Unhealthy dietary habits of children and adolescents are a growing public health problem in many countries. Schools are good operational environments to promote healthy eating habits both via teaching and school meals. Various interventions in school settings have been conducted during previous years with variable results(Reference Pérez-Rodrigo and Aranceta1, Reference Hoelscher, Evans and Parcel2). Many previous interventions have had a focus on overweight and obesity, and have, in addition to diet, concentrated on increasing physical activity(Reference Thomas3, Reference Shaya, Flores and Gbarayor4).
Curriculum and school meal arrangements may differ a great deal between countries, and, therefore, more research is needed to find effective local practices for health promotion. In Finland, the free school lunch prepared according to national recommendations provides a good basis for the diet of adolescents(5, 6). However, it has been observed that not all pupils eat the lunch every day, or at least not all components of it (main course, bread, salad and milk). Despite the recommendations, the selling of unhealthy snacks is still common in Finnish schools(Reference Kankaanpää, Lahti and Manninen7), and in many schools the possibility of leaving the school premises attracts pupils to go to nearby shops to buy unhealthy snacks. Teaching of home economics and health education are good tools to deliver information and skills to secondary-school pupils, but themes of healthy eating may also be incorporated into other lessons.
Many factors influence adolescents’ eating behaviours(Reference Story, Neumark-Sztainer and French8). Individual characteristics include psychosocial factors such as attitudes, beliefs, knowledge, taste and food preferences as well as biological factors (hunger and gender) and lifestyle factors (time and convenience, dieting, etc.). Adolescents’ eating behaviours are also strongly influenced by their social environment that includes family, friends and teachers. The physical environment influences accessibility and availability of foods, e.g. from vending machines at schools. Understanding these factors is important in order to develop effective nutrition interventions to change the eating habits of adolescents(Reference Story, Neumark-Sztainer and French8).
Previous studies have indicated that the intake of fruit and vegetables is low and that the consumption of unhealthy, sugary snacks is common among adolescents in Finland(Reference Roos, Karvonen and Rahkonen9). Therefore, the main objectives of this intervention study were to decrease the intake of sucrose, increase the intake of fibre and consumption of fruit and vegetables. Specifically, the aims were to increase consumption of the full school lunch and to change the snacks eaten during school hours into healthier options.
Subjects and methods
Subjects and study protocol
Altogether twelve schools in three cities agreed to participate. A detailed description of the recruitment of schools has been published in the baseline article(Reference Hoppu, Lehtisalo and Tapanainen10). The study protocol was approved by the ethical committee of the Hospital District of Helsinki and Uusimaa.
There were 1469 seventh grade pupils in these schools and 769 pupils (52 %) agreed to participate in the study. However, since the intervention was implemented as part of normal teaching, it was targeted at all pupils, not just the participants.
The baseline study for the seventh grade pupils was performed in April–May 2007. After the baseline study, the schools were randomised into intervention (IS) and control schools (CS) within the cities. Thus, there were three IS and three CS in Tampere, two IS and two CS in Lahti and one IS and one CS in Mikkeli. Intervention for the eighth graders was started at the beginning of the school year (August 2007) and continued until the follow-up study (April 2008). At baseline, the participants were on average 13·8 years old. The sociodemographic characteristics of participants (parents’ education and employment) differed between single schools but were comparable between IS and CS as a group (data not shown).
Structured questionnaires including questions about eating habits and frequency of food use were sent to pupils’ homes by post. Altogether 659 adolescents returned their questionnaires both in 2007 and 2008. Computer-assisted 48 h dietary recalls were performed for a randomly selected subgroup of participants (altogether 287 in both years). Dietary interviews were performed at schools by trained nutritionists during school hours. A detailed description of the methodology can be found in the baseline article(Reference Hoppu, Lehtisalo and Tapanainen10).
Statistical methods
All the seventh grade pupils in twelve study schools were invited to participate in the study. This cluster sampling was taken into account in all analyses considering intervention effects. In many variables, there were baseline differences between IS and CS. However, these baseline differences were not considered in the analysis.
All analysed questionnaire variables were in ordinal scale. New variables measuring the direction of the change were formed (positive change, no change and negative change). In order to create these new variables we examined the answers from years 2007 and 2008. If the answer was the same in both years there was no change, but if the participant reported, e.g. eating rye bread more frequently in 2008 than in 2007, the change was positive. These new variables were used to test the difference in changes between IS and CS. The difference was tested with Kendall’s τ-rank correlation coefficient. The reliability of the test result in the cluster-sampling framework was evaluated by a permutation test.
The intervention effects in nutritional variables were analysed with ANOVA in which the explanatory variables in the model were intervention group and the interaction intervention group school to take the effect of cluster sampling into account. Analyses were performed with the SAS statistical software package version 8·2 (SAS Institute Inc., Cary, NC, USA).
Intervention
The frame of reference of the intervention was Bandura’s social cognitive theory according to which behavioural, cognitive and other internal factors interact with environmental factors(Reference Bandura11). The intervention consisted of various actions targeted at different actors and they were implemented at different times according to a planned schedule. The intervention actions focused on the development of a healthy food environment as well as on nutritional education. In developing the food environment, the main target groups were headmasters, teachers and school catering personnel who carried out the actions at schools after discussing with the study personnel. Nutritional education was implemented by the teachers during regular lessons.
