CAFÉ: a multicomponent audit and feedback intervention to improve implementation of healthy food policy in primary school canteens: a randomised controlled trial

The implementation of nutrition policies in schools has been recommended as a strategy to improve child dietary intake. Internationally, research suggests that the majority of schools do not implement these policies. In New South Wales (NSW), Australia, the NSW Healthy School Canteen Policy requires that school canteens prohibit the sale of 'red' foods (i.e. foods that are typically nutrient poor and high in energy, such as confectionary and deep-fried foods) and 'banned'drinks (i.e. soft drinks); and that the majority of items on the menu are 'green' (i.e. foods that are good sources of nutrients, such fruits, vegetables and lean meats). This study examined the impact of a multicomponent audit and feedback intervention on schools' implementation of the NSW Healthy School Canteen Policy. A secondary aim was to assess the impact of the intervention on menu composition. This study was a parallel group randomised controlled trial with 72 rural and remote primary schools (36 interventions, 36 controls) located in one region within NSW, Australia. Intervention schools received an initial face to face contact and up to four cycles of audit and feedback (consisting of a menu audit, written feedback report and telephone feedback) over a 12-month period. The primary trial outcomes were the proportion of schools with a canteen menu that had: i) no 'red' foods or 'banned' drinks; and ii) >50% 'green' items, as assessed via standardised menu audits undertaken by trained dietitians. For each primary outcome, between-group differences were assessed using Fisher's exact test under an intention to treat approach. There was insufficient evidence to conclude the intervention had a positive impact on the proportion of intervention schools with no 'red' or 'banned' items on their menu (RR = 2.8; 95% CI: 0.9 to 8.9; p = 0.0895), or on the proportion of intervention schools with more than 50% 'green' items (RR = 1.5; 95% CI: 0.7 to 3.2; p = 0.2568). These findings remained non-significant in the multiple imputation analyses. Intervention schools were significantly more likely to have a lower percentage of 'red' items (p-value: 0.007) and a higher percentage of 'green' items on the menu (p-value: 0.014). This remained statistically significant in the multiple imputation analyses for 'red items' (p-value: 0.0081) but not for 'green' items (p-value: 0.0910). While there was insufficient statistical evidence to suggest that this multicomponent audit and feedback intervention was ef