Sample size calculations for clinical studies allowing for uncertainty about the variance

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmaceutical Research

Background:

• Sample size estimation is critical for pharmaceutical clinical trial design.
• Traditional sample size formulas for superiority trials assume known population variance, which is often an estimate with associated uncertainty.

Purpose of the Study:

• To develop and present a method for calculating clinical trial sample sizes that accounts for the imprecision in the estimated sample variance.
• To highlight the inadequacy of traditional formulas when variance is estimated with limited precision.

Main Methods:

• The study proposes a novel sample size calculation that incorporates the uncertainty of the estimated population variance.
• The methodology addresses the degrees of freedom associated with the variance estimate.

Main Results:

• Traditional sample size formulas yield results that are too small when population variance is estimated.
• The deficiency in sample size is more pronounced with fewer degrees of freedom for the variance estimate.

Conclusions:

• The proposed methodology for sample size calculation is recommended when the estimated sample variance has fewer than 200 degrees of freedom.
• Accounting for variance estimate uncertainty ensures more robust and accurate clinical trial sample sizes.