Multi-arm multi-stage adaptive trials compare several treatment arms to a common control arm over several stages. The treatment arms might originate within the same organization, or several organizations might each contribute a treatment arm to be tested on a common platform. At any stage treatment arms may be dropped either for overwhelming efficacy or futility. The sample sizes of the remaining arms, the number and spacing of the remaining stages and the alpha spending function may then be adaptively altered while nevertheless preserving strong control of the family wise error rate (FWER).
Two methods are available for designing trials that have such a high degree of flexibility.
The stage-wise MAMS approach was historically the first to be developed and remains the standard method for designing inferentially seamless phase II/III clinical trials. In this approach, at each stage, the data from each treatment comparison are summarized by a single multiplicity adjusted p-value. These stage-wise p-values are combined by a pre-specified combination function and the resultant test statistic is monitored with respect to the classical two-arm group sequential boundaries. The cumulative MAMS approach is a more recent development in which a separate test statistic is constructed for each treatment comparison from the cumulative data at each stage. These statistics are then monitored with respect to multiplicity adjusted group sequential boundaries.
In this talk I will introduce each method and discuss how they differ in their approach to FWER control. I will then demonstrate that the cumulative MAMS designs out-perform stage-wise MAMS and recommend that they replace the latter as the method of choice for designing multi-arm multi-stage adaptive designs. The methods will be applied to a multi-arm trial of heart failure in patients with reduced ejection fraction.