In the food environment component, attention was paid to food-related attitudes. Drama workshops dealing with eating and school meals were organised both for teachers and pupils to improve their commitment to the intervention. Discussions and informative meetings with headmasters, teachers and school catering personnel were also arranged to discuss how they could improve the food environments in their own schools. To deliver information to the families, parents’ meetings were organised in each school. Parents were offered a school meal and they were provided with information on the intervention and school meals. Furthermore, a magazine dealing with healthy eating was delivered to all parents. Pupils, parents and teachers also received information on the results of the baseline study concerning their own school.
The food environment component also included improvement of the supply of healthy snacks available at school. Supply of sugary snacks was restricted; for example, vending machines containing sweets and soft drinks were removed, and in schools offering snacks for sale the selection of snacks was improved to include sandwiches, milk and fruit instead of sweet cakes, etc. In two schools not offering snacks, vending machines selling healthy snacks were introduced. Several actions were also implemented during the year to increase the appreciation of the school lunch and to encourage the eating of a balanced lunch daily. In Finnish schools, crisp bread is usually offered daily at lunch and fresh bread approximately once a week. During the intervention, the supply of fresh bread was increased, but the other components of the lunch were provided as usual. Free supply of fruit and vegetables at school was not improved.
For nutritional education, various materials providing information on sugar, fibre, fruit and vegetables were offered for teachers to use, such as illustrative pictures of typical snacks, posters, informative brochures, games and tests. For example, the sugar test included a set of questions by which the pupils could assess their own sugar intake. Teachers were offered ready-planned lessons, but they were also encouraged to use these materials during their normal lessons according to their needs.
The implementation of the intervention was followed by a web-based follow-up system in which the teachers were asked to report the use of teaching material for each class. Because the teachers were allowed to freely choose the material they used, there were some differences between the schools. Overall, the implementation of the intervention was successful and the teaching material was actively used. After the study, the most popular parts of the intervention material were published on the Internet to be used freely in all schools (see http://www.jarkipalaa.fi (Reference Keso, Lehtisalo and Garam12)). For evaluation of the intervention, pupils, parents, teachers and catering personnel were asked their opinions on the intervention and most of the feedback was positive.
Results
Frequency of consumption
The frequency of consumption of rye bread, sweets, fruit and vegetables according to the questionnaire data is reported in Table 1. Among girls, the frequency of consumption of rye bread increased (P = 0·03; Fig. 1) and that of sweets decreased significantly (P = 0·006; Fig. 2) in the IS compared to CS. Among boys, there were no significant differences between the IS and CS.
Table 1 Frequency of consumption during the past week (7 d)
Fig. 1 Change in the frequency of consumption of rye bread among girls (
, intervention school; 
, control school)
Fig. 2 Change in the frequency of consumption of sweets among girls (
, intervention school; 
, control school)
At baseline, the daily consumption of fresh vegetables was infrequent; of the girls ∼ 40 %, and of the boys < 30 %, reported eating them daily (frequencies of 6–7 d/week and more than once a day together). At follow-up, it was found that the frequency of daily consumption of vegetables decreased among boys and remained the same among girls in both school groups. Daily consumption of fruit remained nearly constant among IS pupils but decreased among CS boys (difference not statistically significant).
Snacks at school
At baseline, the most commonly consumed snacks during school hours were sweets and sandwiches. In addition, girls ate cereal bars and fruit and boys drank sugary soft drinks (Table 2) Along with the intervention, there was some improvement in the quality of snacks. Girls in the IS reported a decrease in their consumption of sweets significantly compared with girls in the CS (P = 0·03). Consumption of sugary soft drinks remained constant among boys in the IS, but at the same time increased significantly among boys in the CS (P = 0·02). There were no differences in the changes of consumption of bread and fruit as snacks between IS and CS (Table 2).
Table 2 Snacks and drinks consumed during school hours
Results of the dietary recall
As can be seen in Table 3, energy-adjusted consumption of fruit, including berries (g/MJ) remained constant in IS, whereas it decreased in CS (difference in change, P = 0·04). There was no significant change in the consumption of vegetables.
Table 3 Amount of total vegetables and fruit consumed
*Statistically significant difference in change (girls and boys together, P < 0·05).
The nutrient intakes among pupils in IS and CS and the changes between the years are presented in Table 4. At baseline, the intake of sucrose was higher and the intake of fibre lower than recommended. The intake of sucrose fell significantly among IS pupils (difference in change, P = 0·01). The decrease in sucrose intake in IS was 18 % from the baseline value. There was no change in the intake of fibre. Furthermore, the intake of folate/MJ increased along with intervention (P = 0·04).
Table 4 Intakes of nutrients
*Statistically significant difference in change (girls and boys together, P < 0·05).
