Adjusting Moran's I for population density

Area of Science:

• Epidemiology
• Spatial Statistics
• Biostatistics

Background:

• Traditional disease cluster analysis often overlooks variations in population density.
• Existing methods may lack power in detecting clusters in areas with heterogeneous populations.
• Accurate spatial analysis is crucial for effective public health interventions.

Purpose of the Study:

• To introduce two novel statistics, Ipop and Ipop*, designed to adjust for population density in disease cluster detection.
• To evaluate the performance of these new statistics compared to existing methods.
• To enhance the sensitivity of spatial scan statistics for identifying disease hotspots.

Main Methods:

• Derivation of two new spatial statistics, Ipop and Ipop*, adjusting Moran's I for population density.
• Simulation studies using Lyme disease data from Georgia to assess statistical power.
• Comparison of the proposed statistics against current methodologies for cluster identification.

Main Results:

• The new Ipop and Ipop* statistics demonstrated increased power in detecting disease clusters in simulations.
• These statistics effectively account for varying population densities across different geographical areas.
• Consideration of both spatial patterns and non-binomial variance improved cluster detection efficacy.

Conclusions:

• Ipop and Ipop* offer a more robust approach to spatial epidemiology by incorporating population density.
• These statistics can lead to more accurate identification of disease clusters, aiding public health planning.
• The findings suggest a significant advancement in the statistical toolkit for spatial disease surveillance.