A network approach to understanding obesogenic environments for children in Pennsylvania

Abstract

Network methods have been applied to obesity to map connections between obesity-related genes, model biological feedback mechanisms and potential interventions, and to understand the spread of obesity through social networks. However, network methods have not been applied to understanding the obesogenic environment. Here, we created a network of 32 features of communities hypothesized to be related to obesity. Data from an existing study of determinants of obesity among 1,288 communities in Pennsylvania were used. Spearman correlation coefficients were used to describe the bivariate association between each pair of features. These correlations were used to create a network in which the nodes are community features and weighted edges are the strength of the correlations among those nodes. Modules of clustered features were identified using the walktrap method. This network was plotted, and then examined separately for communities stratified by quartiles of child obesity prevalence. We also examined the relationship between measures of network centrality and child obesity prevalence. The overall structure of the network suggests that environmental features geographically co-occur, and features of the environment that were more highly correlated with body mass index were more central to the network. Three clusters were identified: a crime-related cluster, a foodenvironment and land use-related cluster, and a physical activityrelated cluster. The structure of connections between features of the environment differed between communities with the highest and lowest burden of childhood obesity, and a higher degree of average correlation was observed in the heaviest communities. Network methods may help to explicate the concept of the obesogenic environment, and ultimately to illuminate features of the environment that may serve as levers of community-level intervention.

Leveraging Artificial Intelligence and Participatory Modeling to Support Paradigm Shifts in Public Health: An Application to Obesity and Evidence-Based Policymaking

“…Degree and betweenness are the two node centrality measures commonly applied in the analysis of obesity maps [35]. In particular, previous work found a correlation (R 2 = 0.51) between degree centrality and body mass index [77]. The results show that obesity is the most central, followed by themes pertaining primarily to mental well-being (e.g., weight bias, weight bullying, self-esteem, body image); the prominence of such constructs and the lack of precise physiological or environmental notions reinforce the findings from the previous two analyses.…”

“…From a small number of PM studies using reduced maps, it appears that the number of categories may scale with the number of factors. A map of the obesogenic environment for children in Pennsylvania had 3 clusters for 32 features [77], one on obesity in Manhattan's Chinatown had 4 themes for 39 factors [73], and another in South Africa had 5 domains for 47 factors [74]. In a different domain, a map for HIV in sub-Saharan Africa had 4 or 5 domains for 71 nodes.…”