Abstract
BACKGROUND: Limited access to healthy foods in low-income, racial and ethnic minority communities has been proposed as a critical factor contributing to health disparities. One policy option for improving access is to introduce supermarkets in low-income communities, but evidence increasingly points to null effects. OBJECTIVE: The aim of this study was to examine whether the introduction of a new supermarket in a public housing community, and proximity to it, were associated with improvements in residents' dietary outcomes. It also explored whether these associations may be moderated by access and cost barriers to eating healthy at baseline. DESIGN: A cohort of public housing residents were recruited from Jordan Downs, where the new supermarket was introduced, and from 2 comparison sites. Longitudinal data on outcomes and exposures were collected at baseline (2018-2019) and follow-up (2020-2021). Quasi-experimental variation in residents' distance to the new supermarket was used to examine whether proximity to the supermarket was associated with dietary improvements. PARTICIPANTS/SETTING: Participants included 557 adult residents from 3 public housing sites in Watts, Los Angeles, CA. INTERVENTION: The intervention is the opening of a new supermarket in the Jordan Downs community. The primary exposure measure is an indicator for whether residents lived in Jordan Downs vs in the comparison sites. The secondary exposure measure is residents' driving distance to the new supermarket. MAIN OUTCOME MEASURES: Self-reports of daily consumption of whole grains, dairy, fruits and vegetables, added sugar, red meat, processed meat, and overall healthiness of diet were obtained from surveys administered at baseline and follow-up. STATISTICAL ANALYSES PERFORMED: Descriptive analyses examined households' grocery shopping patterns after the new supermarket's introduction. Analysis of covariance linear regressions models estimated the association between residents' dietary outcomes at follow-up and the exposure measures, conditional on dietary outcomes at baseline and other covariates. RESULTS: Proximity to the new supermarket was associated with a significantly higher probability of shopping there, particularly for households that reported an access or cost barrier at baseline. Overall, there were no statistically significant or practically meaningful differences observed in dietary outcomes between Jordan Downs residents and those in comparison sites, or between residents who lived at varying distances from the supermarket. Exploratory moderation analyses suggested some meaningful dietary benefits for those who had a frequent access barrier at baseline, but not for those without such a barrier, but there was mostly no evidence of moderation by price barrier at baseline. CONCLUSIONS: Supermarket opening in an urban, low-income, and racial and ethnic minority community was not found to be associated with improvements in dietary outcomes for most residents in the first year after its opening. Improving dietary outcomes in such communities may require a broader and nuanced approach that addresses varied barriers faced by residents.