Interpretable Machine Learning-based Prediction of 28-day Mortality in ICU Patients with Sepsis: a Multicenter Retrospective Study

Overview

Journal Front Cell Infect Microbiol

Date 2025 Jan 23

PMID 39844844

Authors

Li Shen

Jiaqiang Wu

Jianger Lan

Chao Chen

Yi Wang

Zhiping Li

Affiliations

Soon will be listed here.

Abstract

Background: Sepsis is a major cause of mortality in intensive care units (ICUs) and continues to pose a significant global health challenge, with sepsis-related deaths contributing substantially to the overall burden on healthcare systems worldwide. The primary objective was to construct and evaluate a machine learning (ML) model for forecasting 28-day all-cause mortality among ICU sepsis patients.

Methods: Data for the study was sourced from the eICU Collaborative Research Database (eICU-CRD) (version 2.0). The main outcome was 28-day all-cause mortality. Predictor selection for the final model was conducted using the least absolute shrinkage and selection operator (LASSO) regression analysis and the Boruta feature selection algorithm. Five machine learning algorithms including logistic regression (LR), decision tree (DT), extreme gradient boosting (XGBoost), support vector machine (SVM), and light gradient boosting machine (lightGBM) were employed to construct models using 10-fold cross-validation. Model performance was evaluated using AUC, accuracy, sensitivity, specificity, recall, and F1 score. Additionally, we performed an interpretability analysis on the model that showed the most stable performance.

Results: The final study cohort comprised 4564 patients, among whom 568 (12.4%) died within 28 days of ICU admission. The XGBoost algorithm demonstrated the most reliable performance, achieving an AUC of 0.821, balancing sensitivity (0.703) and specificity (0.798). The top three risk predictors of mortality included APACHE score, serum lactate levels, and AST.

Conclusion: ML models reliably predicted 28-day mortality in critically ill sepsis patients. Of the models evaluated, the XGBoost algorithm exhibited the most stable performance in identifying patients at elevated mortality risk. Model interpretability analysis identified crucial predictors, potentially informing clinical decisions for sepsis patients in the ICU.

References

Alhamzawi R, Mohammad Ali H . The Bayesian adaptive lasso regression. Math Biosci. 2018; 303:75-82. DOI: 10.1016/j.mbs.2018.06.004. View

Zangmo K, Khwannimit B . Validating the APACHE IV score in predicting length of stay in the intensive care unit among patients with sepsis. Sci Rep. 2023; 13(1):5899. PMC: 10090054. DOI: 10.1038/s41598-023-33173-4. View

Singer M, Deutschman C, Seymour C, Shankar-Hari M, Annane D, Bauer M . The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016; 315(8):801-10. PMC: 4968574. DOI: 10.1001/jama.2016.0287. View

Yang J, Peng H, Luo Y, Zhu T, Xie L . Explainable ensemble machine learning model for prediction of 28-day mortality risk in patients with sepsis-associated acute kidney injury. Front Med (Lausanne). 2023; 10:1165129. PMC: 10232880. DOI: 10.3389/fmed.2023.1165129. View

Cui S, Liang C, Hao Y . Analysis of risk factors affecting the prognosis of patients with sepsis and construction of nomogram prediction model. Eur Rev Med Pharmacol Sci. 2024; 28(6):2409-2418. DOI: 10.26355/eurrev_202403_35748. View

Pollard T, Johnson A, Raffa J, Celi L, Mark R, Badawi O . The eICU Collaborative Research Database, a freely available multi-center database for critical care research. Sci Data. 2018; 5:180178. PMC: 6132188. DOI: 10.1038/sdata.2018.178. View

Ejiyi C, Qin Z, Ukwuoma C, Nneji G, Monday H, Ejiyi M . Comparative performance analysis of Boruta, SHAP, and Borutashap for disease diagnosis: A study with multiple machine learning algorithms. Network. 2024; :1-38. DOI: 10.1080/0954898X.2024.2331506. View

