|
Activity Number:
|
123
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #305341 |
|
Title:
|
Convergence of Best Intention Adaptive Designs in Dose-Finding Experiments
|
|
Author(s):
|
Yuehui Wu*+ and Valerii V. Fedorov
|
|
Companies:
|
GlaxoSmithKline and GlaxoSmithKline
|
|
Address:
|
2517 Condor Dr., Audubon, PA, 19403,
|
|
Keywords:
|
best intention adaptive designs ; adaptive designs ; dose-finding designs ; optimal design
|
|
Abstract:
|
Adaptive model-based dose-finding designs are widely used in practice while the analysis of their convergence to the right dose is often neglected. In this talk we discuss a few adaptive designs such as best intention, modified best intention, optimal adaptive design etc. under a few popular models including first and second order linear models and first and second order probit models via simulation studies to access the convergence of the adaptive procedures. Adaptive designs related to more complicated models such as drug combination model (two dimensional linear model) and multiple endpoints model (bivariate probit model) are also discussed. This exercise provides evidence that some cautiousness is needed in applying adaptive designs. For instance, we have shown that even for some relatively simple models the "best intention" adaptive designs may converge to a wrong dose.
|
