Blog Post

Healthcare Upside/Down: Automated Screening for Diabetic Retinopathy

Upside Down Blog Web

ECG’s radio show and podcast, Healthcare Upside Down, offers unfiltered perspectives on what’s working in US healthcare and what’s not. Hosted by ECG principal Dr. Nick van Terheyden, each episode features guest panelists who explore the upsides and downsides of healthcare in the US—and how to make the system work for everyone.

It’s easy to take for granted the access we have to the internet in this country and the bandwidth that allows us to transfer massive amounts of information. It was not that long ago that limits on bandwidth made providing the best care to remote areas a challenge.

Our guest on episode 52 of Healthcare Upside Down is Jennifer Lim, MD, Distinguished Professor of Ophthalmology at the University of Illinois Chicago.

Listen Now

Bandwidth can still be limited, but it has rapidly increased, allowing for quicker access to resources online. That can allow for the rapid sharing of data, which can help ameliorate patient access issues driven by a lack of skilled resources in some areas. And as Jennifer Lim, MD, tells us, it can improve access for the enormous number of chronic patients suffering from diabetes who are at risk of diabetic retinopathy.

Dr. Lim is a distinguished professor of ophthalmology at the University of Illinois Chicago. On episode 52 of Healthcare Upside Down, she explains how we can screen more patients for diabetic retinopathy while freeing up time for specialists through the use of artificial intelligence. Below are a few excerpts.

Early detection of diabetic retinopathy.

“With diabetes, multiple systems can be affected, and one of them is the thalamic system. And the longer a patient has diabetes, the higher the chance they’re going to have diabetic retinopathy. So it’s really important that a patient with diabetes gets checked, because they may have findings that can result in loss of vision, but not yet know it. It may be completely asymptomatic. It’s treatable, but only 30% to 50% of patients undergo their annual diabetic retinopathy screening examination. And that really is a tragedy, because earlier diagnosis and earlier treatment absolutely can save vision.”

Using AI to diagnose patients in rural and underserved areas.

“We know from studies that if you take a picture of someone’s retina and send it to trained graders, you’re going to pick up more and be able to label it more specifically than an examiner looking at that patient. That’s not to say the examiner is going to miss vision-threatening diabetic retinopathy, but they might miss a little hemorrhage out in the periphery. But if you take a picture, you’re not going to miss it. So more recently, artificial intelligence (AI) has entered the game. And people have figured out how you can take an image and train an AI system to recognize it as having diabetic retinopathy. There are two FDA-approved imaging systems that are autonomous, meaning that they take 45-degree photographs using a person to actually take the picture. The picture is then sent up to the cloud, where the image is read, and a diagnosis is rendered all within one minute.”

Technology will augment, not replace, eye care providers.

“We are not reaching all of those patients who have to be screened, and these machines are helping us screen them. By virtue of that fact, it actually is helping us because we are not using retina specialists or general ophthalmologists to screen patients, but rather to treat and follow those who are at higher risk of progression until they need treatment. And there are a lot of patients in underserved areas, in rural communities, where having such an autonomous system is really going to be helpful. It’s going to help them get to their appointments, because they all they have to do is get to the general office for primary care, get screened, and not have to drive all the way to find a specialist, just to be told that come back in a year.”

On the podcast, Dr. Lim talks about mobile screening units and elaborates on the cutting-edge technology that’s helping detect diabetic retinopathy as early as possible.

Listen Here

Edited by: Matt Maslin