Evaluation of a Food Hub Initiative’s Effect on Food Shoppers’ Perceptions, Shopping Behavior, Diet, and Weight in a Community of Low Income and Low Access to Healthy Food (OR16-06-19)

Abstract

Abstract Objectives The objective was to evaluate the impact of a Food Hub (FH) initiative on food shoppers’ shopping behavior, fruit and vegetable intake, diet quality, perceived local food environment, and body mass index (BMI) compared to a matched community's shoppers. Methods A community-inititated food hub (farmers’ market, grocery store, urban garden) was implemented to enhance healthy food access and community economic development using Healthy Food Financing Initiative and local funds. An external evaluation used a quasi-experimental design. Household food shoppers (n = 265 FH, n = 262 Comparison) completed interviews, including 24-hour dietary recalls, at baseline, post-test and delayed post-test (at 6 months’ and 18 months’ FH exposure, respectively). Community-level matching achieved comparable groups (not significantly different) on baseline dietary, anthropometric, and socioeconomic characteristics. Majority were women, overweight/obese, of low/very low household food security, low income, with high school education or less. There was a small race difference (African Americans = 89% FH, 96% Comparison, P = .01). We analyzed community by time changes using 2-level linear and logit growth models and estimated effect sizes (Cohen's d). Results While Comparison participants had better food environment perceptions at baseline, FH's perceptions improved significantly versus Comparison (b = 0.13, se (b) = 0.05, P = .01, d = .25). There was a small but significantly greater increase in fruit and vegetable intake among Comparison versus FH participants (b = −0.14, se (b) = 0.07 P = .04, Cohen's d = −.23) and very small, non-significant effects for Healthy Eating Index-2010 (d = −0.01), kcals/day (d = .03), BMI (d = .03), and weight (d = .02). Predicted probability of shopping at local farmers’ market across both communities combined increased from baseline to delayed post-test, but the change in probabilities over time did not differ between the two groups (b = 0.06, P = .77). Conclusions Despite FH participants’ comparatively positive improvement in their food environment perception, other outcomes did not change significantly relative to Comparison. Additional strategies to affect food shopping decisions and dietary intake merit consideration in addition to improved spatial access in disadvantaged communities. Funding Sources National Cancer Institute.