Performance on fitness and back related isometric strength tests, as well as the response to a lifestyle questionnaire, were related to the subsequent occurrence of back injuries in 119 nurses. In all, 22% of subjects sustained injuries during the 18-month study. Injured nurses were more likely to be from high-risk wards and to have received worker's compensation pay for past back injuries. Fitness and lifestyle characteristics did not differ significantly between injured and not-injured groups. Using backward stepwise logistic regression, a model was developed that accounted for 41% of the variability between groups and predicted 67% of those injured. Prior compensation pay, smoking status, and job satisfaction were the most useful discriminators. It was concluded, however, that the fitness and lifestyle parameters measured did not effectively predict back injury in nurses.