Area-level socioeconomic status and children’s body weight: examining the role of unhealthy food environments

BACKGROUND: Socioeconomic status (SES) is associated with inequities in children's body weight. Yet little is known about whether features of the urban food environment modify this association. We assessed whether and to what extent the availability of unhealthy food retailers modifies the association between area-level SES and weight status among children, overall and by sex. METHODS: We conducted an ecological study using individual data from a city-wide representative survey of 3505 children, aged 3 to 12 years living in Madrid, Spain. We measured area-level SES at the census-section level using a composite measure and categorized it into tertiles, with T3 representing the highest area-level SES. Our outcome variables were overweight and obesity, measured using anthropometric data on height and weight. As a potential effect modifier, we examined the availability of unhealthy food retailers around children's households. Using Geographic Information Systems, we calculated 400-m street-network buffers and categorized unhealthy retailers' availability into tertiles. We used Poisson regression models with robust standard errors to estimate prevalence ratios (PR) and 95% confidence intervals (CI), adjusting for age, sex, and population density. RESULTS: Overall, the prevalence estimates for overweight and obesity were 26.2% and 16.6%, respectively. Compared with children living in high-SES areas (T3), those living in the lowest-SES areas (T1) showed a higher prevalence of overweight (PR = 1.31, 95% CI: 1.13, 1.51) and obesity (PR = 2.30, 95% CI: 1.75, 3.03). Among children with the lowest exposure to unhealthy food retailers, overweight prevalence was higher for children in low-SES areas than in high-SES areas (PR = 1.44; 95% CI: 1.06, 1.96). For obesity, the prevalence was also higher among children living in low- (T1; PR = 3.96, 95% CI: 2.36, 6.65) and medium-SES areas (T2; PR = 2.40; 95% CI: 1.42, 4.07) relative to those living in high-SES areas (T3). We found a statistically significant interaction between area-level SES and unhealthy food retailer availability for obesity on both additive and multiplicative scales. CONCLUSION: Our findings suggest that the urban food environment modifies the association between area-level SES and childhood obesity, whereas the evidence was less consistent for overweight. The persistence of higher obesity prevalence in lower-SES areas, even where exposure to unhealthy food retailers was lower, indicates that food environment interventions may not be sufficient to reduce socioeconomic inequities unless complemented by other structural approaches that address social disadvantage.