Online Program Home
My Program

Abstract Details

Activity Number: 497
Type: Contributed
Date/Time: Wednesday, August 3, 2016 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistics in Epidemiology
Abstract #320736 View Presentation
Title: From Presence of Pathogens to Etiology of Disease: An Innovative Latent Class Model with Two Latent Variables
Author(s): Nong Shang*
Companies: CDC
Keywords: Disease etiology ; Latent class model ; Mixture model ; Gibss sampler ; Non-parametric Bayesian ; Latent regression

Many infectious diseases are caused by different pathogens. The etiology distribution of the pathogens changes over time, locations and other population characteristics variables. It is of great public health importance to estimate the etiology distribution promptly and accurately. Recent developments in molecular biology provide efficient tests for the presence of these pathogens in patients. However, the presence might only show carriage of the pathogens in general population, rather than etiology of disease. Through considering the carriage rates as false positive Scott and Wu (2015) redirected the presence tests as etiology tests with a linear mixture model approach. However, there are some clear violations to their model assumptions. There are also conceptual and technical issues in extending their model to study covariates. We proposed a new latent class model that introduces a separate latent variable for presence of background pathogens, in addition to the latent variable for disease etiology. This new model avoids many conceptual pitfalls in previous model and provides greater flexibilities to model associations between background pathogens.

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

Back to the full JSM 2016 program

Copyright © American Statistical Association