Online Program Home
My Program

Abstract Details

Activity Number: 258 - SPEED: Causal Inference and Related Methodology
Type: Contributed
Date/Time: Monday, July 30, 2018 : 2:00 PM to 2:45 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #332587
Title: Multivariate Mediation Analysis with a Multi-Categorical Exposure Variable: An Application to Explore Racial and Ethnic Disparities in Obesity
Author(s): Qingzhao Yu* and Lin Zhu and Bin Li
Companies: Louisiana State University Health Sciences Ctr and Louisiana State University Health Sciences Ctr and Louisiana State University
Keywords: mediation analysis; multivariate multiple analysis; neighborhood walkability; physical activity ; obesity; racial/ethnic disparities
Abstract:

Research has shown consistent racial and ethnic disparities in obesity between non-Hispanic whites (NHW), Hispanic whites (HW) and blacks (BL) in US population. Accounting for the disparities is a challenge given the variety of factors, the nature of the association (i.e., linear versus non-linear) as well as the multi-categorical exposure variable (i.e., NHW, HW and BL). A multivariate mediation analysis methodology is developed to address the challenge. Restricted geographic data from the 2003-2004 and 2005-2006 cycles of the NHANES allowed by individual and tract level variables to be used in characterizing the disparities in both continuous (i.e., BMI) and dichotomous (i.e., obesity risk) models. The novel multivariate mediation analysis method is used to account for races and ethnicities as the exposure variables, and for potential nonlinear associations and complex interactions among variables. With the multivariate multiple mediation analysis, the nonlinear models explained about 40% (95% confidence interval: 18.6%, 58.3%) of racial disparity in obesity risk. However, we did not find significant ethnic disparity (NHW vs. HW) in obese.


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

Back to the full JSM 2018 program