Reproducibility of sphero-cylindrical prescriptions*

Purpose: To investigate the reproducibility of the sphero-cylindrical prescriptions provided by 40 optometrists. Methods: Forty registered optometrists were randomly selected from the cities of Oxford and Westminster in the UK to perform a sphero-cylindrical refraction on an asymptomatic 29-year-old male subject. The 95 % limits of reproducibility for each component of refraction were assessed and are presented together with scatterplots, distribution ellipsoids and polar profiles of dioptric power. Results: The mean stigmatic (spherical equivalent) refraction for the right eye was)0.83 D (S.D. = 0.28 D) with 95 % limits of agreement)1.38 and)0.28 D. The 95 % reproducibility limits for stigmatic data [1:96ð ffiffiffi2p ÞðS.D.Þ] was 0.78 D. The average inter-ocular difference in the stigmatic components of refraction was)0.044 D (S.D. = 0.20 D) but estimates ranged from)0.50 to +0.50 D. Mean ortho- and oblique antistigmatic refractions were)0.23 D (S.D. = 0.084 D) and)0.14 D (S.D. = 0.086 D) respectively. Conclusions: The findings of this paper suggest that refractions performed by multiple optometrists on a single eye will differ in their stigmatic component by over 0.78 D on average not more than once in 20 refractions. The reproducibility of refractions reported here, approximately twice as variable as those reported under repeatability conditions, has profound implications for the analysis of refractive data collected by multiple optometrists over the course of replication, longitudinal and epidemiolog-ical studies