Abstract
A comparison was made of the relative power of a brief post-injury neuropsychological assessment and a brain-injury severity measure (PTA) to predict long-term cognitive outcome after severe brain injury. Subjects were 78 patients with severe brain injury. Regression analysis revealed that neuropsychological variables were as predictive of outcome as PTA. A combination of neuropsychological, brain-injury severity, and demographic variables predicted a substantial proportion of variance in both general cognitive ability and memory functioning at outcome. Removal of PTA from the set of predictors had only modest impact on predictive power, suggesting that, in the absence of accurate injury severity data, meaningful prediction about long-term cognitive outcome can still be made.