Predicting Patient Readmission Risk from Medical Text via Knowledge Graph Enhanced Multiview Graph Convolution

Readmissions also put families at higher financial burden and increase healthcare providers' costs. Therefore, it is beneficial for both Unplanned intensive care unit (ICU) readmission rate is an important patients and hospitals to identify patients that are inappropriately metric for evaluating the quality of hospital care. Efficient or prematurely discharged from ICU. and accurate prediction of ICU readmission risk can not only help Over the past few years, there has been a surge of interest in prevent patients from inappropriate discharge and potential dangers, applying machine learning techniques to clinical forecasting tasks, but also reduce associated costs of healthcare. In this paper, such as readmission prediction [12], mortality prediction [6], length we propose a new method that uses medical text of Electronic of stay prediction [14], etc. Earlier studies generally select statistically Health Records (EHRs) for prediction, which provides an alternative significant features from patients' Electronic Health Records perspective to previous studies that heavily depend on numerical (EHRs), and feed them into traditional machine learning models and time-series features of patients. More specifically, we extract like logistic regression [19]. Deep learning models have also been discharge summaries of patients from their EHRs, and represent gaining more and more attention in recent years, and have shown them with multiview graphs enhanced by an external knowledge superior performance in medical prediction tasks.