An explainable machine learning model for comorbidity risk stratification in patients with fractures admitted to the intensive care unit: a multicenter study.
• Combines 3763 MIMIC-IV cases with 558 external validations to enhance clinical generalizability. • Assesses comorbidity risk based on admission indicators.. • Eliminates AI black-box effects via quantifiable feature contribution analysis. • Deploys interactive calculator for real-time risk assess...
| Published in: | Archives of Gerontology & Geriatrics Vol. 141 |
|---|---|
| Main Authors: | , , , , , , , |
| Format: | research Journal Article |
| Published: |
Elsevier B.V.
Feb2026
|
| Online Access: | View this record in EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=189667526&site=ehost-live header: @attributes: shortDbName: ccm uiTerm: 189667526 longDbName: CINAHL Complete uiTag: AN controlInfo: bkinfo: dissinfo: jinfo: jid: 01674943 3HX jtl: Archives of Gerontology & Geriatrics issn: 01674943 maglogo: N pubinfo: dt: Feb2026 vid: 141 pid: 1004 pub: Elsevier B.V. artinfo: ui: 189667526 189667526 189667526 10.1016/j.archger.2025.106082 189667526 ppct: 1 formats: tig: atl: An explainable machine learning model for comorbidity risk stratification in patients with fractures admitted to the intensive care unit: a multicenter study. aug: au: Liang, Xuelong Zhao, Weijie Liufu, Weigui Qian, Jiale Xiang, Nantian Zhang, Xinzhe Zhou, Jihui Cui, Hongwang affil: Trauma Orthopedics Department, Maoming People's Hospital, Maoming, Guangdong 525000, China sug: subj: Fractures Complications Comorbidity Risk Factors Global Burden of Disease Risk Factors Intensive Care Units Electronic Health Records Machine Learning Models, Theoretical Risk Assessment Human Adult Multicenter Studies Scales Survival Analysis Random Forest Machine Learning Algorithms Instrument Validation Descriptive Statistics Discriminant Validity Calibration Physicians Health Resource Allocation Adult: 19-44 years ab: • Combines 3763 MIMIC-IV cases with 558 external validations to enhance clinical generalizability. • Assesses comorbidity risk based on admission indicators.. • Eliminates AI black-box effects via quantifiable feature contribution analysis. • Deploys interactive calculator for real-time risk assessment at bedside. Among traumatic-fracture patients admitted to intensive care units (ICUs), those with substantial chronic comorbidities recover more slowly and die more often than their counterparts without such conditions. The age-adjusted Charlson Comorbidity Index (aCCI) quantifies this burden, yet clinicians still lack a tool that can identify—at the point of ICU admission—which fracture patients are likely to have a high aCCI. To fill this gap, we used a large electronic health-record repository to develop and externally validate an interpretable machine-learning model that predicts severe comorbidity burden in this population. We extracted 3 763 adult fracture cases from MIMIC-IV (2008–2019) and split them 3:1 into training and internal validation sets. High comorbidity (aCCI ≥ 7) was defined as the optimal cut-off derived from one-year survival analysis. Nine key predictors emerged from the intersection of LASSO, SVM-RFE, and random-forest importance. Eleven candidate algorithms underwent grid-search hyperparameter tuning with 10-fold cross-validation, and their performance was compared to identify the optimal model, while SHAP clarified model logic. External validation in two Chinese tertiary centres (n = 558) confirmed generalisability, and the final model was deployed as a bedside Shiny calculator. XGBoost achieved the best internal discrimination (AUROC = 0.84; AUPRC = 0.76) and exhibited excellent calibration and net benefit across clinically relevant thresholds. In external validation, AUROC values were 0.88 (Hainan) and 0.83 (Guangdong). The interactive calculator delivers patient-specific risk explanations in real time. An XGBoost-based, SHAP-interpretable model accurately predicts high aCCI in ICU fracture patients and generalises across institutions. The readily accessible web tool can help clinicians identify high-risk individuals early, personalise management, and allocate resources more efficiently. pubtype: Academic Journal doctype: research Journal Article ougenre: Article language: English refInfo: holdings: @attributes: islocal: N |
|---|