GIScience is the science of collecting, analyzing, and theorizing about geographic information through GISs and geospatial analysis.
There has been increased attention in recent years to the public health applications of GIS, GIScience, and geospatial analysis (e.g., Melnick, 2001; Cromley and McLafferty, 2002; Cromley, 2003). One of the most useful applications of GIS in the public health arena is as a component of exposure and dose assessment. Ideally, exposure to a potential toxin, environmental contaminant, or pollutant would be measured directly at the individual level through monitors that the individual would carry or wear, allowing direct measurement of exposure and subsequent calculation of exposure-response curves. However, such studies are extremely costly, inconvenient, and infrequently carried out for large populations. Instead, exposures may be estimated based on models, and since exposure frequently varies spatially and temporally, a combination of GIS and spatial/temporal models has become an indispensable component of exposure assessment (Nuckols et al., 2004).
A GIS involves the merging of spatially based data—coordinates corresponding to latitude and longitude—with a graphical user interface (GUI). GISs that use data from remote sensing instruments such as aerial photographs and satellite images have become indispensable tools in the development of causal models linking environmental factors and both infectious and noninfectious disease. The spatially referenced database in an epidemiological context usually consists of geocoded (i.e., geo-spatially located) health information, such as the residential locations of people who have contracted a specific cancer, the location of traffic fatalities, or the location of incident cases of myocardial infarction. These data are then superimposed on other data layers, usually geocoded to the same unit as the health data.
Unfortunately, the power of GIS is not always realized in public health applications, or it is misused, because of a lack of understanding of the underlying geographic principles. Just as a person who learns to use statistical software (e.g., STATA, SAS, or SPSS) does not necessarily understand statistics, learning to use GIS software (e.g., ArcGIS, ArcView, MapInfo) does not necessarily ensure an understanding of the underlying principles of geospatial analysis.
A GIS is a powerful tool for the analysis of relationships, including causal relationships, between a broad range of measurable variables from the natural sciences—climatic and weather conditions, surface water characteristics, vegetation and land cover, soil geochemistry, and many others—and public health. It is thus a tool and a set of concepts that bring the