Related Experiment Videos

Gatekeeping strategies for clinical trials that do not require all primary effects to be significant

Alexei Dmitrienko1, Walter W Offen, Peter H Westfall

  • 1Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285, USA. dmitrienko_alex@lilly.com

Insights

This study introduces gatekeeping strategies to manage multiplicity issues in clinical trials. These methods improve statistical power for testing multiple endpoints or dose levels, ensuring primary objectives are met efficiently.

Area of Science:

  • Biostatistics
  • Clinical Trial Design
  • Statistical Methodology

Background:

  • Clinical trials frequently involve multiple endpoints and dose levels, leading to multiplicity issues.
  • Standard statistical methods can be underpowered when addressing these complex trial designs.
  • Controlling for false positive rates is crucial for valid interpretation of trial results.

Purpose of the Study:

  • To develop and present efficient gatekeeping strategies for handling multiplicity in clinical trials.
  • To enhance the statistical power of hypothesis testing in trials with multiple objectives.
  • To provide practical methods for constructing robust testing procedures.

Main Methods:

  • Gatekeeping strategies are developed, designating a primary set of hypotheses as 'gatekeepers'.
  • Secondary and tertiary hypotheses are tested conditionally upon rejection of gatekeeper hypotheses.
  • Methods include weighted Bonferroni, weighted Simes, and weighted resampling tests within a closed testing framework.

Main Results:

  • The proposed gatekeeping strategies were illustrated with examples from trials featuring co-primary endpoints and multiple endpoints in dose-finding studies.
  • Power comparisons demonstrated that gatekeeping methods are more powerful than competing approaches when primary trial objectives must be satisfied.
  • The developed procedures offer a statistically sound and efficient approach to multiplicity adjustment.

Conclusions:

  • Gatekeeping strategies provide a powerful and flexible framework for addressing multiplicity in clinical trials.
  • These methods are particularly beneficial for trials where meeting primary objectives is paramount.
  • The study offers practical tools for statisticians and researchers designing and analyzing complex clinical trials.

Related Concept Videos