Activity Number:
|
188
- Bayesian Application to Biological and Health Sciences
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 4, 2020 : 10:00 AM to 2:00 PM
|
Sponsor:
|
Section on Bayesian Statistical Science
|
Abstract #313415
|
|
Title:
|
A Latent Function Approach for Chemical Mixture Interactions in Risk Assessment
|
Author(s):
|
Debamita Kundu* and Sung Duk Kim and Paul S. Albert
|
Companies:
|
National Cancer Institute and National Cancer Institute and National Cancer Institute
|
Keywords:
|
Chemical mixture;
Interaction;
MCMC
|
Abstract:
|
Analyzing health effects associated with exposure to environmental chemical mixtures in population-based studies is a challenging problem in epidemiology, toxicology and exposure science. Various authors have proposed latent class and machine learning-based approaches that can be used to identify interactions between chemicals in risk assessment. These approaches have their advantages but can suffer from interpretation issues since it may be difficult to directly interpret both individual chemicals and their interactions on disease risk. In this paper we directly model the main and interactions in logistic regression, recognizing that the number of such terms can be very large. Rather than using Lasso regression that shrinks estimated terms to zero to avoid over-fitting, we propose a latent function approach to estimate the main effects along with higher-order interactions by assuming that these effects follow a small group of unobserved latent functions. A Bayesian approach is used for parameter estimation and to obtain shrinkage estimates of individual effects for all potential main and interaction terms. A cancer cohort study is used to illustrate the proposed methodology.
|
Authors who are presenting talks have a * after their name.