Development and Validation of a PSS Criteria-Based Prediction Model for 28-Day Mortality in Pediatric Sepsis

ObjectiveTo construct and clinically validate a prognostic nomogram for 28-day mortality prediction in pediatric sepsis cases meeting PSS diagnostic standards. Methods This study retrospectively enrolled 254 pediatric sepsis patients who met the diagnostic criteria for pediatric sepsis syndrome (PSS) and were admitted to the Pediatric Intensive Care Unit (PICU) of Women and Children's Hospital of Ningbo University from June 2022 to December 2024. The patients were randomly divided into a training set (n=178) and a test set (n=76) at a 7:3 ratio. Demographic data (age, sex), PICU length of stay, infection sites, and routine laboratory parameters within 24 hours of PICU admission were collected, and the comparability of clinical characteristics between the two cohorts was assessed. Based on 28-day outcomes, the patients were classified into survival and non-survival groups. Univariate and multivariate logistic regression analyses were performed on the training set to identify risk factors for mortality and establish a nomogram prediction model. The predictive performance, accuracy, and clinical utility of the nomogram were evaluated in both the training and validation sets using receiver operating characteristic (ROC) curve analysis, calibration curves, and decision curve analysis (DCA). Results In the training cohort, univariate and multivariate logistic regression analyses of the clinical and laboratory parameters identified IL-6, DD , and FIB as independent risk factors for 28-day mortality in pediatric sepsis patients (P < 0.05). A nomogram prediction model constructed using these three variables demonstrated superior predictive performance compared to individual indicators (IL-6, DD, or FIB), with an AUC of 0.834, sensitivity of 0.810, and specificity of 0.801.ROC curve analysis revealed that the nomogram model achieved an AUC of 0.883 (95% CI: 0.802–0.964) in the training set and 0.834 (95% CI: 0.758–0.911) in the test set, indicating good discriminative ability. The calibration curve, assessed by the Hosmer-Lemeshow goodness-of-fit test, showed no significant deviation between predicted and observed outcomes in either the training cohort (P = 0.369) or the validation cohort (P = 0.798), confirming good model fit. Furthermore, decision curve analysis (DCA) demonstrated that the model had favorable clinical utility. Conclusion The nomogram incorporating IL-6, DD, and FIB represents a reliable tool for early risk stratification of 28-day mortality in pediatric sepsis patients meeting PSS criteria, potentially assisting clinicians in optimizing timely interventions and improving prognosis.