The School Meal Project was a non-randomized trial with an intervention and a control group conducted in the southern part of Norway in the school year 2014/2015, and the current study is part of this project. The study sample consisted of children aged 10 to 12 years from two primary schools. Consumption of FV and unhealthy snacks was assessed with items derived from a validated questionnaire used in the Fruits and Vegetables Make the Marks-project (32). The school children answered the food frequency questionnaire at school (approximately 45 minutes) at pre, mid intervention and post intervention. Ethical approval was obtained from the Norwegian Centre for Research Data (NSD) and the ethical committee at the Faculty of Health and Sport Sciences at the University of Agder.
Sample and procedure
Due to practical considerations, a convenience study sample was used. The cook who prepared the school meal every day lived close to the intervention school, making it feasible to deliver the school meal every day. In order to select another control school, school size and characteristics regarding area (e.g. rural) and location in the same county were considered. Two primary schools participated in the project, and they were both located in a rural area in the same county and were of equal size. One school included both an intervention and a control group, and the other school included a control group. Information about the intervention was given to the principals to get their approval for participation. After this approval was obtained, all children were invited to participate in the study. This constituted a total of 219 invited children. One of the parents/care takers of each child was also invited. Parents were given oral and written information about the study through parent meetings and by an invitation letter. An active written consent of participation regarding their child and themselves was needed in order to participate in the study.
Data were collected at baseline (August/September 2014), in January 2015 (Time 1) and in June 2015 (Time 2). Data from both children and parents were collected at all three data collection points. Parents answered a short questionnaire and only the variable regarding level of education is used in the present study.
Intervention
The intervention consisted of a free school meal at lunch time served to children in the intervention group for one school year (33). The school meal was prepared in accordance with current Norwegian dietary guidelines and consisted of whole grain bread (at least 50 % whole grain), FV and several types of healthy spread. The spread included butter, fish spread (mostly mackerel in tomato sauce and smoked salmon, cheeses, different kinds of lean meat, liver paste, Norwegian caviar, and eggs. Tomatoes, cucumber, peppers and lettuce were available to put on the sandwiches, and pieces of fruit and vegetable were served on the side. Natural yoghurt (no added sugar) together with berries were occasionally served. Food high in added sugar and/or saturated fat, like jam and chocolate spread, were not served. No beverages were served as part of the free meal, but children who subscribed to the national milk subscription program received milk, otherwise they were encouraged to drink water. The food was served on trays, and the children helped themselves to the food they preferred. The food was consumed in the classroom, and the children ate together around one or two tables. A teacher was always present during the meals.
Measures
Personal variables
Parents’ level of education was assessed in the parent questionnaire by two items: “What is your highest level of completed education?” with four response options; “primary school (elementary school or lower secondary school)”, “upper secondary school”, “3-4 years of college or university” and “5 or more years of college or university” and “what is your spouse/partner’s highest level of completed education?”. The response options were the same as the previous item, but also included; “I do not have a spouse/partner”. The parents’ educational level was a proxy for socio-economic status (SES). Both scores were combined and dichotomized into “lower SES” (both parents having completed primary school and upper secondary school) and “higher SES” (at least one parent having completed 3–4 years and more than 5 years of college/university) (34)
Intake of Fruits and Vegetables
Intake of FV was assessed by five food frequency items in the child questionnaire. The items were “How often do you eat vegetables with your dinner?”, “How often do you eat vegetables on your sandwiches?”, “How often do you eat other vegetables (e.g. carrots with your lunch)?”, “How often do you eat apple, orange, pear or banana?” and “How often do you eat other types of fruits or berries (other fruits and berries than apple, orange, pear and banana?”. All questions had ten response options (never, less than once a week, once a week, twice a week, three times a week, four times a week, five times a week, six times a week, every day and several times per day). They were recoded to consumption times per week (0, 0.5, 1, 2, 3, 4, 5, 6, 7 or 10) (27).
Intake of unhealthy snacks
The four included unhealthy snack items in this study were assessed by the following questions: “How often do you eat potato chips?”, “How often do you eat candy (e.g., chocolate, mixed candy)”, “How often do you drink fruit squash?” and “How often do you drink soda with sugar?”. All the questions had ten response options (never, less than once a week, once a week, twice a week, three times a week, four times a week, five times a week, six times a week, every day and several times per day). They were recoded to consumption times per week (0, 0.5, 1, 2, 3, 4, 5, 6, 7 or 10). The children were asked to think of their intake during the whole day when they answered these questions.
Statistical analysis
Preliminary analyses consisting of the descriptive statistics of sample characteristics and key variables were conducted using IBM SPSS 25.0. Descriptive analyses were performed to characterize the sample, and to detect differences between the intervention and the control group. Participants’ characteristics at baseline were compared by independent sample t-tests for quantitative variables and by chi-square tests for qualitative variables to detect baseline differences between the control and the intervention group. Socio-demographics (age, gender and SES) and outcome variables at baseline are presented in table 1.
The normality of the key variables was checked. Although some of the outcome variables showed a tendency to a skewed distribution, the distribution of the residuals was acceptable. Therefore, the untransformed outcome variables were used. Multiple linear regression analyses were performed to assess intervention effects on the intake of FV and unhealthy snacks (table 2). All analyses were adjusted for baseline intake. No other covariates were included in the analyses given that they did not significantly associate with the outcomes. All cases with complete data for baseline and follow-ups were included in the analyses.