Randomisation in double-blind multicentre trials with many treatments

Trials of many (>2) treatments are common, particularly double-blind dose finding trials in Phase IIb where trials of four or more treatments are often seen. Such trials will almost certainly be conducted at multiple centres given the sponsor's desire to speed up the development program. When conducting such multicentre trials, the study team will always have to consider the question of whether the randomisation of the trial is to be stratified or balanced in some way by participating centre. In addressing this question there are three considerations: the impact on power, predictability and medication supply requirements. This paper will describe the use of alternative complex list based randomisation solutions that have been recently developed and implemented; two of the solutions have not previously been presented or published. The techniques maintain various degrees of control over treatment allocations at the centre and study as a whole; consequently the methods help with conserving supplies of medication. Simulation evidence will be presented for each technique.