A comparison of sample size methods for the logrank statistic.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Survival Analysis

Background:

• Accurate sample size calculation is crucial for the statistical power of clinical trials.
• Comparing survival curves using the logrank statistic is a common endpoint in many clinical studies.
• Existing methods for sample size determination vary in their applicability and assumptions.

Purpose of the Study:

• To review and compare different methods for sample size calculation in clinical trials.
• To evaluate the performance of these methods under various survival data scenarios.
• To provide guidance on selecting appropriate sample size calculation methods for logrank statistic-based survival comparisons.

Main Methods:

• Literature review of existing sample size calculation methodologies.
• Simulation studies to assess method performance under different hazard assumptions (exponential, proportional hazards, non-proportional hazards).
• Comparative analysis of method advantages and disadvantages.

Main Results:

• The performance of sample size calculation methods differs significantly based on the underlying survival data distribution and hazard proportionality.
• Some methods demonstrate robustness across various scenarios, while others are sensitive to deviations from proportional hazards.
• Simulation results highlight specific method recommendations for different clinical trial settings.

Conclusions:

• No single sample size calculation method is universally optimal for all clinical trial survival analyses.
• Method selection should be informed by the expected hazard characteristics of the study population.
• Understanding the limitations and strengths of each method is essential for appropriate sample size determination in survival trials.