The sensitivity and specificity of the Cognitive Test for Delirium, which was originally developed for use in intensive care units, were tested in a group of patients with traumatic brain injury who were admitted to a neurorehabilitation center. Sixty-five consecutive patients were evaluated weekly by using the DSM-IV criteria for delirium and the Cognitive Test for Delirium. Complete ratings were available for 249 of 304 weekly observations. Analysis of the receiver operating characteristic curve suggested an optimum cutoff score of less than 22 for identification of delirium by using the Cognitive Test for Delirium, with a sensitivity of 72% and a specificity of 71% compared with the DSM-IV diagnosis. The results suggest that the Cognitive Test for Delirium provides an acceptable level of differentiation between delirious and nondelirious patients with traumatic brain injury.