|
Activity Number:
|
303
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 2, 2016 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Mental Health Statistics Section
|
|
Abstract #321051
|
View Presentation
|
|
Title:
|
Biosignatures for Treatment Response: Statistical Methods for Developing Depression Treatment Response Index (DTRI)
|
|
Author(s):
|
Eva Petkova* and Thaddeus Tarpey and Robert Todd Ogden and Adam Ciarleglio and Hyung G. Park
|
|
Companies:
|
New York University and Wright State University and Columbia University and Columbia University and Columbia University
|
|
Keywords:
|
treatment decision ;
linear model ;
effect modifier ;
non-parametric regression
|
|
Abstract:
|
We discuss methods for combining pretreatment covariates into a single variable with the goal of developing an interpretable index for prediction of treatment response that could be used in clinical practice. The problem is motivated by the goals of the Establishing Moderators and Biosignatures of Antidepressant Response for Clinical Care (EMBARC) study sponsored by the National Institute for Mental Health (NIMH) of the National Institute of Health (NIH). We discuss optimal choices for defining a single continuous variable that is "best" (according to some criteria) for indexing a treatment response. We propose approaches for selection of both classical clinical scalar measures and advanced brain imaging functional measures in generating the index. The work is an extension of previously proposed concept of a generated effect modifier (GEM) for linear and non-parametric regression. The properties of the proposed index are studied and an illustration using data from the EMBARC trial is presented.
|
Authors who are presenting talks have a * after their name.
