Construction and Verification of Prediction Model of Hospital Mortality in Sepsis-related Chronic Critically Ill Patients

Yuan Cheng; Ge Wenqi

doi:10.3969/j.issn.1004-5775.2025.12.008

Abstract：Objective:To construct and verify 28-day mortality prediction model for sepsis-related chronic critically ill patients.Methods:Clinical data of 154 sepsis patients admitted to the Department of Critical Care Medicine,The First Affiliated Hospital of Bengbu Medical University from January 1,2022 to November 30,2023 were retrospectively analyzed.Patients were divided into a death group and a survival group based on 28-day outcomes.Differences in indicators between the two groups were compared.The dataset was randomly split into training and validation sets at 6∶4 ratio.Univariate and multivariate logistic regression analysis were used to identify independent risk factors influencing 28-day mortality,and a predictive model was constructed,visualizing as a nomogram.Predictive performance of the model was evaluated using receiver operating characteristic(ROC)curves,and its accuracy was assessed via calibration curves.Results:A total of 154 sepsis patients were included,with 94 survivors and 60 non-survivors at 28 days.Compared with the survival group,neutrophil/lymphocyte ratio(NLR),platelet/lymphocyte ratio(PLR),total bilirubin and APACHE Ⅱ score were significantly increased in the death group,while lymphocyte and albumin were significantly decreased.In the training set,4 predictors,namely APACHEⅡ score(OR=1.146),lymphocyte count(OR=0.220),total bilirubin(OR=1.022)and albumin(OR=0.890),were finally included in the logistic regression analysis,and the prediction model was built based on them.The model demonstrated AUCs of 0.821 in the training set and 0.706 in the validation set,outperforming the APACHE II score alone.Calibration curves confirmed high accuracy,indicating strong clinical utility.Conclusion:The predictive model based on APACHE Ⅱ score,lymphoid cell count,albumin and total bilirubin is of good value in predicting 28-day death in sepsis-related chronic critical patients.