Construction and Validation of a Risk Prediction Model for Sub Delirium Syndrome in ICU Patients

ZHENG Qingyue; GUO Wenfeng; HU Jia; XING Huanmin

doi:10.3969/j.issn.1004-437X.2026.06.010

Abstract：Objective To analyze the risk factors of sub delirium syndrome(SSD)in ICU patients,construct a risk prediction model,and validate it.Methods A total of 214 ICU patients were selected to evaluate the occurrence of SSD using the confusion assessment method for the ICU(CAM-ICU).They were divided into an SSD group of 71 cases and a non SSD group of 143 cases based on the presence or absence of SSD.Binary logistic regression was used to analyze the risk factors for SSD occurrence in ICU patients,and a prediction model was constructed based on the results of binary logistic regression.Results According to binary logistic regression analysis,physical restraint(OR=5.403,P＜0.001),elevated C-reactive protein(CRP)(OR=8.968,P＜0.001),mechanical ventilation(OR=11.145,P＜0.001),use of sedatives(OR=7.060,P＜0.001),use of analgesics(OR=2.979,P=0.030),high critical-care pain observation tool(CPOT)score(OR=1.735,P=0.013),and high acute physiology and chronic health evaluation Ⅱ(APACHE Ⅱ)score(OR=1.543,P=0.013)were risk factors for SSD in ICU patients.Using risk factors to construct a model,Hosmer-Lemeshow fit test showed that,x2=6.205,P=0.624.The area under the curve(AUC)of the logistic regression model predicting SSD in ICU patients was 0.928,with a 95％CI of 0.891-0.965.The sensitivity and specificity were 85.9％and 88.8％,respectively,and the actual application accuracy was 87.9％.By validating the data of the validation group,the receiver operating characteristic(ROC)curve analysis model was used to verify that the AUC of the validation group was 0.914,the 95％CI was 0.845-0.984,and the sensitivity and specificity were 87.5％and 83.0％,respectively.Conclusion The risk prediction model constructed in this article has high predictive value for SSD occurrence in ICU patients.