Abstract 11963: A Digital Heart Failure Coach Intervention for Improving Education, Self-Management and Outcomes in Patients Admitted for Acute Decompensated Heart Failure: Barriers to Effective Implementation

Abstract

Introduction: Heart Failure (HF) education is fundamental for improving outcomes. The use of mobile health (m-health) devices and Artificial Intelligence (AI) are becoming more broadly available. We designed an avatar style digital HF coach (HF-app) to improve HF knowledge and self-care management. Hypothesis: We hypothesized that an avatar style HF-app could improve outcomes by enhancing HF knowledge and improving quality of life and self-care behaviors in patients admitted for Acute Decompensated HF (ADHF). Methods: The Risk-HF trial recruited patients admitted for ADHF who were at high risk for hospital readmission and/or death (defined by a previously developed risk score, ETHELRED trial). Patients were randomized to usual care or intervention, in which the HF-app was a component of a multifactorial intervention aimed at reducing HF readmissions. Our HF-app was designed to engage daily with the patient by monitoring HF symptoms and delivering education over 52-days. Results: Over a period of 16 months (August 2019 up until December 2020), 36 of 74 high risk patients were randomized to the intervention arm and were offered education with our HF-app. Of these, 26/36 (72%) had restricted access to the HF-app ( unable to use a tablet device and/or unwilling to be taught how to use it, n=20; no internet access, n=4; died in hospital n= 2 ) and 10/36 (28%) were enrolled to use the HF-app [ gender male n=7; median age 71.5 years, p<0.01 (IQR: 57.50 - 78.50); mean years of education 11.6(±2.5), p<0.01 ]. Younger patients [ Odds Ratio (OR) 0.89 [95% Confidence Interval (CI): 0.82-0.97], p<0.01 ] and those with a higher education level [ OR 1.58 (1.09-2.28), p=0.03 ] were more likely to enroll. Of those enrolled, only two patients engaged and completed ≥70% of the program. Six of the remaining 8 patients who did not engage were readmitted. Conclusions: Our study provides a note of caution in studies using m-health devices and apps in critically ill and elderly patients with unstable HF. Future research should consider how to use such technology-based interventions with elderly patients as well as to target younger ADHF patients, which may provide stronger evidence of benefits in patient self-care and outcomes.