Dazed and confused: how map projections affect disease map analysis and perception. An echo from GeoVet2019.

Area of Science:

• Spatial epidemiology
• Geographic Information Systems (GIS)
• Public health informatics

Background:

• Disease maps are crucial for public health and spatial epidemiology.
• Map projections, used to represent 3D data on 2D surfaces, can introduce bias.
• The impact of map projection bias on spatial analysis and disease mapping is under-recognized.

Purpose of the Study:

• To demonstrate the influence of map projections on spatial analysis and disease maps.
• To highlight the potential for bias in public health spatial data due to projection choices.

Main Methods:

• Applied various map projections (Lambert, Mercator, Robinson) to case study areas: Israel, North Carolina, and Southern Ontario.
• Assessed the effects of projection changes on distance measures, area calculations, statistical tests (Moran's I), and disease cluster detection.

Main Results:

• Distance measurements in Israel varied by up to 30% with different projections.
• Areal size calculations for Southern Ontario showed variations of nearly 95%.
• Statistical significance (Moran's I) and disease cluster patterns (North Carolina) were altered by projection changes.

Conclusions:

• Visual and analytical bias in disease mapping due to map projections is unavoidable.
• Results underscore the importance of recognizing and reporting the specific geographic projection used for disease maps and spatial analyses.
• Using unprojected geographic coordinates is recommended to prevent analytical bias in spatial epidemiology.