Purchase this article with an account.
S.J. Huffman, K.L. Schimmoeller, P. Schmidt, G. Mitchell; Repeatability and Accuracy of Autorefractor Measurements in Young Children . Invest. Ophthalmol. Vis. Sci. 2004;45(13):2736.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: This investigation establishes repeatability and accuracy of the Nikon Retinomax and the Welch Allyn SureSight 2.04 autorefractor measurements not yet fully documented in young children with significant refractive error. Methods: Children (n=31), age 3–5 years, with significant refractive error (hyperopia>+3.25DS and astigmatism>1.50DC) were identified. Refractive error was separately determined by cycloplegic retinoscopy. Separate, repeated measurements were made with two autorefractors on noncyclopleged eyes of each child. Repeatability is defined by within subject standard deviation (SD). Accuracy is defined as the comparison of autorefractor measurements to that of cycloplegic retinoscopy. ROC curves were used to investigate accuracy as measured by the discriminative ability of each autorefractor. Results: No significant difference was found between autorefractors for the number of measurements required to obtain 4 reliable measurements or the percent of reliable measurements. Retinomax–The within subject standard deviation was 0.38 for sphere and 0.20 for cylinder. The expected difference in pairs of observations was 1.0DS and 0.50DC. The Retinomax offers excellent discrimination between hyperopic and non–hyperopic eyes (area=0.9000). The Retinomax also offers acceptable discrimination between astigmatic and non–astigmatic eyes (area=0.716). SureSight–The within subject standard deviation was 0.69 for sphere and 0.28 for cylinder. The expected difference in pairs of observations was 1.9DS and 0.75DC. The ability of the SureSight to discriminate hyperopic eyes from non–hyperopic eyes appears less than acceptable (area=0.667). The SureSight does offer acceptable discrimination of astigmatic and non–astigmatic eye (area=0.790). Conclusions: Children were testable with both instruments. For both the Retinomax and the SureSight, variability of measurements is not related to the mean magnitude of refractive error. The mean difference of autrefractor measurements from cycloplegic retinoscopy does not depend on the magnitude of refractive error for the Retinomax, but does for the SureSight. The Retinomax offers excellent discrimination between astigmatic and non–astigmatic eyes. The SureSight appears not to offer acceptable discrimination between hyperopic and non–hyperopic eyes, but does between astigmatic and non–astigmatic eyes.
This PDF is available to Subscribers Only