Observational approach to subjects with mild-to-moderate head injury and initial non-neurosurgical lesions.

BACKGROUND The model of care for patients with mild-to-moderate head injury and CT-detected lesions that do not require an immediate intervention is a matter of debate. This study compared the effects on outcome of a model based either on observation in a neurosurgical unit (NSU) or in a peripheral hospital (PH), making use of neurosurgical expertise via a teleradiology system. PATIENTS AND METHODS The investigation reviewed the data that was prospectively collected in 865 cases with mild-to-moderate head injury and positive CT scan, not needing immediate neurosurgical evacuation. Outcome was determined at 6 months. The predictive value of location of observation on outcome was evaluated by logistic regression, after adjustment for the propensity score to the type of observation (calculated on main entry variables). FINDINGS 700 subjects had a mild head injury, 105 had a moderate injury with GCS 13-11 and 60 with Glasgow Coma Scale (GCS) 10-9. Only 152/865 subjects (17.6%) were admitted to a NSU. During observation, neurosurgery was necessary in 117 cases (13.5%), 74/152 (48.7%) NSU-observed patients and 43/713 (6.0%; p<0.001) PH-observed cases. The outcome was unfavourable in 18% of the NSU cases versus 10% of the PH cases (p = 0.143). After correction for propensity, no significant differences were found between models of observation (NSU vs. PH; odds ratio, 0.92; 95% confidence interval, 0.49 to 1.75). INTERPRETATION A model of care based on observation in PH with neurosurgical consult by teleradiology system, repeat CT scanning and transfer time 30-60 min to a NSU is not detrimental for subjects with initial non-neurosurgical lesions after mild-to-moderate head injury.