Test-retest variability of visual acuity in a busy eye clinic
Date
2016
Authors
Khantsi, B
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
OBJECTIVES
The main objective is to demonstrate and to quantify any variability between nurse
measured and doctor measured visual acuity (VA) performed on the same patient
under the same testing conditions in the out patients clinic at the St John Eye Hospital.
DESIGN
A prospective, cross-sectional study using adult patients presenting with visual acuity
of at least 6/60 in one eye as measured on an illuminated Snellen chart.
The study was performed on patients aged 18 years and older.
STUDY METHOD
Patients who have had VA testing by the nurses but have not consulted with the
doctors were randomly selected to have their VA re-tested by a doctor under the same
testing conditions as the initial testing by the nurses. Both tests were performed using
an illuminated Snellen chart.
STATISTICAL METHODS
Using a Bland-Altman analysis, a measurement error between the two observers was
determined. The difference in measurements between observers for each patient was
determined and each value was plotted on the y-axis. The mean difference in
observations was in-turn calculated and plotted on the x-axis. The degree of
disagreement with its 95% confidence interval (CI) was plotted on the graph.
VI
RESULTS
The calculated difference in Early Treatment of Diabetic Retinopathy Study (ETDRS)
letters between the observers ranged between 0 and 55 letters, which equates to 11
Snellen lines.
When the differences are averaged out, the mean difference was only 0.045 letters
with a standard deviation of 9.62 letters and the 95% confidence interval (CI) of -55
to 70.
On performing the Bland-Altman analysis, the limits of agreement between the
observers at the 95% CI were -18.79 letters (at lower end) and 19.70 letters (at the
upper end), a total of 38.49 ETDRS letters.
CONCLUSIONS
The VAs measured at the St John Eye Hospital out patients department are not
reproducible. Variability in measurement as found in the study was 38,49 ETDRS
letters, an equivalent of 7,69 Snellen lines. This is not within the accepted limit of 2
Snellen lines or 10 ETDRS letters.
Description
Registered for: MMed Ophthalmology
Division of Ophthalmology
Department of Neurosciences
University of the Witwatersrand