Incorporating small-area estimation into mediation analyses with areal datasets.

Area of Science:

• Epidemiology
• Biostatistics
• Geospatial Health Analysis

Background:

• Mediation analyses with areal datasets are crucial for understanding geographic variations in health outcomes.
• Existing methods like Calculation before mediation (C-BM) and Small-area estimation before mediation (SAE-BM) have limitations impacting inference accuracy.
• These limitations can lead to flawed conclusions regarding mediation effects in geographically referenced health data.

Purpose of the Study:

• To introduce and evaluate a novel method, Small-area estimation within mediation (SAE-WM), for mediation analysis with areal datasets.
• To demonstrate the advantages of SAE-WM over traditional C-BM and SAE-BM approaches through simulation studies.
• To apply the SAE-WM method to a real-world scenario investigating colorectal cancer incidence in Iowa.

Main Methods:

• Development of the Small-area estimation within mediation (SAE-WM) approach, integrating Bayesian small-area estimation into mediation outcome models.
• Conducting a simulation study to compare the performance of SAE-WM against C-BM and SAE-BM.
• Application of SAE-WM to analyze the mediation effect of healthcare access on the relationship between socioeconomic environment and colorectal cancer incidence at the ZIP code level.

Main Results:

• The simulation study highlights the superior precision of the SAE-WM method in estimating mediation effects compared to C-BM and SAE-BM.
• SAE-WM effectively addresses the limitations inherent in pre-mediation calculation or estimation approaches.
• The application demonstrates SAE-WM's utility in real-world epidemiological research, providing more reliable mediation effect estimates.

Conclusions:

• The Small-area estimation within mediation (SAE-WM) approach offers a more precise and reliable method for conducting mediation analyses with areal datasets.
• SAE-WM overcomes significant limitations of previous methods, leading to improved inference and more robust conclusions in geographic health studies.
• This novel method has broad applicability in public health research for understanding complex relationships between environmental factors and health outcomes.