3.2. Availability of Smart Snacks-Compliant and Non-Compliant Beverages in Schools
Overall, middle school and high school students attended a school with a high rate of compliance with the Smart Snacks beverage standards, on average. At the middle school level, 84.54% of beverages sold were compliant (Table 1
) and 15.46% were non-compliant. At the high school level, 74.11% of the beverages sold were compliant (Table 1
), while 25.89% were non-compliant.
presents data on the weighted percentage of students located in middle and high schools with access to specific Smart Snacks-compliant and non-compliant beverages in à la carte and vending machine locations. Student exposure to compliant beverages was fairly comparable across school levels for à la carte settings with the exception of fat-free/skim flavored milk at breakfast and lunch, 100% fruit/vegetable juice at breakfast, bottled water at breakfast, and low-fat white milk at lunch (with student access higher at the high school levels in each of these instances). Student access to Smart Snacks-compliant beverages in vending machines was also greater at the high school level than at the middle school level, likely due to less availability of vending machines in middle schools (only 45.7% of middle school students had access to vending machines in school as compared to 88.2% of high school students). Non-compliant beverages available in middle and high schools primarily included juice drinks and other sweetened drinks, regular soda, sports drinks, coffees or teas, whole or reduced-fat white or flavored milk, and low-fat flavored milk.
3.4. Association between School Availability of Compliant Beverages and Student Consumption of Non-Compliant Beverages
and Figure 3
and Table A1
and Table A2
present the results of the multi-variable logistic regression models showing the association between the percent of beverages sold in schools that were Smart Snacks compliant and student consumption of non-compliant beverages while in school at the middle and high school levels, respectively, controlling for district policy, school, and student characteristics. (Descriptive statistics on characteristics of students who did and did not consume non-compliant beverages are shown in Table A3
There was not an association between beverage availability and student consumption at the middle school level (Figure 2
; Table A1
). Yet, in the adjusted models (Table A1
), middle school students had lower odds of consuming non-compliant beverages if they were enrolled in a school that had ≥50% Hispanic students (OR: 0.25, 95% CI: 0.08, 0.75) compared to a school that had ≥50% non-Hispanic white students and higher odds of consumption if they were enrolled in a school with a high compared to low FRPL eligibility rate (OR: 4.59, 95% CI: 1.85, 11.38).
At the high school level, however, there was a significant, inverse relationship between the percentage of Smart Snacks-compliant beverages sold in schools and student consumption of non-compliant beverages while in school (see Figure 3
and Table A1
and Table A2
). Specifically, the odds of students consuming Smart Snacks non-compliant beverages in school at the 10 and 60 kcal thresholds was significantly lower in schools with a higher percentage of Smart Snacks-compliant beverages (10 kcal threshold: OR: 0.97, 95% CI: 0.95, 0.99, p
< 0.01; 60 kcal threshold: OR: 0.98, 95% CI: 0.96, 1.00, p
< 0.05). Figure 3
illustrates the adjusted prevalence of high school student consumption of non-compliant beverages at the 10 kcal (panel a) and 60 kcal (panel b) thresholds as the percentage of compliant beverages available in schools increases from 0 to 100%. At the 10 kcal threshold (Figure 3
, panel a), if none of the beverages in high school were Smart Snacks compliant, the adjusted prevalence of student consumption of non-compliant beverages at school is estimated to be nearly 60%; whereas, if 100% of the beverages sold in high school were compliant, the estimated adjusted prevalence of student consumption of non-compliant beverages at school is only 10% (brought from home or from a friend perhaps). Similarly, at the 60 kcal threshold (Figure 3
, panel b), if none of the beverages sold in high school were Smart Snacks compliant, it is estimated that slightly less than 40% of students would consume non-compliant beverages at school; whereas, if 60% of the beverages sold in high school were compliant, slightly less than 20% of students would be estimated to consume non-compliant beverages at school.
The adjusted models (Table A1
and Table A2
) also revealed that high school students had lower odds of consuming non-compliant beverages at the 10 kcal threshold (i.e., diet beverages) if they were in a school with a diverse racial/ethnic mix (compared to majority non-Hispanic white) of students (OR: 0.33, 95% CI: 0.15, 0.73), from a medium-sized school (as compared to a large school) (OR: 0.27, 95% CI: 0.11, 0.65), or from a household with an income that was ≤130% of the federal poverty level (OR: 0.34, 95% CI: 0.14, 0.83) or at >130–185% of the federal poverty level (OR: 0.11, 95% CI: 0.03, 0.44) as compared to a household with an income >185% of the federal poverty level. Similar results were found for the 60 kcal (lower calorie) beverage threshold (albeit with slightly different coefficients and confidence intervals; see Table A2
), except that students enrolled in schools with high rates of FRPL eligibility had higher odds than high school students enrolled in schools with low rates of FRPL eligibility of consuming non-compliant beverages at the 60 kcal threshold (OR: 3.31, 95% CI: 1.20, 9.13).
In additional sensitivity analyses (not shown in tables), we added body mass index (BMI) status as a covariate, based on BMI-for-age percentiles computed based on age, gender, and measured height and weight with the following categories: underweight or healthy weight (<85th percentile); overweight (≥85th–<95th percentile); and obese (≥95th percentile). Results were similar to those in our primary models; for BMI status itself, we found that obese middle school students were less likely to consume non-compliant beverages than underweight or healthy weight middle school students (OR: 0.46, 95% CI: 0.24, 0.90), although no significant differences were found for overweight middle students compared to underweight or healthy weight middle school students or for either category relative to underweight or healthy weight high school students. Due to missing data, these analyses included 489 middle school and 478 high school students.