1992PubMedRequires access

The repeatability of measurement of the ocular components.

Karla Zadnik, Donald O. Mutti, Anthony J. Adams

Open publisher page 313 citations

Abstract

Studies of the ocular components of refraction typically neglect issues of repeatability of measurement methods or analyze method comparison/repeatability data inappropriately using correlation. The authors have examined the repeatability of refractive error measures (retinoscopy, subjective refraction, and Canon R-1 autorefraction, noncycloplegic and cycloplegic), axial dimension measures (Allergan-Humphrey A-scan ultrasound), and corneoscopy (keratometry and KERA photokeratoscopy), and the agreement between different refractive error and corneal measurement methods on 40 pre-presbyopic normal adults. The authors plotted the difference versus the mean of two different measurement occasions (repeatability), or two different methods (agreement), to determine the bias (mean of the differences relative to zero) and 95% limits of agreement of each technique. The most reliable measure of refractive error was autorefraction with cycloplegia, with 95% limits of agreement of +/- 0.32 diopters. Cycloplegic autorefraction had no statistically significant bias compared to cycloplegic subjective refraction. Cycloplegic retinoscopy was the least reliable refractive error measure, with repeat measures on two separate occasions extending over 95% limits of agreement of +/- 0.95 D. Anterior chamber depth was reliable to +/- 0.29 mm, lens thickness to +/- 0.20 mm, and vitreous chamber depth to +/- 0.37 mm. Corneal curvature measures show keratometry to be more reliable (to +/- 0.87 D) than photokeratoscopy (+/- 2.02 D) with a statistically significant bias (paired t-test, P less than 0.0001) of 0.57 D flatter for photokeratoscopy.

About this research paper

What this paper is about

Studies of the ocular components of refraction typically neglect issues of repeatability of measurement methods or analyze method comparison/repeatability data inappropriately using correlation. The authors have examined the repeatability of refractive error measures (retinoscopy, subjective refraction, and Canon R-1 autorefraction, noncycloplegic and cycloplegic), axial dimension measures (Allergan-Humphrey A-scan ultrasound), and corneoscopy (keratometry and KERA photokeratoscopy), and the agreement between different refractive error and corneal measurement methods on 40 pre-presbyopic normal adults. The authors plotted the difference versus the mean of two different measurement occasions (repeatability), or two different methods (agreement), to determine the bias (mean of the differences relative to zero) and 95% limits of agreement of each technique. The most reliable measure of refractive error was autorefraction with cycloplegia, with 95% limits of agreement of +/- 0.32 diopters. Cycloplegic autorefraction had no statistically significant bias compared to cycloplegic subjective refraction. Cycloplegic retinoscopy was the least reliable refractive error measure, with repeat measures on two separate occasions extending over 95% limits of agreement of +/- 0.95 D. Anterior chamber depth was reliable to +/- 0.29 mm, lens thickness to +/- 0.20 mm, and vitreous chamber depth to +/- 0.37 mm. Corneal curvature measures show keratometry to be more reliable (to +/- 0.87 D) than photokeratoscopy (+/- 2.02 D) with a statistically significant bias (paired t-test, P less than 0.0001) of 0.57 D flatter for photokeratoscopy.

Why it matters

OpenAlex reports 313 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Studies of the ocular components of refraction typically neglect issues of repeatability of measurement methods or analyze method comparison/repeatability data inappropriately using correlation. The authors have examined the repeatability of refractive error measures (retinoscopy, subjective refraction, and Canon R-1 autorefraction, noncycloplegic and cycloplegic), axial dimension measures (Allergan-Humphrey A-scan ultrasound), and corneoscopy (keratometry and KERA photokeratoscopy), and the agreement between different refractive error and corneal measurement methods on 40 pre-presbyopic normal adults. The authors plotted the difference versus the mean of two different measurement occasions (repeatability), or two different methods (agreement), to determine the bias (mean of the differences relative to zero) and 95% limits of agreement of each technique. The most reliable measure of refractive error was autorefraction with cycloplegia, with 95% limits of agreement of +/- 0.32 diopters. Cycloplegic autorefraction had no statistically significant bias compared to cycloplegic subjective refraction. Cycloplegic retinoscopy was the least reliable refractive error measure, with repeat measures on two separate occasions extending over 95% limits of agreement of +/- 0.95 D. Anterior chamber depth was reliable to +/- 0.29 mm, lens thickness to +/- 0.20 mm, and vitreous chamber depth to +/- 0.37 mm. Corneal curvature measures show keratometry to be more reliable (to +/- 0.87 D) than photokeratoscopy (+/- 2.02 D) with a statistically significant bias (paired t-test, P less than 0.0001) of 0.57 D flatter for photokeratoscopy.

Key concepts: Retinoscopy, Repeatability, Keratometer, Cycloplegia, Dioptre, Refractive error, Limits of agreement, Medicine

Related papers

Back to paper searchBrowse research topicsOriginal source
The repeatability of measurement of the ocular components. — Research Paper | ScholarLens