The impact of childhood obesity on health and health service use: an instrumental variable approach

In the following paper we estimate the impact of obesity in childhood on health and health service use in England using instrumental variables. We use data on children and adolescents aged 3-18 years old from fifteen rounds of the Health Survey for England (1998-2012), which has measures of self-assessed health, primary care use, prescribed medication use, and nurse-measured height and weight. We use instruments for child obesity using genetic variation in weight. We detect a few potential issues with the validity of the instrument; however further testing does not suggest that this has an effect on our results. We find that obesity has a statistically significant and negative impact on self-rated health and a positive impact on health service use in girls, boys, younger children (aged 3-10) and adolescents (aged 11-18). We detect significant endogeneity, which suggest that previous studies underestimate the impact of childhood obesity on health and health service use. For example, obesity is associated with and increased probability of doctor utilisation of 2%, but the IV results show that obesity increase the probability of use by 10%. This suggests that obesity has consequences for health and health service use when the children are still young.