Care

Adherence to therapy is one of the major determinants of therapy success, while non-adherence leads to worsening of patient condition and increased healthcare costs. The aim of our work is to evaluate therapies recommended by a clinical practice guideline in order to select a therapy that is most suited for a patient’s adherence profile and accounts for patient’s preferences. We define three broad categories of adherence—good, moderate, and poor. Each category is associated with a single adherence profile that defines patient characteristics and thus describes a typical patient population for that category. Moreover, each category is also associated with an adherence model that defines therapeutic characteristics linked to adherence (e.g., complexity of therapy). We assume that each patient has a preference model that defines preferences for specific therapies (e.g., an attitude toward invasiveness of therapy). Adherence and patient preference models are constructed from preferential information elicited using multiple-criteria decision analysis methods, and they are represented as value functions. Once a patient has been associated with an adherence profile, both models are used to evaluate therapies generated from a guideline. We present an illustrative clinical scenario describing a patient with atrial fibrillation to demonstrate our proposed approach.