Abstract Background Myopia is a growing public health concern among school-going children, with increased screen time and reduced outdoor activities contributing to its early onset and rapid progression. In India, the absence of a standardized, population-based dataset reflecting socio-economic and environmental diversity limits the development of accurate prediction tools. This study aims to design and validate an artificial intelligence (AI)-based model to analyze and predict myopia progression among children aged 6 to 18 years in urban and rural areas. Methods This is a prospective, community-based study conducted in two phases across five districts of Bhopal division, India. Phase one involves a cross-sectional survey of approximately 5,000 school children to collect demographic and ocular health data, including visual acuity, refractive error, axial length, corneal parameters, and lifestyle factors. In phase two, a longitudinal cohort comprising 10% of the phase one participants (balanced for existing and non-existing myopia cases) will be followed every 6 months over 2 years. Machine learning algorithms, including linear regression, support vector machines, XGBoost, and deep learning models such as convolutional neural networks, will be trained on this dataset. An AI-based mobile application will be developed to enable field-level prediction and screening. Discussion This study will generate the first large-scale Indian dataset on myopia progression among children, incorporating diverse socio-economic backgrounds. The validated AI model and mobile application will support early identification of at-risk children, optimize resource allocation, and inform national screening strategies. By integrating real-time data analytics with field-l. Trial registration CTRI/2025/07/091243. Registered on 21/07/2025.
Singh et al. (Tue,) studied this question.