Outcome Prediction of Patients for Different Stages of Sepsis Using Machine Learning Models

Abstract

Sepsis is a major challenge in the field of medical science. It affects over a million patients annually and also increases the mortality rate. Generally, sepsis condition is not identified easily. Thus, an intensive analysis of patients is required for identifying sepsis in the Intensive Care Unit (ICU). In this research work, an outcome prediction based machine learning models for identifying different stages of sepsis is proposed. Machine Learning (ML) models can help to predict the current stage of sepsis using existing clinical measurements like clinical laboratory test values and crucial signs in which patients are at high risk. We explore four ML models namely XGBoost, Random Forest, Logistic Regression, and Support Vector Machine by utilizing clinical laboratory values and vital signs. The performance evaluation of the proposed and existing techniques is performed by considering the same dataset. These models achieve an AUC (Area under the Curve) 0.95, 0.91, 0.76, and 0.93, respectively, for recognition of sepsis. Experimental results demonstrate that the XGBoost model with 10-fold cross-validation performs well than other models across all the performance metrics.