AI-Based Point-of-Care Pediatric DR Screenings May Improve Diagnosis Rates

Optician concentrating using Optalmoscope focusing on school girls eye, while doing an eye test.
Research shows that point-of-care diabetic retinopathy screenings that provide immediate results and are likely to lead to an in-office exam for patients who need it.

Diabetic retinopathy (DR) can lead to vision loss, yet pediatric diabetic patients are not consistently getting screened for the condition. Researchers recently examined if autonomous artificial intelligence (AI) DR screenings were more cost-effective than standard eye care screening examinations performed by general practitioners.

The prevalence of DR among children with type 1 and type 2 diabetes ranges from 4% to 13%, according to the study published in JAMA Ophthalmology. While the American Diabetes Association and the American Academy of Ophthalmology recommend yearly

screenings, youth DR screening adherence ranges from 35% to 72%.

According to the researchers, patients are more likely to consent to examination with point-of-care (POC) screenings than with traditional referral systems. Further, patients who have positive findings on POC screenings are more likely to seek an examination from an ophthalmologist.

Point-of-care DR screening using autonomous AI provides immediate results in clinic settings. To assess the cost-effectiveness of detecting and treating DR and its sequelae among children with type 1 and type 2 diabetes using AI DR screening, researchers conducted an economic evaluation using a decision-analysis model.

In the model, the number (proportion) of DR cases discovered was defined as “effectiveness” and patient and family out-of-pocket cost was defined as the “cost.” The expected true-positive proportions for standard ophthalmologic screening by an ophthalmologist were 0.006 for type 1 and 0.01 for type 2, and the expected true-positive proportions for autonomous AI were 0.03 for type 1 and 0.04 for type 2. The base case scenario of 20% adherence estimated that the use of autonomous AI would result in a higher mean patient payment ($8.52 for type 1 and $10.85 for type 2) than conventional screening ($7.91 for type 1 and $8.20 for type 2). The study found an incremental cost-effectiveness ratio of $31 for type 1 diabetes and $95 for type 2 diabetes for each additional case of diabetic retinopathy identified, compared with standard practice.

Researchers concluded that when more than 23% of patients adhere to DR screening

recommendations, autonomous AI screenings are the preferred strategy—and save money for patients and their families.

Limitations of the study include not assessing downstream costs of complications, focusing only on DR as a reason to see an ophthalmologist, and not evaluating the costs of implementing technology.

Researcher Michael D. Abramoff, MD, PhD, disclosed that he is the founder, executive chairman, director, and an investor in Digital Diagnostics and has a patent assigned to Digital Diagnostics and the University of Iowa that is relevant to the content of this study.


Wolf R, Channa R, Abramoff MD, Lehmann HP. Cost-effectiveness of Autonomous Point-of-Care Diabetic Retinopathy Screening for Pediatric Patients With Diabetes. JAMA Ophthalmol. Published online September 3, 2020. doi:10.1001/jamaophthalmol.2020.3190