Ultra-processed food consumption by Brazilian adolescents in cafeterias and school meals

This cross-sectional study utilized the National School Health Survey 2015 database to assess the association between school cafeterias; the meals offered by the Brazilian School Food Program (PNAE); and the consumption of industrialized/ultra-processed salty foods, sweets, and soft drinks among Brazilian adolescents. A sample of 102,072 adolescents, aged 11–19 years, who were enrolled in the 9th grade completed the survey. The evaluated outcome was the consumption of industrialized/ultra-processed salty foods, sweets, and soft drinks. A Poisson regression model-based multivariate analysis was performed. The effect measure was the prevalence ratio (PR) with its respective 95% confidence intervals (CIs). The results indicated that Brazilian adolescents who attended schools without meals offered through the PNAE had a higher probability of regularly (≥5 times/week) consuming ultra-processed salty foods [PR = 1.06, CI = 1.01–1.11] and soft drinks [PR = 1.08, CI = 1.03–1.14] compared to those who attended schools that offered PNAE meals. Moreover, the presence of a school cafeteria was associated with a higher probability to consume industrialized/ultra-processed salty foods [PR = 1.05, CI = 1.02–1.08], sweets [PR = 1.09, CI = 1.07–1.11], and soft drinks [PR = 1.10, CI = 1.07–1.13]. School meals appear to be associated with the consumption of ultra-processed foods by Brazilian adolescents, indicating areas for health promotion programs.

Being overweight is a global public health problem and a key risk factor for the development of chronic, non-communicable diseases [1][2][3] . Worldwide, approximately 20-25% of adolescents (individuals aged 10-19 years) 4 are overweight 1 . In Brazil, 23.7% of adolescents are overweight 5 , which is associated with poor eating habits such as high consumption of ultra-processed foods and low consumption of minimally processed and/or fresh foods [6][7][8][9][10] . Ultra-processed foods are generally defined as those manufactured industrially through several stages of processing 6,7,11 . Accordingly, such foods contain laboratory-synthesized industrial substances (e.g., emulsifiers, colorants, flavorings, flavor enhancers, and thickeners) and large amounts of calories, trans fats, sugars, sodium, and chemical additives 6,7,11 .
For adolescents, school is where messages promoting health, disease prevention, and healthy eating habits can be disseminated [12][13][14] , including the beneficial effects of healthy meals on learning and school performance [15][16][17][18][19] . However, the presence of a food outlet or cafeteria that sells food either within or near a school is associated with the consumption of foods with low nutrition and high calories, as well as with low vegetable consumption (seen in United States, and the Republic of Ireland) 20,21 . In Australia, for example, there is food with good sources of nutrients; however, it is more expensive than is food that lacks the adequate nutritional value 22,23 . Browne and colleagues 20 evaluated three different types of school lunches: sourced from home, cafeterias, and take-out from food outlets in the Republic of Ireland. They found that canteens and take-out purchased lunches had worse nutritional value than did packed lunches from home, as well as a greater amount of total fat and added sugars.
Some countries have established policies regarding the sale of foods in schools, which limit foods considered unhealthy (e.g., sweets and soft drinks) 22,[24][25][26][27][28] . Positive results were found in France when evaluating the consumption of school canteen lunches when compared with other lunches including the consumption of more fruits, vegetables, fish, and dairy products, and less soft drinks and chocolate 27 . However, in Australia, an (2019) 9:7162 | https://doi.org/10.1038/s41598-019-43611-x www.nature.com/scientificreports www.nature.com/scientificreports/ evaluation of the policy guidelines to promote nutritious food sales in school canteens revealed unsatisfactory results 22,23 .
In Brazil, school meals are considered a basic human right that plays a key role in student development. Therefore, public schools offer free meals through the Brazilian School Food Program (Programa Nacional de Alimentação Escolar; PNAE), which aims to provide healthy meals for students 29,30 . This comprehensive and long-lasting program has been recognized nationally and internationally by global food and nutrition security organizations 29,30 . Additionally, Brazilian school cafeterias, particularly those in the private school system, also sell food on school premises 5 .
Despite the widespread application of the PNAE in Brazil, studies examining the effects of school meals via the PNAE and school cafeterias on the consumption of ultra-processed foods by adolescents remain scarce, especially at the national level 31 . In this context, it is vital to prevent students from becoming overweight and unhealthy eating habits and identify the influence of food offered by the PNAE and/or sold by school cafeterias on school premises on the consumption of ultra-processed salty foods, soft drinks, and sweets. Therefore, we assessed the association between school cafeterias and the meals offered by the PNAE and the consumption of industrialized/ ultra-processed salty foods, sweets, and soft drinks among Brazilian adolescents.

