Exploring incomplete variants of stepped wedge cluster randomised trial designs

This thesis considers how to make group-based clinical trials more efficient. In standard clinical trials, individual participants are randomised to interventions; however, in group-based clinical trials, entire groups of individuals are randomised to interventions. These groups of individuals could be all patients in a hospital, or all students in a school, for example. These group-based trials are often expensive and time-consuming to conduct. This research developed statistical methods to assist researchers in selecting less burdensome and less expensive designs for conducting these trials, investigated the statistical properties of these designs, and offered practical guidance on how they should be described.