Related Experiment Video
Updated: Jan 14, 2026
07:42
A Data-Driven Approach to Quantifying Immune States in Sepsis
Published on: February 7, 2025
476
Streamlined machine learning model for early sepsis risk prediction in burn patients
Marius Drysch1, Felix Reinkemeier2, Flemming Puscz2
1Department of Plastic Surgery, BG University Hospital Bergmannsheil, Ruhr University Bochum, Bochum, Germany. marius.drysch@bergmannsheil.de.
NPJ Digital Medicine
|October 21, 2025
Summary
Early sepsis detection in burn patients is crucial. A new machine learning model uses six admission factors for accurate sepsis risk prediction, improving critical care outcomes.
Area of Science:
- Critical Care Medicine
- Computational Biology
- Burn Surgery
Background:
- Sepsis is a primary cause of death in burn patients.
- Early sepsis identification is challenging due to inflammation and physiological changes.
Purpose of the Study:
- To develop a machine learning model for early sepsis risk prediction in burn patients.
- To utilize readily available admission data for accessible risk stratification.
Main Methods:
- A Random Forest machine learning model was developed.
- Trained on data from 6,629 burn patients across 11 centers (German Burn Registry).
- Utilized six admission variables: age, burn surface area, burn depth, inhalation injury, and hypertension.
Main Results:
- The model achieved an Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.91.
- Demonstrated high sensitivity (0.81), specificity (0.85), and negative predictive value (0.98).
- Enabled reliable early risk stratification upon Intensive Care Unit (ICU) admission.
Conclusions:
- A streamlined machine learning model accurately predicts sepsis risk in burn patients using only admission data.
- This interpretable model facilitates timely interventions and can improve patient outcomes.
- Supports early sepsis detection in critical care settings for burn survivors.
