Bayesian Models to Generate Small Area Estimates of Population Health: Tutorial for Using Rate Stabilizing Tools and...

Area of Science:

• Public Health Data Analysis
• Geospatial Statistics
• Health Informatics

Background:

• High-quality local-level population health data are crucial for community health initiatives.
• Traditional small area estimation methods, often using Bayesian techniques, are computationally complex and inaccessible to many public health professionals.
• There is a need for accessible tools to generate reliable small area estimates.

Purpose of the Study:

• To introduce and demonstrate two user-friendly tools, the Rate Stabilizing Toolbox (RSTbx) and the Rate Stabilizing Tool (RSTr), for calculating small area estimates.
• To showcase the benefits of these tools in improving the reliability and utility of local health data.
• To provide a tutorial for public health professionals on using these tools for data analysis.

Main Methods:

• Development of the Rate Stabilizing Toolbox (RSTbx) as an ArcGIS Pro plugin.
• Development of the Rate Stabilizing Tool (RSTr) as an R package.
• Demonstration of tool usage with census tract-level mortality data from North Carolina and hospitalization data from Rhode Island.

Main Results:

• The tools decrease the number of geographic units with suppressed estimates.
• Users can flexibly set thresholds for statistical reliability.
• Credible intervals generated by the tools aid in identifying statistically significant differences between geographic units.
• Both tools include built-in age-standardization capabilities.

Conclusions:

• The Rate Stabilizing Toolbox and Rate Stabilizing Tool for R are powerful, accessible resources for generating high-quality local-level health data.
• These tools can inform public health programs and tailor health promotion activities to specific community needs.
• The tools facilitate the enhancement of community health across the country through improved data analysis.