Spatial Modeling of Human Risk of Exposure to Vector-Borne Pathogens Based on Epidemiological Versus Arthropod Vector Data

Abstract

Understanding spatial patterns of human risk of exposure to arthropod vectors and their associated pathogens is critical for targeting limited prevention, surveillance, and control resources (e.g., spatial targeting of vaccination, drug administration, or education campaigns; use of sentinel sites to monitor vector abundance; and identifying areas for most effective use of pesticides). Vector-borne disease risk can, in many cases, be modeled with high predictive accuracy by using geographic information system approaches because abundances of vectors and pathogen reservoirs often are associated with environmental factors. Spatial risk models for human exposure to vector-borne pathogens, which ideally should have high accuracy for predicting areas of elevated risk without overestimating risk coverage, can be constructed based on epidemiological data or abundance of vectors or infected vectors. We use five bacterial or viral vector-borne diseases occurring in the United States and with pathogen transmission by fleas (plague), ticks (Lyme disease and tularemia), or mosquitoes (dengue and West Nile virus disease) to 1) examine how spatial risk of human exposure to vector-borne pathogens typically is presented to the public health community and public and 2) evaluate the utility of basing spatial risk models on epidemiological data relative to data for arthropod vectors or infected vectors. Recommended future directions for vector-borne disease risk modeling include development of subcounty level spatial risk models combining epidemiological and vector data and the use of simulation or analytical models to assess critical vector abundance thresholds required for enzootic pathogen maintenance.