We propose a new class of adaptive randomized trial designs aimed at gaining the advantages of wider generalizability and faster recruitment, while mitigating the risks of including a population for which there is greater a priori uncertainty. Our designs use adaptive enrichment, i.e., they have preplanned decision rules for modifying enrollment criteria based on data accrued at interim analyses. For example, enrollment can be restricted if the participants from predefined subpopulations are not benefiting from the new treatment. To the best of our knowledge, our designs are the first adaptive enrichment designs to have all of the following features: the multiple testing procedure fully leverages the correlation among statistics for different populations; the familywise Type I error rate is strongly controlled; for outcomes that are binary, normally distributed, or Poisson distributed, the decision rule and multiple testing procedure are functions of the data only through minimal sufficient statistics. The advantage of relying solely on minimal sufficient statistics is that not doing so can lead to losses in power. Our designs incorporate standard group sequential boundaries for each population of interest; this may be helpful in communicating our designs, since many clinical investigators are familiar with such boundaries, which can be summarized succinctly in a single table or graph. We demonstrate these adaptive designs in the context of a Phase III trial of a new treatment for stroke, and provide user-friendly, free software implementing these designs.



Media Format


Included in

Biostatistics Commons