Impact of the Food-Labeling and Advertising Law Banning Competitive Food and Beverages in Chilean Public Schools, 2014–2016

Abstract

Objectives. To evaluate the impact of a national law banning sales of competitive food and beverages (CF&B) in schools on the availability of CF&B sold at school kiosks.

Methods. This study was uncontrolled before and after study. We evaluated public schools in Santiago de Chile (n = 21; 78% response rate) in 2014 and 2016 (6 months after the law came into force). Trained personnel collected data on calories, total sugars, saturated fat, and sodium from food labels. The outcome was the percentage of foods exceeding the cutoff levels defined in the law and the mean difference between 2014 and 2016.

Results. Foods exceeding any cutoffs decreased from 90.4% in 2014 to 15.0% in 2016. Solid products had a substantial reduction in calories, sugar, saturated fat, and sodium. Liquid products had a reduction in calories, total sugar, and saturated fat, whereas sodium increased. This was a result of changes in product mix.

Conclusions. A ban on sales of CF&B reduced the availability of CF&B at Santiago’s school kiosks. Further research should examine the impact of this ban on food intake and health outcomes.

Childhood obesity levels are rising worldwide and have become a major public health issue, with negative health, social, and economic impacts.¹ About 337 million children and adolescents are overweight or obese.¹ Obese children have an increased risk of type 2 diabetes, cardiovascular diseases, and cancer; shorter life expectancy; and lower quality of life.^2–4 Obesity is also associated with lower self-esteem, higher risk of bullying, and lower academic achievement.^4,5

School interventions are among the most cost-effective interventions after food advertising regulation and fiscal measures.⁶ Evidence from randomized controlled trials has shown a modest decrease in body mass index, adiposity, and blood pressure.⁷ School food environments have thus become a focal point for actions aimed at reducing childhood obesity.

Competitive foods and beverages (CF&B)—food and beverages high in calories, sugar, and saturated fats offered in competition with government school meals—have received increasing attention as the subject of school food policies.^8,9 First, CF&B interventions have shown positive effects on school food environments, reducing on-campus availability, access, and, consequently, purchase and consumption of unhealthy foods and beverages; nevertheless results on overall intake have been inconclusive.^9–11 Furthermore, CF&B policies have been shown to improve nutritional status outcomes. Studies in the United States have reported stabilization of overweight prevalence trends after CF&B policies were implemented, with greater effects in socioeconomically advantaged communities and states where policies included stronger laws.^12,13 Similar initiatives that reduce children’s exposure to unhealthy foods during school hours have been implemented in some US states and some European countries.^14,15 Bans on sales of CF&B in schools have also been reported in Mexico, France, and Canada.^16–18

Despite wide implementation, the impact of CF&B policies has been evaluated only in some states of the United States and France.^10,11 These policies have been on the basis of establishing nutritional standards or introducing partial bans on the sales of sugar-sweetened beverages (SSB) and vending machines.^13,15 Evidence of a total ban of sales of CF&B is still lacking.

Chile has one of the highest rates of obesity in the Americas.¹⁹ Prevalence of obesity among children aged 6 years has doubled from 12.0% in 1993 to 23.2% in 2010.²⁰ In 2015, 26.9% of Chilean children aged 6 years were overweight and 24.2% obese; among adolescents aged 14 years, 32% were overweight and 12.5% obese.²¹

As a response to the obesity epidemic, Chile enacted and implemented the Food-Labeling and Advertising Law (FLAL) in 2016. The FLAL (Law 20 606/2012) was enacted in 2012 and entered into force on June 26, 2016. The law includes 4 components: (1) a compulsory front-of-package warning sign for packaged foods that exceed the established cutoffs, (2) a total ban on sales in schools of all packaged foods exceeding any cutoff, (3) marketing restrictions for these products, and (4) nutritional education. The norm established cutoffs for calories, total sugar, saturated fat, and sodium in packaged foods, with different levels for solid and liquid products.²² An expert committee appointed by the Ministry of Health proposed the cutoffs.²³ Following the release of the proposal from the expert committee, the food industry exerted great pressure, resulting in a final norm that included gradually decreasing cutoffs until 2019. Table A (available as a supplement to the online version of this article at http://www.ajph.org) shows the cutoffs at the beginning of implementation (June 2016) and the final cutoffs (starting in June 2019).

