You are here:
Modeling Human Exposure Risk to Nontuberculous Mycobacteria in Central North Carolina
Citation:
Murphy, M., K. Messier, AND E. D. HILBORN. Modeling Human Exposure Risk to Nontuberculous Mycobacteria in Central North Carolina. Presented at ESRI International User Conference, San Diego, CA, July 11 - 15, 2011.
Impact/Purpose:
The US-EPA Environmental Public Health Division is investigating reports of NTM isolations during 2006 -2011 among residents of three central North Carolina counties to determine environmental risk factors for NTM.
Description:
Nontuberculous mycobacteria (NTM) are a broad group of soil-and water-borne bacteria. Some species are pathogenic and may cause serious infections in the lungs, soft tissues, bones and skin. Infections in humans are associated with environmental exposures to contaminated soil, aerosols and a variety of water sources, including public drinking water. The US-EPA Environmental Public Health Division is investigating reports of NTM isolations during 2006 -2011 among residents of three central North Carolina counties to determine environmental risk factors for NTM. Residents known to have NTM isolation will be linked to residential parcels. The parcels will then be characterized based upon environmental risk factors that may put residents at risk of exposure to NTM: public versus private drinking water source; proximity to hydric and acidic soils, and wetlands. Data and Methods NTM isolation reports were acquired in collaboration with the North Central Tuberculosis Control Region and geocoded using a parcel-based composite address locator. Block-level Census 2010 data were downloaded from the US Census Bureau and aggregated to cadastral and environmental boundaries using a dasymetric mapping approach. Acidic and hydric soil features were extracted from the Soil Survey Geographic Database (SSURGO) using USDA's Soil Data Viewer. Public drinking water supply service areas were created using cadastral data and water distribution lines. Additional data were acquired from the National Wetlands Inventory (NWI). Environmental risk will be modeled as a function of distance between parcel centroids and environmental risk factor variables. Four metrics of interest are: 1) Euclidean distance to environmental risk variable boundaries. 2) The total area of an environmental risk variable within a radial buffer. 3) The percentage area of an environmental risk variable within a radial buffer. 4) The percentage of all combined environmental risk variables within a radial buffer. Metrics 2 -4 will rely on univariate regression analysis of nested distance hyperparameters to assist in selecting variables for the overall predictive model. We will evaluate the utility of these 4 methods in logistic models to estimate risk of human NTM exposure. This abstract does not necessarily reflect U.S. EPA policy.