A hierarchical model for spatially clustered disease rates

Area of Science:

• Biostatistics
• Epidemiology
• Spatial Analysis

Background:

• Disease maps are crucial for understanding spatial disease patterns and identifying clusters.
• Bayesian and empirical Bayes methods are established for smoothing disease rate maps.
• Current popular models incorporate both spatially correlated and unstructured random effects to capture data clustering.

Purpose of the Study:

• To propose an alternative statistical model for disease mapping.
• To replace spatially structured random effects with fixed clustering effects for specific areas.
• To develop and describe a reversible jump Markov chain Monte Carlo (RJMCMC) algorithm for posterior inference.

Main Methods:

• Development of a novel statistical model for disease rate analysis.
• Implementation of a reversible jump Markov chain Monte Carlo (RJMCMC) algorithm.
• Application of the model to the New York leukaemia dataset.

Main Results:

• The proposed model effectively analyzes regional disease rates by incorporating fixed clustering effects.
• The RJMCMC algorithm facilitates robust posterior inference for the new model.
• The model's utility is demonstrated using a well-known public health dataset.

Conclusions:

• The proposed fixed clustering effects model offers a viable alternative to traditional spatially structured random effects in disease mapping.
• This approach simplifies the analysis of spatial disease patterns and cluster identification.
• The model provides a valuable tool for epidemiological research and public health surveillance.