Background: Coccidioidomycosis (valley fever) is caused by the inhalation of spores from Coccidioides immitis and C. Posadasi, diamorphic ascomycete fungi that grows in soils of warm, semi-arid climates of the Americas. They are the most virulent of the primary fungal pathogens of humans and other animals and are listed as select agents of bioterrorism. The endemic regions of the fungi have well-known general ecologic characteristics; however little is known about the specific ecologic niche required for Coccidioides ssp. to flourish. Assessment of risk factors is likely to be confounded by the poor characterization of exposure to pathogenic spores because of the poor under-standing of the pathogen's occurrence beyond a regional scale. Current techniques for culturing and isolating the pathogen from soil and air samples are expensive, difficult, and have poor sensitivity.
Methods: A landscape ecology approach was used in selecting the primary sampling units (90) that are based on 2000 Census blockgroups, three landscape strata, and two demographic strata. A probability proportional to size household survey design was used to screen for cases and identify potential controls by telephone. Case and control data were collected through personal interviews. Soil samples from case and control residences and air samples were collected, cultured in a BL-3 laboratory and analyzed using PCR. Canine case data from a veterinary diagnostic laboratory were also mapped and analyzed using a GIS to evaluate association between soil type and diseases.
Results: Landscape types, urbanization, and demographic composition are major factors that explain the current epidemic in southern Arizona USA. Other risk factors will also be discussed.
Discussion: Landscape ecology, GIS, and analysis techniques for cluster sampling provide powerful tools for improving assessment of risk factors of valley fever by reducing classification bias and confounding. This landscape ecological approach to exposure assessment can serve as an analytical bridge between the continuum of environmental properties and the categorical nature of epidemiological analysis. By better classifying the continuum of potential risk factors, information bias and residual confounding can be reduced. Improved predictive models will provide a basis from which to distinguish disease clusters from a background of disease that is spatially and temporally variable.
Acknowledgements: This research is made possible through funding provided by the Arizona Disease Control Research Commission and by the Centers for Disease Control and Prevention through the Association of Teachers of Preventive Medicine. This presentation made possible through additional support provided by the Mel and Enid Zuckerman College of Public Health and the Office of Arid Lands Studies, College of Agriculture and Life Sciences at the University of Arizona.