Diagnostic accuracy and agreement between visual acuity charts for detecting significant refractive errors in preschoolers

Background: Preschool vision screenings are considered to be cost-effective methods to identify children with vision disorders. The children of this age group are poor at communicating their symptoms and hence screening is mandated. This study is aimed at estimating the diagnostic accuracy and agreement of Lea, HOTV and E visual acuity charts for detecting significant refractive errors.

Methods: A cross-sectional study was conducted, in which monocular unaided vision assessment of each study participant was performed with Lea, HOTV and E charts. Stereo acuity was measured with the Randot Preschool Test and a comprehensive eye examination including dilatation was performed. Significant refractive error was defined as hyperopia > 3.25 D, myopia > 2.00 D, astigmatism > 1.50 D, anisometropia if interocular difference > 1.00 D for hyperopia, > 3.00 D for myopia or > 1.50 D for astigmatism. Sensitivity, specificity, positive and negative predictive values were estimated. Bland-Altmann plots were generated to help identify the level of agreement between the vision charts.

Results: A total of 256 eyes were analysed. Lea, HOTV and E had sensitivities of 87.8 per cent, 90.2 per cent and 90.2 per cent, respectively. Specificity and positive predictive values were better for HOTV (77.3 per cent, 65.5 per cent) and Lea (75 per cent, 62.6 per cent), compared to E chart (69.8 per cent, 58.7 per cent). Negative predictive values for Lea, HOTV and E charts were 92.8 per cent, 93.8 per cent and 93.8 per cent, respectively. Bland-Altmann analysis showed good agreement between Lea and HOTV, Lea and E, and HOTV and E visual acuity charts. The acuity difference was least between Lea and HOTV charts (0.1 logMAR). Eighty-five (33.2 per cent) eyes had significant refractive errors. Eighty (94.1 per cent) eyes were astigmatic.

Conclusion: The diagnostic accuracy of the visual acuity charts was high for the identification of significant refractive errors in preschool children. There was very good agreement between the visual acuity charts.