Abstract
Abstract: :
Purpose: Significant refractive error, hyperopia, anisometropia, and astigmatism,in early childhood is the most frequent cause of amblyopia. To compare refracting methods for detect significant refractive error when receiver operating characteristic (ROC) curve analysis is used to maximize performance. Methods: 48 preschool children (Mean Age: 4.4 yrs + 1.0 yr) had refractive error measured in a pre–determined order by cycloplegic (C) retinoscopy, noncycloplegic (NC) retinoscopy (Ret), and autorefraction (Nikon Retinomax K–plus (Retmax) and 2 Welch Allyn (WA) autorefractors [Prototype (WAP) & SureSight (SS) 1.22 Child Mode]). All testers were masked; C retinoscopy was the gold standard. ROC curve analysis was conducted for each refracting method. For each refractive error type, specificity (Sp) levels were set at 0.85, 0.90, and 0.95; sensitivity (Se) and cut points (CP) were determined. Results: Hyperopia– Sp=0.85, Se: 0.42 (WAP) – 1.00 (NC Ret); CP=3.44D, 1.37D. Sp=0.90, Se: 0.62 (NC Retmax) – 1.00 (NC Ret); CP=1.37D, 1.37D. Sp=0.95, Se: 0.17 (WAP) – 0.70 (SS 1.22); CP=3.95D, 3.87D. Anisometriopia– Sp=0.85, Se:0.0 (SS 1.22) – 0.50 (NC Ret); CP=1.13D, 0.62D. Sp=0.90, Se:0.0 (SS 1.22) – 0.50 (NC Ret); CP=1.38D, 0.88D. Sp=0.95, Se: 0.0 (SS 1.22, Retmax) – 0.25 (NC Ret, WAP); CP=1.25D – 2.06D. Astigmatism– Sp=0.85, Se:1.00 for all instruments; CP=0.88D – 1.63D. Sp=0.90, Se:1.00 for all instruments; CP=1.13D – 1.88D. Sp=0.95, Se: 1.00 for all instruments; CP=1.75D – 2.50D. Conclusions:Hyperopic cut points and sensitivity vary across methods. Anisometropia cut points vary; sensitivity is poor. Astigmatic cut points and sensitivity are similar across methods. ROC curve analysis enhances the potential usefulness of NC retinoscopy, Retinomax, and SureSight 1.22 method of screening preschool children for significant refractive error.
Keywords: screening for ambylopia and strabismus • hyperopia • astigmatism