Abstract: Objective To compare the accuracy of common visual indexes, uncorrected distant visual acuity(UCDVA), axial length/mean corneal redius(AL/CR), refractive state with cycloplegia and non-cycloplegia in school-age large sample myopia for screening, and to put forward a suitable method to replace cycloplegic refractometry. Design Cross-sectional study based on school-age children. Participants School-age children in Suning Country, Hebei Province  were selected. Those students who had the informed consent and had no ocular diseases took part in this study. There were 232 children (123 boys and 109 girls), aging from 6 to 12 years old enrolled. Methods All the students received examinations including UCDVA with standard logarithmic visual acuity chart, <0.9 defined as visual impairment; axial length (AL) and corneal redius (CR) with Optical coherence biological measuring instrument (IOL Master), AL/CR>3 defined as suspicious of myopia; refractometry with automatic refractor (Nissan RM-8000) both in the condition of cycloplegia and non-cycloplegia(≤-0.50 D defined as myopia). 1% cyclopentolate was applied for cycloplegic refraction. The sensitivity, specificity and Youden index of each method was analyzed using the cycoplegic refractometry as gold standard. Main Outcome Measures UCDVA, AL/CR, diopter by refraction with cycloplegia and non-cycloplegia; sensitivity, specificity and Youden index. Results The sensitivity, specificity and Youden Index of UCDVA and non-cycloplegic refractometry in myopia screening was 86.8%, 59.3%, 0.46, and was 94.3%, 66.6%, 0.61 respectively, while those of AL/CR were 90.6%, 81.3%, and 0.72. Conclusion In a large-scale myopia screening, AL/CR could be used as an effective and accurate method instead of cycolplegic refractometry. （Ophthalmol CHN, 2012, 21: 327-330）

Key words: myopia screening, sensitivity, specificity, Youden index