Estimating significance level and power comparisons for testing multiple endpoints in clinical trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Significance

Background:

• Traditional clinical trials prioritize a single primary outcome, with secondary outcomes ranked by importance.
• Recent research indicates a need for alternative approaches, particularly when multiple key outcomes exist with a preferred order.

Purpose of the Study:

• To describe and compare two novel methods for testing hierarchically ordered primary and secondary endpoints in clinical trials.
• To evaluate methods for controlling the overall significance level when multiple key outcomes are present.

Main Methods:

• Described two sequential testing methods: a global test for combined endpoints and a partial Bonferroni correction.
• Employed simulation studies to compare the performance of the two methods.

Main Results:

• Both the global test and partial Bonferroni correction methods effectively control the overall significance level.
• The partial Bonferroni correction demonstrated comparable or superior performance to the global test in most simulated scenarios.

Conclusions:

• The partial Bonferroni correction offers a viable and efficient alternative for analyzing clinical trials with hierarchically structured endpoints.
• These methods provide robust statistical frameworks for trials where multiple key outcomes are assessed sequentially.