Abstract
IMPORTANCE: The establishment and renovation of supermarkets may promote healthy diet practices among youth by increasing retail infrastructure for fresh foods. OBJECTIVE: To estimate the association between the Food Retail Expansion to Support Health (FRESH) program and the weight status of children and adolescents. DESIGN, SETTING, AND PARTICIPANTS: Using a difference-in-differences (DiD) design and including 12 months before and after a FRESH supermarket opened, data were analyzed for residentially stable public school students in kindergarten through 12th grade with objectively measured height and weight data from the academic years 2009 through 2016. Of the 8 FRESH-subsidized supermarkets in residential neighborhoods in New York City, New York, 5 were new and 3 were renovation projects between December 2011 and June 2014. Data were analyzed from June 2021 to January 2022. INTERVENTIONS: The treatment group included students who resided within 0.50 miles of a FRESH-subsidized supermarket and had at least 1 body mass index (BMI) measurement within 12 months before and 3 to 12 months after the month a FRESH supermarket opened (n = 22 712 student-year observations). A 2-stage matching-weighting approach was used to construct a control group of students who resided more than 0.50 miles from a FRESH supermarket in a FRESH-eligible area (n = 86 744 student-year observations). MAIN OUTCOMES AND MEASURES: BMI z score was calculated using objectively measured height and weight data from FITNESSGRAM, an annual, school-based, standardized fitness assessment of every New York City public school student. Obesity was defined as 95th percentile or greater of the BMI z score using Centers for Disease Control and Prevention growth charts. RESULTS: The treatment group in the analytic sample had 11 356 students (22 712 student-year observations), and the control group had 43 372 students (86 744 student-year observations). The students were predominately Black (18.8%) and Hispanic and Latino (68.5%) and eligible for free or reduced-priced lunch (84.6%). There was a significant decrease in BMI z score among students who resided within 0.50 miles of a FRESH supermarket (vs control group students) in the 3- to 12-month follow-up period (DiD, -0.04; 95% CI, -0.06 to -0.02). This was true for those exposed to supermarkets that were either new (DiD, -0.07; 95% CI, -0.11 to -0.03) or renovated (DiD, -0.03; 95% CI, -0.06 to -0.01). A statistically significant decrease was also observed in the likelihood of obesity (DiD, -0.01; 95% CI, -0.02 to -0.002). CONCLUSIONS AND RELEVANCE: Government-subsidized supermarkets may contribute to a small decrease in obesity risk among children residing near those supermarkets, if part of a comprehensive policy approach.