Preservation of Type I Error for Partially-Unblinded Sample Size Re-Estimation.

Sample size re-estimation (SSR) at an interim analysis allows for adjustments based on accrued data. Existing strategies rely on either blinded or unblinded methods to inform such adjustments and, ideally, perform these adjustments in a way that preserves Type I error at the nominal level. Here, we propose an approach that uses partially-unblinded methods for SSR for both binary and continuous endpoints. Although this approach has operational unblinding, its partial use of the unblinded information for SSR does not include the interim effect size, hence the term 'partially-unblinded.' Through proof-of-concept and simulation studies, we demonstrate that these adjustments can be made without compromising the Type I error rate. We also investigate different mathematical expressions for SSR under different variance scenarios: homogeneity, heterogeneity, and a combination of both. Of particular interest is the third form of dual variance, for which we provide additional clarifications for binary outcomes and derive an analogous form for continuous outcomes. We show that the corresponding mathematical expressions for the dual variance method are a compromise between those for variance homogeneity and heterogeneity, resulting in sample size estimates that are bounded between those produced by the other expressions, and extend their applicability to adaptive trial design.

Duke Scholars

Author Kevin J. Anstrom Biostatistics & Bioinformatics, Division of Biostatistics