Investigating effect of air pollution on health via Spatial-Resolution Varying Coefficient Models

We focus on observational studies of environmental epidemiology where the goal is to estimate the effect of an exposure variable, such as air pollution, on a health outcome using spatial area-level data. We describe a novel framework introduced in [1] seeking estimation of the exposure effect at different spatial resolution and show how these methods behave under different types of conditionally autore- gressive spatial models defined by the user. We illustrate the methods in a study on the association between COVID-19 mortality and air pollution.