A logrank test-based method for sizing clinical trials with two co-primary time-to-event endpoints

Area of Science:

• Biostatistics
• Clinical Trial Design
• Survival Analysis

Background:

• Clinical trials often evaluate interventions using multiple co-primary endpoints.
• Time-to-event endpoints are common in many therapeutic areas, such as oncology and cardiology.
• The correlation between co-primary endpoints can significantly impact study power and sample size requirements.

Purpose of the Study:

• To develop and present methods for sample size determination in clinical trials with two correlated co-primary time-to-event endpoints.
• To provide a framework for accounting for endpoint correlation in sample size calculations.
• To evaluate the performance and behavior of the proposed sample size methods.

Main Methods:

• Utilized copula families to model the joint distribution of bivariate time-to-event endpoints.
• Specified a correlation structure for the bivariate logrank statistic to account for endpoint correlation.
• Employed simulation studies to assess the proposed sample size calculation methods and their sensitivity to correlation.
• Focused on a two-arm randomized trial design, the most common scenario.

Main Results:

• The proposed methods provide a means to calculate the necessary sample size when dealing with correlated co-primary time-to-event endpoints.
• Simulation results demonstrate the performance of the developed methods and illustrate the impact of correlation on required sample sizes.
• The study highlights the importance of explicitly modeling endpoint correlation for accurate sample size estimation.

Conclusions:

• Accurate sample size calculation is essential for the successful design of clinical trials with multiple correlated endpoints.
• The presented methods offer a practical approach for biostatisticians and clinical trial designers.
• Future research could extend these methods to more complex trial designs or endpoint structures.