Optimal Spatial Design for Environmental Health Research

Duke University, Durham NC

Investigators

Linked publications & trials

Abstract

[unreadable] DESCRIPTION (provided by applicant): Recent advances in the theory and application of spatial statistics, as well as new kinds of data management techniques and analytical approaches made possible through the development of GIS software, hold tremendous promise for ex ante enhancing study design in environmental health sciences. [unreadable] In this R21 proposal, we seek to devise a general approach for collecting environmental and biological samples that incorporates optimal spatial design. Specific aims are to: [unreadable] 1. Develop a spatially-based sampling procedure that improves upon traditional random or categorical sampling approaches. [unreadable] 2. Incorporate optimality into the procedure with regard to information gain concerning the nature of the relationship between contaminant and explanatory variables as well as the spatial pattern in the contaminant levels. [unreadable] 3. Develop a system for sampling sequentially and adaptively to take full advantage of the information made available through the sampling process. [unreadable] 4. Demonstrate innovative spatial design approaches by collecting environmental samples in the field. [unreadable] 5. Assess the importance of the spatial resolution at which analyses and sampling protocols are undertaken (i.e., small changes in location may in fact equate to large changes in exposure). [unreadable] We will use mercury as a prototype contaminant for exploring how to advance these new methods. We are concerned primarily with developing a general framework for applying these methods to optimized sampling design across a wide variety of contaminants. This research will help researchers improve estimates of exposures, sample more strategically, update models more efficiently, and draw better and more meaningful links between environmental contaminants and health endpoints. [unreadable] [unreadable]