Development of a Machine Learning-based Compliance Scoring System for Glaucoma Patients using Electronic Health Records
CoI: Dr. Wejdan Almutairi (King Khaled Eye Specialist Hospital)
Description: This study aims to develop and internally validate a machine-learning model to predict future medication non-adherence among glaucoma patients in Saudi Arabia. Using retrospective EHR and pharmacy refill data from approximately 300 KKESH patients, the study will measure adherence through the Proportion of Days Covered (PDC) and analyze demographic, clinical, medication, healthcare-utilization, and behavioral factors. Several models will be evaluated based on discrimination, calibration, and clinical performance. Key outputs will include a validated prediction model, a 0–100 risk score, important predictors of non-adherence, a prototype decision-support concept, documented code, and a peer-reviewed publication. The project will provide a reproducible, locally validated research prototype to support future interventions and data-driven glaucoma care.