Two-stage designs with small sample sizes.

Area of Science:

• Biostatistics
• Clinical Trial Design

Background:

• Group-sequential designs in clinical trials often use approximate critical values assuming normal distributions.
• This approximation can inflate the type I error rate, particularly with small sample sizes, due to the actual t-distribution of test statistics.

Purpose of the Study:

• To present novel two-stage clinical trial designs that accurately control the type I error rate.
• To offer flexible design modifications and evaluate sample size savings through futility stopping and design optimization.

Main Methods:

• Developed and evaluated six alternative two-stage designs for clinical trials with one interim analysis.
• Compared the characteristics of these designs, focusing on type I error rate control and sample size efficiency.
• Incorporated futility stopping rules and design optimization strategies.

Main Results:

• The proposed two-stage designs effectively control the type I error rate.
• Significant sample size savings were demonstrated by incorporating futility stopping and optimizing design parameters.
• These designs offer a valuable alternative for clinical trials with limited sample sizes, such as those for rare diseases.

Conclusions:

• Optimal two-stage designs with futility stopping provide a more efficient and accurate approach than classical group-sequential designs for small sample sizes.
• These methods are particularly beneficial in rare disease research where sample size is often a constraint.
• Accurate critical boundary computation is essential for maintaining statistical integrity in group-sequential trials.