Artificial Intelligence-Enhanced Telerehabilitation in Post-Acute Coronary Syndrome: A Narrative Review of Opportunities, Evidence, and Future Directions

TL;DR

AI-enhanced telerehabilitation offers personalized, adaptive care for patients after heart attacks, but more research is needed to confirm its effectiveness.

Contribution

This paper reviews how AI can improve cardiac telerehabilitation by enabling personalized and adaptive care models.

Findings

AI-enhanced telerehabilitation shows potential in adaptive risk prediction and personalized exercise modulation.

Systems using AI can adjust exercise protocols in real-time and predict rehospitalization risks.

Most AI-based interventions are evaluated in small-scale studies, not large randomized trials.

Abstract

Cardiac telerehabilitation has become a promising alternative to traditional programmes for preventing acute coronary syndrome (ACS) in the secondary phase. However, current implementations are still reactive and standardised, lacking personalisation and flexibility in clinical settings. By integrating artificial intelligence (AI), it may be possible to overcome these limitations and provide intelligent, scalable, and patient-centred care. Methods: We conducted a structured literature review across PubMed, Scopus, the Cochrane Library, and Web of Science, targeting English-language studies published from January 2015 to May 2025. Inclusion criteria included adult populations with a history of ACS or high cardiovascular risk, assessing interventions based on AI, telerehabilitation, or their combination. Studies are needed to report clinical, functional, behavioural, or technological outcomes. A thematic narrative synthesis was utilised. Results: AI-enhanced telerehabilitation demonstrates potential advantages over conventional digital care in selected domains, including adaptive risk prediction, personalised exercise modulation, and adherence support. Several systems report real-time adjustment of exercise protocols, early dropout detection, and predictive analytics for rehospitalisation. AI integration may also contribute to personalised behavioural feedback and psychosocial monitoring. Nevertheless, the overall level of evidence remains preliminary and heterogeneous, with most AI-based interventions evaluated in pilot, feasibility, or modelling studies rather than large-scale randomized trials. Conclusions: The integration of AI into telerehabilitation represents a promising evolution in post-ACS care, shifting from predominantly reactive monitoring toward more adaptive and data-driven support models. While early-phase studies suggest feasibility and potential clinical benefit, robust multicentre randomized controlled trials and cost-effectiveness analyses are required before definitive conclusions regarding superiority or widespread implementation can be drawn.

Figures