Supporting the continuum of care in coronary artery disease using patient-centered eHealth and artificial intelligence: outcome of the TIMELY multicenter randomized controlled trial

Abstract Background eHealth solutions may optimize personalized care and support disease management for patients with coronary artery disease (CAD) throughout the course of the condition. TIMELY is the first artificial intelligence (AI)-driven eHealth approach based on cardiac rehabilitation (CR) components and integrated with Internet of Things (IoT) devices. Purpose To report the primary outcome of the TIMELY RCT. Methods TIMELY is a multicenter RCT (Germany, Spain, The Netherlands) involving 360 CAD patients. Patients randomized to the intervention group (IG, n=180) received a 6-month, app-based, behavior change program, as part of their CR aftercare (phase III CR). Patients in the control group (CG) received care as usual. Assessments were performed at baseline, 6, and 12 months. The TIMELY platform and app provide support for behavior change and self-management. TIMELY integrates clinical and psychosocial data, evaluates risk, and supports healthy behaviors via app-based feedback and prompts. Engagement in physical activity is supported by a chatbot and personalized exercise prescription based on weekly activity profiles. Ecological Momentary Assessment (EMA) is used. Patients were equipped with an activity tracker, BP monitor, and a 3-channel Holter monitor connected to the integrative platform using a cloud computing environment. AI in TIMELY is primarily used to optimize prediction of CAD risk and behavior. Data are provided in a dashboard to case managers supporting patients. Primary outcomes were changed 1) mortality risk (via composite biomarker score), 2) 6-minute walking test (6MWT) distance. Secondary outcomes included physical fitness, physical activity, dietary habits, body weight, smoking cessation, medication adherence, and psychological stress. Results Patients’ mean age was 60.5±9.2 years, 34% had experienced MI, ~21% were women. At 6 months, 161 (89%) and 156 (87%) patients were assessed for the primary outcome in the IG and CG, respectively. Severe adverse events (n=28) were unrelated to the intervention. In the intention-to-treat analysis, no significant difference between the groups was seen for the tested outcomes. The per-protocol (PP) analysis revealed that the IG sustained a significant ~7% better mean performance in the 6MWT over 12 months (IG=567.5±90.3 m, CG=529.6±111.7 m; p0.05). The IG also showed larger improvements in physical fitness (PP, time×group p=0.027), with higher values at 6 months (IG=144.5±39.6 W, CG=130.9±44.0 W) and 12 months (IG=141.9±45.2 W, CG=130.3±45.2 W). A significant group difference was also seen for medication adherence (MARS-5 Score) which remained higher in the IG at 12 months (PP, time×group p=0.0071). No other significant differences in secondary outcomes were detected. Conclusions The AI-driven eHealth intervention TIMELY offers personalized support for CAD patients during phase III CR and shows potential specifically in terms of improving physical exercise capacity.TIMELYFor image description, please refer to the figure legend and surrounding text.