NIHR Obesity Policy Research Unit at UCL


Modelling the impact of obesity policy interventions across the range of deprivation


New simulated intervention techniques allow us to estimate the effect of interventions from cohort data without having trial data. While these data provide a lower level of evidence than randomised trials, careful use of causal inference techniques provide more confidence in these estimates. Importantly, these new techniques also allow us to estimate effects of interventions across the range of deprivation, and therefore assess whether an intervention will reduce inequalities and if so, by how much. Robustly assessing the impact of interventions across a range of deprivation is likely to have meaningful implications for policy. 


Nearly a third of children aged two to 15 are overweight or obese and the burden falls disproportionately on children from low socioeconomic groups. There is a clear need to develop evidence-based intervention programs to prevent overweight and obesity in children. 

Aims and methodology 

The aim of this project is to estimate the impact on childhood obesity prevalence of policy-relevant interventions relating to known childhood obesity risk factors. Six policy-relevant interventions will be simulated; these are likely to be three diet-related interventions and three interventions related to other obesity risk factors. The impact of each intervention will be assessed across the whole population, and by level of deprivation. The project will use marginal structural models to study the effects of potential interventions on a range of obesity-related risk factors, including: known early life risk factors (e.g. breastfeeding, maternal obesity, maternal smoking during pregnancy, paternal factors); specific aspects of diet; and, other risk factors, such as stress or sleep. For this work, a 

number of longitudinal cohorts will be used, depending on particular risk factor/intervention. These are likely to include: the Millennium cohort study; the ALSPAC cohort; the Southampton Women's Survey; and the Gemini study. 


  • First round modelling: June 2018 - March 2019 
  • Briefing paper of first round findings - April 2019 
  • Second round modelling: May 2019 - March 2020 
  • Briefing paper of second round findings - April 2020