Published data from human volunteer testing and real world collision studies were assessed with a survival analysis to predict injury symptom duration based on collision parameters. This metanalysis considered the effect of age, occupant position, sex, impact direction, type of impact (real world versus staged) and change in velocity (V) as predictor variables. A log-logistic model survivor function was defined with V as the only predictor variable. All other parameters were found not to be significant in predicting injury symptom duration.