ABSTRACT
When investigating the associations between human exposure to air pollution and various adverse health outcomes, the majority of past epidemiologic and biostatistical studies have relied on the use of ambient air pollution concentrations measured from a set of stationary air pollution monitors located across a study region to estimate exposure. Ambient air pollutant levels can vary within cities, which can introduce exposure misclassification when a central monitor or small number of monitors are used to estimate exposure for a health-based study. The name “land use regression” is a bit of a misnomer in that it only partially covers the type of information incorporated; other variables are often included such as traffic data, weather and other factors. The land use regression (LUR) method uses ambient measures of air pollution as a dependent variable and land use and the other factors as independent variables in a regression model, often through a geographic information system (GIS).
