Related Experiment Videos
Comparison of small-area analysis techniques for estimating county-level outcomes
Haomiao Jia1, Peter Muennig, Elaine Borawski
1Department of Community Medicine, Mercer University School of Medicine, Macon, Georgia 31207, USA. haomia@yahoo.com
American Journal of Preventive Medicine
|May 29, 2004
Summary
The regression method offers the most accurate county-level health estimates when state-level data is used. Local health departments should use this method for tracking health trends.
Area of Science:
- Public Health
- Biostatistics
- Health Services Research
Background:
- County-level health data is often unavailable, necessitating state-level data for local policy.
- Small-area estimation techniques are used to infer local health needs from broader datasets.
- The optimal small-area estimation technique for health data remains undetermined.
Purpose of the Study:
- To evaluate the validity and precision of three small-area estimation techniques.
- To identify the most reliable method for estimating county-level health indicators.
Main Methods:
- Validated severe work disability measures using the Behavioral Risk Factor Surveillance System (BRFSS) and Census 2000 data.
- Applied synthetic, spatial smoothing, and regression methods to 2000 BRFSS data.
- Assessed the predictive accuracy of each method for county-level disability prevalence.
Main Results:
- The regression method demonstrated the highest validity and precision for county-level disability prevalence estimates.
- This finding holds true when using a single year of data across numerous counties.
Conclusions:
- Local health departments and policymakers should adopt the regression method for tracking county-level health trends.
- Direct estimation is an alternative only for counties with sufficient data for the specific health outcome.