Method of balanced adjustment in testing co-primary endpoints.

Area of Science:

• Clinical Trials
• Biostatistics
• Statistical Methods

Background:

• Multiple co-primary endpoints in clinical trials can inflate Type II error rates, necessitating larger sample sizes.
• Existing methods for adjusting significance levels (e.g., Patel, 1991; Chuang-Stein et al., 2007) have limitations, particularly when treatment effects are small or individual hypothesis significance is not paramount.

Purpose of the Study:

• To introduce a novel compromise testing approach for clinical trials with multiple co-primary endpoints.
• To control the maximum joint false positive rate within a restricted null space.
• To offer an alternative when individual statistical significance for each endpoint is not strictly required.

Main Methods:

• A compromise testing strategy is proposed where significance levels are adjusted upward for co-primary endpoints.
• Adjustment is contingent on demonstrating high statistical significance for one or more other co-primary endpoints.
• The method incorporates endpoint correlations, requiring larger adjustments for smaller correlations, and is applicable to restricted null spaces.

Main Results:

• The proposed approach effectively controls the maximum joint false positive rate over the restricted null space.
• This method offers a way to manage Type II error inflation without necessarily demanding excessively large sample sizes.
• The adjustment mechanism is sensitive to the correlation structure among co-primary endpoints.

Conclusions:

• The compromise testing approach provides a viable statistical framework for clinical trials with multiple co-primary endpoints.
• It offers a flexible alternative to traditional methods, especially when seeking to balance statistical power and sample size.
• This method enhances the management of statistical errors in complex trial designs.