Abstract
INTRODUCTION: Pediatric traumatic brain injury (PTBI) is linked to significant disability and mortality. This study aimed to identify risk factors for unfavorable outcomes in patients with PTBI and develop a predictive risk model. METHODS: A retrospective analysis was conducted on patients with PTBI treated at the 900th Hospital from September 2021 to June 2023. Univariate and multivariate regression analyses identified risk factors for adverse outcomes and facilitated the creation of a nomogram. The model's predictive accuracy was assessed using Receiver Operating Characteristic (ROC) curves, calibration curves, and Decision Curve Analysis (DCA). External validation was performed with patients with PTBI from Fujian Children's Hospital. RESULTS: Key findings indicated that a Glasgow Coma Scale (GCS) score ≤8, subdural hematoma, subarachnoid hemorrhage, and coagulopathy were independent risk factors. The nomogram achieved an area under the ROC curve of 0.947 in the development cohort and 0.834 in the external validation cohort, demonstrating a good fit. DCA results confirmed that the nomogram enhanced the prediction of unfavorable outcomes. CONCLUSIONS: This risk prediction model offers high accuracy for early identification of adverse outcomes, enabling timely interventions to improve the quality of life for patients with PTBI.