Addressing Food Insecurity: An Evaluation of Factors Associated with Reach of a School-Based Summer Meals Program

Background One in six US households with children experiences food insecurity, with higher rates in the summer. Approximately 3 million children receive free meals each summer weekday through the US Department of Agriculture's Summer Nutrition Programs. However, participation in these programs has been declining in recent years and is lower than participation in programs that serve free or reduced-price meals during the school year. Objective To identify school and site characteristics associated with greater reach by school-based free summer meals program sites. Design This observational study combined program data, public school data, and Google Maps data to determine factors associated with site reach. Participants/setting LunchStop Summer Meals Program sites (N=100) and schools at which they were based during summer 2018 in Chicago, IL. Main outcome measures Reach of sites was measured by the mean daily meals served at each site throughout summer 2018. Statistical analyses performed Associations between site reach and each independent variable were evaluated using Mann-Whitney tests and simple linear regressions. Variables significantly associated with site reach in bivariate analyses (P<0.05) were included in a multivariate linear regression. Results In bivariate analyses, sites with greater reach were significantly more likely to be based at schools with higher attendance percentages, higher percentages of Hispanic/Latino students, larger student populations, and locations in a network of southwest Chicago schools. Those with greater reach were also significantly more likely to have continuity in program staff and more years of site operation (all P values <0.01). The last four factors remained significant in multivariate analysis. Conclusions School-based summer meals programs may be able to reach more vulnerable children by taking into account continuity among sites and staff. Further research is needed to clarify whether the relationships between these variables and increased reach is causal.