Practical and Optimal Crossover Designs for Clinical Trials

• Author / Creator

  Su Hwan Kim

• This thesis investigates various issues arising from crossover designs, which received great attention for their efficiency. Crossover designs gain advantage over parallel designs for efficient estimation of treatment effects and smaller required sample size as the within-subject variability is in general maller than between-subject variability. We investigate optimal crossover designs under various assumptions and objectives.

  The within-subject comparisons allow subjects to serve as their own controls. Such within-subject comparisons remove nuisance subject effects. The direct treatment effects and carryover effects, which are portion of effects being carried from one period to the next, are defined through various assumptions. Often, washout periods are applied between periods to minimize the carryover effects. However, it is often difficult to precisely determine or practically implement the sufficiently long washout period. For this reason, optimal designs were constructed with carryover effects.

  First, we investigate the effect of unequal treatment variances. Traditionally, the optimal designs were constructed with an assumption that all treatments being tested have equal variances. However, this assumption may be too naive to describe designs and experiments that are increasingly becoming more complex. Therefore, we investigate how the unequal treatment variability affects the existing optimal designs and construct appropriate optimal designs with unequal treatment variability.

  Second, we investigate the assumption that the carryover effects are proportional to the direct treatment effects. When a constant washout period is applied and failed to remove the carryover effects completely, the existing carryover effects may be described as a proportion of the direct treatment effect, assuming that the proportion may be similar for all treatments being compared. Under this model, we investigate the benefits of adding baseline measurements where a portion of the direct treatment effects remain and can be described by another proportion.

  Lastly, we investigate response adaptive crossover design with two different objective functions. Adaptive designs were designed to allow clinical trials to respond to the information acquired during the trials to achieve specific objectives, which could include but not limited to allocate more subjects to superior treatments, improve statistical efficiency, reduce the sample size for cost consideration, increase the sample size to maintain pre-specified statistical power, or include covariates. In this chapter, we apply a multiple objective function to find balance between treatment effects and statistical efficiency
  and propose a new allocation method that achieves balance between the two objectives.

• Subjects / Keywords

  • Optimal Design
  • Unequal Variance
  • Crossover Design
  • Proportional Model
  • Response Adaptive Design

• Graduation date

  Spring 2020

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-99tj-3x43

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.