Spatial downscaling of multivariate disease risk.

Area of Science:

• Spatial epidemiology
• Geostatistics
• Public health analytics

Background:

• Accurate disease risk assessment requires downscaling areal health data to finer resolutions.
• Understanding spatial disease patterns aids in identifying risk factors and developing targeted interventions.
• Existing methods often struggle with population heterogeneity and varying spatial unit characteristics.

Purpose of the Study:

• Introduce Area-to-Area (ATA) and Area-to-Point (ATP) Poisson cokriging for downscaling spatial disease risks.
• Address challenges including inter-disease correlation, population heterogeneity, and spatial unit variability.
• Enhance multivariate disease mapping for improved public health insights.

Main Methods:

• Developed and applied ATA and ATP Poisson cokriging methodologies.
• Incorporated correlation between multiple diseases.
• Adjusted for population heterogeneity and varying spatial entity shapes/sizes.
• Validated through simulation studies and application to COVID-19 and asthma data in Bogota, Colombia.

Main Results:

• ATA and ATP Poisson cokriging outperformed univariate methods in simulations.
• Achieved lower mean squared prediction errors and preserved small-scale spatial variations.
• Revealed detailed disease hotspots/coldspots and refined COVID-19 risk estimates by leveraging asthma correlation.
• Demonstrated improved small-area estimation and spatial pattern understanding.

Conclusions:

• The proposed ATA and ATP Poisson cokriging methods offer significant advantages for downscaling spatial disease risks.
• These methods enable more accurate risk assessment and enhanced understanding of multivariate disease patterns.
• Provide valuable insights for targeted public health interventions and resource allocation.