Harnessing Geospatial Artificial Intelligence (GeoAI) for Environmental Epidemiology: A Narrative Review

Area of Science:

• Environmental epidemiology
• Geospatial health research
• Public health informatics

Background:

• Geospatial analysis is crucial for understanding environmental exposures and their impact on chronic diseases and behaviors.
• Emerging interest in machine learning and artificial intelligence (AI) for environmental epidemiology.

Purpose of the Study:

• To provide an overview of recent advances in geospatial analysis and AI for environmental epidemiology.
• To highlight the integration of geospatial data with machine learning and AI.

Main Methods:

• Review of novel statistical prediction methods combining geospatial analysis with machine learning and AI (GeoAI).
• Leveraging data from smartphones and wearables with global positioning systems (GPS) and other sensors for passive data collection.
• Integration with geographic information systems (GIS) for fine-scale spatial and temporal exposure assessment.

Main Results:

• GeoAI enables scalable geospatial exposure assessment in large population health databases.
• Smartphones and wearables allow for passive data collection, enhancing spatial and temporal resolution of exposure assessment.
• Applications include refining air pollution models, identifying populations vulnerable to water pollution, and using deep learning for greenspace assessment.

Conclusions:

• GeoAI offers potential for more rapid, higher-quality objective exposure measures in environmental health.
• Challenges include participant privacy, data representativeness, and validation set curation for GeoAI algorithms.
• Epidemiologists must critically assess measurement accuracy and design validity when using these new tools.