Go to  Advanced Search

Procedures for multiple outcome measures with applications to multiple sclerosis clinical trials

Show full item record

Files in this item

Files Size Format Description   View
ubc_1998-0020.pdf 5.895Mb Adobe Portable Document Format   View/Open
Title: Procedures for multiple outcome measures with applications to multiple sclerosis clinical trials
Author: Guh, Payhsuan Daphne
Degree Master of Science - MSc
Program Statistics
Copyright Date: 1997
Abstract: In planning clinical trials in many subject areas, researchers often find it difficult to designate one single outcome measure as the primary endpoint to describe treatment efficacy. When a disease affects a patient's functions in multiple dimensions, expecting one outcome measure to assess treatment efficacy in a comprehensive way may not be realistic. Multiple sclerosis (MS) is one such complex disease. The topic addressed in this thesis concerns approaches for the design and analysis of clinical trials where a multidimensional outcome measure is used to measure treatment efficacy. The most common approach is to select a single primary endpoint for formal statistical testing with all other outcome measures considered as secondary. This thesis is concerned with the situation where agreement on a single primary endpoint is not possible so that methods based on multiple endpoints are required. Five methods, Bonferroni adjustment, Hotelling's T2, O'Brien's OLS and GLS statistics and disjunctive outcome measures are examined and compared through power and sample size calculations. Our discussion of these methods is focused on two-armed (placebo and treatment) randomized clinical trials based on continuous outcome measures. We assume that the data to be analyzed are the changes in the responses from the baseline to the end of the trial and the underlying distribution of the multiple outcome measures can be approximated as multivariate normal. Our investigation is focused on the features of the configuration of the standardized differences in the underlying population means and the correlation structure among the multiple outcome measures. Specifically, several special cases are examined to highlight the main differences among the statistical properties of these methods. We also apply the methods considered to two MS clinical trial data sets for a more focused comparison of these methods for actual MS patient populations.
URI: http://hdl.handle.net/2429/7612
Series/Report no. UBC Retrospective Theses Digitization Project [http://www.library.ubc.ca/archives/retro_theses/]
Scholarly Level: Graduate

This item appears in the following Collection(s)

Show full item record

All items in cIRcle are protected by copyright, with all rights reserved.

UBC Library
1961 East Mall
Vancouver, B.C.
Canada V6T 1Z1
Tel: 604-822-6375
Fax: 604-822-3893