Researchers at the Wilmer Eye Institute, Johns Hopkins Medicine found that African American patients were more likely to receive a diabetic eye exam referral when screened by an AI-assisted diagnostic tool.
The peer-reviewed study published in npj Digital Medicine examined whether referrals generated through a U.S. FDA-approved AI-assisted diagnostic screening program increased adherence to recommended annual diabetic eye exams. Investigators focused on African American patients and patients covered by Medicaid, two historically disadvantaged groups.
Diabetic retinopathy patients may not experience symptoms early and is a leading cause of blindness globally, the researchers note, increasing the importance of timely diagnosis and treatment.
Study Design and Results
Led by T.Y. Alvin Liu, MD, principal investigator and founding director of the James P. Gills Jr., MD, & Heather Gills Artificial Intelligence Innovation Center, the retrospective analysis included 3,745 adult patients with diabetes who visited Wilmer Eye Institute for a diabetic retinopathy evaluation between August 2020 and September 2022.
Of those patients, 3,352 received referrals from primary care providers, while 393 received a recommendation from the AI-assisted screening tool. Patients screened with the AI tool had retinal images captured and analyzed in real time during their primary care appointment. They received same-day referral guidance if diabetic retinopathy was detected.
Researchers found that a higher percentage of African American patients received an eye exam referral when the AI tool was used (64.9%) compared with standard referral methods (44.4%).
Impact on Care Access
In evaluating the referrals, researchers found that people who opted for the AI-assisted tool and attended their diabetic retinopathy evaluation were 15% more likely to be African American. Medicaid coverage did not impact patient appointment attendance whether a referral was received.
“A referral [from a primary care provider] doesn’t guarantee people will attend a diabetic eye exam, even if it’s needed,” says Dr. Liu.
Dr. Liu makes the distinction that the patients evaluated with the AI tool were given a test result on the spot rather than being asked to attend an appointment because they may have something wrong.
“Other obstacles may limit whether patients can attend the screenings,” he says. “But we were able to see that they are more convinced they need care if they’re given immediate results with clear instructions on what to do.”
Future Research
Dr. Liu adds further work is needed to determine whether improved screening access translates into long-term vision health outcomes.
“Ultimately, AI tools are not meaningful unless you can demonstrate that their real-world deployment positively impacts patient lives,” he says. “With future work, we want to examine how patients continue to interact with these AI tools over time and how that translates to specific eye health outcomes.”