Treatment allocation for nonlinear models in clinical trials: the logistic model

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Modeling

Background:

• Binary outcome variables are common in clinical trials.
• Logistic regression is frequently used for covariate adjustment in these trials.
• Optimal designs for nonlinear models often result in unbalanced treatment allocations.

Purpose of the Study:

• To evaluate the efficiency of balanced sequential allocation schemes.
• To compare these schemes against sequential Ds-optimal designs for logistic models.
• To assess the impact of different allocation methods on trial efficiency.

Main Methods:

• Efficiency measurement of balanced sequential allocation schemes.
• Comparison with sequential Ds-optimal designs for the logistic model.
• Utilized completed trials from the Eastern Cooperative Oncology Group and systematic simulations.

Main Results:

• Stratified, balanced designs demonstrate general efficiency.
• Complete randomization frequently results in inefficient treatment allocation.
• Some trials may experience significant inefficiency due to complete randomization.

Conclusions:

• Balanced sequential allocation schemes are efficient for logistic regression models.
• Stratified balanced designs offer a robust approach to treatment allocation.
• Careful consideration of allocation methods is crucial for efficient clinical trial conduct.