Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 322 - Novel Statistical Methods and Applications in Precision Mental Health
Type: Topic Contributed
Date/Time: Tuesday, August 9, 2022 : 2:00 PM to 3:50 PM
Sponsor: Mental Health Statistics Section
Abstract #323463
Title: Predicting Antidepressant Response Rate Profiles from High-Dimensional and Multimodal Data Using Ensemble Methods
Author(s): Adam Ciarleglio*
Companies: George Washington University
Keywords: ensemble machine-learning; high-dimensional data; latent class models; response trajectories; treatment effect; depression

In most studies of antidepressant treatment effect, response data are collected at multiple timepoints, yielding a treatment response trajectory for each subject. It is often reasonable to assume that these trajectories arise from latent classes that can be characterized by the timing or quality of the response. Characterizing response in this way may be more relevant than simply utilizing the value of the response at the end of the study, as is commonly done. Using data from a randomized controlled trial comparing sertraline to placebo in subjects with major depressive disorder, we sought to investigate if any pre-treatment characteristics are predictive of membership in these response trajectory classes and/or prescriptive of an optimal treatment. We employ a two-stage approach. The first stage uses a latent class mixed effects model to identify trajectory groups. The second stage employs an ensemble learning approach for predicting trajectory group membership from a collection of high-dimensional measures. This two-stage approach is evaluated via a simulation study and applied to data from the aforementioned study.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2022 program