The present study investigated the contribution of functional ratings to prediction of employment outcome after traumatic brain injury (TBI). Previous studies have suggested that functional ratings obtained at a significant time post-injury can supplement neurologic, pre-injury, neuropsychologic, and other post-injury variables in predicting long-term employment outcome. Functional ratings studied were patients' needs for physical, cognitive, and behavioural supervision. This investigation also addressed the issue of predicting long-term outcome for the select group of TBI patients who receive post-acute brain injury rehabilitation. Subjects were 76 patients with TBI. The mean age (25th, 50th, and 75th percentiles) was 32 (22, 28, 39) years and mean premorbid education level was 13 (12, 12, 14) years. Predictors studied were severity of injury, premorbid education level, pre-injury substance use, and needs for physical, cognitive and behavioural supervision at discharge from post-acute rehabilitation. Supervision needs ratings were obtained an average of 9.6 (4.2, 5.9, 11.2) months post-injury. Productivity status was assessed an average of 22.5 (12.6, 20.7, 30.5) months post-injury and 12.9 (4.9, 12.4, 16.6) months post-discharge from treatment. Spearman correlation coefficients revealed that premorbid educational level, pre-injury substance use, and needs for physical and behavioural supervision were related to long-term functional outcome (p < 0.05). However, multiple logistic regression analysis revealed that only level of pre-injury substance use was predictive of long-term productivity outcome once adjusted for the effects of the other predictors. Patients with no history of pre-injury substance use were more than eight times as likely to be employed at follow-up as those with a history of pre-injury substance abuse (p < 0.01).