Second-order analysis of spatial clustering for inhomogeneous populations

Area of Science:

• Epidemiology
• Spatial Statistics
• Biostatistics

Background:

• Growing interest in spatial disease clustering, particularly for rare conditions.
• Need for robust statistical methods to assess disease distribution patterns.
• Existing methods may not fully capture the nuances of rare disease spatial aggregation.

Purpose of the Study:

• To develop and present a novel approach for assessing spatial clustering of rare diseases.
• To utilize second-moment properties of labelled point processes for this assessment.
• To provide a diagnostic tool for evaluating clustering effects and their scale.

Main Methods:

• Application of labelled point process theory.
• Analysis based on second-moment properties.
• Development of a diagnostic plot for spatial clustering.
• Utilisation of Monte Carlo tests for statistical significance.

Main Results:

• A new method for assessing spatial clustering is proposed.
• A diagnostic plot is introduced to estimate the nature and scale of clustering.
• The method is applicable to stationary spatial point processes with two event types.
• Monte Carlo tests for significance are available.

Conclusions:

• The developed approach offers a statistically sound method for rare disease spatial clustering analysis.
• The diagnostic plot provides valuable insights into disease distribution patterns.
• The methodology is demonstrated using real-world data on childhood leukaemia and lymphoma.