Early prediction of sepsis in the ICU: a comparative analysis of multiple machine-learning algorithms using the MIMIC-III database

Xiuli Shao, Ruiling Li, Yaqi Lan, Xinyue Gao, De-Zong Liu, Rui Li, Jianpei Niu · 2026

Read Early prediction of sepsis in the ICU: a comparative analysis of multiple machine-learning algorithms using the MIMIC-III database by Xiuli Shao, Ruiling Li, Yaqi Lan, Xinyue Gao, De-Zong Liu, Rui Li, Jianpei Niu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13386605.txt.

Description

Shao and colleagues race nine machine learning algorithms to predict which intensive care patients will develop sepsis after the first day, in 2026 informatics research. Sepsis strikes up to 30 percent of ICU patients and rewards early treatment. Using the MIMIC-III database, the team excluded patients septic within 24 hours, leaving 1,634 patients, of whom 349, or 21.4 percent, developed sepsis. Features were selected automatically, class imbalance was corrected in training, and models were judged on discrimination, calibration, and clinical usefulness.

The gradient boosting model XGBoost-DART won the comparison, reaching an accuracy score of 0.881 on the discrimination scale. It posted the best accuracy, at 0.847, and specificity of 0.897 among the nine algorithms. Rival models' scores ranged from 0.794 to 0.881 across the evaluation. Decision-curve analysis showed the model delivered the greatest net benefit over the broadest span of risk thresholds. Shao concludes models can flag deteriorating patients a day ahead reliably enough to guide monitoring intensity.

Catalog record

CREATORXiuli Shao, Ruiling Li, Yaqi Lan, Xinyue Gao, De-Zong Liu, Rui Li, Jianpei Niu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE46 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC13386605

Browse the full library · About the Tower of Babel