[The use of randomisation in clinical studies in rehabilitation medicine: basics and practical aspects]

New therapies in rehabilitation medicine have to be evaluated with clinical trials. For drug approval the methodology of clinical trials is standardized world wide and the results of these studies are widely accepted. This standard should be achieved in clinical trials in rehabilitation research, too. One of the standards is the existence of a control group, comparing the effect of the new intervention against controls. In addition, the investigational and control groups must be equal in terms of the structure of possible confounders. Randomisation is the best possibility to distribute the patients to the therapy-groups, confounders will be equally distributed by chance. Other procedures for assignment to the study groups can result in confounding and lead into biased results. In spite of these advantages, randomisation is not generally accepted in rehabilitation research up to now. There are some reservations, mostly ethical, organisational and methodological ones. However, randomised clinical trials should be conducted in rehabilitation research in order to obtain more convincing results. Our intention is to bring some input in this debate and to present basics and practical aspects of randomisation.