Supplementary file 1_Artificial intelligence driven transformation of pediatric eye health education based on bibliometric analysis and a cross-sectional survey.docx

BackgroundPediatric eye diseases, especially myopia and other refractive errors as well as binocular vision disorders, have evolved into a pressing global public health concern. Effective caregiver-targeted eye health education is crucial for the early prevention and long-term management of these conditions. However, conventional educational approaches often struggle to deliver actionable and accessible guidance. Emerging artificial intelligence (AI) technologies, particularly large language models (LLMs), present novel opportunities to address this gap.
MethodsA mixed-methods design was conducted in this study. On one hand, a bibliometric analysis was conducted on publications related to pediatric eye health education from 2015 to 2025, using the Web of Science Core Collection and PubMed databases, with visualization implemented via Bibliometrix and CiteSpace software. On the other hand, a cross-sectional online questionnaire survey was administered to guardians of children aged 1–6 years to systematically assess their key concerns about pediatric eye diseases, knowledge levels, information sources, and preferences for educational formats.
ResultsA total of 111 publications were included in the bibliometric analysis, which revealed a significant surge in research output after 2020. Traditional themes such as vision screening and patient education remained central, while AI- and LLM-related topics have emerged as recent research hotspots. China and the United States led in terms of publication volume, whereas several countries with lower output demonstrated higher average citation impact. Among 328 valid survey responses, guardians showed high concern for myopia, astigmatism, and strabismus but lacked sufficient practical knowledge regarding preventive pharmacological strategies and correct eye drop administration. Additionally, child-friendly, visualized, interactive, and online educational formats were strongly preferred.
ConclusionBy integrating bibliometric trends with the actual needs of caregivers, this study demonstrates that the growing role of AI and LLMs in pediatric eye health education reflects both technological advancement and unmet demands for personalized, actionable educational content. These findings provide an evidence-based foundation for the development of AI-driven pediatric eye health education tools, emphasizing the necessity of clinical oversight, ethical governance, and equitable access in future practice.