A new artificial intelligence algorithm improves cardiovascular risk prediction

Ernest Vinyoles, a researcher at IDIAPJGol, participates in the development of a model that, according to study results, outperforms traditional tools

A study involving IDIAPJGol researcher Ernest Vinyoles has developed a model for predicting cardiovascular mortality risk using artificial intelligence to interpret blood pressure readings taken during consultations and data detected by 24-hour ambulatory blood pressure monitoring equipment. According to the article published in the journal Hypertension, this model surpasses traditional tools like SCORE-2 and Framingham in accuracy. Its application in primary care could help identify high-risk patients and better tailor medical treatments.

The study demonstrates that ambulatory monitoring data adds additional information to blood pressure readings, improving predictions.

The study involved 59,124 patients included in the Spanish Ambulatory Blood Pressure Monitoring Registry. Using machine learning techniques such as logistic regression, decision trees, and deep neural networks, the research team analyzed over 60 clinical variables. Cross-validation and key feature selection ensured the model's robustness and applicability.

 

One of the main risk factors
Hypertension is a chronic disease that can damage arteries and increase the risk of circulatory system diseases, primarily stroke and acute myocardial infarction. It is the most significant cardiovascular risk factor and one of the leading risk factors for premature mortality globally. In Catalonia, it affects one in four people over the age of 15.

Incorporating ambulatory blood pressure monitoring data into prediction tools could improve care for hypertensive patients, especially in settings where blood pressure is difficult to control.

Reference to the article
Guimarães P, Keller A, Böhm M, Lauder L, Fehlmann T, Ruilope LM, Vinyoles E, Gorostidi M, Segura J, Ruiz-Hurtado G, Staplin N, Williams B, de la Sierra A, Mahfoud F. Artificial Intelligence-Derived Risk Prediction: A Novel Risk Calculator Using Office and Ambulatory Blood Pressure. Hypertension. 2025 Jan;82(1):46-56. doi: 10.1161/HYPERTENSIONAHA.123.22529. Epub 2024 Apr 25. PMID: 38660828.