Related Experiment Videos
Accounting for the correlation between fellow eyes in regression analysis
Archives of Ophthalmology (Chicago, Ill. : 1960)
|March 1, 1992
Summary
Ophthalmology research should use advanced regression models that account for paired eye data. These methods improve statistical power and data interpretation compared to traditional approaches.
Area of Science:
- Ophthalmology
- Biostatistics
- Statistical Modeling
Background:
- Regression techniques are underutilized in ophthalmology despite statistical advances.
- Existing methods often fail to account for the correlation between fellow eyes.
Purpose of the Study:
- To review and illustrate advanced regression models for ophthalmologic research.
- To highlight the advantages of these models over traditional statistical approaches.
Main Methods:
- Applied general linear models and polychotomous logistic regression (Rosner).
- Utilized the Liang-Zeger estimating equation approach for linear and logistic regression.
- Demonstrated methods using datasets on retinitis pigmentosa visual acuity and glaucoma visual field impairment.
Main Results:
- Advanced models offer enhanced statistical power and precision.
- Regression coefficients are more interpretable.
- Models show less sensitivity to missing data compared to separate eye analyses.
Conclusions:
- Recommended increased adoption of these paired-eye regression models in ophthalmologic studies.
- Provided guidelines for selecting appropriate models based on research needs.