|
Activity Number:
|
217
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #304339 |
|
Title:
|
An Estimating Equations Approach for Latent Transition Models with Latent Class Predictors in Drug Use Epidemiology
|
|
Author(s):
|
Beth A. Reboussin*+ and Nicholas Ialongo
|
|
Companies:
|
Wake Forest University School of Medicine and Johns Hopkins Bloomberg School of Public Health
|
|
Address:
|
Department of Biostatistical Sciences, Winston-Salem, NC, 27157,
|
|
Keywords:
|
estimating equations ; GEE ; latent transition ; latent class ; robust variance ; drug use
|
|
Abstract:
|
We present a latent transition model (LTM) to guide our understanding of drug use progression while accounting for measurement error in self-report data and extend the LTM to include a latent class predictor. We begin by fitting two separate latent class analysis (LCA) models using second-order estimating equations: (1) a longitudinal LCA model to define stages of drug use, and (2) a cross-sectional LCA model to define latent class predictor subtypes. The LTM model parameters describing the probability of transitioning between the LCA-defined stages of drug use and the influence of the LCA-defined subtypes on these transition rates are then estimated using a set of first-order estimating equations given the LCA parameter estimates. A robust estimate of the LTM parameter variance that accounts for the variation due to the estimation of the two sets of LCA parameters is proposed.
|
