Using methods from the data-mining and machine-learning literature for disease classification and prediction: a case study examining classification of heart failure subtypes

• Published in: Journal of clinical epidemiology 2013-04, Vol.66 (4), p.398-407
• Main Authors: Austin, Peter C, Tu, Jack V, Ho, Jennifer E, Levy, Daniel, Lee, Douglas S
• Format: Article
• Language: English
• Subjects: Abbreviations; Accuracy; Algorithms; Artificial Intelligence; Bagging; Biological and medical sciences; Boosting; Cardiology. Vascular system; Classification; Classification trees; Data Mining; Epidemiology; Feedback; Forests; General aspects; Heart; Heart attacks; Heart failure; Heart Failure - classification; Heart failure, cardiogenic pulmonary edema, cardiac enlargement; Humans; Internal Medicine; Logistics; Medical sciences; Methods; Prediction; Predictive Value of Tests; Random forests; Regression Analysis; Regression methods; Regression trees; Support vector machines
• ISSN: 0895-4356; 1878-5921
• DOI: 10.1016/j.jclinepi.2012.11.008

Abstract Objective Physicians classify patients into those with or without a specific disease. Furthermore, there is often interest in classifying patients according to disease etiology or subtype. Classification trees are frequently used to classify patients according to the presence or absence of a disease. However, classification trees can suffer from limited accuracy. In the data-mining and machine-learning literature, alternate classification schemes have been developed. These include bootstrap aggregation (bagging), boosting, random forests, and support vector machines. Study Design and Setting We compared the performance of these classification methods with that of conventional classification trees to classify patients with heart failure (HF) according to the following subtypes: HF with preserved ejection fraction (HFPEF) and HF with reduced ejection fraction. We also compared the ability of these methods to predict the probability of the presence of HFPEF with that of conventional logistic regression. Results We found that modern, flexible tree-based methods from the data-mining literature offer substantial improvement in prediction and classification of HF subtype compared with conventional classification and regression trees. However, conventional logistic regression had superior performance for predicting the probability of the presence of HFPEF compared with the methods proposed in the data-mining literature. Conclusion The use of tree-based methods offers superior performance over conventional classification and regression trees for predicting and classifying HF subtypes in a population-based sample of patients from Ontario, Canada. However, these methods do not offer substantial improvements over logistic regression for predicting the presence of HFPEF.