遵义医科大学学报2026，Vol.49Issue(4)：380-389,10.
危重症患者ICU后疲劳发生风险预测模型的构建与验证
Development and validation of a predictive model for post-ICU fatigue risk in critically ill patients
摘要
Abstract
Objective To construct a risk prediction model for post-ICU fatigue in critically ill patients and vali-date its predictive efficacy.Methods A convenience sample of 236 critically ill patients admitted to the general ICU of a tertiary-level hospital in Guizhou Province from January to December 2024 was selected as the modeling cohort.Telephone follow-up using the Fatigue Severity Scale was conducted one month after ICU discharge.Pa-tients were categorized into a fatigue group(n=155)and a non-fatigue group(n=81)based on the occurrence of post-ICU fatigue.Univariate and multivariate logistic regression analyses identified independent risk factors for post-ICU fatigue.Univariate analysis combined with logistic regression was employed to identify independent risk factors for post-ICU fatigue.A risk prediction model was constructed and a nomogram was developed.The mod-el's accuracy and validity were evaluated using receiver operating characteristic(ROC)curves,calibration curves,Hosmer-Lemeshow tests,and decision curve analysis(DCA).A validation cohort of 114 critically ill patients admitted to the same ICU at the same hospital between January and March 2025 was selected.Model validation was performed using ROC curves and DCA.Results The incidence of post-ICU fatigue in critically ill patients was 65.68％in the training cohort and 60.52％in the validation cohort.Age-adjusted Charlson Comor-bidity Index(ACCI score),mechanical ventilation duration,ICU length of stay,Acute Physiology and Chronic Health Evaluation Ⅱ score(APACHE Ⅱ score),restraint use,anemia,Chinese version Critical Care Pain Ob-servation Tool(CPOT score),and Activities of Daily Living(ADL)score were identified as independent risk factors(P＜0.05).The Hosmer-Lemeshow test for the modeling set yielded x2=6.511,P=0.590,with an ar-ea under the ROC curve of 0.936(95％CI:0.903-0.970),sensitivity of 81.5％,and specificity of 92.6％.The Hosmer-Lemeshow test for the validation set yielded x2=10.528,P=0.230,with an AUC of 0.894(95％CI:0.828-0.960).The model demonstrated a sensitivity of 91.7％and specificity of 78.6％.Conclusion This risk prediction model demonstrates good predictive performance and can serve as a reference for healthcare providers in assessing the risk of post-ICU fatigue in critically ill patients.关键词
危重患者/ICU后综合征/疲劳/危险因素/风险预测模型/列线图/模型验证Key words
critically ill patients/post-intensive care syndrome/fatigue/risk factors/risk prediction model/nomogram/model validation分类
医药卫生引用本文复制引用
辜甜田,陈俊希,王朝平,袁薇薇,王本金,胡汝均..危重症患者ICU后疲劳发生风险预测模型的构建与验证[J].遵义医科大学学报,2026,49(4):380-389,10.基金项目
国家自然科学基金资助项目(NO:72464040). （NO:72464040）
