Using claims data linked with electronic health records to monitor and improve adherence to medication

Poor adherence to medication is a serious problem in the United States, leading to complications and preventable hospitalizations, particularly for patients with chronic diseases. Interventions have been proposed as a means to improve adherence to medication, but the optimal time to perform an intervention has not been well studied. We provide a use case for how claims data linked with electronic health records (EHRs) can be used to monitor patient adherence to medication and provide a source of information to help decide when to perform an intervention. We propose a Markov decision process (MDP) model to determine when to perform adherence-improving interventions based on a patient’s EHR. We consider the the societal perspective where we trade off maximization of time to first adverse health event and minimization of cost of interventions, medication, and adverse events. We use our model to evaluate the costs and benefits of implementing an EHR-based active surveillance system for adherence-improving interventions in the context of cardiovascular disease management for patients with type 2 diabetes. We also provide some theoretical insights into the structure of the optimal intervention policy and the influence of health risks and costs on intervention decisions.