County level estimates of childhood obesity using the 2020 and 2021 National Survey of Children’s Health

Obesity is common, serious, costly and varies significantly across geographic regions in the US. Local data is needed to inform program interventions but traditional survey methods for obtaining local estimates of obesity nationwide are too costly. We estimated the prevalence of childhood (10 – 17 years) obesity prevalence for every U.S. county using previously developed methods and the 2020-2021 National Survey of Children’s Health (NSCH). We first constructed a multilevel mixed-effects logistic regression model to evaluate the influence of child demographic characteristics and area level characteristics (block group, county, and state) on childhood obesity. We then estimated the obesity risk for a child in each census block group based on this multilevel model and obtained county-level obesity estimates using a post stratification approach. No surveillance system provides direct estimates of childhood obesity for each county in the US. In response to a recent congressional inquiry, we used population level statistics generated to identify areas of high need and where grantees from CDC’s High Obesity Program may want to collaborate with local and state practitioners to implement programs to reduce obesity.