Discussion
This intervention study succeeded in decreasing the intake of sucrose among secondary-school pupils. High sucrose intake has been associated with poor quality of diet(Reference Ruottinen, Niinikoski and Lagström13). If the decrease in the consumption of sugary drinks persists longer, it can have positive health effects both on dental health(Reference Touger-Decker and van Loveren14) and weight(Reference Vartanian, Schwartz and Brownell15). A focus in this intervention was the snacks eaten at school, but other sources of sucrose, such as cereals and yoghurts commonly eaten for breakfast at home, should also be targeted to further decrease sucrose intake. Although the supply of fresh bread was increased at school lunch and the girls in the IS reported eating rye bread more often, the intake of fibre did not change. It is possible that the consumption of other sources of fibre has decreased concomitantly. It has also been reported that dietary fibre intake (g/MJ) seems to stay fairly constant during adolescence(Reference Alexy, Kersting and Sichert-Hellert16).
The intervention did not succeed in its aim of increasing the consumption of fruit and vegetables. However, the intake of fruit remained stable in the IS, whereas it decreased in the CS. This can be regarded as a good result, since it has been observed that the consumption of fruit and vegetables seems to decrease with age(Reference Lien, Lytle and Klepp17, Reference Larson, Neumark-Sztainer and Hannan18). In this intervention, education material on fruit and vegetables was used, but the supply of fruit and vegetables at schools could not be improved. It is not surprising that increasing knowledge is not easily turned into action if the supply of fruit and vegetables is not improved both at school and at home(Reference Bere, Veierod and Bjelland19). Other interventions, which have included the distribution of fruit or vegetables at school, have shown small positive changes(Reference Reinaerts, de Nooijer and Candel20), but this may not result in sustainable changes. A combination of various personal and social factors has been observed to be related to fruit and vegetable intake, and therefore, multi-level intervention strategies are suggested(Reference Knai, Pomerleau and Lock21, Reference De Bourdeaudhuij, te Velde and Brug22).
School food policies have been shown to have an impact on adolescents’ food habits(Reference Kubik, Lytle and Hannan23–Reference Vereecken, Bobelijn and Maes25). Improvement in the supply of healthy foods at school, both as school lunch and snack options, has been the focus of many recent interventions and projects(Reference Suarez-Balcazar, Redmond and Kouba26). In this intervention, changes in the supply of snacks at school resulted in some improvements in pupils’ snack choices. The selling of unhealthy snacks at school is in contrast with the health education aims of the curriculum and puts pupils’ potential choices in conflict. It is the responsibility of all the adults at school, headmasters, teachers and catering personnel, to create a supportive environment for healthy eating and to make healthy choices easy for pupils.
The age group of 14–15-year old is challenging, as the changes related to puberty are rapid at this age. For example, among boys the rapid growth spurt may cause a considerable increase in energy needs(Reference Deheeger, Bellisle and Rolland-Cachera27). It seems that along with age, eating habits may change in a more unhealthy direction(Reference Lien, Lytle and Klepp17), and, therefore, this trend should be stopped as early as possible. The intervention period was also short, at less than one school year. To achieve more permanent results, a longer intervention time starting at a younger age would be desirable. The strength of this intervention was that it was wide, with many actions targeted at many actors. However, a limitation is that it is not possible to identify the single actions that were the most effective.
Some differences between boys and girls were observed in the results of the intervention. As in previous studies, the results have often been found to be better among girls(Reference Shepherd, Harden and Rees28, Reference Haerens, de Bourdeaudhuij and Maes29). This may partly be explained by different taste preferences of boys and girls; for example, girls have been found to like vegetables more than boys(Reference Cooke and Wardle30). Thus, it is necessary to find new and effective intervention methods targeted specifically at boys.
This intervention was only targeted at eighth grade pupils. In order to get the whole school community to be more involved, interventions targeted at all pupils at school would be desirable. In Finnish comprehensive schools, grades 7–9 are usually located in the same building, so at least all these classes could be included. More involvement of the participants themselves in the planning and implementation of the intervention would also be an advantage. Good results have been observed in peer-led health education(Reference Mellanby, Rees and Tripp31). The interest of adolescents in answering traditional questionnaires may be limited and new attractive data collection methods, e.g. via the Internet, should also be developed in the future(Reference Matthys, Pynaert and de Keyzer32). The limitation of the present study is the low participation rate, but other school studies requiring active parental consent have often had low participation rates(Reference Esbensen, Melde and Taylor33).
Parents are responsible for providing healthy foods at home and they are also important role models for their children(Reference Pearson, Biddle and Gorely34). However, some parents seem not to have the time for or much interest in the food habits of their children, and are not willing to participate in meetings at school. Improving parental involvement in interventions is a challenge and new communication strategies to reach the parents should be sought out.
In conclusion, this intervention study succeeded in decreasing the sucrose consumption of Finnish secondary-school pupils. The feedback on the intervention was positive both from the teachers and pupils. Participation of pupils in the planning and implementation of the intervention should be improved in the future. The experiences gained in the present study offer encouragement for continuing to develop school-based healthy eating promotion for adolescents.aaa
Acknowledgements
The authors thank all the schools and pupils for their participation and SITRA (Finnish Innovation Fund) for funding the study.
