Accuracy of noncycloplegic retinoscopy, retinomax autorefractor, and SureSight vision screener for detecting significant refractive errors

Invest Ophthalmol Vis Sci. 2014 Mar 6;55(3):1378-85. doi: 10.1167/iovs.13-13433.

Abstract

Purpose: To evaluate, by receiver operating characteristic (ROC) analysis, the ability of noncycloplegic retinoscopy (NCR), Retinomax Autorefractor (Retinomax), and SureSight Vision Screener (SureSight) to detect significant refractive errors (RE) among preschoolers.

Methods: Refraction results of eye care professionals using NCR, Retinomax, and SureSight (n = 2588) and of nurse and lay screeners using Retinomax and SureSight (n = 1452) were compared with masked cycloplegic retinoscopy results. Significant RE was defined as hyperopia greater than +3.25 diopters (D), myopia greater than 2.00 D, astigmatism greater than 1.50 D, and anisometropia greater than 1.00 D interocular difference in hyperopia, greater than 3.00 D interocular difference in myopia, or greater than 1.50 D interocular difference in astigmatism. The ability of each screening test to identify presence, type, and/or severity of significant RE was summarized by the area under the ROC curve (AUC) and calculated from weighted logistic regression models.

Results: For detection of each type of significant RE, AUC of each test was high; AUC was better for detecting the most severe levels of RE than for all REs considered important to detect (AUC 0.97-1.00 vs. 0.92-0.93). The area under the curve of each screening test was high for myopia (AUC 0.97-0.99). Noncycloplegic retinoscopy and Retinomax performed better than SureSight for hyperopia (AUC 0.92-0.99 and 0.90-0.98 vs. 0.85-0.94, P ≤ 0.02), Retinomax performed better than NCR for astigmatism greater than 1.50 D (AUC 0.95 vs. 0.90, P = 0.01), and SureSight performed better than Retinomax for anisometropia (AUC 0.85-1.00 vs. 0.76-0.96, P ≤ 0.07). Performance was similar for nurse and lay screeners in detecting any significant RE (AUC 0.92-1.00 vs. 0.92-0.99).

Conclusions: Each test had a very high discriminatory power for detecting children with any significant RE.

Keywords: children's vision; refractive error; vision screening.

Publication types

  • Comparative Study
  • Research Support, N.I.H., Extramural

MeSH terms

  • Child, Preschool
  • Equipment Design
  • Female
  • Humans
  • Male
  • ROC Curve
  • Refraction, Ocular*
  • Refractive Errors / classification
  • Refractive Errors / diagnosis*
  • Refractive Errors / physiopathology
  • Reproducibility of Results
  • Retinoscopy / methods*
  • Vision Screening / instrumentation*
  • Visual Acuity*