ORIGINAL RESEARCH article
Front. Med.
Sec. Intensive Care Medicine and Anesthesiology
Integrating dynamic SOFA changes and Age to Predict 28-Day Mortality in ICU 1 Patients:A Nomogram and Machine Learning Validation Study
Provisionally accepted
- 1Longgang Central Hospital of Shenzhen, Shenzhen, China
- 2Department of Critical Care Medicine，Hebei Medical University Third Hospital, Shijiazhuang, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
BACKGROUND: 11 The Sequential Organ Failure Assessment (SOFA) score is widely used to predict 12 prognosis in critically ill patients, but the prognostic value of dynamic SOFA changes 13 (Δ SOFA) and their integration into prediction models remains unclear. 14 METHODS: 15 This retrospective study included 665 ICU patients from the Third Hospital of 16 Hebei Medical University between July 2022 and December 2023. Initial and daily 17 SOFA scores (day 1–3) and demographic data were collected. Patients were stratified 18 by SOFA 1 scores (4–7, 8–11, ≥12). A nomogram combining SOFA1, Δ SOFA 3–1, 19 and age was developed, and its discriminative ability and calibration were evaluated. 20 Additionally, an XGBoost model using the same predictors was constructed to explore 21 the potential value of machine learning. External validation was performed using the 22 MIMIC-IV database. 23 RESULTS: 24 Overall 28-day mortality was 18.9%. Mortality increased with higher SOFA 1 25 scores and Δ SOFA 3–1. The nomogram showed high discriminative ability (C-index: 26 0.852 for SOFA 1=4–7; 0.845 for SOFA 1=8–11) and good calibration. The optimized 27 XGBoost model exhibited excellent discriminative performance in the internal 28 training cohort, achieving an AUC of 0.833. In the independent internal test cohort, 29 the AUC was 0.863.and 0.671 during external validation. SHAP analysis identified Δ 30 SOFA 3–1 as the most influential predictor across datasets. 31 CONCLUSION: 32 Dynamic changes in SOFA scores (Δ SOFA 3–1), especially in patients with 33 moderate baseline SOFA 1 scores (4–11), significantly improve prognostic accuracy 34 when combined with age. The nomogram provides an intuitive bedside tool for early 35 risk stratification, while the XGBoost model demonstrates the potential value of 36 machine learning. External validation highlights the need for further multicenter 37 studies to enhance model generalizability.
Keywords: dynamic, SOFA, Segment, nomogram, XGBoost
Received: 17 Sep 2025; Accepted: 09 Dec 2025.
Copyright: © 2025 Yu, Man, Kang, Jing and ying. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence:
Xu Yu
Guo Jian ying
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.