Previous studies of predictors of long-term productivity outcome following traumatic brain injury have identified preinjury, neurologic, neuropsychologic, and postinjury variables that are associated with patients' productivity status at follow-up. While some of these studies have investigated prediction of outcome of patients treated in postacute brain rehabilitation programs, most have not been restricted to this subject population. Given the selection factors that determine admission to such programs, it is possible that factors predictive of outcome within a broader population of traumatic brain injury survivors will not be predictive in this selected population. To address this issue, we studied predictors of productivity outcome in a sample of 73 traumatic brain injury survivors (mean age = 33.3 years + 13.0, mean education = 12.6 years + 2.4, mean time since injury = 12.5 months + 11.7). Possible predictors studied were preinjury employment status, preinjury education, preinjury substance use, severity of brain injury, Rancho level at admission to postacute rehabilitation, and need for physical, cognitive, and behavioral supervision at discharge from postacute rehabilitation (these were 3 separate scores). Correlational analyses revealed that preinjury substance use and the three supervision ratings were associated with long-term productivity outcome (mean interval from discharge to follow-up = 7.0 months + 4.6). Discriminant analysis using these 4 variables correctly predicted the employment status of 70% of the sample. Results indicated that in this population, preinjury adjustment and postinjury functional status are more powerful predictors of long-term productivity outcome than neurologic severity of injury.