PS2-35: Associations of Body Mass Index with Measures of the Built and Social Environments in Children and Adolescents: The Environmental Health Institute Child and Adolescent Obesity Study

BACKGROUND/AIMS: No prior studies in children have evaluated how age may modify relations of the built and social environments with obesity outcomes, nor evaluated the range of scales and contexts over which these influences may operate. METHODS: We performed a cross-sectional, multilevel analysis of body mass index (BMI) data from over 47,000 Geisinger primary care patients 5 to 18 years old, residing in a 31-county region of Pennsylvania. Individual-level data were obtained from the clinic's electronic health record. A total of 33 environmental measures were derived to characterize land use (e.g., population density, intersection density, average block size), land use change (e.g., change in population and population density 1990 to 2000), physical activity (e.g., density and diversity of and accessibility to nine categories of physical activity establishments), and social environments (e.g., indexes of community socioeconomic deprivation and social disorganization). Associations of environmental measures with BMI z-scores were evaluated in various contexts and at a range of scales (e.g., 0.5 mi network buffers, census tracts, minor civil divisions [townships, boroughs, cities], a mixed definition of place [townships, boroughs, census tracts in cities], counties), first in all children and also in three age strata (5-9, 10-13, and 14-18 years), controlling for individual-level covariates. RESULTS: In all children, only lower community socioeconomic deprivation and higher diversity of physical activity establishments were associated with lower BMI. There were some differences in associations of environmental measures in the age strata, especially when comparing associations in the various scales and contexts. For example, higher population density was associated with lower BMI in older children, and this effect appeared to operate most clearly in the larger contexts. Similarly, lower county sprawl was also associated with lower BMI in older children. CONCLUSIONS: There was some evidence of effect modification by age. This suggests that no single intervention to change the built environment is likely to be effective across the childhood age range. The observation that surrogates for compact development may operate over larger contexts can be a challenge in many regions of the U.S. as zoning and planning are generally done on smaller scales.