gms | German Medical Science

With growing interest in personalized medicine and targeted therapies the focus on adaptive trial designs for subgroup analyses is increasing as well. Here we take a look at a combination of a blinded sample size recalculation procedure in an internal pilot study and subgroup selection at interim analyses [1], [2]. The increasing complexity of such an adaptive enrichment design with sample size review comes with an increasing number of nuisance parameters that have to be estimated or reestimated, e.g. prevalences of the subgroups, variances in the subgroups, timepoint of the sample size review and timepoint of the interim analysis. Additionally, for the final analysis the different stages of the design have to be combined while still controlling the familywise error rate. Accounting for this we will present two adaptive testing strategies, the combination test [3] and the conditional error function approach [4]. For nested as well as non-overlapping subgroups with normally distributed endpoints we explore the effect of estimating the variances in the subpopulations and the varying timepoints of the sample size review and interim analysis. We give exact results using a multivariate t-distribution where possible and provide approximations otherwise. The performance of the proposed methods are analysed via simulations.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.