Novel metrics that quantify each Census unit's contribution to overall disparity between groups

A substantial literature has demonstrated disparities between demographic groups in various measures of health or environmental exposure, sometimes based on surveys of individuals. However, much analysis in public or environmental health relies on Census data -- demographic data and health or exposure data by Census unit, not by individual. Such data can be used to estimate overall disparity between groups in a given health or exposure metric. To date, many analysts have relied on GIS visualizations or spatial analysis to identify hotspots or clusters, but these methods do not characterize each location in terms familiar to risk analysts or epidemiologists. In particular, it would be useful to quantify how much each Census unit contributes to the overall disparity between groups. This would provide a way to set priorities across locations using metrics that can be interpreted in terms of relative risk or excess population risk (the number of cases of disease in a group in excess of what would be expected if the group's risk were as low as the rest of the population's risk).

The US Environmental Protection Agency (EPA) recently launched a new screening and mapping tool for environmental justice (EJ), called EJSCREEN. Development of EJSCREEN included construction of an EJ Index to quantify the combination of environmental and demographic indicators for each location. Several methods of combining demographic and health or exposure data were explored, and a class of indicators can now be described to quantify and localize drivers of disparity or inequality.

Indicators are presented that quantify how much each Census unit contributes to overall excess population risk, to overall relative risk, or to other widely understood metrics of disparity. Formulas for these new metrics are presented, as well as their interpretation and practical applicability to public health, environmental justice, and disparity research.