Heart failure (HF) is among the most common and costly chronic illnesses in older adults in the United States. Medication adherence is a critical component of long-term self-management in HF and is associated with improved symptom management, physical functioning, and the recurrence of complications. Despite the well- established evidence that medication adherence improves outcomes in HF, only half of patients with HF achieve adequate medication adherence. Although clinical guidelines emphasize the long-term benefits of medication adherence for HF outcomes, we lack critical knowledge of actionable time point(s) to effectively promote medication adherence during the course of illness. To date, current studies of adherence in HF patients have largely ignored heterogeneous patterns of adherence over time and are agnostic to the classes of medication. Indeed, there is evidence to suggest that differences in short- and long-term patterns of adherence of certain medications may be associated with important patient characteristics. Therefore, there are urgent needs to accurately classify longitudinal patterns of adherence based on medication class and to understand the factors associated with distinct patterns of adherence in patients with HF. This information is crucial for developing and implementing tailored interventions to improve adherence in this vulnerable population. To address this gap in knowledge, we propose to carry out a series of analyses that use a novel method?group-based trajectory models?and leverage the strengths of two national datasets: (a) Medicare claims data, and (b) the Health and Retirement Study (HRS). In doing so, our overall objectives for the current proposal are twofold. First, linking Medicare claims to the HRS data, we will first classify the medication adherence trajectories of the guideline- recommended classes of medications (angiotensin-converting enzyme inhibitors [ACEI] /angiotensin II receptor blockers [ARB], and Beta blockers) separately in patients with HF. We will then simultaneously examine the longitudinal patterns of adherence across the classes of medications (i.e. group-based multi trajectories). Second, guided by the World Health Organization model of adherence, we will first examine how patients? demographics, socioeconomic status, patient-, condition-, therapy-, and healthcare system-related characteristics at the time of HF diagnosis are associated with trajectory typologies of medication adherence that we identified in Aim 1. We will then assess how changes in these factors are related to the assignment of a patient to certain trajectories of medication adherence. We will examine factors that are associated with trajectories for each class of medication as well as factors that are contributing to a patient?s multi-class trajectories. Results from this study will provide an important scientific foundation for a future large-scale proposal to develop tailored strategies to improve adherence according to identified, predictable time points in the illness trajectory.
Medication adherence is a critical component of heart failure (HF) self-care. Despite the well-established evidence that medication adherence improves outcomes in HF, only half of patients with HF achieve adequate medication adherence. We propose to leverage the strengths of two national datasets and group-based multi trajectory modeling to identify distinct trajectories of medication adherence and the associated factors in older adults diagnosed with HF to provide a greater understanding of how patients? medication adherence changes over the course of their illness.
