Visual acuity screening versus noncycloplegic autorefraction screening for astigmatism in Native American preschool children.

INTRODUCTION Visual acuity screening (VAS) is less reliable in preschoolers than in school-aged children as a means of detecting significant refractive error. We wished to compare the effectiveness of VAS with the effectiveness of an objective method, noncycloplegic autorefraction screening (NCARS), in detecting the presence of significant astigmatism warranting spectacle correction. METHODS We examined 245 Native American Head Start registrants aged 3 to 5 years. We attempted to obtain uncorrected visual acuity using Lea Symbols logMAR Chart (Precision Vision Inc, Villa Park, Ill), noncycloplegic autorefraction using the Nikon Retinomax K-plus (Nikon Corp, Melville, NY), and cycloplegic refraction (CR) on each eye. The VAS failure criterion was either a 2-line acuity difference between eyes or acuity worse than 20/40 in either eye. The NCARS and CR failure criterion was the spectacle correction threshold exceeding the 50th percentile on the basis of a survey of AAPOS members. RESULTS We completed VAS in 96% of children and NCARS and CR in 100% of children. There was high prevalence (31%) of significant astigmatic refractive error in this sample. Ten subjects who did not permit bilateral visual acuity measurements were scored as having a positive test result. The sensitivity and specificity of VAS were 90% and 44%, respectively. NCARS had sensitivity and specificity of 91% and 86%, respectively. NCARS becomes cost-effective after 1044 children are screened, assuming that the cost of the autorefractor is 300 times the cost of the referral examination. CONCLUSION VAS offers high sensitivity but suffers from poor specificity. NCARS greatly reduces the number of unnecessary referrals. In this population, NCARS becomes cost-effective after approximately 1000 children are screened.