Machine Learning-Based Predictive Modeling of Surgical Intervention in Glaucoma Using Systemic Data From Electronic...

Area of Science:

• Ophthalmology
• Medical Informatics
• Machine Learning

Background:

• Primary open-angle glaucoma (POAG) is a leading cause of irreversible blindness.
• Predicting the need for surgical intervention is crucial for managing POAG progression.
• Electronic health records (EHRs) contain vast amounts of patient data, including systemic information.

Purpose of the Study:

• To predict the need for surgical intervention in primary open-angle glaucoma (POAG) patients.
• To utilize systemic data available in electronic health records (EHRs) for this prediction.
• To develop and evaluate machine learning models for risk stratification.

Main Methods:

• Development and evaluation of machine learning models including logistic regression, random forests, and artificial neural networks.
• Utilized structured EHR data from 385 POAG patients.
• Performed leave-one-out cross-validation and calculated performance metrics like AUC, sensitivity, and specificity.

Main Results:

• Multivariable logistic regression achieved the highest discrimination (AUC = 0.67).
• Higher systolic blood pressure significantly increased the odds of needing glaucoma surgery (OR = 1.09).
• Specific medication classes (ophthalmic, non-opioid analgesics, anti-hyperlipidemics, macrolides, calcium blockers) were associated with decreased odds of requiring surgery.

Conclusions:

• Systemic EHR data possesses predictive value for identifying POAG patients requiring surgical intervention, even without ocular data.
• Blood pressure metrics and certain medication classes are key predictors of glaucoma progression.
• This approach offers potential for automated, EHR-integrated risk prediction to support clinical decision-making.