Sample size calculation for the weighted rank statistics with paired survival data

Area of Science:

• Biostatistics
• Clinical Trial Design
• Survival Analysis

Background:

• Accurate sample size calculation is crucial for the validity and efficiency of clinical trials.
• Paired survival data analysis presents unique challenges due to the dependency between subjects within a pair.
• Existing methods may not adequately address the complexities of paired data with common censoring times.

Purpose of the Study:

• To introduce a novel sample size calculation method specifically for weighted rank test statistics in paired two-sample survival data.
• To provide a practical formula that accounts for joint survival and censoring distributions in paired study designs.
• To facilitate robust sample size determination for clinical trials involving matched or paired subjects.

Main Methods:

• Developed a sample size formula requiring specification of joint survival and censoring distributions for paired data.
• Utilized a paired exponential survival distribution model, defined by marginal hazard rates and a dependency measure.
• Incorporated practical trial settings, including common censoring times for paired subjects over defined accrual and follow-up periods.

Main Results:

• The proposed sample size formula accurately estimates required sample sizes under realistic study conditions.
• Simulations demonstrated the formula's reliability across various practical settings for paired survival data.
• The method effectively integrates key design parameters such as error probabilities, hazard rates, correlation, and study durations.

Conclusions:

• The introduced sample size calculation method offers a valuable tool for researchers designing studies with paired survival data.
• The formula's ability to handle common censoring times enhances its applicability in real-world clinical trials.
• This method contributes to more precise and efficient study designs, ultimately improving the reliability of research findings.