An intelligent systems framework for closed-loop chronic glaucoma management

Background Glaucoma, the world’s leading cause of irreversible blindness, highlights chronic disease management challenges: fragmented health records, poor adherence, and limited privacy safeguards. While artificial intelligence (AI) shows promise, its clinical application remains confined to isolated algorithmic tasks. Methods To bridge the gap, GlauHelper, an internet-based platform integrating multiple AI modules, is developed. The dual-end (physician/patient) platform integrates a comprehensive clinical workflow: (1) Secure Data Acquisition, featuring age-friendly interfaces, home-based tonometer, and an image privacy system; (2) Automated Processing Pipeline, which cleans and structures real-world uploads through quality scoring, cascaded classification, and data extraction, with human verification; and (3) Clinical Application, utilizing structured data to drive personalized follow-up, medication tracking, and a domain-specific Large Language Model (LLM) Q&A assistant (“Xiaoqing”) optimized for glaucoma education. Results Real-world deployment across 30 hospitals and over 10,000 users in China demonstrated high adoption. Experimental validation confirmed the system’s ability to extract structured clinical data, safeguard patient privacy, assess and classify medical images, and deliver educational content. Conclusion GlauHelper demonstrates the technical feasibility of integrating secure data acquisition, automated information processing, and patient–clinician interaction within a unified glaucoma management workflow. These findings support its preliminary usability and real-world deployability, while its effects on treatment adherence, disease progression, and long-term clinical outcomes require further prospective evaluation.