Disease models implicit in statistical tests of disease clustering

Area of Science:

• * Public Health
• * Biostatistics
• * Spatial Epidemiology

Background:

• * Health departments face increasing disease cluster allegations.
• * Selecting appropriate statistical methods for spatial analysis is challenging.
• * Existing methods often have differing, implicit disease models.

Purpose of the Study:

• * To review issues in statistical analysis of spatial disease patterns.
• * To describe recently proposed methods for detecting increased disease rates.
• * To provide a basis for comparing clustering methods using explicit disease models.

Main Methods:

• * Review of statistical methods for spatial disease pattern analysis.
• * Explicit description of underlying disease models for each method.
• * Comparison of clustering methods based on their disease models.

Main Results:

• * Identified challenges in comparing spatial analysis methods due to differing disease models.
• * Described several novel statistical methods for disease cluster detection.
• * Demonstrated how explicit disease models aid in method comparison.

Conclusions:

• * Explicitly defining disease models is crucial for comparing spatial clustering methods.
• * Understanding underlying models improves the selection of appropriate statistical tools for public health investigations.
• * This work provides a framework for evaluating and choosing methods for disease surveillance.