Accelerated Summer Weight Gain in a Low-Income, Ethnically Diverse Sample of Elementary School Children in Massachusetts

Background: Several studies have found that children in the United States gain weight at a faster rate in the summer than in the school year, but little is known about the prevalence of this problem, its effect on high-risk subgroups, or its determinants. This study compares school year and summer weight change in a low-income, ethnically diverse sample of school-age children in Massachusetts and explores differences by race/ethnicity, weight status, and exposure to school year physical activity (PA) programming. Diet and PA are examined as potential mediators of summer weight gain. Methods: Children participating in a school-based PA program evaluation (in which weight change was not a primary outcome) had their height and weight measured three times between October 2015 and September 2016 to capture a school year and summer interval. Diet and PA patterns were assessed mid-school year and mid-summer in a subsample of children. Mixed linear models were used to estimate the effect of season (school year vs. summer) on change in BMI and to examine the influence of race/ethnicity, weight status, and program (walk/run, classroom activity breaks, or control) on any observed effects. Structural equation models were used to explore diet and PA as mediators of seasonal weight change in a subsample of participants. Results: Of 769 participants, 53% were non-Caucasian, 40% were overweight or obese, and 58% were eligible for free or reduced-price school meals. BMI increased in both the school year and summer but increased more rapidly in the summer (0.046 kg/m 2 more per month, p  = 0.007). Of the three tested interactions, statistical significance was only observed between season and program (χ 2  = 14.90, p  < 0.001); on average, children exposed to a school year walk/run program did not gain weight more rapidly during the summer, whereas children in the control group and a classroom activity breaks program did. Poorer diet and PA patterns were observed in the summer, but neither diet nor PA was statistically significant mediators of BMI change. Conclusions: Children in this high-risk sample gained weight at a faster rate during the summer than during the school year, with no discernable demographic differences. However, this phenomenon was not observed in the subgroup exposed to a school year walk/run program. More research is needed to clarify the determinants of summer weight gain and understand how school year programming and its effects can be transferred to the summer months.