Investigation of spatial clustering from individually matched case-control studies

Area of Science:

• Epidemiology
• Spatial Statistics
• Biostatistics

Background:

• Assessing spatial clustering is crucial for understanding disease etiology.
• Existing methods for spatial analysis are often not directly applicable to matched case-control study designs.
• Matched case-control studies are frequently used in epidemiology to investigate disease risk factors.

Purpose of the Study:

• To adapt existing spatial clustering theory for second-moment properties of labelled point processes to matched case-control studies.
• To introduce a novel null hypothesis for spatial analysis in matched case-control settings.
• To evaluate the performance of the proposed spatial clustering test against existing methods.

Main Methods:

• The study adapts second-moment properties of labelled point processes for spatial clustering assessment.
• A new null hypothesis is proposed: each case is a random sample from the set of itself and its k matched controls.
• The proposed test is compared with other spatial clustering tests.

Main Results:

• The adapted theory provides a valid framework for assessing spatial clustering in matched case-control studies.
• The proposed hypothesis offers a more appropriate null model for matched data compared to random sampling from the superposition of cases and controls.
• Empirical application demonstrates the utility of the method.

Conclusions:

• The proposed method effectively adapts spatial clustering analysis for matched case-control studies.
• This approach enhances the understanding of spatial disease patterns, particularly in epidemiological research.
• The study highlights the importance of appropriate null hypotheses in spatial statistical modeling for case-control data.