A conditional error function approach for adaptive enrichment designs with continuous endpoints.

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Pharmacostatistics

Background:

• Adaptive enrichment designs offer efficient clinical trial strategies for demonstrating treatment efficacy in specific patient subpopulations.
• Controlling the family-wise Type I error rate is crucial for robust statistical inference in these designs.
• Common testing strategies include the combination test and the conditional error function approach.

Purpose of the Study:

• To extend the conditional error function approach for adaptive enrichment designs with multiple subgroups.
• To develop and evaluate new statistical methods for analyzing normally distributed endpoints in nested and nonoverlapping subgroups.
• To address the impact of estimating variances within subpopulations.

Main Methods:

• Derivation of exact t-distribution-based methods for equal variances across subpopulations using multivariate t-distributions.
• Development of improved approximations compared to normal approximations for potentially varying variances.
• Simulation studies to assess the performance of proposed conditional error function approaches against the combination test.

Main Results:

• New exact and approximate statistical methods were derived for adaptive enrichment designs with multiple subgroups.
• The proposed t-distribution-based methods offer improved accuracy over normal approximations, especially when subgroup variances differ.
• Simulation results demonstrate the effectiveness of the conditional error function approaches.

Conclusions:

• The extended conditional error function approach provides a flexible and statistically sound framework for adaptive enrichment designs with multiple subgroups.
• The newly derived t-distribution-based methods enhance the reliability of treatment effect estimation in complex subgroup settings.
• These methods are applicable to clinical trial optimization, as illustrated by an example in pulmonary arterial hypertension.