|
Activity Number:
|
109
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 4, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #301969 |
|
Title:
|
Preference-Based Estimates of Treatment Efficacy Among Compliant Patients in a Randomized Clinical Trial with a Continuous Outcome
|
|
Author(s):
|
Afisi S. Ismaila*+ and Stephen Walter
|
|
Companies:
|
McMaster University and McMaster University
|
|
Address:
|
Dept of Clinical Epidemiology and Biostatistics, Hamilton, ON, L8N 3Z5, Canada
|
|
Keywords:
|
compliance ; efficacy ; randomized controlled trials ; treatment preference
|
|
Abstract:
|
When there are departures from assigned treatments in a randomized controlled trial (RCT), researchers are faced with the problem of estimating treatment benefits in compliant patients while respecting randomization. Traditional intention-to-treat (ITT) analysis avoids biases associated with the alternative approaches like per protocol (PP) and as-treated (AT) analyses but it provides incomplete information about the expected treatment effect among patients who are committed to taking the treatment. In this talk, we examine four patterns of departures from assigned treatments and present the preference-based (PB) approach to estimating treatment benefits by extending the work of Walter et al. (2006) on binary data. Specifically, we develop new PB estimators for continuous outcomes for each pattern and address the statistical properties of these estimators.
|
