Related Experiment Videos
Comparison of methods to detect visual field progression in glaucoma
K Nouri-Mahdavi1, L Brigatti, M Weitzman
1Department of Ophthalmology and Visual Science, Yale University, New Haven, Connecticut, USA.
Ophthalmology
|August 1, 1997
Summary
New statistical methods, specifically multivariate regression analyses with fixed effects on panel data, can accurately assess visual field changes over time in glaucoma patients. These advanced techniques show good agreement with expert human observers, improving glaucoma progression monitoring.
Area of Science:
- Ophthalmology
- Biostatistics
Background:
- Evaluating visual field progression in open-angle glaucoma is crucial for patient management.
- Traditional methods may not fully capture the nuances of visual field changes over time.
Purpose of the Study:
- To develop and compare novel statistical approaches for analyzing visual field series trends.
- To assess the agreement of these statistical methods with human expert interpretation.
Main Methods:
- Retrospective analysis of visual field data from 83 open-angle glaucoma patients.
- Application of pointwise univariate regression, glaucoma change analysis, univariate regression on indices, and multivariate regression analyses (pointwise and clusterwise) with fixed effects.
- Comparison of statistical method agreement with three experienced observers using Cohen's weighted kappa.
Main Results:
- Multivariate regression analyses with fixed effects demonstrated good agreement with human observers (kappa = 0.52-0.55).
- These methods showed superior agreement compared to glaucoma change analysis (kappa = 0.41).
- Univariate regression analysis on visual field indices was less effective in detecting progression.
Conclusions:
- Multivariate regression analyses with fixed effects on panel data represent a robust statistical model for evaluating visual field series.
- This approach offers a reliable and objective tool for assessing glaucoma progression, complementing expert clinical judgment.