Use of satellite imagery to identify a target population for recruitment of households within a large rural tribal area

摘要

Introduction: Study recruitment is increasingly challenging. In rural and tribal communities, which may lack traditional street addresses. Strategies are needed to identify and randomly select participants. The Hopi Environmental Health Project is seeking recruitment of households representative of housing type, fuel source and village. Hopi lands are >1.5 million acres, with 12 villages. Methods: Census block vectors for tribal lands were overlaid onto satellite imagery in Google Earth Pro 7.1. Potential household structures were marked with a geotag to record coordinates and an identifier describing block number, housing type, and region. Metadata were copied to an associated .KLM file and included block number, housing type, region, latitude, and longitude. Recruitment goals were set to be proportional to population size, based on the number of houses identified within village vectors. 'Pins' were randomized by village/region; latitude and longitude were used to generate web-addresses which were used by Hopi to locate dwellings and print maps. Results: The process identified 2512 potential 'houses' within 400 block vectors. This number is similar to US Census 2010. Initial recruitment of the first 30 target households identified several issues: some 'houses' were multi-family dwellings, or vacant, outbuildings of an occupied house, public building or a kiva, all of which were ineligible. Recruiters are tracking all households for eligibility; reasons for ineligibility are type of structure, probable fuel type, and response. Discussion: Some researchers suggest that a random recruitment process will not be effective on native lands. Our approach provides a baseline list that is "field-truthed" prior to recruitment. Recruitment statistics generated using this approach will provide insight to response rates by region and housing type. The Hopi project provides an opportunity to evaluate the described approach.