Street-view-measured greenspace components and childhood myopia: A population-based cross-sectional and prospective cohort study

BACKGROUND: Greater greenspace exposure may be associated with a lower risk of childhood myopia. However, most studies rely on satellite-derived greenness indices that cannot capture ground-level visual exposure or differentiate greenspace components. Whether specific greenspace components are differentially associated with myopia remains unclear. OBJECTIVES: To examine associations between street-view-measured greenspace components and childhood myopia. METHODS: This cross-sectional and prospective cohort study included children aged 6 to 8 years from a population-based study in Hong Kong (2015-2021). Incident myopia was assessed among children without myopia at baseline who completed 3-year follow-up. Applying deep learning segmentation to street-view imagery, we quantified four specific greenspace components (trees, grass, plants, and fields) within 500-m residential and school buffers. A time-weighted "home-school" metric was generated to represent cumulative daily exposure. Myopia was defined as cycloplegic spherical equivalent refraction ≤-80.50 diopters. Associations were estimated using multivariable logistic regression and Cox proportional hazards regression models with school-level cluster-robust standard errors, adjusting for sociodemographic and behavioral covariates. RESULTS: Among 20,422 children, 5,623 (27.5%) had myopia at baseline. Cross-sectionally, higher cumulative exposures to home-school-trees (OR, 0.94; 95% CI, 0.90-0.97), home-school-grass (OR, 0.95; 95% CI, 0.90-0.99), and home-school-plants (OR, 0.93; 95% CI, 0.89-0.97) were associated with lower myopia prevalence, while home-school-fields increased myopia odds (OR, 1.06; 95% CI, 1.01-1.10). In prospective analyses of 2,667 children, only home-school-plants consistently associated with reduced incident myopia (HR, 0.89; 95% CI, 0.83-0.96), whereas home-fields increased incident myopia risk (HR, 1.05; 95% CI, 1.01-1.09). Protective effects of grass and plants were more pronounced among children with higher near-work time and those from lower-income families (both P for interaction < 0.05). CONCLUSION: Specific street-view greenspace components were differentially associated with childhood myopia. These findings suggest the potential importance of considering specific greenspace components in urban planning strategies relevant to childhood myopia prevention.