Farmers' markets (FMs) present opportunities for health promotion. In New York City (NYC), the Just Say Yes to Fruits and Vegetables Stellar FM Program (Stellar) provides free nutrition and cooking demonstrations for adults from July-November at select FMs coupled with distribution of one $2 Health Bucks (HB) coupon, redeemable toward fresh produce at all NYC FMs.
Evaluate the relative impact of Stellar and HB distribution on initial and longer-term changes in produce consumption.
Study Design, Setting, Participants
This quasi-experimental, longitudinal evaluation recruited the following groups from NYC FMs located in neighborhoods with low incomes from July-August 2016: Stellar (n = 110): Stellar attendees from 2 FMs; HB (n = 330): shoppers at 2 non-Stellar FMs who received $6 HB at recruitment, and $6 HB/month for 3 months; Controls (n = 325): shoppers at 2 other non-Stellar FMs. Baseline surveys were conducted at recruitment. Phone/mail follow-up surveys were conducted from August-September 2016 (Wave 2) and February-April 2017 (Wave 3). Participant retention at Waves 2 and 3, respectively, were: Stellar 67%/49%; HB 37%/33%; Controls 38%/37%.
Daily fruit and vegetable (FV) consumption were measured using the NHANES Dietary Screener Questionnaire. Linear regression models accounting for participant-level clustering and fixed effects for wave, group, wave*group, education, race/ethnicity, and income-assessed adjusted average intake and within-group changes from baseline (α = 0.05).
Baseline total FV intake averaged 2.81 times/day (Controls), 2.80 (HB), and 2.63 (Stellar). At Wave 2, all groups had non-significant increases (Controls: 0.04, P = .734; HB: 0.22, P = .088; Stellar: 0.28, P = .114). At Wave 3, Controls and HB had non-significant changes from baseline by −0.22 (P = .131) and −0.27 (P = .054), respectively, while Stellar significantly increased by 0.41 times/day (P = .041).
These findings indicate that Stellar programming, which includes HB, may support sustained increased FV consumption more than HB distribution alone. This study is limited by unexpectedly small sample sizes and seasonal variation between survey waves.