BACKGROUND: A number of child/parental factors have been shown to be significant predictors of childhood overweight, although a better understanding of possible contextual influences of neighbourhood-level characteristics might provide new insights leading to tailored, targeted interventions. The aim of this study was to explore the impact of neighbourhood purchasing power and its relationship with other known risk factors related to childhood overweight in a prospective birth cohort.

METHODS: A prospective, population-based, birth-cohort study was conducted in south-western Sweden, comprising 2,666 infants born in 2007-2008. Childhood overweight was assessed by body mass index (BMI) data from follow-up examinations at four years of age (n = 2,026) and overweight defined according to the International Obesity Task Force. Using logistic regression analysis, the influential child/parental predictors were identified from the candidate predictors, viz. child's gender, as well as birth weight adjusted for gestational age and parental factors at recruitment, including maternal smoking status, maternal BMI (before pregnancy), paternal BMI and parental educational level. The children's residential parishes at follow-up were stratified by parish-level household purchasing power (<10 %, 10-19.9 %, 20-29.9 % and ≥30 % of all resident families with low purchasing power) and the "contextual" influence was analysed. In each such neighbourhood stratum, the adjusted overweight ratio (AOR), i.e. the ratio between the observed number of overweight children and the expected number, taking account of the influential child/parental predictors, was estimated.

RESULTS: The prevalence of overweight at four years of age was 11.9 %. In the economically strongest neighbourhoods (i.e. <10 % of resident families with low purchasing power), the AOR was 0.60 (95 % confidence interval (CI): 0.34-0.98). The corresponding empirically Bayes-adjusted AOR was 0.73 (95 % CI: 0.46-1.02; 97 % posterior probability of AOR <1). In the other neighbourhood strata, the statistical evidence of a deviant AOR was weaker.

CONCLUSION: The economically strongest neighbourhoods had a lower prevalence than expected of overweight at four years of age. This finding should prompt studies to acquire more knowledge of potentially modifiable factors that differ between neighbourhoods and are related to childhood overweight, providing a basis for tailored, targeted interventions.