Diagnostic accuracy and variability of autorefraction by the Tracey Visual Function Analyzer and the Shin‐Nippon NVision‐K 5001 in relation to subjective refraction

Abstract

To determine the accuracy of distance autorefractions obtained by two 'open field' devices, the Tracey Visual Function Analyzer and the Shin-Nippon NVision-K 5001, by comparison with subjective refraction. Both eyes of 50 healthy phakic participants underwent subjective refraction. Autorefractions were then performed on undilated pupils using the Tracey and a modified Shin-Nippon autorefractor and these were repeated within 50 days. Agreement with subjective refraction was calculated for sphere, mean spherical equivalent (MSE) and cylindrical vectors J(0) and J(45). Intratest and intertest variability were also evaluated. The mean age of the participants was 37.4 years. Subjective refraction MSE ranged from -6.25 D to +3.62 D, mean -0.49 D +/- 1.79 D. Bias between subjective refraction and Tracey was -0.001 D, +0.045 D, +0.017 D, and -0.015 D for sphere, MSE, J(0) and J(45) respectively; these were not significant. Bias between subjective refraction and Shin-Nippon was +0.004 D, +0.033 D, +0.106 D, and -0.021 D; only the J(0) vector was significantly different (p < 0.0001) although this difference was small. Intratest variability for Tracey was low, measured at 0.189 D for sphere and 0.178 for MSE, and for the Shin-Nippon 0.099 D and 0.086 D respectively. Tracey intertest variability revealed small, statistically significant bias for sphere and MSE (+0.071 D and +0.070 D, p = 0.011, 0.013). Shin-Nippon reproducibility showed no significant bias. Autorefraction measurements captured by both the Tracey and Shin-Nippon devices agree well with subjective refraction. The Shin-Nippon shows lower intratest variability.