OBESITY, SPATIAL PROPENSITY

Propensity score methods are applied to observational studies to assess causal impacts. We consider an ecological (area) application of propensity scores for continuous exposures, namely measures of obesogenicity and urban sprawl in US counties, with obesity prevalence as the outcome. Reflecting the spatial application, adjustment is made for interference (spatial overspill) in impacts of the exposures. We then seek to estimate the effect of obesogenicity and sprawl, adjusting for confounders (e.g. poverty) via local and overspill propensity scores. To this end, we develop parsimonious causal outcome models incorporating propensity scores. Inter alia, we seek to obtain a true measure of the impact on obesity of obesogenic environments, assessing the contribution of spatial overspill to total environmental impacts.