We evaluated the impact of a total ban on CF&Bs on the availability of foods and beverages sold at public school kiosks in Santiago, Chile, between 2014 and 6 months after the law came into force in December 2016. We examined the effects of the policy using both the gradual and final cutoffs.

METHODS

The study took place in the Municipality of Santiago, Chile—the capital municipality of Chile and one of the 52 municipalities in the metropolitan region. With approximately 400 000 residents, it has a large transient population of 1.8 million people. The Municipality of Santiago is responsible for the administration of 44 public schools, including 23 primary schools (first to eighth grade in the Chilean school system), 16 secondary schools, 1 school for children with disabilities, and 1 school inside a jail. Altogether, the schools have close to 30 000 students.

Food sales in Chilean schools can happen both in school canteens and school kiosks. Both are regulated by the sanitary code and are subject to regulation by the Ministry of Health. In the specific case of public schools in Santiago, sales occur only in school kiosks.

In addition to national efforts to improve school environments, the Municipality of Santiago has implemented an intersectoral health promotion program using the “health in all policies” approach. A healthy kiosk initiative started in 2014, providing FLAL training to kiosk administrators and improving the contracts between kiosk administrators and the municipality. In 2015, a voluntary program enrolled 10 kiosks in 10 schools to reach a 30% availability of healthy product, as preparation for the implementation of the FLAL at the national level in 2016.

Study Design and Data Sources

Our study had a quasiexperimental design with an uncontrolled before and after setup of the data collection. All public schools that have school kiosks and are administered by the Municipality of Santiago were eligible to participate in the study.

We collected data in December 2014 and December 2016 using similar data collection procedures each time. In 2014, 2 trained dietitians from the Municipality of Santiago conducted audits in the school kiosks. In 2016, 3 trained professionals collected the data using the same procedure as in 2014. At both time points, audits were unannounced to prevent the school kiosk administrator from modifying the availability of products. All food products available for sale were audited using either photographs or voice records.

We considered food products unique if they differed in brand, flavors, or portion sizes. We extracted nutritional information on calories, total sugars, saturated fat, and sodium from packaged products from nutritional labels. These labels are mandatory by law and are expressed per portions and 100 grams or milliliters. We used the information per 100 grams or milliliters to be able to compare food composition. We recorded calories in kilocalories, saturated fat and total sugars in grams or milliliters, and sodium in milligrams. Food products with less than 1 gram or milliliter of total fat per 100 grams or milliliters are not required to report saturated fats; we used total fat instead in those cases. We also classified packaged products into food categories. We took these categories from the Chilean Sanitary Code²⁴ and included 12 on the basis of their availability in school kiosks.

We also registered and classified nonpackaged food items, such as fruits, sandwiches, nuts, and dried fruits, into their corresponding food categories. However, nutritional composition was not available because food labeling is not mandatory for nonpackaged items. We decided not to impute nonpackaged products, as there is no official (and up-to-date) food composition database currently available in Chile—the most recent available source is from 1990.²⁵ As a result, we included only nonpackaged products in the analysis by food category.

Outcomes

The primary outcome was the proportion of CF&B available for sale in 2014 and 2016. We defined CF&B as any food product with levels of calories, total sugar, saturated fat, or sodium above the cutoffs established in the FLAL. As the main interest was in the availability at the school level, we included all food products available in the analysis, even if the same ones were sold at several kiosks.

