Landscape-epidemiological study design to investigate an environmentally based disease

J Expo Sci Environ Epidemiol. Mar-Apr 2011;21(2):197-211. doi: 10.1038/jes.2009.67. Epub 2010 Mar 3.

Abstract

Cost-effective approaches for identifying and enrolling subjects in community-based epidemiological studies face many challenges. Additional challenges arise when a neighborhood scale of analysis is required to distinguish between individual- and group-level risk factors with strong environmental determinants. A stratified, two-stage, cross-sectional, address-based telephone survey of Greater Tucson, Arizona, was conducted in 2002-2003. Subjects were recruited from direct marketing data at neighborhood resolution using a geographic information system (GIS). Three geomorphic strata were divided into two demographic units. Households were randomly selected within census block groups, selected using the probability proportional to size technique. Purchased direct marketing lists represented 45.2% of Census 2000 households in the surveyed block groups. Survey design effect (1.6) on coccidioidomycosis prevalence (88 per 100,000 per year) was substantially reduced in four of the six strata (0.3-0.9). Race-ethnicity was more robust than age and gender to compensate for significant selection bias using poststratification. Clustered, address-based telephone surveys provide a cost-effective, valid method for recruiting populations from address-based lists using a GIS to design surveys and population survey statistical methods for analysis. Landscape ecology provides effective methods for identifying scales of analysis and units for stratification that will improve sampling efficiency when environmental variables of interest are strong predictors.

Publication types

  • Research Support, Non-U.S. Gov't
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Arizona / epidemiology
  • Coccidioidomycosis / diagnosis
  • Coccidioidomycosis / epidemiology*
  • Coccidioidomycosis / ethnology
  • Coccidioidomycosis / transmission
  • Cross-Sectional Studies
  • Demography
  • Environmental Exposure / analysis*
  • Epidemiologic Studies
  • Female
  • Geographic Information Systems*
  • Humans
  • Male
  • Risk Factors
  • Telephone