Mapping disease and mortality rates using empirical Bayes estimators

Area of Science:

• Epidemiology
• Biostatistics
• Geographic Information Systems (GIS) for Health

Background:

• Accurate estimation of regional disease rates is crucial for public health surveillance and resource allocation.
• Traditional methods may lack precision in areas with small populations or sparse data.
• Mapping disease patterns requires robust statistical approaches to account for spatial variations.

Purpose of the Study:

• To propose and evaluate a novel empirical Bayes estimator for regional disease rate estimation.
• To compare the performance of the new local shrinkage estimator against existing iterative methods.
• To demonstrate the application of the proposed method using infant mortality data.

Main Methods:

• Development of a local shrinkage estimator that shrinks crude disease rates towards a neighborhood average.
• Parameter estimation using the method of moments for the empirical Bayes estimator.
• Comparative analysis through Monte Carlo simulations and a real-world case study.

Main Results:

• The proposed local shrinkage estimator demonstrated improved performance in simulations compared to iterative alternatives.
• The empirical Bayes approach effectively stabilized disease rate estimates, particularly in areas with limited data.
• Application to Auckland infant mortality data provided a refined spatial disease map.

Conclusions:

• The local shrinkage empirical Bayes estimator offers a valuable tool for disease mapping and regional health analysis.
• This method enhances the reliability of disease rate estimations, aiding in targeted public health interventions.
• The findings support the use of advanced statistical modeling for understanding geographic health disparities.