Background: To determine which short-term changes on clinical scales including the Expanded Disability Status Scale (EDSS), Timed 25-Foot Walk (T25FW), 9-Hole Peg test (9HPT) and Guy's Neurological Disability Scale (GNDS) are most predictive of long-term outcome of disability as rated by the EDSS in progressive multiple sclerosis (MS). Methods: From a longitudinal database, all progressive patients, both primary (PP) and secondary (SP), were selected on the basis of at least two complete examinations being available within a time interval of 1-2 years (short-term change). All patients who fulfilled the selection criteria were invited for a third visit after an interval of at least 3 years (long-term outcome). We used ordinal logistic regression to see which early changes were most predictive of the long-term EDSS. Results: 181 patients fulfilled the selection criteria. Early change on EDSS and T25FW were the best predictors of long-term EDSS; both were significant predictors in a 'single predictor' model. Early EDSS change was a slightly stronger single predictor (R2 0.38, Wald χ2 42.65, p < 0.001) compared with early T25FW change (R2 0.27, Wald χ2 12.35, p < 0.001). Adding early T25FW change to early EDSS change in a 'combined predictor' model improved prediction (p = 0.036). Conclusion: Both early change on EDSS and T25FW predict long-term EDSS with comparable strength. Early change on T25FW adds significant independent information and improves the prediction model with early EDSS change only. Therefore we support the use of early T25FW examinations in future clinical trials in progressive MS. © 2012 SAGE Publications.