Epidemiology using a GIS: The need for caution

Abstract This paper argues that a more cautionary and critical approach needs to be adopted in applied Geographic Information System (GIS) research. The substantive area for discussion lies in epidemiology but practitioners from other areas of application may identify with the mainpoints. Health is an emotive area. Researchers must base their work on clean data and present clear results based on theoretically sound and applicable techniques. Only then can their findings and recommendations for either health care service planning or disease control be truly appropriate. Many of the issues discussed here are not inherent to GIS research but merely reflect the kinds of difficulties researchers usually encounter in their work. The present euphoria surrounding GIS provides a seed-bed where basic mistakes could be generated which could lead to erroneous findings and policy recommendations. Such events would deflect attention away from the relevant research currently being pursued and the true potential which GIS seems to offer the scientific research community. Garbage in, garbage out must be avoided and the integrity of our work secured. There must be respect for problems which arise in three areas of a research project: input, analysis and output, or as they are referred to here as data issues, statistical issues, and mapping issues.