Beyond absolute space: modeling disease dispersion and reactive actions from a multi-spatialization perspective

Area of Science:

• Geographic Information Science (GIScience)
• Public Health
• Spatial Epidemiology

Background:

• Traditional GIScience approaches often use absolute spatialization, which may oversimplify complex dynamic phenomena like disease transmission.
• There is a growing need to incorporate multiple spatial perspectives to better understand health risks.

Purpose of the Study:

• To investigate health risks by considering multiple spatializations: relational, relative, and mental space.
• To estimate health risks associated with influenza-like illness symptoms in different spatial contexts.
• To assess the influence of perceived health threats on individuals' reactive actions.

Main Methods:

• Utilized a multi-spatialization framework including relational (home, workplaces), relative (serviceplaces), and mental (perception) spaces.
• Employed a two-layer Bayesian network model to analyze health risks and disease transmission.
• Applied the SHAP (SHapley Additive exPlanations) model to understand the impact of perceived threats on actions.

Main Results:

• Relational space remains crucial for disease transmission, as confirmed by findings.
• Relative space (serviceplaces) presented significant health risks, often exceeding those in relational space, with variations across serviceplace types.
• Perceived health threats in mental space were found to be effective motivators for reactive health behaviors.

Conclusions:

• A multi-spatialization approach provides a comprehensive view of health risks, integrating proximity, social relations, relative contacts, and perception.
• This framework enhances the understanding of dynamic geographical phenomena beyond traditional spatial analysis.
• The study's methodology can be extended to analyze various geographical phenomena in both theoretical and empirical contexts.