Sequential analysis of censored survival data from three treatment groups

Area of Science:

• Biostatistics
• Clinical Trials
• Survival Analysis

Background:

• Comparing multiple treatments in clinical trials is complex, especially with survival data.
• Sequential testing offers an adaptive approach to efficiently analyze accumulating data.

Purpose of the Study:

• To propose and analyze a sequential procedure for comparing three treatments using survival data.
• To identify the best treatment while minimizing patient exposure to inferior options.
• To simplify the design and analysis of such procedures for censored survival data.

Main Methods:

• The proposed procedure concatenates two sequential tests.
• The first test detects an overall treatment effect (global test).
• If significant, the least effective treatment is eliminated, and a second test compares the remaining two.

Main Results:

• The procedure allows for the identification of the best treatment among three.
• Information time scales for pairwise comparisons can be approximated by a single scale under certain conditions.
• This approximation simplifies the analysis of censored survival data, drawing parallels with instantaneous normal data.

Conclusions:

• The developed sequential procedure offers a computationally efficient method for comparing three treatments with survival data.
• It provides a simplified approach to designing and analyzing clinical trials focused on identifying the optimal treatment.
• The method reduces the need for extensive simulations, making it more accessible for practical applications.