Abstract
Current approaches to predicting a cardiovascular disease (CVD) event rely on conventional risk factors and cross-sectional data. In this study, we applied machine learning and deep learning models to 10-year CVD event prediction by using longitudinal electronic health record (EHR) and genetic data. Our study cohort included 109, 490 individuals. In the first experiment, we extracted aggregated and longitudinal features from EHR. We applied logistic regression, random forests, gradient boosting trees, convolutional neural networks (CNN) and recurrent neural networks with long short-term memory (LSTM) units. In the second experiment, we applied a late-fusion approach to incorporate genetic features. We compared the performance with approaches currently utilized in routine clinical practice – American College of Cardiology and the American Heart Association (ACC/AHA) Pooled Cohort Risk Equation. Our results indicated that incorporating longitudinal feature lead to better event prediction. Combining genetic features through a late-fusion approach can further improve CVD prediction, underscoring the importance of integrating relevant genetic data whenever available.
Highlights
Electronic health records (EHRs) contain a wealth of detailed clinical information and provide several distinct advantages for clinical research, including cost efficiency, big data scalability, and the ability to analyze data over time
Machine learning and deep learning approaches are suited to exploiting such big data for individual outcome prediction, especially when EHRs can be linked to genetic data[9,10]
Consistent with previous reports[11,15], we initially observed a benchmark predictability for the ACC/AHA equation of AUROC 0.732
Summary
Electronic health records (EHRs) contain a wealth of detailed clinical information and provide several distinct advantages for clinical research, including cost efficiency, big data scalability, and the ability to analyze data over time. The recent convergence of two rapidly developing technologies— high-throughput genotyping and deep phenotyping within EHRs – presents an unprecedented opportunity to utilize routine healthcare data and genetic information to accelerate the healthcare. Machine learning and deep learning approaches are suited to exploiting such big data for individual outcome prediction, especially when EHRs can be linked to genetic data[9,10]. A recent study from the United Kingdom (UK) applied machine learning to conventional CVD risk factors on a large UK population and improved the prediction accuracy by 4.9%11. We examined: i) the performance of machine learning and deep learning on longitudinal EHR data for the prediction of 10-year CVD event, compared to a gold standard achieved by (ACC/AHA) Pooled Cohort Risk Equations, and ii) the benefits of incorporating extra genetic information
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have