Clinical performance of NICE recommendations versus NCWFNS proposal in patients with mild head injury

Abstract

Recommendations on treatment of patients with head injury were recently proposed by the National Institute for Clinical Excellence (NICE). We tested the clinical performance of NICE variables versus the proposal of the Neurotraumatology Committee of the World Federation of Neurosurgical Societies (NCWFNS). Over a 5-year period, the clinical data of 7,955 adolescent and adult patients with mild head injury were prospectively collected and patients were managed according to the NCWFNS proposal. Outcome measures were (a) any post-traumatic lesion; (b) need for neurosurgical intervention; (c) unfavorable outcome (death, permanent vegetative state, severe disability) after 6 months. The predictive value of NICE variables was tested by logistic regression analysis. Three hundred fifty-four patients (6.8%) had intracranial lesions on computed tomography (CT) scan; neurosurgical intervention was needed in 108 patients (1.3%), and an unfavorable outcome occurred in 54 patients (0.7%) at 6-month follow-up. NICE variables were less sensitive than NCWFNS (93.5%; 95% confidence interval 91.0-95.2; vs. 97.8%; 96.1-98.7; p < 0.001), but far more specific (70.0%, 69.0-71.0, vs. 45.9%, 44.8-47.0; p < 0.001) for predicting intracranial lesions. NICE variables were also more specific (66.5%, 65.5-67.5, vs. 43.5%, 42.4-44.6; p < 0.001) in the prediction of neurosurgical intervention. 99.1% of unfavorable outcomes were predicted by both protocols. The CT order rate of NICE was much lower (34.1% vs. 57.1%; p < 0.001). In sum, the variables selected by NICE recommendations, when applied to a typical broad sample of emergency medicine, are a reliable, clinically sensible tool in predicting significant outcomes in patients with mild head injury and are resource saving.