Statistical Analysis

Frequencies and means are reported for all products examined. We used both gradual and final cutoffs to understand the impact of a gradual policy in changing school availability of CF&B. We analyzed solid and liquid food products separately. Because nearly 90% of the food products available in 2014 and in 2016 were not the same ones, we treated data as a repeated cross-sectional design. To account for clustering in schools, we carried out a 3-level multilevel analysis with food products (level 1) nested in school cohort (level 2) in school (level 3). The model contained a fixed effect variable for year (2014 or 2016). For continuous variables (Table 1), we fitted multilevel mixed-effects linear regression and used marginal effects to obtain predicted means and mean difference between 2014 and 2016. For dichotomous variables, we fitted a multilevel mixed-effects logistic regression and used marginal effects to obtain predicted percentages and 95% confidence intervals (CIs) in 2014 and 2016. We used quantile–quantile plots to check whether random effects were normally distributed.²⁶ We performed sensitivity analyses, eliminating products available at both time points and using only unique products. Neither of the analyses showed relevant differences with the main analyses. We used Stata/SE 14.2 (StataCorp LP, College Station, TX) in all analyses.

TABLE 1—

Changes in Nutritional Composition of Food Products High in Calories, Saturated Fat, Sugar, and Sodium Per Kiosks in 21 Public Schools: Santiago de Chile, 2014–2016

Food Product	2014, No. or Mean (SE)	2016, No. or Mean (SE)	Difference (95% CI)
Solid
No. of products	976	579	−397
Calories per product	443.4 (5.7)	265.8 (6.5)	−177.6 (−161.6, −193.7)
Total sugar per product	35 (1.2)	13.7 (1.3)	−21.3 (−17.9, −24.8)
Saturated fat per product	7.8 (0.2)	2.2 (0.3)	−5.6 (−4.9, −6.2)
Sodium per product	310.6 (14.1)	137.8 (17.1)	−172.8 (−132.1, −213.5)
Liquid
No. of products	441	691	+138
Calories per product	40.0 (1.6)	17.2 (1.4)	−22.8 (−18.7, −26.9)
Total sugar per product	9.1 (0.3)	3.6 (0.3)	−5.5 (−4.6, −6.4)
Saturated fat per product	0.2 (0.1)	0.05 (0.1)	−0.1 (−0.02, −0.2)
Sodium per product	14.5 (1.0)	19.9 (0.9)	+5.3 (2.8, 7.9)

Note. CI = confidence interval. The analyses include only packaged items. Calories are expressed as kcal per product per 100 g or mL. Total sugar and saturated fat are expressed as g per product per 100 g or mL. Sodium is expressed as mL per product per 100 g or mL. Means are marginal averages from mixed-effects multilevel regression.

RESULTS

Of 44 public schools, 12 did not meet the inclusion criteria in 2014 and 17 in 2016. We did not evaluate 5 schools in 2014 and 2 schools in 2016 because the kiosk was closed at the time of the visit or the administrator refused to include it (participation rates were 84.4% and 92.6% in 2014 and 2016, respectively). There were comparable data from 2014 and 2016 on 21 schools and 24 school kiosks; 1 school had 4 kiosks (Figure 1).

FIGURE 1—

Flowchart of Selection of School Kiosks: Santiago de Chile, 2014–2016

Figure 2 compares changes in the proportion of CF&B products meeting the standards established by law using both the gradual and final cutoffs. Using the gradual cutoffs, availability of food products (both solid and liquids) exceeding any cutoff decreased from 90.4% in 2014 to 15.0% in 2016. For solid products, availability of products exceeding any cutoffs went from 97% in 2014 to 14% in 2016, whereas liquid products went from 75.8% in 2014 to 14.3% in 2016 (Figure A [available as a supplement to the online version of this article at http://www.ajph.org]). Using the stricter final cutoffs, CF&B available in school kiosks that did not meet the standards decreased from 91.4% to 50.0% between 2014 and 2016. Availability of solid and liquid products experienced a similar reduction as in the gradual cutoffs, but the magnitude of change was much smaller.

FIGURE 2—

Percentage of Competitive Food and Beverages Available According to (a) Gradual Cutoffs and (b) Final Cutoffs: Santiago de Chile, 2014–2016

Note. Percentages are marginal proportion from mixed-effects multilevel regression. The analyses include only packaged items.

