|
Activity Number:
|
338
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #310163 |
|
Title:
|
Semiparametric Sequential D-Optimal Design
|
|
Author(s):
|
Joseph Warfield*+ and Anindya Roy
|
|
Companies:
|
Johns Hopkins University and University of Maryland, Baltimore County
|
|
Address:
|
11100 Johns Hopkins Rd, Laurel, MD, 20723,
|
|
Keywords:
|
Constrained optimal design ; Isotonic Smoothing spline ; Link function ; Sequential Design ; Phase I Clinical Trial
|
|
Abstract:
|
Sequential D-optimal designs for binary regression models are commonly used in Phase I clinical trials and efficient estimation quantiles of the dose-response curve given by the link function. However, in most applications the link function is assumed to be known up to a location and a scale. We perform an empirical study to illustrate how severely the sequential design and the dose-response quantile estimation maybe affected when the link function is mis-specified. We propose nonparametric estimation of the link function via isotonic smoothing spline estimator under monotonicity constraint and incorporate the estimator into the sequential allocation scheme. This makes the procedure more robust against model misspecification and at the same time maintains the objective of efficient estimation of the quantiles. The methodology is applied to data from a phase I clinical trial.
|
