[Application of parametric statistics to visual acuity data]

Area of Science:

• Ophthalmology
• Biostatistics
• Statistical analysis of visual acuity data

Context:

• Visual acuity measurements in clinical studies often exhibit non-normal distributions.
• Parametric statistical methods are preferred for their power but require normally distributed data.
• Tukey's power transformations offer a method to normalize skewed data.

Purpose:

• To identify the visual acuity notation most suitable for parametric statistical methods using Tukey's power transformations.
• To determine the optimal transformation exponent (λ) for normalizing visual acuity data.

Summary:

• Tukey's power transformations were applied to visual acuity data from multiple large studies (CIEMS, ETDRS, AREDS2) and original measurements.
• The optimal transformation exponent (λ) consistently favored decimal notation (λ=-1.0) over logarithmic notation (λ=0).
• Despite transformations, normality was not achieved, but decimal notation provided the closest distribution to normal.

Impact:

• Recommends using decimal notation for statistical processing and presentation of visual acuity data.
• Enhances the applicability of powerful parametric statistical methods in ophthalmological research.
• Improves the statistical rigor and interpretability of visual acuity findings in clinical studies.