An AI-assisted screening tool may help improve diabetic eyecare referral patterns for some underserved patients, according to investigators at the Wilmer Eye Institute at Johns Hopkins Medicine.
In an exploratory study conducted across community-based primary care sites, African American adults with diabetes were more likely to receive a referral for a diabetic eye exam when screened with the AI tool than when referred through standard primary care pathways.
Methodology
Researchers examined whether referrals generated through an FDA-approved AI-assisted diagnostic screening program were associated with greater adherence to recommended annual diabetic eye exams. The study focused on two historically disadvantaged groups: African American patients and patients covered by Medicaid.
In a retrospective analysis, the team identified 3,745 adults with diabetes who visited the Wilmer Eye Institute for diabetic retinopathy evaluation between August 2020 and September 2022. Of those patients, 3,352 were referred by primary care providers and 393 received a recommendation through the AI-assisted screening tool.
Patients screened with the AI system had retinal images taken with a specialized camera during a primary care visit. The images were analyzed in real time. If diabetic retinopathy was detected, patients were informed during the same visit and referred either to the Wilmer Eye Institute or to another eyecare specialist of their choice.
Results
The researchers found that a larger share of African American patients received an eye exam referral when the AI-assisted tool was used compared with standard primary care referral methods: 64.9% versus 44.4%.
Referral rates for patients insured through Medicaid were similar regardless of referral source, at 0.8% for the AI-assisted tool and 0.6% for primary care provider referral.
The study also found that when the AI-assisted tool was used, patients with hypertension were more likely to receive an eye exam referral than those without the condition, 89.6% versus 82.6%. A similar pattern was seen among patients with chronic kidney disease, 26.2% versus 20.9%.
When the researchers looked at follow-through on care, they found that patients who both chose AI-assisted screening and attended a diabetic retinopathy evaluation were 15% more likely to be African American. Medicaid coverage was not associated with differences in appointment attendance whether the referral came from a primary care provider or the AI tool.
What this means for eyecare providers
For eyecare providers, the findings suggest that AI-assisted screening in primary care settings may influence who gets referred for diabetic eye evaluation and who follows through on that care. In this study, the tool was associated with higher referral rates among African American patients. Additionally, referral and attendance patterns among Medicaid-covered patients did not materially differ by referral method.
The study does not establish long-term effects on vision outcomes. The authors said additional research is needed to determine whether expanded access to AI-assisted screening leads to sustained improvements in eye health. Still, the results indicate that same-visit screening, immediate results and clear next steps may affect how some patients engage with recommended diabetic eye care.
Read more AI in Eye Care news here
Author
