Purpose: The purpose of this paper is to determine the repeatability of visual acuity measurement in keratoconus and to describe the impact of measurement repeatability on sample size.
Methods: Approximately 10% of a 1209 patient sample in the Collaborative Longitudinal Evaluation of Keratoconus (CLEK) Study were selected randomly for a Repeat CLEK Study Visit. Patients were tested at the 15 CLEK Participating Clinics. The test-retest sample consisted of 1 34 keratoconus patients who met the entry criteria of the CLEK Study. High and low contrast Bailey-Lovie visual acuity was measured with the patient's habitual visual correction (entrance visual acuity monocularly and binocularly), and with the best correction monocularly (either the patient's rigid contact lens and over-refraction, or with a CLEK Study trial lens and appropriate over-refraction) at two visits separated by a median of 90 days (range 22 to 268 days).
Results: The mean absolute differences between the number of letters correct at the two visits ranged from a low of 3.24 +/- 3.1 for entrance high contrast binocular acuity to a high of 5.48 +/- 5.1 for best corrected low contrast monocular acuity. Intraclass correlation coefficients ranged from 0.757 to 0.853. The visual acuity score was somewhat higher at the Repeat Visit than at the Baseline Visit when the examiners were different between visits.
Conclusions: Given the variable vision reported by keratoconus patients, visual acuity in this sample was very repeatable. Repeatability was slightly poorer when different examiners tested visual acuity at the Baseline and Repeat Visits.