Measurement of the relationship between children's body mass index and food environments: A feasibility pilot study

BACKGROUND/OBJECTIVE: While the obesity rates rise, children's environmental experience includes an overabundance of high-calorie foods. This pilot study (1) examined the relationship between children's body mass index (BMI) z-scores and food store, restaurant, and home food environments; (2) assessed the feasibility of related measures for use in a larger study. METHODS: Home visits were conducted with 11 to 12 year-old children that attended the same school in a suburban area (pilot N=12). BMI z-score was calculated as [BMI=weight/height²] and converted to sex and age-appropriate z-scores. The Nutrition Environment Measurement Survey (NEMS) was used to measure food availability and facilitators/barriers to healthful eating in two most frequented food stores and restaurants identified by each participant. The NEMS was also adapted to measure the home food environment. In addition to the observational measures, home food availability (fruits/vegetables, grains, meat, dairy, sweet snacks, salty snacks), home food accessibility (fruits/vegetables), and food purchasing questionnaires were administered to parents. RESULTS: The average BMI z-score in the sample was 0.06±0.29 (range=-1.44-1.87, median=0.275); one child was considered overweight (≥95th percentile of sex and age-adjusted BMI distribution). Variability in store NEMS scores was limited (range=0.62-0.72, mean=0.68±0.03, median=0.70). Not including school meals, away-from-home meals averaged 5.42±5.52/month (range=0-20, median=3). BMI z-scores and NEMS scores were not associated (store r=-0.01, p=0.97; restaurant r=0.04, p=0.91; home r=-0.01, p=0.98). BMI z-scores and parent-reported home food availability trended in predicted directions for the following food categories: fruits/vegetables (ρ=-0.49, p=0.11), dairy (ρ=-0.42, p=0.18), sweets (ρ=0.41, p=0.19). Accessibility of fruits/vegetables was inversely associated with BMI z-scores (ρ=-0.60, p=0.04). CONCLUSION: Families tended to purchases foods from large chain supermarkets and ate out rather infrequently. While the NEMS tool has been validated for comparing food environments at the community level, it may not be relevant in studying BMI z-score variation between children with homogenous store environments and low restaurant exposure. Though requiring further validation (considering the low pilot N), the bivariate correlations between BMI z-scores and parent-reported home food availability and accessibility are higher than those previously reported.