Neuroimaging Rates for Closed Head Trauma in a Community Hospital.

Abstract

We aimed to characterize the utility of neuroimaging for head trauma in a suburban community hospital and determine whether imaging practices conform to most recent pediatric guidelines. The electronic medical record was surveyed for computed tomographic and magnetic resonance imaging head scans on patients aged 1 to 18 years who were evaluated for trauma. The query included the following: date, sex, type of scan (computed tomography or magnetic resonance imaging), age, patient location, reason for scan, Glasgow Coma Scale (GCS) score (if entered), result, and text from physician's notes. A total of 2679 patients were identified. Within this cohort was a maximum of 29 surgical patients, of whom 8 required a surgical procedure but not neurosurgery among the 592 patients who had a GCS score of 14-15 entered, 2 were confirmed/possible neurosurgical patients, giving a neurosurgical rate of 0.34%. When the GCS 3-13 patient group was analyzed, the relative risk of requiring neurosurgery climbed to 52. Using an established algorithm for pediatric head trauma imaging would have reduced the number of scanned patients to 533. The individual cost of identifying the 29 surgical patients in our population exceeded $31,000. Our rate of serious lesions in GCS 14-15 patients was identical to a larger prospective study in urban teaching hospitals. Using their previously described algorithm might have reduced the number of patients scanned by more than 70% and saved close to $750,000 for the study period.