The best model for predicting vocational outcome at year 1 included premorbid employment, FIM Motor and FIM Cognitive at discharge, age, the interaction of premorbid employment*age, and education level. The goodness-of-fit test indicated that the model fit quite well. The employment*age interaction suggested that age is a factor for individuals who were not employed pre-injury, with younger individuals being more likely to enter the workforce post-injury. Age is not a factor if the individual was premorbidly employed. Similar results were found with Year 2, however FIM Cognitive was no longer found to be a significant contributor to the model. At Year 3, the best predictive model was composed of premorbid employment and FIM Motor. These results have important implications for rehabilitation. Older individuals who were unemployed pre-injury are at a higher risk for a negative vocational outcome than younger individuals, and should be specifically targeted for vocational rehabilitation.