A Decision Tree Model for Predicting Heart Failure Patient Readmissions

摘要

In this research endeavor, we investigate the use and implementation of a predictive model aimed at identifying and stratifying patients at high-risk for readmission for Heart Failure. Though commonly used, statistical techniques have been aimed at identifying those factors which lead to a higher relative risk. There are large limitations on the usefulness of the determined relationships due to the underlying nature of humans to vary widely. While these techniques provide for simple relations, there is a drawback in accuracy. This can lead to either Type I or Type II error for the case management and care coordination processes, low effectiveness, or low efficiency, respectively. Through the use of psychological, sociological, demographic and care related information, we utilized data mining techniques to develop a decision tree based predictive model which is not only accurate, but can readily be used throughout the case management and care coordination process. By identifying each patient's risk, we can then distribute the care that is required to the patients who need it most.