Artificial intelligence-assisted personalized cardiac rehabilitation: integrating wearable technology with physiotherapist-guided exercise prescription

Cardiac rehabilitation is a central component of secondary cardiovascular prevention, supporting improvements in exercise capacity, cardiovascular risk-factor control, functional independence, and quality of life. However, conventional centre-based programmes may not provide continuous insight into patients’ physiological responses and daily activity outside supervised sessions. Limited access, travel requirements, workforce constraints, and inconsistent adherence further restrict participation. Artificial intelligence (AI), wearable sensors, and digital health platforms offer opportunities to extend rehabilitation beyond the clinic through continuous monitoring, predictive analysis, individualized feedback, and adaptive exercise prescription. Wearable devices can capture heart rate, rhythm, physical activity, sleep, oxygen saturation, and other physiological or behavioural indicators, while AI models may identify trends, estimate exercise tolerance, detect abnormal responses, and support clinical decision-making. When integrated with physiotherapist-guided care, these technologies could help tailor exercise intensity, duration, frequency, and progression to an individual's changing clinical condition. Nevertheless, AIassisted cardiac rehabilitation should not be understood as an autonomous replacement for professional assessment. Device measurements may be inaccurate, algorithms may produce biased or poorly calibrated predictions, and automated exercise recommendations may be unsafe when cardiovascular symptoms, medication changes, or comorbidities are not adequately considered. This review discusses the physiological and clinical foundations of cardiac rehabilitation, wearable technologies, AI methods, data integration, physiotherapist-led exercise prescription, remote and hybrid delivery, safety, ethics, implementation, and future research. It proposes a clinician-supervised model in which wearable data are interpreted alongside symptoms, functional testing, cardiovascular risk, patient-reported outcomes, and professional judgement. Recent guidance supports the use of digital sensors and smartphone applications as adjuncts to cardiac rehabilitation and recognizes a potential role for AI in early event detection and clinical decision support, while emphasizing the need for validation and responsible implementation. The central principle is that technology should strengthen the therapeutic relationship between the patient and rehabilitation team, enabling more responsive, accessible, and individualized cardiovascular care.