AI heart disease prediction tools show promise but are not ready for clinical use: review
The findings are particularly relevant for India, where cardiovascular disease accounts for nearly one-third of all deaths and often affects people at younger ages than in many other countries
Artificial intelligence (AI) holds potential for earlier identification of individuals at risk of cardiovascular disease (CVD) in India, allowing for more personalized prevention strategies. However, AI tools are not yet ready for routine clinical decision-making, according to a systematic review by researchers from the Indian Institute of Science, M.S.
Ramaiah University of Applied Sciences, and the London School of Hygiene and Tropical Medicine. The review, published in BMC Medical Informatics and Decision Making, analyzed 30 studies on AI-based CVD prediction models published since 2017. Most models were based on datasets from the U.S., U.K., and South Korea and used machine learning algorithms such as Random Forests, Support Vector Machines, and neural networks.
While AI models generally performed as well as or slightly better than conventional risk calculators like the Framingham Risk Score, the evidence is still not adequate for widespread clinical use. Rigorous independent validation, calibration, and assessment of clinical usefulness are necessary before AI tools can guide long-term treatment decisions.
A major concern is the lack of evidence showing whether the predicted risk accurately reflects what happens in real-world populations. Most studies assessed discrimination, but none examined calibration or decision-curve analyses. India's unique risk factors and disease patterns should be taken into account, as models developed in other countries may not perform similarly.
The researchers recommend prospective validation of AI-based cardiovascular risk models in diverse populations, including India, before their integration into routine healthcare.
Written by urgent.news from The Hindu Health's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
