183 A Clinical Risk Score for Managing Children With Glasgow Coma Scale 13 to 15 Head Injuries and Intracranial Injury.

INTRODUCTION: Although rigorous decision tools exist to evaluate the need for commuted tomography (CT) imaging in children with Glasgow Coma Scale (GCS) score 13 to 15 head injuries, the appropriate management in children with intracranial injury (ICI) on CT remains unclear. The purpose of this study was to develop a risk score to predict neurosurgical intervention (NI), a concrete measure of neurological impairment or decline requiring intensive care unit observation.

METHODS: We included patients younger than 18 years with GCS score 13 to 15 that were enrolled in 1 of 25 hospitals participating in the prospective Pediatric Emergency Care Applied Research Network (PECARN) head injury study and had ICI identified on CT. We used multivariable logistic regression on 10 imputed data sets to identify independent predictors of NI, and used the model β-coefficients to develop a clinical risk score.

RESULTS: Among 42 735 children with GCS 13 to 15 head injuries, 15 162 received a head CT scan and 846 had ICI. Among these patients, 8.3% required NI and no patients died. The most common type of ICI was subdural hematoma (24% of patients). Clinical variables independently associated with NI included palpable depressed skull fracture (odds ratio [OR] = 2.3) and GCS score of 13 (OR = 2.9) or 14 (OR = 1.2); radiological predictors included midline shift (OR = 7.2), epidural hematoma (OR = 1.3), and any skull fracture on CT (OR = 2.1). The model c-statistic was 0.81. Using the model's β-coefficients, the Children's Intracranial Injury Decision Aid (CHIIDA) score was created, which ranged from zero (2% risk of NI) to 27 (96% risk of NI) points.

CONCLUSION: Midline shift, epidural hematoma, any radiological skull fracture, palpable depressed skull fracture, and GCS score accurately predict the risk of NI in children with GCS 13 to 15 head injury and ICI. Based on these results, the CHIIDA score is a novel clinical decision tool to aid physicians caring for these patients.