BACKGROUND: Campylobacter causes an estimated 1.3 million illnesses in the United States annually and is the second most common infection reported to the Foodborne Diseases Active Surveillance Network (FoodNet). We developed an analytical method to evaluate census tract-level analyses to assess the poorly understood relationships between Campylobacter infection rates and demographic and environmental factors.
METHODS: We examined Campylobactercases reported to FoodNet, a population-based active surveillance system conducted in 10 U.S. sites during 2010–2014. We utilized 2010 U.S. census data for race, income, education, and population density, and Homeland Security data for enumeration and classification of farm types. Negative binominal regression (NBR) was conducted to select variables for inclusion in geographically-weighted regression (GWR) models that adjust for location. We conducted cluster analyses to summarize GWR model coefficients to create site-specific risk factor profiles.
RESULTS: Factors significantly associated with Campylobacter incidence varied by site, resulting in selection of different variables for each site-specific GWR model. A total of 52 risk-profile areas were identified with an average of 5 areas (range 4–8) in each site. Each risk-profile area quantified where factors were more or less influential. Among the 10 sites increased Campylobacterincidence adjusted by census tract was significantly associated with an increased percentage of population with college education (7/10 sites), higher median income (4/10), and residing in an area of higher population density (4/10). Farm related variables were also associated with increased Campylobacter incidence in most sites (7/10 sites).
CONCLUSIONS: The incidence of Campylobacter infection varies geographically and our methods provide insights into region-specific factors associated with infection. The relationship between Campylobacter incidence and demographic and environmental factors varied by strength and region within a site. Census tract-level analyses that link multiple data sources can help characterize subpopulations at higher risk for Campylobacter infection and identify areas for targeted studies.