Systematic review of artificial intelligence for personalized oral anticoagulation management in cardiovascular care
Artificial intelligence and machine learning are increasingly used to personalize oral anticoagulation. We systematically reviewed patient-level studies in adults actively receiving warfarin, vitamin K antagonists, or direct oral anticoagulants. PubMed, Europe PMC, and Scopus were searched from inception to 15 August 2026. Of 984 records, 54 studies met eligibility criteria after duplicate removal, title and abstract screening, and full-text assessment. Applications included dose optimization and anticoagulation control, prediction of bleeding and thromboembolic outcomes, and individualized treatment selection. Warfarin dominated dose-optimization research, whereas direct oral anticoagulants were more prominent in risk and comparative-treatment modelling. Across the evidence base, limited external validation, inconsistent calibration, and sparse prospective evaluation restricted clinical translation. AI-enabled anticoagulation is progressing from static prediction toward longitudinal and treatment-specific decision support, but robust multicentre validation remains essential.
Authors
- Rosnani Hashim (ORCID: https://orcid.org/0000-0003-0890-4031)
- Nurul Adilla Hayat Jamaluddin (ORCID: https://orcid.org/0000-0003-4099-9472)
- Farizan binti Abdul Ghaffar
- Ahmed Alaqab
- Maali Mujahid Mustafa¹
Institutions
- Hospital Kuala Lumpur (MY)
- University of Cyberjaya (MY)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-18
- DOI
- https://doi.org/10.21203/rs.3.rs-11296432/v1
- Primary Topic
- Atrial Fibrillation Management and Outcomes
- Type
- preprint