Using behavioural insights to reduce sugar in schools – pilot study in Brighton and Hove City Council

This pilot project focused on developing an intervention to reduce sugar consumption in schools. The intervention we tested with Varndean School in Brighton was a series of personalised ‘sugar swap’ messages (e.g. messages suggesting swapping a food item for a different item with lower sugar content).

Efficiency and income generation

The challenge
Children in the UK today are consuming almost three times more sugar than they should be according to nutritional guidelines - an alarming figure, considering the widely documented link between sugar intake and obesity. Children and young people consume a significant proportion of their daily calories at school, and for disadvantaged students this may be the only hot meal of the day. Food consumption at school does not only have a significant impact on children’s health in the here and now, but shapes future eating habits. At the same time, within the rules of the school food standards, schools are at liberty to develop their own cafeteria menus and change the offerings on a regular basis. This means that school catering services can play an essential role in helping to tackle the growing rates of obesity among school-aged children.

The solution
Based on our review of the literature and fieldwork, we decided to focus on an intervention that could help students develop self-monitoring skills. Since students’ food choices at Varndean appear to be heavily driven by their habits, the idea was that providing them with feedback about their purchases together with swap suggestions could help students to notice what they were buying and encourage them to try something different than usual.

Changing the ‘choice architecture’ to improve dietary choices have been successfully tested in a variety of settings (including one of our previous trials). Given the very busy canteen environment, we decided to trial a novel intervention outside of this setting.

We aimed to test whether providing students with feedback about their previous purchases could prompt them to reduce the sugar content of their subsequent purchases. This was done by sending students a series of personalised ‘sugar swap’ messages (e.g. messages suggesting swapping a food item for a different item with lower sugar content).

The impact
It appears that students who received the intervention made more purchases in the school canteen (they purchased about 0.6 items per week more than students who did not receive the intervention) – particularly of savoury items and swaps. This explains why we observe more calories and more sugar purchased in the intervention group (as these items also contain some sugar). We cannot tell how this affected students’ overall food consumption (as we only have data on food purchased in the canteen).

Based on the results from the limited efficacy testing, we therefore think it is likely that the intervention was successful in encouraging students to try savoury items and the suggested swaps, but not successful in reducing purchases of sweet treats.

However, as we were unable to randomly allocate students to receive the intervention, it is important to note that the differences we observe could simply be due to underlying differences between the students whose parents consented for them to participate in the study and those whose parents did not (though we try to control for these differences as much as possible).

How is the new approach being sustained?
While we think the intervention we tested has promise, further evidence would be needed before we would recommend implementation at scale.

Designing and implementing the email intervention was very labour intensive, as it required weekly data transfers of purchasing data from the school. While the process of generating the email messages was semi-automated, it still required manual checking of each message. For the intervention to be scaled, a more automated system would need to be in place.

Although not an issue in this study, if this intervention were to be scaled, all participating schools would need to have access to accurate student purchasing data.

Lessons learned
Among participating students, feedback was positive in terms of the sentiment towards both the suggested swaps and receiving feedback on their purchasing history. 

However, unfortunately there was low uptake for the intervention, as participation required written parental consent - only approximately 12% of eligible students participated. Parents were more likely to sign up if their children were younger, ate more sweet treats prior to the study commencing, and if their children were not eligible for free school meals. For future interventions using direct communication to students, it would be good to find out whether uptake could be increased (for example, by seeking consent at the start of the school year or at parents’ evenings).

We were not able to track email opening rates (the school’s email system did not allow for read-receipts to be generated), so it is unclear if and how regularly students were opening the emails. Students may not have received the intervention (or opened the emails at a time when the information would be most effective). Depending on schools’ policies on mobile phone use in school, a text-message based intervention might be able to provide more timely information. The results of the study and recommendations for next steps will be taken forward for discussion with Brighton’s Healthy Weight Programme Board.


Katie Cuming

Hannah Behrendt

Links to relevant documents

Message bank