Controlling Multiplicity in Confirmatory Clinical Trials Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 20, 2019
  • Sun, Hengrui
    • Affiliation: Gillings School of Global Public Health, Department of Biostatistics
  • Multiplicity is an important statistical consideration that arises in clinical trials when the efficacy of the test treatment is evaluated in multiple ways. The scope of multiplicity includes multiple efficacy endpoints, multiple inferential subgroups, multiple treatments or doses, etc. The major concern for multiplicity is that insufficiently controlled multiple assessments lead to an inflated family-wise (or experiment-wise) Type I error rate (FWER) and they thereby undermine the integrity of the statistical inferences. Therefore, a sound statistical strategy that controls FWER in a strong sense for the multiple assessments, without excessive loss of power (or Type II error) from over-control, is crucial for the success of the trial. Chapters 1 and 2 explore strategies for multiplicity issues that come from only one source, which may involve strictly ordinal response outcomes, stratified design, baseline imbalances, and missing values. Realistic case examples are provided, and solutions are proposed to address those issues comprehensively. Multiplicity can also come simultaneously from two or more sources in a confirmatory clinical trial. For example, a clinical trial is designed to make inferences either on the overall population or some pre-specified sub-populations, while multiple endpoints need to be evaluated for each population. For this complex multiplicity problem, we propose a strategy based on multi-way averages in Chapter 3. We apply this strategy to a cardiovascular clinical trial, and discuss a simulation study based on this trial to compare the power of the proposed strategy to some more well-known alternative approaches. More than two sources of multiplicity can exist in one confirmatory clinical trial concurrently. Chapter 4 presents an illustrative example of a confirmatory clinical trial that is designed to have efficacy assessed for two or more primary endpoints, for multiple dose groups, and at two or more post baseline visits. The proposed strategies with multi-way averages manage the higher dose and the lower dose equally and enable evaluation for the multiple endpoints at multiple visits collectively. Simulation studies show that the proposed strategy can comprehensively meet most of the primary objectives of the trial effectively with reasonably high overall power.
Date of publication
Resource type
Rights statement
  • In Copyright
  • Koch, Gary
  • Preisser, John
  • Ivanova, Anastasia
  • Heiss, Gerardo
  • Herring, Amy
  • Davis, Sonia
  • Doctor of Public Health
Degree granting institution
  • University of North Carolina at Chapel Hill Graduate School
Graduation year
  • 2016

This work has no parents.