You are here:
Physical inactivity and obesity: Using a novel environmental quality measure to control confounding
Citation:
Gray, C., S. Grabich, L. Messer, K. Rappazzo, J. Jagai, AND D. Lobdell. Physical inactivity and obesity: Using a novel environmental quality measure to control confounding. Presented at Society for Epidemiologic Research, Seattle, WA, June 24 - 27, 2014.
Impact/Purpose:
This abstract demonstrates the use of the Environmental Quality Index as a confounding variable in health outcome analysis. This is one of the functions of the EQI... to control environmental factors in health outcome research.
Description:
Physical inactivity is well-established as a contributor to obesity prevalence in the US. Many aspects of the ambient environment (e.g., air pollution, food deserts, neighborhood socioeconomics) have also been associated with obesity. Yet, controlling for the overall ambient environment in studies examining physical inactivity effects on obesity has been limited due to the sheer number of simultaneously occurring exposures. A novel county-level environmental quality index (EQI) was developed for all US counties from 2000-2005 representing 5 environmental domains: air, water, land, built, and sociodemographic. We extracted county-level rates of age-adjusted physical inactivity and adult obesity prevalence from the 2006 Behavioral Risk Factor Surveillance System and linked the EQI for all US counties (N=3,141). We used random intercept multi-level linear regression with clustering by state to estimate fixed effects of physical inactivity on county obesity rates while controlling for overall environmental quality using the EQI. The crude prevalence difference (PD) (95% confidence interval) was 0.51 (0.48, 0.53); adjusted PD=0.47 (0.44, 0.50). To examine further, we stratified by 4 rural-urban continuum codes (RUCC) ranging from metropolitan urbanized (RUCC1) to rural (RUCC4). In all strata, physical inactivity was crudely associated with obesity (RUCC1: PD=0.55 (0.51, 0.58); RUCC2: PD=0.48 (0.41, 0.56); RUCC3: PD=0.47 (0.52, 0.51); RUCC4: PD=0.46 (0.40, 0.52)). Adjusting for the EQI, the magnitude of association was reduced by roughly 10% in each RUCC (8% in RUCC1, 8% in RUCC2, 11% in RUCC3, and 11% in RUCC4). The effect of physical inactivity on obesity rates at the county level is influenced by overall environmental quality, regardless of the extent of urbanization. The EQI may be a useful tool in controlling for the overall ambient environment. This abstract does not necessarily reflect EPA policy.