Authors: Connor Donegan*, University of Texas at Dallas, Yongwan Chun, University of Texas at Dallas, Daniel A. Griffith, University of Texas at Dallas
Topics: Health and Medical, Spatial Analysis & Modeling, Geography and Urban Health
Keywords: Mortality rates, Bayesian inference, Health inequality
Session Type: Virtual Paper
Start / End Time: 3:05 PM / 4:20 PM
Room: Virtual 8
Presentation File: No File Uploaded
Epidemiologists and health geographers routinely use survey estimates as covariates to model areal and even individual health outcomes. American Community Survey (ACS) estimates are accompanied by standard errors but it is not standard practice to use them for evaluating or modeling data reliability. Failure to model probable observational error may have substantial impact on the large bodies of literature relying on small-area estimates including inferential biases and over-confidence in results. As we illustrate, ACS data quality varies systematically by region, neighborhood, and socioeconomic characteristics.
This paper examines the use of hierarchical Bayesian modeling (HBM) to meet the inferential challenges posed by data reliability and spatial autocorrelation. We show how to incorporate models of observational error into a workflow for Bayesian spatial data analysis. To illustrate, we follow Krieger et al.'s (2018) call to routinely use the Index of Concentration at the Extremes (ICE) to monitor spatial inequalities in health and mortality. We construct standard errors for the ICE, use visual diagnostics to evaluate our observational error model for the ICE, and then estimate an ICE-mortality gradient by incorporating the latter model into our model of sex-specific midlife (ages 55-64) all-cause U.S. county mortality rates.
We urge researchers to consider data quality as a criterion for variable selection prior to modeling and to incorporate data reliability information into their models when possible. Our methodology will be implemented in the geostan R package, providing a user-friendly interface to spatial HBMs.