A hierarchical aggregate data model with spatially correlated disease rates

Area of Science:

• Epidemiology
• Biostatistics
• Spatial Analysis

Background:

• Aggregate data studies estimate exposure effects using population disease rates and survey covariates.
• Geographical variation in disease rates, particularly for rare chronic conditions like breast cancer, often exhibits spatial correlation.
• Existing aggregate data models do not fully account for this spatial component.

Purpose of the Study:

• To extend the aggregate data study design to incorporate residual spatial correlation among disease rates.
• To develop an intuitive approach for modeling spatial effects in aggregate data analysis.
• To enable correct inference regarding exposure effects in the presence of spatial autocorrelation.

Main Methods:

• Combined aggregate data regression models with Bayesian disease-mapping approaches.
• Developed a model to account for unexplained geographical variation with a spatial component.
• Utilized simulation studies to evaluate the proposed model's performance.

Main Results:

• The proposed model successfully integrates spatial correlation into aggregate data analysis.
• The approach allows for accurate inference on exposure effects even with spatial dependencies.
• Simulation studies provided insights into the model's utility and limitations.

Conclusions:

• The enhanced aggregate data model provides an intuitive framework for analyzing spatial patterns in disease rates.
• This method improves the estimation of individual-level exposure effects in epidemiological studies, especially for rare chronic diseases.
• Guidelines for the application of the proposed model are suggested based on simulation findings.