PNS13 IDENTIFYING INDIVIDUALIZED MEDICATION ADHERENCE INTERVENTIONS BASED ON PREDICTIVE MODELS OBTAINED THROUGH A GROUP-BASED MULTI-TRAJECTORY APPROACH

Abstract

BACKGROUND Medication non-adherence is a major obstacle to improving healthcare outcomes in long-term therapies for chronic diseases. Previous research has used group-based trajectories methods (GBTM) to identify longitudinal adherence trajectories. However, medication adherence is not an isolated behavior. Instead, it’s influenced by factors that many current interventions fail to address. Current measures like PDC and MPR are well established proxy-measures of medication adherence but have several limitations. For example, they are insensitive to temporal changes and classify patients dichotomously (e.g., PDC > 80%). An analysis of medication adherence should consider the longitudinal complexity of adherence behavior, as well as the characteristics of the patient, the disease, etc. GBTM has been used to identify clusters of patients who follow similar longitudinal behavior patterns. However, no previous study has investigated the longitudinal variances of the predictors of medication adherence. If adherence can be fluid over time, so can the patient and contextual characteristics that predict that behavior. This study will determine longitudinal group trajectories and its adherence predictors, such as healthcare services utilization, other patient and disease information. The analysis will be based on a multi-trajectory GBTM. Data from the Medical Expenditure Panel Survey data from 2016-2017 will be used. Economic burden of disease will be calculated for each trajectory group. A multi-trajectory group-based model will produce a matrix-like graph, that will demonstrate the degree to which each predictor variance influences medication adherence. The economic burden of non-adherence will be estimated for each trajectory group, as well as for indicators of healthcare resource utilization. The resulting model will allow forecasting the natural evolution of medication adherence behavior, as well as the identification of groups of patient profiles and which factors most influence adherence behavior. The findings from this study will inform pharmacy practitioners for developing targeted medication adherence interventions.