Comparing the survival of two groups with an intermediate clinical event

Area of Science:

• Biostatistics
• Clinical Trial Design
• Survival Analysis

Background:

• Clinical trials often measure time-to-event endpoints, but intermediate events can alter survival distributions.
• Existing statistical methods may not fully account for the impact of these intermediate events.

Purpose of the Study:

• To develop and evaluate novel statistical models for analyzing survival data in the presence of intermediate clinical events.
• To compare the performance of these new models against established statistical tests.

Main Methods:

• Developed semi-Markov and time-dependent mixture models for one-sample (detecting change) and two-sample (comparing distributions) problems.
• Conducted simulation studies to assess the power of the new statistical tests.
• Applied the methods to a real-world clinical trial data.

Main Results:

• The newly developed statistical tests demonstrated uniformly greater power compared to log rank, stratified log rank, and landmark tests.
• The models effectively analyze survival distributions influenced by intermediate clinical events.

Conclusions:

• The proposed statistical models provide a robust framework for analyzing time-to-event data with intermediate events.
• These methods enhance the ability to detect therapy benefits in clinical trials, as demonstrated in an AIDS Clinical Trial Group (ACTG) study.