acemachine learning algorithms predicting cardiovascular disease risk compared with Framingham Risk Score in healthy adults
Source article: Future promise, current clinical ambiguity: a systematic review of machine learning algorithm outputs predicting risk of cardiovascular disease
Abstract: Objective To examine whether the outputs of machine learning algorithms designed to predict risk of cardiovascular disease (CVD) address known deficiencies of the Framingham Risk Score (FRS) and improve risk estimates. Methods For this critical review, Medline, Embase and IEEE were searched from inception to 1 January 2025. Included were studies describing machine learning algorithms designed to specifically compare output of cardiovascular risk assessment with the FRS. Commentaries, letters, unpublished work or…
Contested: both sides are scored from claims and sources, not community votes.
Hospital Universitari Doctor Peset, València 10 by 19Tarrestnom65. CC BY-SA 4.0 · https://creativecommons.org/licenses/by-sa/4.0
A systematic review published 12 September 2026 searched three databases to 1 January 2025 and included 29 studies that directly compared machine learning CVD risk predictions with the Framingham Risk Score in healthy adults. Twenty-three studies reported improved predictive ability, often by adding sociodemographic predictors absent from FRS or costly diagnostics such as CT angiography.
The findings matter because improved statistical prediction does not automatically translate to patient benefit. The review highlights that overestimation of at-risk proportions raises overdiagnosis concerns, inconsistent outcome definitions limit comparability, and cost-benefit and screening-principle considerations remain unresolved for clinical adoption.
- Systematic review searched Medline, Embase and IEEE to 1 January 2025 and included 29 studies comparing ML outputs to FRS.
- 23 of 29 studies reported improved predictive ability relative to FRS.
- Most datasets included sociodemographic predictors not in FRS inputs; some added costly tests like CT angiography.
- Review noted statistical significance was generally taken as proxy for clinical significance and definitions of hypertension or myocardial infarction did not always adhere to FRS values.
Systematic review found most ML algorithms improved CVD risk prediction compared with Framingham Risk Score by incorporating sociodemographic predictors and modeling non-linear interactions.
Some algorithms overestimated the number at risk versus FRS without addressing overdiagnosis risk, while treating statistical significance as clinical significance.
The rundown
The review followed PRISMA guidelines with two independent reviewers screening 707 retrieved studies, ultimately including 29 that directly compared ML outputs to FRS for healthy adults.
Thematic analysis examined strengths, added value, potential harms, unintended consequences and equity implications, finding that while technological merging of sociodemographic and biologic variables is feasible, clinical translation requires better adherence to screening principles and evaluation of costly inputs like CT angiography.
Review notes remaining gaps for patient benefit including need for clinical insight, adherence to screening principles, and cost-benefit assessment, plus inconsistent definitions of inputs and outcomes.
Sources
- Peer-reviewedOpen Heart2026-09-12
How should this claim be treated?
The debate