Increasing the sample size during clinical trials with t-distributed test statistics without inflating the type I...

Area of Science:

• Statistics
• Clinical Trials
• Biostatistics

Background:

• Sample size determination in clinical trials using t-tests is critically dependent on accurate estimation of variance.
• Utilizing variance estimates from prior studies can lead to misspecification and underpowered trials.
• Adaptive sample size re-estimation offers a solution to refine trial planning based on accumulating data.

Purpose of the Study:

• To introduce a flexible methodology for sample size re-estimation in clinical trials employing t-distributed test statistics.
• To enable sample size adjustments at any point during a trial without a pre-defined rule.
• To maintain the efficiency of the t-test design even when no sample size extension is ultimately required.

Main Methods:

• Development of a novel, flexible method for sample size re-calculation applicable to fixed-sample and group-sequential trials.
• The method accommodates the re-estimation of variance using all available data collected during the trial.
• Ensures that the type I error rate is rigorously maintained throughout the adaptive process.

Main Results:

• The proposed method allows for timely and accurate sample size adjustments based on current data.
• It offers an advantage over other adaptive methods by preserving the t-test design efficiency when extensions are not needed.
• Demonstrated preservation of the type I error rate, ensuring trial integrity.

Conclusions:

• This flexible approach enhances the adaptability of clinical trial designs by allowing dynamic sample size adjustments.
• The method provides a robust way to handle uncertainty in variance estimation, improving trial efficiency.
• It offers a valuable tool for statisticians and researchers conducting clinical trials with t-tests.