Vasco M. Pontinha, PharmD
Trajectories of Medication Non-adherence with Time-Varying Predictors and Association with Clinical and Economic Outcomes: A Comparison of Classical Statistical Methods with Machine-Learning Algorithms
This study aims to 1) identify trajectories of medication adherence of chronic diseases treated with oral medications, and 2) distinguish the predisposing, enabling, need, and provider/care characteristics factors that determine trajectory membership using group-based trajectory modelling. Additionally, this study will investigate the association between adherence trajectories and economic and health outcomes. This association will be investigated by deploying two alternative predictive methods, one based on classic logistic regression and the other based on machine-learning algorithms. It is hoped that the findings of this study will elicit longitudinal medication adherence trajectories, the factors that determine trajectory membership, as well as establish optimal medication adherence trajectories based on the association with outcomes. Lastly, the conclusions of this study will allow health care professionals to identify patients at risk and payers to develop new value-based payments schemes based on medication adherence.
This award provides me the opportunity to pursue a research project that I am truly passionate about. The contribution from the PhRMA Foundation will help identify the developmental perspective of how patients establish their medication adherence behavior patterns, while also identifying the life-changing events or time-varying predictors of medication adherence that can help health care professionals design targeted interventions.