Dietary consumption patterns in the UK and substitution between food groups over time
The aims of the project are to understand what foods are eaten together and predict what we would eat if we made changes to our diet in response to public health interventions.
In the UK, on average, we consume too much sugar, saturated fat and not enough fruit and vegetables. This is a leading cause for rising obesity and will lead to long-term health complications. Public Health strategies are needed to improve our diets.
We are developing a tool which will estimate the long-term health impact, NHS cost savings, and health inequalities of dietary policies. The simulation will be able to investigate a broad range of policies in the UK from those that stop us buying chocolate on the work commute, to those that warn us how many cubes of sugar are in our cereal. In order to understand how these policies impact on our health it is important to understand how our whole diet would change. For example, if we wanted to eat less chocolate, would we eat more crisps or fruit?
Our project will look at data from the UK Biobank, which asked individuals to record what they ate over 24 hours on four occasions over a year. We will use the data to look at what types of food are eaten together, i.e. meat and potatoes and what food might replace each other, i.e. cakes and biscuits. Looking at diet across individuals is useful but can be affected by the other factors such as our education, culture, tastes. The UK Biobank data have questionnaires from the same individual at different times so we can look at dietary patterns within an individual over time.
The overall project will end in March 2022, and the statistical analysis of the Biobank data is anticipated to last for 4 months.
Public Health Impact
Policymakers often don't have information about the likely effects of these policies on health, life expectancy, costs to the NHS and health inequalities, and this may act as a barrier to implementing effective policies that will improve the health of the UK population. This type of information is particularly difficult to generate when these interventions affect large groups or populations, such as junk food advertising or "Change for life" TV adverts. Our statistical model will help policymakers estimate the benefits of interventions by accounting for what food substitutions individuals are likely to choose.