Unfortunately, in most practices, less than 50% of people with diabetes make it to an eye care specialist for a critical, sight-saving eye exam for diabetic retinopathy (DR). Too many individuals with diabetes lose vision needlessly from DR, primarily because they are not diagnosed and treated in time.
With new advances in technology, primary care physicians can now administer an eye exam for DR at the point of care, taking back control of this essential item on their diabetes care checklist. Patients can enjoy the convenience and affordability of receiving the exam during their diabetes check-up.
There are currently two options for DR testing in primary care: telemedicine and autonomous artificial intelligence (AI) diagnostic solutions. Telemedicine options have existed for many years, whereas AI diagnostics only recently became available for use in patient care.
Telemedicine has failed to become widely used at scale despite availability for many years. Some healthcare systems find it difficult to secure qualified professionals for these services. In the case of the tele-retinal exam, an eye care specialist or trained reader is required to interpret the images, which can present challenges to screening patients at scale.
Because primary care physicians must wait for a human to interpret the images, there can be a delay of hours to days in receiving results, which are rarely delivered while the patient is in the office and thus require staff to follow up with the patient on the results.
Even more problematic is that most telemedicine solutions lack meaningful image quality feedback, which may result in the provider conducting the reading of low-quality images. This may require the primary care staff to contact the patient to return to recapture images, or the telemedicine specialist may make a diagnosis based on a borderline image. If image quality is insufficient, signs of disease may go undetected, resulting in vision loss.
Autonomous AI Diagnostic Systems
With the FDA’s recent clearance of IDx-DR (IDx Technologies Inc.)—the first autonomous AI system for the detection of diabetic retinopathy—primary care physicians now have the option to test for DR in their office without requiring a specialist to interpret images. It is important to note that IDx-DR is intended for use to automatically detect more than mild diabetic retinopathy (mtmDR) in adults aged 22 or older diagnosed with diabetes who have not been previously diagnosed with diabetic retinopathy. With immediately provided results, physicians can discuss results while consulting on diabetes care and management.
Autonomous AI systems can also provide immediate image quality feedback to operators, ensuring that the images captured are of sufficient quality to make a definitive diagnosis. “In addition, there are studies that suggest AI systems may deliver more consistent and reliable output than human experts. In a pivotal clinical trial for IDx-DR, the system achieved 87% sensitivity and 90% specificity. While IDx-DR was not directly compared to clinicians, a number of studies report sensitivities as low as 33% in a representative sample of board-certified ophthalmologists when using the same reference standard.”
An Automated Future
Telemedicine and AI diagnostics represent two of the most promising routes to expand DR testing in the primary care setting, which is critical given the soaring rates of diabetes across the country. Indeed, diabetes prevalence in the U.S. is projected to increase to 21% of the adult population by 2050.
All patients with diabetes are at risk for developing DR, and approximately 10% will develop a vision-threatening form of the disease. What’s more, 24,000 Americans go blind annually due to diabetic retinopathy (Figure). Primary care physicians can take back control of this important diabetes care metric by testing for DR at the point of care. Doing so may increase the quality, affordability, and accessibility of care for patients with diabetes.
Abramoff M, Lavin P, Birch M, Sha N, Folk J. Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices. Npj Digital Medicine. 2018;39. Available at www.nature.com/articles/s41746-018-0040-6.
CDC: Vision Health Initiative | Economic Studies. Available at www.cdc.gov/visionhealth/projects/economic_studies.htm.
Rowley W, et al. Diabetes 2030: Insights from Yesterday, Today, and Future Trends. Popul Health Manag. 20(1):6-12.
World Health Organization. Prevention of Blindness from Diabetes Mellitus. Available at www.who.int/blindness/Prevention%20of%20Blindness%20from%20Diabetes%20Mellitus-with-cover-small.pdf.
CDC. Keep an Eye on Your Vision Health. Available at www.cdc.gov/features/healthyvision/index.html.
Pugh J, et al. Screening for diabetic retinopathy. The wide-angle retinal camera. Diabetes Care. 1993;16:889-895.
Boyle J, Thompson T, Gregg E Barker L, Williamson D. Projection of the year 2050 burden of diabetes in the US adult population: dynamic modeling of incidence, mortality, and prediabetes prevalence. Popul Health Metr. 2010;8:29.
The International Agency for the Prevention of Blindness. Diabetic Retinopathy. Available at www.iapb.org/knowledge/what-is-avoidable-blindness/diabetic-retinopathy/.