Measuring the obesogenic environment of childhood obesity

Obesity prevalence has accelerated over the last two decades and is predicted to continue to rise, bringing with it increased morbidity and mortality as well as rising dramatically health care costs. Obesogenic environments are one of the explanations for the rising prevalence. Accordingly this thesis investigates the obesogenic environment factors, as well as obesogenic behaviour factors, associated with the increased prevalence of childhood obesity, using familiar geographical techniques in novel ways. These results were then applied to develop a targeted childhood obesity prevention policy for Leeds to reduce the risk of childhood obesity in different populations. In this ecological study body mass index in Leeds for children aged 3 to 13 years old was examined to measure variations in childhood obesity. Spatial microsimulation modelling was utilised to give synthetic individual estimates of obesogenic covariates (e. g. obesogenic environment variables such as socio-economic characteristics and perceived social capital; individuals' behavioural variables such as dietary variables and physical activity levels) at the micro level. Additionally two demographic indices based on the 2001 Census were employed. The relationship between childhood obesity and the obesogenic covariates were considered at the home and school level using a combination of spatial statistical techniques. Spatial microsimulation modelling was shown to be a robust method to estimate obesogenic covariates at the micro-level. In the design of a spatial microsimulation model using a deterministic re-weighting algorithm, the input variables must be strongly correlated with the output variables to be able to accurately simulate micro-area estimates. Also this thesis has highlighted that there is considerable advantage to analysing health data at a small scale, otherwise micro-level differences are simply "averaged" away and missed. As well as showing that individuals' behaviours are important in determining risk of childhood obesity, this study adds to the increasing evidence of the existence of "obesogenic environments": features of the local environment in Leeds may affect childhood obesity by changing health behaviours. There was significant variation in childhood obesity across Leeds, with "hot spots" in both deprived and affluent areas. Further, relationships between obesogenic covariates and childhood obesity were not uniform across Leeds, highlighting the need for tailored, multifaceted public health policies that are based on locally relevant evidence.