Article Text
Abstract
Background Structural interventions to improve access to healthy food in deprived communities are a current policy priority. In 2010, the Obama administration unveiled the Healthy Food Financing Initiative to promote interventions that expand access to nutritious foods. The initiative provided more than $400 million in funding to bring grocery stores and health food retailers to deprived communities. However, no evidence for the effectiveness of such policy interventions exists. This paper is a first attempt to evaluate the effect of this intervention on perceptions of the food environment, fruit and vegetable intake and body mass index (BMI) in Philadelphia, USA.
Methods Data from the Philadelphia Neighbourhood Food Environment Study were used. This study utilised a prospective quasi-experimental design comparing baseline and follow up data in an ‘intervention’ community with a matched ‘comparison’ community. The intervention was the introduction of a large food supermarket in an underserved area. Outcomes were BMI, weekly fruit and vegetable consumption (Block FFQ), and perception of the neighbourhood food environment. Data were analysed using an intention-to-treat (ITT) and adopters vs non-adopters (ANA) approach. Difference-in-differences analyses (DID) assessed whether significantly different changes over time existed between intervention and comparison groups. Models were adjusted for age, sex, income, race, education, household composition and employment status.
Results In ITT and ANA analyses no significant difference-in-differences in mean BMI and mean weekly fruit and vegetable consumption were found. ITT analyses demonstrated a significant improvement in perceptions of the neighbourhood food environment in the intervention versus comparison neighbourhood (DID=1.71, p=0.001). Using ANA, a similar result was found for adopters, with those who used the grocery store as their primary shop (DID=1.57, p=0.05) and those who used it for any food shopping (DID=2.34, p<0.001) having significantly better perceptions of the neighbourhood food environment compared to non-adopters. After adjustment, ITT analyses remained significant (DID=1.65, p=0.002). For ANA, significant DID results were attenuated for primary adopters (DID=1.30, p=1.105) but remained significant for those undertaking any food shopping (DID=2.18, p<0.001).
Conclusion The introduction of a new grocery store had a positive effect on how local residents perceive the quality of their neighbourhood food environment. However, this did not have a significant effect on fruit and vegetable consumption or BMI. Improvement in perceptions of neighbourhood food environment quality did not translate into improved dietary behaviour or reductions in BMI.