[The stepped wedge design]

Abstract

Not all questions concerning therapeutic effects in medical research can be answered satisfactorily by standard trials. The stepped wedge cluster design is a special form of randomised study in which an intervention at group level is implemented in stages. The design can be used in situations where randomization at the patient level is inappropriate or impossible and where a stepped implementation of the experimental intervention is important for ethical, logistic or financial reasons. The time before the experimental intervention is introduced in a cluster is used as control period. The statistical analysis of cluster trials should account for the fact that individuals within clusters may show many similarities. Ethical aspects of stepped wedge design should be considered carefully, especially when obtaining individual informed consent is not possible.