Abstract
Computed tomography (CT) scans are useful for confirming head injury diagnoses. However, there is no standard clinical decision rule (CDR) for determining the need for CT scanning in pediatric patients with head injuries. We developed a CDR and conducted a retrospective cohort study to evaluate its diagnostic accuracy in identifying children with clinically important traumatic brain injury (ciTBI). We selected predictors based on three existing CDRs: CATCH, CHALICE, and PECARN. Of the 2569 eligible patients, 645 (439 (68%) boys, median age: five years) were included in this study. In total, 59 (9%) patients showed ciTBI, and 129 (20%) were admitted to hospital. The novel CDR comprised six predictors of abnormal CT findings. It had a sensitivity of 79.5% (95% confidence interval (CI): 65.5-89.0%) and a specificity of 50.9% (95% CI: 48.9-52.3%). The area under the receiver-operating characteristic curve (0.72, 95% CI: 0.67-0.77) was non-inferior to those of CATCH, CHALICE, and PECARN (0.71, 95% CI: 0.66-0.77; 0.67, 95% CI: 0.61-0.74; and 0.69, 95% CI: 0.64-0.73, respectively; p = 0.57). The novel CDR was statistically noninferior in diagnostic accuracy compared to the three existing CDRs. Further development and validation studies are needed before clinical application.