Mapping health data: improved privacy protection with donut method geomasking.

Area of Science:

• Spatial epidemiology
• Geographic Information Systems (GIS)
• Public Health Informatics

Background:

• Protecting patient privacy in health data mapping is crucial for spatial surveillance and outbreak identification.
• Existing methods like random displacement and aggregation present challenges in balancing privacy and spatial resolution.
• The need for advanced geomasking techniques that guarantee a minimum level of geoprivacy is evident.

Purpose of the Study:

• To introduce and evaluate a novel adaptive geomasking technique, the donut method.
• To compare the donut method against random perturbation and aggregation for privacy protection and cluster detection.
• To assess the effectiveness of the donut method in maintaining spatial resolution while enhancing geoprivacy.

Main Methods:

• The donut method involves relocating geocoded addresses randomly within a specified minimum and maximum distance.
• Multiple disease field simulations were masked using the donut method, random perturbation, and aggregation.
• Performance was evaluated using measures of privacy protection and cluster detection accuracy across various parameters.

Main Results:

• Both the donut method and random perturbation outperformed aggregation in cluster detection.
• The donut method demonstrated at least a 42.7% improvement in geoprivacy measures compared to random perturbation.
• Cluster detection performance with the donut method was only marginally lower (less than 4.8%) than random perturbation.

Conclusions:

• The donut method offers a significant enhancement in geoprivacy protection for health data.
• This technique achieves a high level of privacy with a minimal reduction in the ability to detect disease clusters.
• The donut method is particularly effective in safeguarding geoprivacy in high-risk areas.