Table 1 shows the mean differences in calories, sugar, saturated fat, and sodium between 2014 and 2016. Solid products experienced a significant reduction in calories (–178 kcal/100 g), sugar (–21 g/100 g), saturated fat (–5.6 g/100 g), and sodium (–173 g/100 g). In liquid products, reductions were observed in calories (–23 kcal/100 mL), total sugar (–6 g/100 mL), and saturated fat (–0.1 g/100 mL), whereas there was an increase in sodium (+5 mg/100 mL). The total number of solid products available decreased, whereas the availability of liquid products increased between 2014 and 2016.

Table 2 shows the changes in availability by food categories. Overall, a 5.1% reduction in the total number of items available was observed. In 2014, bakery products, candy, and drinks were the most common food categories and accounted for 87.3% of food products. In 2016, the availability of bakery products decreased substantially, whereas drinks, cereals, ice cream, and milk and dairy products were available more frequently. The proportion of fruits and vegetables available increased from 0.7% to 3.2% between 2014 and 2016. A small proportion of items were available at both time points, reflecting a change in the mix of products rather than a reformulation of the same product. More information on the reduction in mean calories by food category can be found in Table B ([available as a supplement to the online version of this article at http://www.ajph.org]).

TABLE 2—

Availability of Food Products in 2014 and 2016 by Food Category in 21 Public Schools: Santiago de Chile

Category	Only Available in 2014, No. (%)	Only Available in 2016, No. (%)	Available in Both 2014 and 2016, No. (%)
Bakery products	532 (38.5)	61 (4.7)	8 (6.9)
Canned	0 (0.0)	1 (0.1)	0 (0.0)
Cereal	39 (2.8)	122 (9.3)	17 (14.6)
Candy	339 (24.5)	267 (20.4)	22 (19.0)
Drinks	336 (24.3)	537 (41.1)	56 (48.2)
Eggs	1 (0.1)	4 (0.33)	0 (0.0)
Fruits and vegetables	11 (0.8)	45 (3.5)	0 (0.0)
Ice cream	25 (1.8)	102 (7.8)	4 (3.5)
Meals and prepared dishes	74 (5.4)	52 (3.9)	1 (0.9)
Milk and dairy products	24 (1.7)	71 (5.4)	5 (4.3)
Stimulants and fruitives	1 (0.1)	14 (1.2)	1 (0.9)
Sugars and honey	1 (0.1)	28 (2.3)	2 (1.9)
Total	1383 (100.0)	1306 (100.0)	116 (100.0)

Note. Includes packaged and nonpackaged products. Food categories were classified according to Chilean Sanitary Code.²⁴ Availability refers to whether the product was available in each school kiosk. Meals and prepared dishes refer to sandwiches, hamburgers, pizzas, and wraps, among other prepared meals. Sugars and honey refer to jelly and liquid honey, among others. Stimulating and fruitive includes tea, coffee, and energy drinks.

DISCUSSION

To the best of our knowledge, this is the first impact evaluation of the FLAL across Chilean public schools. Results suggest that the law has resulted in substantial reductions in the availability of CF&B in public schools in Santiago after 6 months of implementation. The largest reductions were in calories and total sugars, showing that changes were primarily made in the mix of products available (i.e., primarily a reduction in the bakery category), rather than in the reformulation of previously CF&B classified products to products that would now comply with the nutrient cutoffs established in the law.

