Heart risk equation outperformed machine learning in adults
The equation scored 0.822 on a measure of how well it distinguished cardiovascular deaths from other outcomes, outperforming machine-learning models in 5,690 adults.
Our read
Biomarkers- Studied:
- Study in people
- Why it’s aging research:
- It bears on staying healthy in midlife and later life.
Editor of record: Boris Djordjevic · AgingResearch.News
In 5,690 adults aged 30 to 79 without prior cardiovascular disease, an observational study compared methods for predicting cardiovascular death over ten years. Participants came from the National Health and Nutrition Examination Survey. During a median follow-up of 11.8 years, 178 participants died from cardiovascular causes and 579 from other causes.
The American Heart Association’s official cardiovascular risk equation distinguished people who died from cardiovascular causes from other participants better than the tested machine-learning models. Those models used either the same health information or an expanded set of inputs. The official equation also performed better at sorting participants into risk categories. The analysis accounted for deaths from other causes, which could occur before a cardiovascular death.
Why it matters
Cardiovascular risk prediction informs efforts to prevent disease as people age. The authors supported continued use of the official equation in this population.
Caveats
The comparison assessed cardiovascular deaths, not nonfatal cardiovascular disease, and included only adults aged 30 to 79 without prior cardiovascular disease who had complete risk-equation inputs. The findings were reported in a conference abstract.
The paper
Cooper Univ Hosp
AHA Scientific Sessions 2026 Abstract, Nov 2, 2026
Presented at AHA Scientific Sessions 2026, abstract 10446
Comments
Be the first to comment.