Application of AI Risk Prediction Tools for Cardiovascular Care in Hong Kong: A Systematic Review with Insights from Clinicians

Background: Cardiovascular diseases (CVD) present a substantial burden on global and local healthcare systems, particularly within Hong Kong's overstretched public secondary care settings coupled with an aging population. The Personalized CARdiovascular DIsease risk Assessment for Chinese (P-CARDIAC), a locally developed and validated artificial intelligence (AI) powered cardiovascular risk assessment tool, offers a transformative opportunity for enhancing care through systematic risk prediction and a multidisciplinary care approach. Implementation strategies for its application in secondary care remain undefined. Methods: We gathered local insights from one expert meeting and thirteen semi-structured interviews with cardiologists and pharmacists, matched these with barriers and respective facilitators identified in a PRISMA-based systematic review from PubMed, EMBASE, and Scopus databases in September and October 2025. Findings were synthesised into an evidence-to-action framework. Results: We identified 1696 records from the systematic review, and included 16 primary and 6 non-primary studies. We synthesized these findings and outlined strategies in three levels. At the individual level, strategies focus mainly on building trust and promoting holistic, personalized care via effective cardiologist-pharmacist communication and targeted training led by leading specialists. The technical level requires leveraging partnerships for continuous data validation and program evaluation to demonstrate the impact of P-CARDIAC on risk stratification and service prioritization. Successful application at the organizational level hinges on seamless workflow integration, active stakeholder engagement, and institutional support through the development of evidence-based consensus and guidance. Conclusion: Promulgating these multilevel strategies is core to realizing P-CARDIAC's full potential, enabling it to overcome utilization challenges and optimize healthcare delivery through territory-wide application.