Combining covariate adjustment with group sequential, information-adaptive designs to improve randomized trial...

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Methods

Background:

• Group sequential designs (GSDs) enable early trial stopping for ethical and efficiency reasons.
• Covariate adjustment improves statistical precision and is recommended by regulatory bodies.
• Combining GSDs with covariate adjustment offers potential dual benefits but presents methodological challenges.

Purpose of the Study:

• To address challenges in combining group sequential designs with covariate adjustment.
• To develop methods ensuring the validity of stopping rules with adjusted estimators.
• To propose adaptive strategies for handling uncertainty in covariate adjustment's precision gains.

Main Methods:

• Applied a linear transformation to adjusted estimators to achieve independent increments for GSDs.
• Generalized existing GSD theory to accommodate regular, asymptotically linear estimators.
• Proposed information-adaptive designs to manage uncertainty in covariate prognostic value.

Main Results:

• Developed a novel sequence of estimators with independent increments, maintaining or improving precision.
• The proposed methods ensure the validity of standard stopping boundaries for GSDs with adjusted estimators.
• Information-adaptive designs allow for efficient trials without compromising validity or power.

Conclusions:

• The study provides valid and efficient methods for integrating covariate adjustment into group sequential trial designs.
• These advancements facilitate more precise and robust clinical trial planning and execution.
• The proposed approaches enhance the ethical and scientific conduct of clinical research through improved statistical power and efficiency.