This was a cross-sectional study of a sample of general EDs in the Nationwide Emergency Department Sample (NEDS; 2006-2010). The authors included visits by patients <19 years of age with International Classification of Diseases, 9th Revision, Clinical Modification, diagnosis codes for head trauma and determined head CT use via Current Procedural Terminology (CPT), Fourth Edition, codes. Crude and risk-adjusted proportions of visits with CT for each hospital were calculated using multilevel mixed effects logistic regression. The association between hospital-level characteristics and head CT were evaluated by constructing multivariable negative binomial regression models.
There were 324,435 pediatric head trauma visits to 848 EDs in the sample. Median patient age was 8 years (interquartile range [IQR] = 2 to 15 years) and 62% of visits were by males. A minority of patients (0.7%) were severely injured, and only 4.2% were admitted to the hospital. Most EDs (79%) were nonteaching institutions, and 84% were nontrauma centers. Risk-adjusted median CT use was 56.0% (IQR = 46.4% to 64.7%). In multivariate analysis, nontrauma centers were 9% (95% confidence interval [CI] = 4% to 15%) less likely to use head CT for pediatric head trauma patients and among discharged patients, EDs within nonteaching hospitals were 8% more likely to perform CT imaging (95% CI = 2% to 14%). There was no change in CT use from 2006 through 2010 (p = 0.31).
The objectives were to evaluate general emergency department (ED) variation in head computed tomography (CT) use for pediatric head trauma, hospital factors associated with CT use, and recent secular trends in CT utilization for pediatric head trauma.
There is significant variability among general EDs in CT use for pediatric head trauma, indicating the need for strategies to reduce variation and improve ED imaging practices for this population.