The present study compares the clinical outcome, as expressed by the Glasgow outcome scale, of two groups of severely head injured patients treated before (49 patients) and after (72 patients) the establishment of a neurosurgical intensive care (NIC) unit at the department of neurosurgery. Uppsala University Hospital. The number of "good recoveries" increased significantly after the establishment of the NIC. This was confirmed by univariate analysis (p less than 0.05) and by multivariate analysis using the logistic regression model to adjust for differences between the two groups of patients (p less than 0.05, p less than 0.005). The most striking improvement was found in patients with a Glasgow coma motor score (GCSM) greater than or equal to 4 on admission. In this subgroup of patients the "good recoveries" increased from 15% to 52%. The object of NIC is basically to prevent or minimize secondary brain damage, and it seems logical that the effect of such care is most obvious in GCS M greater than or equal to 4 patients in whom a good outcome can be anticipated if secondary damage can be prevented or minimized. In conclusion, the present study shows that improved clinical outcome after severe head injury can be achieved by organizing an NIC unit with a well trained staff capable of providing this care 24 hours a day using established methods of surveillance and treatment.