Abstract
Childhood obesity remains a public health crisis because of its alarming prevalence and potential for costly long-term health consequences, especially among rural children. Schools are considered natural loci for policies to combat obesity because children spend most of their active hours and consume a substantial share of their calories, at school. Recent state policy efforts have involved measuring children's BMI and/or fitness to notify parents or inform surveillance efforts, but the empirical evidence to date is far from definitive. This study leverages plausibly exogenous assignment of military families as a natural experiment to assess the association of such polices with children's BMI and obesogenic behaviors. The sample is stratified by urbanicity because of the likely differences in obesity prevalence and in environments necessary to support healthy lifestyles. Data were collected in 2013-2014 and analyzed in 2018-9. The policies were associated with lower odds of overweight (OR: 0.422; CI: 0.251-0.708) and at-risk of overweight (OR: 0.360; CI: 0.161-0.801) among children in rural, but not urban, schools. The policies were also associated with greater activity levels and less frequent intake of unhealthy foods, particularly among children attending rural schools. Results were robust to stratification based on urbanicity of the installation versus school, omission of family covariates, and other sensitivity analyses. Falsification checks on children's height, parental BMI outcomes, home food environments and neighborhood environments indicate findings are not driven by selection. BMI assessment policies were associated with children's BMI outcomes and health behaviors in rural, but not urban schools.