Health statistics by congressional district: A foundation for political epidemiology to inform health policy

Introduction: While the US has an excellent system for producing health statistics, data are not aggregated at the fundamental political level, the US Congressional District. Failure to do so denies policymakers important information for representation of their constituents. Health statistics are recorded at the county level while congressional districts are aggregates of census blocks which are often politically gerrymandered. Most congressional districts do not follow county boundaries. Methods: We construct a system to reapportion US deaths recorded at the county level by race, gender, and cause to census blocks. We aggregate the block data to provide approximation of mortality rates for congressional districts which are ranked and mapped using GIS. Mortality rate disparities by congressional district are characterized, compared to whole county rates, and correlated with racial proportions in each district. Results: Mapping reveals that mortality rates for congressional districts vary significantly from rates for the counties from which they are extracted, as well as variation from rates of adjacent districts. Significant correlation of mortality rates with variation in racial proportions of the congressional districts is evident. Discussion: There is substantial variation in death rates across congressional districts. These data can inform public policy decisions about amounts and allocations of resources and empower individual members of congress to advocate for their constituents. This approximation of mortality by congressional district argues strong support for the recommendation of the National Committee on Vital and Statistic's to geocode ongoing data sets that feed health statistics at the census block group level.