A randomised controlled trial of an intervention to increase the implementation of a healthy canteen policy in Australian primary schools: study protocol

In 2005, the NSW government launched the Fresh Tastes @ School NSW Healthy School Canteen Strategy as a key component of its action plan to prevent childhood obesity [16]. The strategy used a traffic light system to classify foods and beverages sold by schools as `red’, `amber’ or `green’ based on their nutritional properties [8]. The strategy was adopted as part of the Department of Education and Communities policy in NSW requiring all government schools to remove unhealthy foods and beverages—those classified as ‘red’—from regular sale at school canteens. Furthermore, schools were encouraged to `fill the menu’ with items classified as `green’ and not let items classified as `amber’ dominate the menu.

Items classified as `red’ based on the healthy canteen policy are those that lack adequate nutritional value and can contribute to excessive energy intake [8]. They are high in saturated fat and/or added sugar and/or salt and include sugar-sweetened drinks (banned), chocolate-coated premium ice creams, confectionary and deep-fried foods. Commercial hot foods and snacks, pastries, ice creams, savoury snack foods, sweet biscuits and cakes can also be classified as `red’ items if they do not meet the nutrition criteria described in Table 1. `Amber’ items are considered to have some nutritional value and moderate levels of saturated fat and/or added sugar and/or salt [8]. If consumed in large amounts, however, these foods can contribute to excess energy intake. `Amber’ items include full fat dairy products, processed meats, sauces, spreads, refined breakfast cereals, most flavoured waters, diet soft drink and ≥99% fruit juice over 200 ml. Amber items are also commercial hot foods and snacks, pastries, ice creams, savoury snack foods, sweet biscuits and cakes that do meet the criteria described in Table 1. `Green’ items are foods that are considered to be good sources of nutrients and contain low amounts of saturated fat and/or added sugar and/or salt [8]. They include fruits, vegetables and legumes, reduced fat dairy products, wholegrain cereals, grains such as pasta, rice and bread, lean red meat, fish, skinless poultry and alternatives, as well as small serves of ≥99% fruit juice.

The study will employ a randomised controlled trial (RCT) design. Primary schools with a canteen will be randomised to either an intervention group or a no intervention comparison group. The effectiveness of the intervention will be determined by comparing post-intervention differences between groups in i) the proportion of schools with a canteen menu that does not sell foods or beverages (`red’ and `banned’) restricted for sale according to the healthy canteen policy and ii) the proportion of schools where healthy canteen items (`green’) comprise more than 50% of products listed on the menu as recommended by the healthy canteen policy. Data will be collected at baseline and immediately following the implementation intervention. The trial will be conducted in the Hunter region, a geographically and socioeconomically diverse region in NSW, Australia. Children in the region frequently purchase foods high in fat, salt and sugar from their school canteen [17].

Publically available Department of Education and Communities’ lists of primary schools in the Hunter region will serve as the sampling frame. There are over 300 schools in the study region. Schools will be randomly selected from the list and approached to participate. Schools with a canteen will be eligible if they have either at least one food or beverage menu item which is restricted (`red’ or `banned’) for sale; or have less than 50% menu items classified as healthy foods or beverages (`green’ items). Schools from the Catholic Schools Office and Association of Independent Schools will be ineligible, as the healthy canteen policy is not mandated in these schools. Schools with both primary and secondary students (i.e. central schools) and schools catering exclusively for children requiring specialist care will also be ineligible. Recruitment procedures will be adapted from a review of recommendations for engaging schools in research trials [18] which has been used previously in the setting to achieve participation rates of schools of between 80–95% [19],[20].

Schools will be randomly allocated after baseline data collection in a 1:1 ratio to either an intervention or control group using a computerised random number function in Microsoft Excel (see Figure 1). A block randomisation procedure will be employed to ensure group allocation is equal. The procedure will be stratified based on the socioeconomic status of a school locality given evidence that the locality may have a potential impact on the trial outcome [17].

The intervention will seek to increase school implementation of the healthy canteen strategy, which is supported by a government policy.

The delivery of all intervention components will be controlled by the research team and will not be provided to control group schools. During the trial period, teachers from either intervention or control group schools will be able to access NSW Government-run programs directed at supporting school promotion of healthy eating and physical activity generally [35]. Data regarding schools’ exposure to such programs and potential sources of contamination for control schools will be collected by the research team during the follow-up survey with principals and the canteen managers. Where evidence of potential bias due to contamination is apparent, sensitivity analyses will be conducted to assess the potential impact on trial outcomes.

A total of 35 intervention and 35 control schools will be recruited. Primary trial outcomes will be assessed by comparing between group differences at follow-up regarding i) the proportion of schools with a canteen menu that does not contain foods or beverages (`red’ and `banned’) restricted for sale under the healthy canteen policy and ii) the proportion of schools where healthy canteen items (`green items’) represent >50% of products listed on the menu. Analyses will be performed under an intention to treat framework. Intervention effectiveness will be assessed using logistic regression models adjusting for baseline values and with all available data. Multiple imputations (including baseline observation carried forward) will be performed as part of sensitivity analysis for schools not providing follow-up data [39]. Based on previous recruitment experiences of the research team in this setting, it is anticipated that 80% of participating schools will be retained at follow-up [19]. Assuming a prevalence of 15% at follow-up in the comparison group for both primary trial outcomes, the sample will be sufficient to detect an absolute difference of 34% with 80% power and an alpha of 0.05. Subgroup analyses will be performed by school size (number of students) and measure of the socioeconomic status of the school’s geographic locality.

The trial is currently in the implementation stage of the intervention and has not initiated follow-up data collection.