Testing equality of proportions for correlated binary data in ophthalmologic studies

Area of Science:

• Ophthalmology
• Biostatistics
• Statistical Methods

Background:

• Bilateral correlated data are common in ophthalmologic studies, involving measurements from paired organs like eyes.
• Existing methods, such as the adjusted chi-square approach, have limitations for analyzing such data.

Purpose of the Study:

• To investigate and derive alternative statistical testing procedures for analyzing bilateral correlated data in ophthalmology.
• To compare the performance of proposed methods with existing approaches.

Main Methods:

• Derivation of three alternative statistical testing procedures.
• Simulation studies to evaluate Type I error control and statistical power.
• Application of methods to illustrative examples from ophthalmologic studies.

Main Results:

• The score testing procedure demonstrated satisfactory Type I error control in simulations.
• The score testing procedure generally exhibited higher statistical power compared to other methods.
• Proposed methods were effectively illustrated using real-world ophthalmologic data.

Conclusions:

• The score testing procedure is a robust and powerful method for analyzing bilateral correlated data in ophthalmology.
• The study recommends the score testing procedure for its superior performance characteristics.
• The findings provide valuable statistical tools for researchers in ophthalmology and related fields.