Sensitivity of Screening Tests for Detecting Vision In Preschoolers-targeted Vision Disorders When Specificity Is 94%

Abstract

To compare the sensitivity of 11 preschool vision screening tests administered by licensed eye care professionals for the detection of the 4 Vision in Preschoolers (VIP)-targeted vision disorders when specificity is 94%. This study consisted of a sample (n = 2588) of 3- to 5-year-old children enrolled in Head Start programs, 57% of whom had failed an initial Head Start vision screening. Screening results from 11 tests were compared with results from a standardized comprehensive eye examination that was used to classify children with respect to the four VIP-targeted vision disorders: amblyopia, strabismus, significant refractive error, and unexplained reduced visual acuity (VA). With overall specificity set to 94%, we calculated the sensitivity for the detection of each targeted vision disorder. With the overall specificity set to 94%, the most accurate tests for detection of amblyopia were noncycloplegic retinoscopy (NCR) (88% sensitivity), the SureSight Vision Screener (80%), and the Retinomax Autorefractor (78%). For detection of strabismus, the most accurate tests were the MTI Photoscreener (65%), the cover-uncover test (60%), the Stereo Smile II stereoacuity test (58%), the SureSight Vision Screener (54%), and the Retinomax Autorefractor (54% in year 1, 53% in year 2). The most accurate tests for detection of significant refractive error were NCR (74%), the Retinomax Autorefractor (66%), the SureSight Vision Screener (63%), and the Lea Symbols VA test (58%). For detection of reduced VA, the most accurate tests were the Lea Symbols Distance VA test (48%), the Retinomax Autorefractor (39%), and NCR (38%). Similar to the previously reported results at 90% specificity, the screening tests vary widely in sensitivity with specificity set at 94%. The rankings of the sensitivities for detection of the 4 VIP-targeted vision disorders are similar to those with specificity set to 90%.