Results
Overall, 102,072 adolescents were included in the sample and the response rate for all questions was higher than 99.5% for each outcome (industrialized/ultra-processed salty foods, sweets, and soft drinks). The data regarding the consumption of ultra-processed foods, and the results of the bivariate analysis are presented in Tables 1-3.
The regular consumption of salty, ultra-processed foods and the consumption of sweets were associated with residing in a capital, private school attendance, female sex, age, absence of PNAE, and presence of a cafeteria (Tables 1 and 2). Regular soft-drink consumption was also associated with residing in a capital, male sex, age, and presence of a cafeteria (Table 3).
Multivariate analyses results are shown in

Discussion
This study assessed the association between school cafeterias; the meals offered by the PNAE; and the consumption of industrialized/ultra-processed salty foods, sweets, and soft drinks among Brazilian adolescents. The results indicate that 9th graders who attended schools covered by the PNAE have a lower probability of consuming ultra-processed salty foods and soft drinks, whereas the probability to consume ultra-processed salty foods, sweets, and soft drinks was higher for those who attend schools with cafeterias. Given the paucity of research related to the influence of meals offered by the PNAE and school cafeterias on the consumption of ultra-processed food by adolescents, our study represents a significant contribution to this field. Students attending schools covered by the PNAE reported a lower consumption of salty, ultra-processed foods and soft drinks. This is consistent with previous findings that national food programs, including the PNAE, tend to prioritize the provision of a healthy diet, defined as a greater quantity of fresh or minimally processed foods such as fruits, vegetables, and milk, and reduced levels of sodium and simple sugars [17][18][19][32][33][34] . In the Philippines, a school meal program developed nutritious meals for adolescents to ensure higher intakes of micronutrients (p < 0.001), fiber (p < 0.001), and the intake of higher-quality fats (p < 0.001) as compared with basal meals 35 . In France, school meals were associated with the consumption of fresh vegetables and fruits (p < 0.001) and family income (p < 0.001). Specifically, the diet provided in France increased the intake of vegetables and fruits among adolescents with a low socioeconomic status, which reduced the differences between them and their peers with a higher socioeconomic status 34 .
In Brazil, the evaluation of school meals through the PNAE in public schools remains challenging because adherence rates are relatively low (38-62%) and institutions tend not to follow the program guidelines strictly 31,36,37 . Despite these limitations, offering school meals via the PNAE is associated with an increased consumption of fresh foods and a lower consumption of ultra-processed foods 31,38 . This may be attributable to PNAE restrictions on the use of ultra-processed foods, such as processed meats, canned goods, sugary beverages, and concentrated and/or prepackaged foods 32,39 .
By contrast, the presence of a school cafeteria was associated with an increased probability that students would regularly consume industrialized/ultra-processed salty foods, sweets, and soft drinks in this study. Although school meals encompass all food offered in the school environment, regardless of origin, the school cafeteria is often decontextualized in relation to healthy eating 18,23,25,28,[40][41][42] . Despite policies related to healthy food supply from school cafeterias in several countries, the implementation and enforcement of such policies remains limited, even in developed countries such as the United States, New Zealand, United Kingdom, and Australia 18,23,25,28,41,42 . In Australia, for example, even with the implementation of policy guidelines for the promotion of healthy food sales in 2005, an evaluation of government school menus found that most school cafeterias (especially in unmonitored schools) did not comply with the guidelines and continued to offer prohibited foods such as sweets and www.nature.com/scientificreports www.nature.com/scientificreports/ soft drinks 28 . In Brazil, no national law currently prohibits or restricts the sale of ultra-processed foods within the school environment, despite the PNAE legislation defining school meals as any food offered in the school environment. Nevertheless, there are laws prohibiting the sale of certain foods, especially sweets and soft drinks in some states (e.g., São Paulo, Santa Catarina, Amazonas, and Goiás) and municipalities (e.g., Florianópolis and Rio de Janeiro) 32,39 .
Interventional studies aiming to change the profile of available foods in school cafeterias by reducing the availability of ultra-processed and processed foods and increasing that of fresh or minimally processed foods and/or meals have yielded positive results concerning improved nutritional profiles and adherence to a healthy diet among adolescents 23,35,41,43,44 . These changes in available foods yielded net reductions in the caloric, sugar, and sodium contents of the offered meals 41 . However, the significant cost difference between a healthy diet and ultra-processed foods must be considered 22,44 .
We found that adolescents enrolled in private schools had a greater probability of regularly consuming ultra-processed salty foods compared to those in public schools. This may be attributable to the presence of PNAE-provided school foods in the latter, which protects against the regular consumption of ultra-processed salty foods 31,45 . Similarly, students who attended schools in capital cities had a greater probability of regularly consuming ultra-processed salty foods and soft drinks than did students who attended other schools, likely because of the greater availability of food outlets or cafeterias inside or near the schools 36,44,46 . However, public school students had a greater probability of regularly consuming soft drinks. This may be due to the fact that PNAE-provided school foods do not exclude the consumption of soft drinks when adolescents have access to these beverages in the school environment 47 . Moreover, a study from 124 Brazilian municipalities noted that there are food outlets or cafeterias inside or near the schools in 58.6% and 47.7% of public and private schools, respectively 45 .
Additional barriers to the initial and adequate implementation of existing policies and/or policy making in countries that have not yet legislated school meals are related to inadequate resources, limited legislation regarding consumption quantities, lack of knowledge about healthy foods, and costs 44,48 . Public policies must be continuously updated, implemented, and monitored to ensure a healthy food environment that will promote good health and appropriate physical, psychological, and social development of adolescents.
The main limitation of this study was its cross-sectional design; thus, we cannot infer causality. Moreover, using a restricted sample composed of 9th-grade students, who may not be inclined to accept foods offered at school, was a limitation. Additionally, we cannot state that these data reflect adolescents' usual food consumption since the information was self-reported; therefore, over-and under-estimation are both possible. Lastly, some significant results should be interpreted with caution because of the large sample size. www.nature.com/scientificreports www.nature.com/scientificreports/ In conclusion, school meals appear to be associated with the consumption of ultra-processed foods by Brazilian adolescents. Specifically, the provision of school meals via the PNAE was associated with less consumption of industrialized/ultra-processed salty foods and soft drinks, whereas the presence of a school cafeteria was related to a higher consumption of industrialized/ultra-processed salty foods, sweets, and soft drinks. These findings address an existing gap in the literature and provide data indicating areas for potential improvement.
Our results inform intervention strategies to promote healthy eating patterns and prevent the development of chronic, non-transmissible diseases among adolescents and young adults. The consumption of ultra-processed food is associated with several health concerns obesity, cardiovascular diseases, diabetes, and cancer 3,10,49,50 , compromising individuals' quality of life during adolescence and later in adulthood. Therefore, schools have a crucial role in promoting students' healthy eating 51 .

