Examining the non-linear association between the built environment around residence and the risk of obesity among children and adolescents at different school stages in Guangzhou and Shenzhen, China

Abstract

Recent evidence suggests that the built environment around residence is linked to the risk of obesity among children and adolescents. However, most previous studies focus on the linear relationships between the built environment around home and the risk of obesity among adults, failing to explore potential non-linear relationships among children and adolescents at different school stages. Here, we use XGBoost to investigate the non-linear relationships between neighborhood built environment factors and the risk of obesity among 41,220 children and adolescents in Guangzhou and Shenzhen, two metropolitan cities in South China. Results from XGBoost indicate that the built environment plays an important role in predicting the risk of obesity among primary school students, middle school students, and high school students, and the relative importance of the built environment increases with age. Obesogenic food environment and population density are found to be leading predictors for all three school stages. Subway accessibility is an important predictor for primary school students, while bus accessibility is an important predictor for secondary school students. Interestingly, proximity to greenspaces is not important in predicting obesity risk. Non-linear relationships are found between several built environment predictors and obesity risks. Our findings may inform policymakers, planners, and designers on how to create an anti-obesogenic environment for children and adolescents in high-density Chinese cities.