Predicting Australian Adults at High Risk of Cardiovascular Disease Mortality Using Standard Risk Factors and Machine Learning.

摘要: Effective cardiovascular disease (CVD) prevention relies on timely identification and intervention for individuals at risk. Conventional formula-based techniques have been demonstrated to over- or under-predict the risk of CVD in Australian population. This study assessed ability machine learning models predict mortality population compare performance with well-established Framingham model. Data is drawn from three cohort studies: North West Adelaide Health Study (NWAHS), Diabetes, Obesity, Lifestyle study, Melbourne Collaborative Cohort (MCCS). Four predicting 15-year were developed compared 2008 Machine performed significantly better model when applied cohorts. based improved prediction by 2.7% 5.2% across In an aggregated cohort, up 5.1% (area-under-curve (AUC) 0.852, 95% CI 0.837–0.867). Net reclassification improvement (NRI) was 26% models. also showed stratified sex diabetes status. Results suggest a potential improving using