Using Autonomous AI for Eye Exams to Detect Diabetic Retinopathy in Youth Ages 8 to 21 with Diabetes

What this Trial Is About

Researchers are evaluating whether using a nonmydriatic fundus camera with autonomous artificial intelligence AI software at the point of care can increase the number of underserved youth with diabetes who are screened for diabetic retinopathy. The study also aims to assess how accurately the AI system detects diabetic retinopathy from retinal images in young people with type 1 or type 2 diabetes. This research focuses on youth aged 8 to 21 years and includes an enriched group with known diabetic retinopathy to better evaluate diagnostic accuracy. Participants will have a diabetic eye exam performed at the point of care using autonomous AI software on a non-mydriatic fundus camera. The AI system provides immediate exam results. Those with abnormal findings will be referred to an eye care provider for a dilated eye exam. The study enrolls up to 500 participants, including 398 youth to determine if the AI increases screening rates among minority and underserved populations, and a subgroup with known diabetic retinopathy to compare AI results to expert grading. During the study, participants receive the AI-based eye exam and immediate results. Researchers will assess the proportion of youth screened, agreement between AI and retinal specialists, and the AI systems sensitivity and specificity in detecting diabetic retinopathy. The main outcome is the proportion screened at day 1. The study involves a single visit and is conducted under the lead sponsor Johns Hopkins University, with completion expected by September 2026.

ACCESS 2: AI for pediatriC diabetiC Eye examS Study 2