Answers are grounded in this paper and cite their sources.
First screen-detected myopia and progression after detection have different implications for prevention and management. We evaluated their temporal patterns separately in linked school-screening data, without quantifying their relative contributions to changes in prevalence. This retrospective longitudinal study included 31,222 students in grades 1–9 from 17 public schools in suburban Shanghai, China, from 2020 to 2025. Among grade 1 baseline nonmyopes in 10 common schools, complementary log-log models compared first screen-detected myopia over a common 1-year horizon. Sensitivity analyses used bias-reduced cluster-robust (CR2) inference with Satterthwaite degrees of freedom at the school level and joint covariate adjustment. Generalized estimating equations assessed within-grade refractive contrasts and annual within-child progression. Axial changes were evaluated in two school divisions with complete biometry. Next-year follow-up among grade 1 baseline nonmyopes was 97.1%. Standardized 1-year risk was 26.1% in the 2020 entry cohort and 19.6% in the 2024 cohort; the hazard ratio was 0.718 (student-clustered 95% confidence interval [CI], 0.616–0.836; school-clustered CR2 CI, 0.519–0.993; P = 0.046). After joint adjustment for age, sex, baseline spherical equivalent, examination month, and school, the school-clustered CI included unity (hazard ratio, 0.721; 95% CI, 0.518–1.004; P = 0.052). Within-grade contrasts varied by grade (interaction P = 0.004). Refractive progression was nonmonotonic but 0.117 D/year slower in 2024–2025 than in 2020–2021 (95% CI, 0.100-0.134). Axial growth was 0.097 mm/year slower (95% CI, 0.081–0.112). The latest interval showed slower refractive and axial progression among children already myopic. The latest entry cohort also had lower estimated first-detection risk, although its precision depended on school-level inference and adjustment. These descriptive patterns support continued surveillance without identifying a causal driver or establishing a population-wide decline.
No takes yet. Share an insight, caveat, or question.
Qiao et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: