Measuring the obesogenic food environment
Embargo until
Date
2015-03-12
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
Given the concern about persistent high rates of overweight and obesity, interest in the local food environment has increased exponentially. This dissertation’s three aims explore different dimensions of the local food environment, specifically, associations between the food environment and obesity, the consequences of food environment measurement heterogeneity, and the change in the food environment over time.
For the first aim, we conducted a systematic review of the association between local food environments and obesity. We included 71 studies and found that overall associations between food outlet availability (e.g. supermarkets, fast food restaurants) were predominantly null, although we saw a trend toward inverse associations between supermarket availability and obesity in adults (22 negative, 4 positive, 67 null) as well as direct associations between fast food restaurant availability and obesity in adults (29 positive, 6 negative, 71 null) and in low income children (12 positive, 7 null). Limiting analyses to the few available higher quality studies did not affect results.
The second aim used a validated dataset of food stores and restaurants in Baltimore City, Maryland to evaluate the impact of different measurement strategies on the identification of both high-risk food environments and food environment disparities. We varied food environment measures by (1) whether they captured healthy food availability or unhealthy food availability, (2) geographic measure such as outlet density or proximity to the nearest outlet, and (3) data source. Overall, we found substantial heterogeneity between different types of measures. Further, we found socio-economic disparities in the availability of unhealthy food, but not in the availability of healthy food. Heterogeneity across geospatial measures impacted the associations of neighborhood income and racial composition with both supermarket and fast food availability.
The third aim assessed whether healthy food availability changed between 2006 and 2012 in a sample of 118 food stores in Baltimore City and examined potential associations of change in healthy food availability with neighborhood socioeconomic status, racial composition or a 2009 policy change to the Special Supplemental Nutrition Program for Women, Infants and Children (WIC). Overall healthy food availability scores increased by a significant 1.25 points, roughly equivalent to having a store sell one additional type of healthy food. Change in healthy food availability was greater in corner stores compared to convenience stores and in census tracts with >60% black residents compared to those with >50% white residents. Results were suggestive of an impact of the WIC program policy change as we saw a significant increase in measured WIC-approved foods and WIC-authorized stores had a greater increase in milk availability than stores not participating in the WIC program. However, there was no overall difference in the change in healthy food availability by WIC authorization.
In conclusion, we found encouraging signs that the food environment in Baltimore City is improving. Further research including rigorous longitudinal designs that control for neighborhood selection and that use standardized food environment measures would provide valuable evidence in support of a causal relationship between the food environment and obesity. In the absence of this, policy makers will need to rely on other types of evidence to support further action to improve food environments.
Description
Keywords
Food environment, obesity