Data sources and study population. This cross-sectional study was conducted using the National School
Health Survey (Pesquisa Nacional de Saúde do Escolar; PeNSE) 2015 database 52 , through a partnership between the Ministry of Health and the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografa e Estatística; IBGE). The study results were made available by the IBGE in 2016 53 . The PeNSE was approved by the National Commission on Ethics in Research (Comissão Nacional de Ética em Pesquisa; CONEP) of the National Health Council, which regulates and approves health research involving human participants (CONEP resolution no. 1,006,467; March 30, 2015).
This study involved adolescents aged 11-19 years who were classified as 9th graders in both public and private schools throughout Brazil from April to September 2015. This sample of adolescents adequately represents Brazil, including the 27 federative units (26 states with its capitals and municipalities, and the Federal District; IBGE) 53 .
The sample size allowed for an estimation of the parameters for each of the 26 states and the Federal District, including capitals and municipalities from the interior of Brazil. Brazil is conventionally divided into five regions (North, Northeast, Southeast, South, and Midwest) 53 .
The samples of the geographic levels comprising capitals and municipalities were random and equiprobabilistic; they were calculated using the following parameters: 0.03% maximum error, 95% confidence level, and prevalence of 0.5. Overall, 120,122 students who were enrolled in and attended one of 4,159 classes across 3,040 schools were included in the 2015 sample. Of these, 102,072 students completed the survey on the sampling day.  www.nature.com/scientificreports www.nature.com/scientificreports/ As all students in the sampled classes were invited to respond to the survey questionnaire, there was a sample loss of approximately 15% 53 .
All students that agreed to participate voluntarily provided written, informed consent. Students were told that they could leave the study at any time if they chose not to participate in any of the procedures 53 . Additionally, this study was verified in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist 54 . outcome measures. The PNAE is a national program based on a public food policy 29,30 . Accordingly, this program provides free meals to all basic education students in the public education network (municipal, state, and federal) and covers 200 school days per year. The PNAE aims to foster students' growth and development, encourage the formation of healthy eating habits, and provide food and nutrition education 29,32 . As noted in the Introduction, school cafeterias also sell food on the premises of educational institutions, especially private schools.
Data collection, involving a validated self-administered survey 55 , was performed using smartphones distributed to the students who were in class on the day of the interview by the IBGE technician. The technician explained to students how to use the device to complete the questionnaire 53 . We considered the consumption of ultra-processed food as an outcome for which three different categories were set according to food and/or food groups: industrialized/ultra-processed salty processed foods (salty, canned, and instant foods), sweets (sweets, candies, chocolate, chewing gum, and lollipops), and soft drinks.
These variables were evaluated using the following question: "In the last 7 days, on how many days did you eat/ drink…?" The eight available response options were "I did not eat/drink in the last 7 days", "I ate/drank on 1 (2, …, 6) of the last 7 days", and "I ate/drank on all of the last 7 days". For data analysis purposes, the food groups were categorized as regularly consumed (at least 5 of the previous 7 days) or not regularly consumed 37,56-58 . statistical analyses. Data were analyzed using descriptive statistics and the Wald chi-square test of association (bivariate analysis) for the following outcome groups: industrialized/ultra-processed salty foods, sweets, and soft drinks. The following factors were considered explanatory group variables: sex, age, municipality type, school type, availability of food offered by the PNAE, and presence of a school cafeteria. All variables are included in Poisson regression model-based multivariate analysis and the effect measured was the prevalence ratio (PR) with the respective 95% confidence intervals (CIs) 59 . Methodological studies support the inclusion of variables with theoretical grounds in the final multivariable statistical analyses 60,61 . Statistical analyses were performed using SPSS 20.0 (IBM, Armonk, NY, USA).