Maintenance Notice

Due to necessary scheduled maintenance, the JMIR Publications website will be unavailable from Wednesday, July 01, 2020 at 8:00 PM to 10:00 PM EST. We apologize in advance for any inconvenience this may cause you.

Who will be affected?

Accepted for/Published in: JMIR Medical Informatics

Date Submitted: Apr 15, 2025
Open Peer Review Period: Apr 23, 2025 - Jun 18, 2025
Date Accepted: Oct 6, 2025
(closed for review but you can still tweet)

The final, peer-reviewed published version of this preprint can be found here:

Prognostic Value of the Charlson Comorbidity Index for Mortality and Machine Learning–Based Prediction in Critically Ill Patients with Paralytic Ileus: Retrospective Cohort Study

Feng H, Zhou F, Shen Y, Wang Z, Yuan Y, Jing W, Zheng Z, Peng H, Yu Q

Prognostic Value of the Charlson Comorbidity Index for Mortality and Machine Learning–Based Prediction in Critically Ill Patients with Paralytic Ileus: Retrospective Cohort Study

JMIR Med Inform 2025;13:e76003

DOI: 10.2196/76003

PMID: 41139400

PMCID: 12554354

Prognostic value of the Charlson Comorbidity Index for mortality and machine learning–based prediction in critically ill patients with paralytic ileus

  • Hui Feng; 
  • Fuhai Zhou; 
  • Yi Shen; 
  • Zhen Wang; 
  • Yiyang Yuan; 
  • Wenshan Jing; 
  • Zhou Zheng; 
  • Hui Peng; 
  • Qingsheng Yu

ABSTRACT

Background:

The burden of paralytic ileus (PI) in the intensive care unit (ICU) remains high, and the Charlson Comorbidity Index (CCI) is strongly associated with the prognosis of several acute and chronic diseases. However, there is no literature on the clinical value of CCI as a prognostic assessment tool for critically ill patients with PI in the ICU.

Objective:

The aim of this study was to investigate the relationship between CCI and clinical prognosis in critically ill patients with PI.

Methods:

In this study, data from the Critical Care Medical Information Marketplace IV 2.2 database were used to determine the optimal cutoff value of CCI for predicting mortality in patients with PI using receiver operating characteristic (ROC) curves, and the relationship between CCI and mortality was evaluated using Cox regression and restricted cubic spline analysis. A machine learning (ML) prediction model was then constructed to predict hospital mortality by combining CCI and other clinical characteristics.

Results:

The study included 863 patients with PI (median age 65.4 years [interquartile range 54.6-75.5 years], 66.6% male). The ROC curve identified an optimal cut-off value of 4.5 for CCI. Multivariate Cox regression analysis showed that compared to the lowest CCI quartile, patients with elevated CCI levels were more likely to have elevated hospital (Q4: HR 2.447, 95% CI 1.210-4.951), 28-day (Q4: HR 3. 891, 95% CI 1.956-7.740) and 90-day (Q4: HR 3.994, 95% CI 2.224-7.173) all-cause mortality were significantly associated with elevated CCI levels; however, the association with ICU mortality (Q4: HR 1.892, 95% CI 0.653-5.480) was weak. Among the 11 ML models, the LightGBM model performed best, with internal validation results showing an area under the curve of 0.811, a G-mean of 0.670, and an F1 score of 0.895.

Conclusions:

The CCI is an important predictor of hospital, 28-day, and 90-day all-cause mortality in critically ill patients with PI, and the optimal threshold is 4.5. ML models including the CCI show high accuracy in predicting hospital mortality, and the CCI occupies an important position in the model. This suggests that the CCI helps to identify high-risk patients, supports clinical decision making, and improves prognosis. Clinical Trial: NO


 Citation

Please cite as:

Feng H, Zhou F, Shen Y, Wang Z, Yuan Y, Jing W, Zheng Z, Peng H, Yu Q

Prognostic Value of the Charlson Comorbidity Index for Mortality and Machine Learning–Based Prediction in Critically Ill Patients with Paralytic Ileus: Retrospective Cohort Study

JMIR Med Inform 2025;13:e76003

DOI: 10.2196/76003

PMID: 41139400

PMCID: 12554354

Download PDF


Request queued. Please wait while the file is being generated. It may take some time.

© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.