Estimation of a secondary parameter in a group sequential clinical trial

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Inference

Background:

• Group sequential tests are crucial for efficient clinical trials.
• Estimating secondary parameters at trial endpoints can introduce bias.
• Subgroup analysis requires careful consideration of sampling proportions.

Purpose of the Study:

• To investigate and correct bias in secondary parameter estimation within group sequential tests.
• To develop an unbiased estimator for a subgroup mean in normal distribution models.
• To evaluate the performance of the unbiased estimator against naive methods.

Main Methods:

• Theoretical derivation of an unbiased estimator for the subgroup parameter and its variance.
• Simulation studies to compare mean squared errors of unbiased and naive estimators.
• Application of estimation methods to a real-world clinical trial data.

Main Results:

• The bias of the naive estimator is significantly influenced by subgroup sampling proportions at the stopping time.
• An unbiased estimator for the subgroup parameter and its variance was successfully derived.
• Simulations demonstrated negligible differences in mean squared error between unbiased and naive estimators.

Conclusions:

• The derived unbiased estimator effectively addresses bias in secondary parameter estimation.
• The choice of sampling proportion critically impacts bias in naive estimation methods.
• The proposed methods are applicable to real-world group sequential clinical trials, such as The Beta-Blocker Heart Attack Trial.