Background: To evaluate the stability and predictive ability of uncorrected visual acuity (UCVA), non-cycloplegic refraction (NCR), and axial length (AL) as indicators in non-cycloplegic school vision screening for myopia.
Methods: This retrospective cohort study is based on the Shanghai Child and Adolescent Large-scale Eye Study (SCALE). Participants included students who failed school screenings and were referred for follow-up cycloplegic refractions at eye hospitals within three months. We evaluated the differences in UCVA, spherical equivalent (SE), and AL between school screenings and hospital re-evaluations. Furthermore, we assessed the validity of using UCVA in combination with NCR as predictive metrics for myopia referral.
Results: Among the 8,492 children, 4,357 (51.3%) were boys, with a mean age of 8.26 years (SD = 2.77). AL was identified as a reliable myopia screening indicator across all age groups (ICC = 0.981; 95% CI: 0.978-0.984), outperforming UCVA (ICC = 0.791) and SE (ICC = 0.806). The estimated prevalence of myopia using UCVA and NCR was 76.10%, significantly higher than 58.37% observed with cycloplegic testing. Sensitivity was 96.93% and specificity was 53.13% (Youden index = 0.5). In preschoolers, myopia rates decreased from 50.17% before to 19.82% after cycloplegia, while school-aged children exhibited better consistency. Decision curve analysis showed that the full model incorporating AL did not significantly benefit school-aged children, but may offer greater net benefits for preschoolers.
Conclusions: AL should be integrated into screening programs for preschoolers. For school-aged children, the combination of UCVA and NCR suffices for myopia prediction, potentially eliminating the need for cycloplegia.
Keywords: Decision analysis; Myopia prediction; Referral criteria; School vision screening.
© 2025. The Author(s).