Leveraging Sight Outcomes Research Collaborative (SOURCE) Data to Predict Initial Presentation With Advanced Open-Angle Glaucoma

Abstract

Purpose: To investigate factors associated with initial presentation at tertiary health systems with advanced primary open-angle glaucoma (OAG) in either eye.

Design: Cross-sectional study.

Subjects: Adults >40 years old diagnosed with OAG initially presenting at one of eleven SOURCE consortium sites between 2016 and 2023.

Methods: OAG severity was dichotomized as "non-advanced" (suspect, mild, moderate) or "advanced," based on the most severe stage in either eye within one year of presentation using ICD-10 codes. Logistic regression identified sociodemographic and neighborhood factors linked to "advanced" OAG presentation, and ordinal models assessed predictors by stage. Predictors included sociodemographic characteristics, health insurance type, income, education, urbanicity, and neighborhood-level affluence (Distressed Communities Index [DCI]). Associations with intraocular pressure (IOP) and cup-to-disc ratio (CDR) were modeled for all eyes using linear regression with generalized estimating equations.

Main outcome measures: Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) for factors associated with presenting with "advanced" OAG.

Results: Among 59 409 individuals, 7560 (12.7%) presented with advanced OAG in at least one eye. Mean age was 65.8(SD 12.7) years; 44.6% male, 56.7% Caucasian. Older age per decade (aOR: 1.74, 95% CI: 1.70-1.78), male sex (aOR: 1.59, CI: 1.51-1.68), Black race (aOR: 1.67, CI: 1.56-1.80), less affluent communities (aOR: 1.04 per 10-unit DCI increase, CI: 1.03-1.05), and non-metropolitan residence (OR range: 1.42-1.99) were associated with advanced presentation. Protective factors included Medicare insurance (aOR: 0.90, CI: 0.90-0.96), higher education (aOR range: 0.74-0.83), and diabetes mellitus (aOR: 0.39, CI: 0.36-0.43). Results were consistent across secondary analyses. Presenting IOP and CDR increased with disease severity and were similarly associated with sociodemographic and neighborhood factors.

Conclusion: Identifying sociodemographic predictors of late presentation using large population-based datasets may help inform targeted glaucoma screening efforts, since delayed initial presentation is a major risk factor for vision loss from advanced OAG.