Objective: To compare a decision rule with clinician judgment for identifying children at risk of traumatic brain injury (TBI) after blunt head trauma.
Methods: We performed an observational study of children with blunt head trauma. Emergency department physicians documented suspicion for TBI before cranial computed tomography (CT), rating suspicion as very low, low, intermediate, or high. Our outcome variable was TBI on CT. We compared clinician judgment (very low vs. higher suspicion) for TBI on CT with a decision rule derived from the same database.
Results: Of 1865 children enrolled for whom physician suspicion was recorded, 1168 (62.6%) underwent CT and comprised the study population. Eighty-nine (7.6%; 95% confidence interval [CI], 6.2% to 9.3%) of the 1168 had TBIs on CT. The decision rule had a sensitivity of 88 (98.9%) of 89 versus 84 (94.4%) of 89 for clinician judgment (difference, 4.5%; 95% CI, -0.9% to 9.9%). The specificity of the decision rule was 288 (26.7%) of 1079 versus 329 (30.5%) 1079 for clinician judgment (difference, 3.8%; 95% CI, 0.5% to 7.1%). Application of the decision rule to the study population would have resulted in 289 (24.7%) fewer CT scans, although missing 1 child with a TBI (who was discharged home from the emergency department).
Conclusions: A decision rule trended toward greater sensitivity than clinician judgment for identifying children with TBI on CT after blunt head trauma but was less specific. Because decisions to order cranial CT did not strictly follow clinician judgment, however, use of the decision rule would have resulted in less frequent use of CT.