The efficacy of a multi-strategy choice architecture intervention on improving the nutritional quality of high school students' lunch purchases from online canteens (Click & Crunch High Schools): a cluster randomized controlled trial.

Background: High school canteens are an ideal setting for public health nutrition intervention, and choice architecture strategies that facilitate the purchase of healthier foods and beverages from school canteens are recommended by the World Health Organization. The rapid uptake of online lunch ordering within school canteens provides a unique opportunity to implement choice architecture strategies that support healthier food choices with high fidelity. Despite this, no trial has tested the efficacy of choice architecture strategies within an online lunch ordering system on improving the nutritional quality of high school student lunch purchases. The objective of this study was to assess the impact of embedding choice architecture strategies into an online lunch ordering system on the nutritional quality of the school canteen lunch purchases of high school students (aged 12-19 years).

Methods: A cluster randomized controlled trial was conducted with nine high schools in one Australian state. Schools were randomized to receive either a 2-month choice architecture intervention (involving menu labelling, prompts, item positioning, and feedback), or usual online ordering. Nutrient quality of online canteen lunch purchases was assessed using routine data collected by the online ordering system. Primary outcomes were the proportion of 'Everyday', 'Occasional', and 'Should not be sold' items purchased, categorized using the state healthy canteen policy. Secondary outcomes were the mean energy, saturated fat, sugar, and sodium content of purchases and the mean weekly revenue from online lunch orders. Linear mixed models were analyzed to assess outcomes.

Results: Analysis of the student cohort (Intervention: 4 schools, 656 students; Control: 5 schools, 675 students) showed significant between group differences over time for the intervention group for the mean percentage of online lunch items per student that were 'Everyday' (+ 5.5%; P < 0.001) and 'Should not be sold' (- 4.4%; P < 0.001). There were no between group differences over time in the mean percentage of online lunch items that were 'Occasional'; the average energy, saturated fat, sugar, or sodium content of lunch orders. There was also no difference in mean weekly revenue from high school student online lunch orders (P = 0.23).

Conclusions: These findings suggest that a low intensity, choice architecture intervention embedded within an online ordering system can increase the purchase of healthier food items for high school students in one Australian state without any adverse impact on canteen revenue.

Trial registration: This trial was prospectively registered on Open Science Framework on 23rd October 2020 as osf.io/h8zfr.