Comparison With Previous Studies

This study reports a substantial decrease in CF&B products after the implementation of the law. This is in line with previous findings from several US states. State policies that banned all SSBs have been associated with lower access to SSBs and weekly prevalence of SSB use.¹¹ In Massachusetts, compliance with nutritional standards increased from 13% to 69% in middle schools and 13% to 54% in high schools.²⁷ In California, state policies have improved the availability of foods meeting the nutritional standards established in the law.²⁸

The results of this study can be attributed to joint policy efforts at the national and local levels, and it is likely that these may have created synergistic effects. For example, the impact of setting nutritional standards in Connecticut was found to be stronger in conjunction with a scheme to incentivize schools to support change.²⁹

The reduction of mean calories, total sugars, saturated fat, and sodium reported in this study are consistent with findings in the United States and France. In Texas, for example, a public school nutrition policy resulted in a reduction in the energy density of available foods.³⁰ In France, a total ban on vending machines in schools resulted in a reduction of sugar intake of 10 grams from morning snacks at school.¹⁷

Our results suggest that changes in food availability have been primarily from changes in the product mix, rather than a reformulation of existing products. No study, to our knowledge, had examined this yet. Food labels in the United States—where most of the studies come from—are compared by portion (and not per 100 g), which makes between-product comparisons of nutritional composition difficult.

Availability of drinks, cereals, and ice cream experienced the largest increase between 2014 and 2016, compared with only a minor increase in fruits and vegetables. Our results suggest that new products complying with the law replaced existing ones, which is in line with Chilean food industry claims that indicate 18% of food products have been reformulated because of FLAL.³¹ This was more prominent for liquid products, in which total sugar reductions could be achieved by adding artificial or natural sweeteners. Recent evidence shows that nonsugar sweeteners could also be linked to increased risk of obesity and diabetes³²; therefore, nonintended consequences of adding additives warrant further investigation.

The impact of the changes in the availability of CF&B on food intake and obesity is still unclear.^9,33 Research on overall effects on intake and obesity will have to look beyond school kiosks, as children are exposed to unhealthy food and marketing outside schools.¹⁷ The Chilean law is relatively comprehensive in addressing front-of-package labels and restricting marketing to children, which is supported by an increase in 5 percentage points in an excise tax on SSBs and a second law restricting marketing to children.³⁴ However, it does not address the availability of unhealthy food in school surroundings and does not restrict marketing to children in public spaces. Evidence from Mexico suggests that children are exposed to high levels of marketing around schools.³⁵

Strengths and Limitations

We used objective measures of food composition derived from the package’s nutritional labeling. Trained personnel assessed all available products using a similar data collection procedure at both time points. Therefore, the data has high internal validity regarding kiosks in Santiago’s public schools.

However, some limitations must be noted. We were able to visit 27 of 32 eligible schools in 2016 and 25 of 27 eligible schools in 2016, which could have led to some selection bias. Nonetheless, we believe this did not have a major impact on the results of the study. The external validity of the study is limited, as it was conducted in public schools in 1 Chilean municipality. Local efforts were carried out in parallel to support the implementation of the law and, consequently, the results should be regarded as a best-case scenario. The evaluation was carried out only 6 months after the implementation of the law. The start of the implementation received a lot of media attention, and results could be a consequence of a honeymoon effect (i.e., a transitory positive effect).

It should also be noted that we collected the data during a period of warm weather (December), which might have influenced the food products available in the kiosks. We were not able to assess the nutritional composition of nonpackaged items. However, nonpackaged items represent a minor fraction of the available products, and we believe our analytical procedure did not influence the main results. We were not able to document changes in sales, consumption patterns, or body mass index or other health outcomes. And lastly, the study is purely descriptive in nature, and we did not adjust for confounding variables that may have partly explained the changes in availability between 2014 and 2016.

Public Health Implications

These results support the implementation of national policies regarding CF&B, establishing cutoff levels for them, and banning their sales in schools. The impact of such policies could be even larger and more sustained when implemented in conjunction with other cost-effective policies, including introducing taxes on SSBs, front-of-package labeling, bans on the marketing of unhealthy food, and primary care counseling. Policies addressing school surroundings may help to protect children from substitution effects.

Conclusions

A total ban on sales of CF&B high in calories, sugar, saturated fat, and sodium resulted in substantial changes in the availability of such foods in school kiosks in Santiago de Chile. Further research should explore the impact of the policy with a longer follow-up and examine the effects in children’s food intake and health outcomes.