AI in cardiology significantly improved clinical workflow, patient engagement, and cardiovascular outcomes in a major new systematic review of randomised controlled trials.
AI is increasingly integrated into cardiovascular medicine, from diagnostic imaging to decision-support systems. However, whether these tools deliver measurable benefits in real-world patient care has remained uncertain. The new analysis provides some of the strongest evidence to date that data-driven AI can improve efficiency and clinical outcomes when embedded into routine cardiovascular practice.
Artificial Intelligence Improved Workflow Efficiency
Researchers reviewed 32 randomised controlled trials evaluating machine-learning and deep-learning interventions across cardiovascular care. Rule-based systems were excluded to focus specifically on data-driven AI applications.
The study found that AI consistently improved workflow efficiency, classified as Tier A evidence under the National Institute for Health and Care Excellence (NICE) framework.
Diagnostic workflow times were significantly reduced (standardised mean difference: −0.71; 95% confidence interval [CI]: −1.04 to −0.39), translating to diagnostic processes that were approximately 30–120 seconds faster and reductions in hospital length of stay ranging from 1.0–4.2 days.
AI Enhanced Patient Engagement and Medication Adherence
AI also demonstrated benefits for patient engagement and health promotion. Behavioural nudging systems designed to support medication-taking behaviour significantly improved adherence rates (risk ratio: 1.59; 95% CI: 1.01–2.50).
These findings are particularly relevant given that poor medication adherence remains a major contributor to recurrent cardiovascular events and preventable hospital admissions worldwide.
Clinical Decision Support Reduced Mortality
The most clinically significant findings were observed in AI-powered clinical decision-support systems. Tier C outcomes showed that these implementations reduced all-cause mortality by 16% (risk ratio: 0.84; 95% CI: 0.75–0.94), with low statistical heterogeneity across studies.
Researchers noted several limitations, including restricted blinding and the lack of sham-AI controls in many trials. Nonetheless, the findings suggest that AI in cardiology may offer measurable clinical benefits beyond operational efficiency alone.
The authors concluded that integrating actionable AI decision support into cardiovascular care pathways could help shape future governance frameworks, implementation strategies, and digital health policy development.
Reference
Lin YE et al. Impact of artificial intelligence on cardiovascular workflow, engagement, and outcomes: a systematic review. 2026. npj Digit Med. 2026;DOI:10.1038/s41746-026-02690-7.
Featured image: elenabsl on Adobe Stock
