Integrating Clinical Physiological Knowledge at the Feature and Classifier Levels in Design of a Clinical Decision Support System for Improved Prediction of Intensive Care Unit Outcome
This chapter is concerned with the application of machine learning in prediction of patient mortality in the intensive care unit (ICU) using physiological measurements collected during the first 48 hours of ICU admission. Even though the patient acuity scoring systems have been widely used in ICUs for mortality prediction, they are still far from being accurate and are unable to detect patients with high risk of sudden critical deterioration. We discuss steps used in developing a clinical decision support system (CDSS) for ICU morbidity prediction that is based on a machine learning approach. First, we categorize the collected data based on the physiological organ that they represent. We then extract clinically relevant features from each data category and then rank these features based on their mutual information with the outcome. Then, we design an artificial neural network to serve as a classifier to detect patients at high risk of critical deterioration. The results from the classification show that the newly designed CDSS outperforms the widely used ICU acuity scoring systems. The F-score classification result of our developed algorithms is 42%, while the F-score results for Sequential Organ Failure (SOFA) and Simplified Acute Physiology Score (SAPS) II are 26% and 29% respectively.