Childhood Obesity Disparities: A Case Study of Hierarchical Spatial Models

• Activity Number: 526
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Jonggyu Baek*+ and Brisa N Sánchez and Emma V. Sanchez-Vaznaugh
• Companies: University of Michigan and University of Michigan School of Public Health and San Francisco State University
• Keywords: Hierarchical spatial model ; Generalized linear mixed model ; Intrinsic conditional autoregressive model

Abstract

With the advent of fast computational methods for Bayesian inference, models to examine spatial patterning of health outcomes have become increasingly common. However, these models have rarely been used in the substantive literature to examine whether spatial patterning of health disparities exists. If any, applications of these models to examine disparities typically stratify data according to race/ethnic group to produce race/ethnic specific maps, but do not rigorously test whether the differences in spatial patterning is more than would be expected by chance alone. We use generalized linear mixed models(GLMM) with intrinsic conditional autoregressive(ICAR) formulation to demonstrate how inferences regarding differences in spatial patterning of health outcomes across groups can be obtained. As a case study, we apply the methods to data regarding childhood obesity disparities in the state of California school-age children. Data on individual-level obesity status is combined with covariates at multiple levels of aggregation (e.g., individual- and school- levels), is combined to examine covariate-adjusted variation(and clustering) in obesity rates and disparities in California.