Machine learning prediction of early hypothermia in sepsis patients.

INTRODUCTION

Sepsis is a systemic inflammatory response syndrome caused by infection and remains a leading cause of mortality worldwide. Abnormal body temperature, especially hypothermia (body temperature < 36°C), is a key clinical feature in sepsis patients and is closely associated with disease severity, impaired immune function, and poor prognosis. Early prediction of hypothermia is crucial for timely intervention and improving prognosis.

MATERIAL AND

METHODS

This study used machine learning algorithms to train and validate a prediction model for early temperature changes in critically ill sepsis patients. Data were extracted from the MIMIC-IV database and five models were established: XGBoost, LR, SVM, KNN, and ANN.

RESULTS

The XGBoost model demonstrated the best predictive performance with AUC values of 0.92 in the training cohort and 0.98 in the validation cohort.

CONCLUSIONS

This model can assist clinicians in identifying sepsis patients at high risk for early hypothermia and implementing early intervention to reduce mortality.