Computer-Generated Dot Maps as an Epidemiologic Tool: Investigating an Outbreak of Toxoplasmosis

Epidemiologists have traditionally used maps to examine the spatial distribution of disease incidence.Computer-generated dot maps have facilitated identification of case clusters (1), formulation of hypotheses about the source of infection or spatially distributed risk factors (2,3), and analysis of data.Because most populations are served by identifiable water systems, waterborne disease outbreaks lend themselves to being plotted on dot maps (1,4-6).We describe an automated address-matching and base map system in a geographic information system structure used to assess information related to an outbreak of toxoplasmosis associated with a municipal water system.The Capital Regional District (population 321,585) is located on Vancouver Island, British Columbia, Canada.The district includes the City of Victoria, surrounding municipalities and districts, and the Gulf Islands.In March 1995, an outbreak of toxoplasmosis was suspected when 15 residents of this district were identified as having acute infection with Toxoplasma gondii.Investigation of these and subsequent cases confirmed an outbreak but identified no common food, beverage, or event source.A hand-drawn dot map of the initial 47 cases showed clustering in the Greater Victoria area.The municipal water supply system was considered a possible explanation for this spatial distribution.To examine this hypothesis, computer-based geographic mapping was used to study the distribution of all outbreak-related acute cases and data collected from a populationbased serologic screening program to detect T. gondii infection in women who were or had been pregnant (7). Classification of Cases and the Water Distribution SystemsPatients were classified as having acute, equivocal, or nonacute cases or as never infected on the basis of serologic tests performed at the Provincial Laboratory (British Columbia Centre for Disease Control) and the Toxoplasma Serology Laboratory, Research Institute, Palo Alto Medical Foundation (8-11).Cases were further classified on the basis of clinical symptoms, outbreak relatedness, patient's residence, and pregnancy status (pregnant, nonpregnant) (7).