Development of a Nomogram Model for Predicting Mortality Risk in Critically Ill Children with Influenza

Background: Influenza virus commonly infects children. Children with critical infection may develop influenza-associated encephalopathy or encephalitis, conditions rarely seen in adults and associated with high mortality. This results in growing healthcare resource consumption and family burden. Objective: This study aimed to identify the predictors for death in critically ill children with influenza, so as to facilitate early recognition of high-risk patients, guide clinical interventions, and reduce mortality. Methods: We retrospectively analyzed the clinical data of critically ill children with influenza admitted to the Children’s Hospital of Soochow University between January 2018 and January 2026. L1-penalized binomial logistic regression (LASSO) was used for variable selection. Variables selected by LASSO were entered into a Firth penalized logistic regression model to identify independent predictors of mortality, based on which a nomogram was constructed to estimate mortality risk. Model discrimination was assessed using the receiver operating characteristic (ROC) curve and area under the curve (AUC). Bootstrap resampling was performed for internal validation and optimism correction. Model calibration was assessed using calibration curves, and decision curve analysis (DCA) was performed to evaluate the clinical net benefit of the model. Results: A total of 113 critically ill children with influenza were included, of whom 34 died, corresponding to a mortality rate of 30.1%. LASSO regression was used for variable selection, followed by Firth penalized logistic regression to develop a mortality risk prediction model. The final model included five variables: Hgb, PLT, UREA, ALB, and confusion. A mortality risk nomogram was developed based on these five predictors and internally validated using bootstrap resampling. The ROC curve showed good discrimination of the nomogram in internal validation. The calibration curve demonstrated good agreement between the predicted probabilities and observed outcomes. DCA showed that the nomogram provided a higher net benefit across a certain range of threshold probabilities, indicating its potential clinical utility. Conclusions: Hgb, PLT, UREA, ALB, and confusion are important predictors of mortality in critically ill children with influenza. The nomogram developed based on these predictors showed good discrimination and calibration, as well as potential clinical net benefit, and may serve as an adjunctive tool for early identification of mortality risk and clinical decision-making in critically ill children with influenza.