Related Experiment Video
Updated: Jul 21, 2025
07:42
A Data-Driven Approach to Quantifying Immune States in Sepsis
Published on: February 7, 2025
249
Machine Learning for Early Prediction of Sepsis in Intensive Care Unit (ICU) Patients
Abdullah Alanazi1,2, Lujain Aldakhil1,2, Mohammed Aldhoayan1,2
1Department of Health Informatics, College of Public Health and Health Informatics, King Saud Ibn Abdulaziz University for Health Sciences, Riyadh 11481, Saudi Arabia.
Medicina (Kaunas, Lithuania)
|July 29, 2023
Summary
Early sepsis detection is vital. This study used machine learning on ICU patient data, finding time, lactic acid, and temperature are key predictors for survival analysis models.
Area of Science:
- Medical Informatics
- Machine Learning in Healthcare
- Critical Care Medicine
Background:
- Early sepsis detection is critical for patient survival but remains challenging.
- Machine learning offers potential for improved sepsis prediction accuracy.
- Analyzing clinical data and vital signs can aid in early sepsis identification.
Purpose of the Study:
- To investigate a novel machine learning approach for early sepsis detection.
- To predict and identify initial sepsis signs in intensive care unit (ICU) patients.
- To analyze survival rates and predict outcomes using various predictive models.
Main Methods:
- Utilized data mining algorithms and proportional hazards models.
- Analyzed data from the BESTCare database (KAMC) for adult ICU patients (≥14 years).
- Included 1182 sepsis-diagnosed patients admitted between April and October 2018.
Main Results:
- A regression model with survival analysis demonstrated moderate predictive ability.
- Time to outcome, lactic acid, and temperature were significant predictors (p < 0.05).
- Other data mining algorithms showed limitations due to independence assumptions.
Conclusions:
- Continuous improvement in sepsis prediction models is essential.
- Meticulous data cleaning and attribute selection are foundational for future advancements.
- Machine learning holds promise for enhancing accuracy in sepsis prediction.
