PROJECT SUMMARY Heart failure (HF) is the most common hospital discharge diagnosis among older adults in the United States. Strikingly, 2 in 5 patients are readmitted within 1-year following their first HF admission. This results in significant potentially avoidable costs to our already strained healthcare system, since hospitalizations result in 70% of yearly HF management costs. The most common causes of readmission are failure to recognize clinical worsening and dietary nonadherence. As a result, HF management is rapidly evolving from the traditional model of face-to-face follow-up visits toward a proactive real-time technological model of assisting patients with monitoring and self-management while in the community. While several HF studies have implemented technologically-driven disease management programs, results have been mixed. These systems effectively deliver monitoring data and alerts to healthcare providers. However, their effectiveness in delivering behavioral interventions to patients and modifying patient behavior, a crucial factor in HF self-management, is unknown. Therefore, it is imperative to develop and test patient- centered technologies that deliver behavioral interventions to promote self-management for the most common causes of readmissions with an overall goal of reducing HF readmissions. This proposed project will determine the effectiveness of two interventions within a mobile application and builds on our work from previous research. Our central hypothesis is that a patient-centered mobile application with contextual just-in-time interventions about self-management during a clinical worsening and dietary sodium will improve the health status of HF patients. The rationale for this project, which is supported by our preliminary data, is that a new model for disease management - placing patients in control of their condition - will have a substantial positive impact on HF outcomes. Our objectives are to: (1) determine the impact of two unique adaptive mobile application interventions on HF readmission and health-related quality of life (HRQOL) in HF patients, (2) establish the effect the interventions have on proximal outcomes and that the proximal outcomes mediate the intervention’s impact on HF readmission and HRQOL, and (3) develop data-driven machine learning models that can predict episodes of clinical worsening.