In this thesis, we consider interim sample size adjustment in clinical trials with multiple co-primary continuous endpoints. We aim to answer two questions: First, how to adjust a sample size in clinical trial with multiple continuous co-primary endpoints using adaptive and group sequential design. Second, how to construct a test in order to control the family-wise type I error rate and maintain the power, even if the correlation ρ between endpoints is not known. To answer the first question, we conduct K different interim tests, each for one endpoint and each at level α/K (i.e. Bonferroni adjustment). To answer the second question, either we perform a sample size re-estimation in which the results of the interim analysis are used to estimate one or more nuisance parameters, and this information is used to determine the sample size for the rest of the trial or the inverse normal combination test type approach; or we conduct a group sequential test where we monitor the information, and the information is adjusted to allow the correlation ρ to be estimated at each stage or the inverse normal combination test type approach. We show that both methods control the family-wise type I error α and maintain the power and that the group sequential methodology seems to be more powerful, as this depends on the spending function.
Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:635617 |
Date | January 2014 |
Creators | Ntambwe, Lupetu Ives |
Publisher | University of Warwick |
Source Sets | Ethos UK |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Source | http://wrap.warwick.ac.uk/66339/ |
Page generated in 0.0021 seconds