Predicting Future Cardiovascular Events in Patients With Peripheral Artery Disease Using Electronic Health Record Data

Abstract

Background: Patients with peripheral artery disease (PAD) are at risk of major adverse cardiac and cerebrovascular events (MACCE). There are no readily available risk scores that can accurately identify which patients are most likely to sustain an event, making it difficult to identify those who might benefit from more aggressive intervention. Thus, we aimed to develop a novel predictive model -using machine learning methods on electronic health record (EHR) data -to identify which PAD patients are most likely to develop MACCE.Methods and Results: Data were derived from patients diagnosed with PAD at two tertiary care institutions. Predictive models were built using a common data model (CDM) that allowed for utilization of both structured (coded) and unstructured (text) data. Only data from time of entry into the health system up to PAD diagnosis were used for modeling. Models were developed and tested using nested cross-validation. A total of 7,686 patients were included in learning our predictive models. Utilizing almost 1,000 variables, our best predictive model accurately determined which PAD patients would go on to develop MACCE with an area under the curve (AUC) of 0.81 (95% Confidence Interval, 0.80-0.83). Conclusions:Machine learning algorithms applied to data in the EHR can learn models that accurately identify PAD patients at risk of future MACCE, highlighting the great potential of EHR to provide automated risk stratification for cardiovascular diseases. Common data models that can enable cross-institution research and technology development could potentially be an important aspect of widespread adoption of newer risk-stratification models.

The Role of Circulating Biomarkers in Peripheral Arterial Disease

“…Regarding CV risk stratification, Ross EG et al reported that state of the art ML algorithms outperformed stepwise logistic regression models for the identification of PAD and the prognostication of mortality risk in this population [ 144 ]. More recently, they have demonstrated that the application of ML to electronic health records can generate learning-based models that accurately identify PAD patients at risk of future major adverse cardiac and cerebrovascular events [ 145 ]. In addition, ML algorithms might be useful to predict not only PAD medical burden, but also its associated financial cost.…”

Classification of SARS-CoV-2 and non-SARS-CoV-2 using machine learning algorithms

“…The second set contained 30% (475 samples) from both classes SARS-CoV-2 (189 samples) and non-SARS-CoV- 2 (286 samples) and was used as a testing dataset with the trained model. The classifiers were validated via ten-fold CV, and the best model was selected based on the F-measure [ 43 ], which balances the recall and precision of the model. In addition, the efficacy of the model was evaluated on the accuracy matrix via the corrected 10x10 fold CV paired t-test [ 44 ].…”

Personalized prediction of live birth prior to the first in vitro fertilization treatment: a machine learning method

“…In contrast, unsupervised machine learning leaves the algorithms to discover on their own the underlying structure within unlabeled data. The machine learning based algorithms with strong data processing ability have become a promising methodology for clinical decision making and medicine study, including clinical prediction, radiology, surgery, drug discovery and pharmacokinetic prediction [13–15]. In the field of reproduction science, machine learning has been applied in areas including embryo scoring and prediction of implantation rate after blastocyte transfer [16, 17].…”