Development and Validation of a Nomogram Model to Predict the Risk of Severe Pneumonia in Children with Pneumococcal Infection.

Background: Methods: A retrospective cohort of 485 children diagnosed with pneumococcal pneumonia (August 2018-August 2023) was randomly divided into a training set (n=339) and a validation set (n=146) in a 7:3 ratio. Independent predictors of PICU admission were identified using univariate and multivariate logistic regression. A nomogram was constructed based on the training set and evaluated using ROC curves, calibration curves, Hosmer-Lemeshow tests, decision curve analysis (DCA), and SHAP analysis. Results: Multivariate analysis identified nine independent predictors: cardiovascular abnormalities, electrolyte disturbances, elevated neutrophil percentage, prolonged wheezing duration, decreased albumin, decreased hemoglobin, and elevated CT score were risk factors, while prolonged fever and cough duration were protective factors. The nomogram achieved an AUC of 0.92 (95% CI: 0.89-0.95) in the training set and 0.87 (95% CI: 0.81-0.93) in the validation set. Calibration was satisfactory on the Hosmer-Lemeshow test, and DCA demonstrated net clinical benefit across a 5%-95% threshold probability range. SHAP analysis identified cough duration, albumin, and cardiovascular abnormalities as the top contributing features. Conclusion: This nine-variable nomogram demonstrates high accuracy, good calibration, and strong interpretability, providing clinicians with a practical tool for early identification of children at high risk for PICU admission, supporting risk stratification and treatment decisions.