Prediction rules help in evaluating head injuries

Computed tomography (CT) is widely used for evaluating children with blunt head trauma (BHT) but it is not without risks. Recently developed prediction rules could help in identifying children at very low risk of clinically important traumatic brain injury, for whom CT can safely be avoided, said Nathan Kuppermann, MD, MPH, professor, Departments of Emergency Medicine and Pediatrics, and the Bo Tomas Brofeldt Endowed Chair, Department of Emergency Medicine, University of California, Davis School of Medicine, Sacramento.

At a session on Sunday, October 12, titled “Evaluation and management of minor closed head injuries,” he explained that for children with overt signs of BHT there is little controversy about evaluation. For those with few or subtle signs, however, the value of CT is controversial and practice varies considerably. The vast majority of children with BHT who are evaluated in the emergency department and undergo CT have minor injuries (Glasgow Coma Scale [GCS] 14–15) and do not require surgery.

Because CT for children has risks, particularly radiation-induced malignancies, Kuppermann and his colleagues undertook the PECARN (Pediatric Emergency Care Applied Research Network) Head Injury Study to derive a clinical decision rule that could accurately identify children at near-zero risk of clinically important traumatic brain injury after BHT and obviate the need for CT. They enrolled 42,412 patients with GCS scores of 14 or 15 and obtained CT scans on 14,969 patients, 376 of whom had clinically important traumatic brain injuries. Separate prediction rules were developed for children aged younger and older than 2 years based on variables such as loss of consciousness, scalp hematoma, palpable skull fracture, headache, emesis, and GCS (14 vs 15).

For children aged younger than 2 years, the prediction rule had a negative predictive value for clinically important traumatic brain injury of 100% and a sensitivity of 100%. For older children, the negative predictive value was 100% and the sensitivity was 97%.  

Kuppermann said that the clinical prediction rules can help clinicians optimize the decision-making process and decrease diagnostic, therapeutic, and medicolegal uncertainties. He suggested effective ways to translate research into practice.