Analysis of patient data from a new neuroscience intensive care unit (NSICU) permitted evaluation of whether such a specialty ICU favorably altered clinical outcomes in critically ill neuroscience patients, and whether such a care model produced an efficient use of resources. A retrospective review was performed to compare (1) the clinical outcomes, as defined by percent mortality and disposition at discharge, between patients with a primary diagnosis of intracerebral hemorrhage treated in 1995 in medical or surgical ICUs and those treated in the same medical facility in an NSICU in 1997; and (2) the efficiency of care, as defined by length of ICU stay, total cost of care, and specific resource use, between patients treated in the NSICU and national benchmark standards for general ICUs during the 1997 fiscal year (FY). In the latter, extracted patient population data on neurosurgery patients requiring ICU treatment during FY 1997 were used with the following adjacent-disease related group (A-DRG)-coded diseases: craniotomy with and without coma or intracerebral hemorrhage, and skull fracture with and without coma lasting longer than 1 hour. Outcome measures of percent mortality and disposition at discharge in patients with intracerebral hemorrhage were significantly improved (P < .05), compared with those in a similar cohort treated 2 years earlier in a general ICU setting. Also, patients treated in the NSICU had shorter hospital stays (P < .01 ) and lower total costs of care (P < .01) than a national benchmark. The data suggest that a neuroscience specialty ICU arena staffed by specialty-trained intensivists and nurses is beneficial.