Dankl D, Rezar R, Mamandipoor B, Zhou Z, Wernly S, Wernly B . Red Cell Distribution Width Is Independently Associated with Mortality in Sepsis. Med Princ Pract. 2022; 31(2):187-194. PMC: 9209973. DOI: 10.1159/000522261. View

Yang S, Guo J, Kong Z, Deng M, Da J, Lin X . Causal effects of gut microbiota on sepsis and sepsis-related death: insights from genome-wide Mendelian randomization, single-cell RNA, bulk RNA sequencing, and network pharmacology. J Transl Med. 2024; 22(1):10. PMC: 10763396. DOI: 10.1186/s12967-023-04835-8. View

10.

Pierce C, Larson D . Inflammatory cytokine inhibition of erythropoiesis in patients implanted with a mechanical circulatory assist device. Perfusion. 2005; 20(2):83-90. DOI: 10.1191/0267659105pf793oa. View

11.

Kim M, Choi E, Choi E . Evolving Paradigms in Sepsis Management: A Narrative Review. Cells. 2024; 13(14. PMC: 11274781. DOI: 10.3390/cells13141172. View

12.

Wang Z, Lan Y, Xu Z, Gu Y, Li J . Comparison of Mortality Predictive Models of Sepsis Patients Based on Machine Learning. Chin Med Sci J. 2022; 37(3):201-209. DOI: 10.24920/004102. View

13.

Gattinoni L, Vasques F, Camporota L, Meessen J, Romitti F, Pasticci I . Understanding Lactatemia in Human Sepsis. Potential Impact for Early Management. Am J Respir Crit Care Med. 2019; 200(5):582-589. DOI: 10.1164/rccm.201812-2342OC. View

14.

Baysan M, Arbous M, Steyerberg E, van der Bom J . Prediction of Inhospital Mortality in Critically Ill Patients With Sepsis: Confirmation of the Added Value of 24-Hour Lactate to Acute Physiology and Chronic Health Evaluation IV. Crit Care Explor. 2022; 4(9):e0750. PMC: 9444407. DOI: 10.1097/CCE.0000000000000750. View

15.

Hotchkiss R, Moldawer L, Opal S, Reinhart K, Turnbull I, Vincent J . Sepsis and septic shock. Nat Rev Dis Primers. 2017; 2:16045. PMC: 5538252. DOI: 10.1038/nrdp.2016.45. View

16.

S S . Assessment of INR to Albumin Ratio in Predicting Outcome During Hospital Stay in Patients with Cirrhosis of Liver with Sepsis. J Assoc Physicians India. 2022; 70(4):11-12. View

17.

Maurya N, Kushwah S, Kushwaha S, Chawade A, Mani A . Prognostic model development for classification of colorectal adenocarcinoma by using machine learning model based on feature selection technique boruta. Sci Rep. 2023; 13(1):6413. PMC: 10115869. DOI: 10.1038/s41598-023-33327-4. View

18.

Zhou J, Qian C, Zhao M, Yu X, Kang Y, Ma X . Epidemiology and outcome of severe sepsis and septic shock in intensive care units in mainland China. PLoS One. 2014; 9(9):e107181. PMC: 4167333. DOI: 10.1371/journal.pone.0107181. View

19.

Li M, Han S, Liang F, Hu C, Zhang B, Hou Q . Machine Learning for Predicting Risk and Prognosis of Acute Kidney Disease in Critically Ill Elderly Patients During Hospitalization: Internet-Based and Interpretable Model Study. J Med Internet Res. 2024; 26:e51354. PMC: 11097053. DOI: 10.2196/51354. View

20.

Friedman J, Lopez M, Fleming M, Rivera A, Martin F, Welsh M . SOD2-deficiency anemia: protein oxidation and altered protein expression reveal targets of damage, stress response, and antioxidant responsiveness. Blood. 2004; 104(8):2565-73. DOI: 10.1182/blood-2003-11-3858. View