Statistical strategies to assess reliability in ophthalmology

Area of Science:

• Ophthalmic research
• Statistical methodology
• Measurement science

Background:

• Reliability of measurements and measurers is crucial for trusting recorded data in ophthalmology.
• Current statistical techniques in ophthalmic literature often fail to appropriately assess measurement or measurer reliability.
• Inappropriate statistical methods hinder the evaluation of inter-measurement or inter-measurer consistency.

Purpose of the Study:

• To review statistical techniques used in ophthalmic reliability studies.
• To identify and describe appropriate statistical methods for assessing reliability with both continuous and categorical data.
• To highlight and discourage the use of inappropriate statistical methods commonly found in ophthalmic literature.

Main Methods:

• Systematic review of statistical techniques employed in ophthalmic reliability studies.
• Analysis of methods for continuous data reliability assessment.
• Analysis of methods for categorical data reliability assessment.

Main Results:

• Many studies in ophthalmology utilize statistically inappropriate methods for reliability analysis.
• Specific appropriate statistical techniques are identified for continuous and categorical data.
• Commonly used but inappropriate methods are highlighted for their limitations in reliability assessment.

Conclusions:

• There is a significant need for improved statistical practices in assessing measurement and measurer reliability in ophthalmology.
• Adoption of appropriate statistical methods will enhance the validity and trustworthiness of ophthalmic research.
• Discouraging the use of inadequate techniques is essential for advancing the field.