This article is more than 5 years old. Information may no longer be current.
BLOG: Artificial intelligence corrals diabetic maculopathy in the developing world
Click Here to Manage Email Alerts
In the developing world, shortcuts are taken everywhere in health care. Surgeons operate without gloves. Instruments are dipped in alcohol between use in two or more patients in the same operating room. Heat-sterilized cigarette filters are cut at an angle and used as sponges. In Armenia, a former Soviet republic the size of Maryland and located among Turkey, Georgia, Azerbaijan and Iran, I’ve witnessed all of these examples over the years. But in the same country a resourceful nonprofit is also showing the world the future of public health with innovation and artificial intelligence.
Diabetes, which is discussed in the cover story in this issue of OSN, is seeing the most progress. This disease is an even bigger problem in Armenia than here in the U.S., where as much as one-third of our population is estimated to have diabetes or its precursor. In the U.S., diabetes ranks as the seventh leading cause of death. In Armenia, it is third, following heart disease and stroke. The Armenian EyeCare Project is addressing screening and treatment for diabetic retinopathy in a way that has been lauded as both novel and highly effective by the U.S. Agency for International Development. For 20 years, the project, under the direction of southern California ophthalmologist Roger Ohanesian, has created a tertiary care referral center in the capital city of Yerevan. About 15 years ago, a fully functioning traveling eye hospital in an 18-wheeler truck was deployed. And in 2017, five regional hospitals were created in the most remote and poor cities to provide nearby and free care to the most underprivileged in the population. This network of access points aims to eliminate preventable blindness, propelled by an award-winning public relations effort to raise awareness about screening for loved ones.
The project recently purchased or had donated cameras for each of these facilities made by Canon, Optovue and Remedied. Images obtained on these non-mydriatic cameras go to the cloud, where they’re analyzed by an artificial intelligence system called Eyenuk. Eyenuk analyzes each picture for patterns suggesting either diabetic retinopathy or proliferative disease. Photos are categorized as being low, medium or high risk. This information is relayed to the caregivers (technicians or non-surgical eye doctors) in the remote locations, and patients are instructed immediately as to their next steps. For most, follow-up in 6 months or 1 year is appropriate. For others, treatment may be possible in a regional clinic. Otherwise, a visit is arranged to the mobile hospital or referral center in Yerevan with funding provided for both bus travel and care.
For diabetic macular edema, patients can receive injections, but Avastin (bevacizumab, Genentech) is outlawed in this country — a ruling being challenged by the Armenian EyeCare Project. Only Lucentis (ranibizumab, Genentech) is permitted for diabetic use. While the maker of this expensive drug, Genentech, has so far been generous in providing supplies, supplies are limited. Steroid injections are also used appropriately; Ozurdex (dexamethasone intravitreal implant, Allergan) and triamcinolone have been employed quite successfully. Many undergo laser treatment for both macular edema and proliferative diabetic retinopathy using an Iridex yellow light laser, which is thought to offer similar efficacy, less collateral tissue damage and less pain than traditional argon green lasers.
To attract patients, public service announcements are delivered through television, Armenia’s most effective electronic portal. But word of mouth still reigns as the leading driver of patients, and the project partners with local doctors, priests, and town officials and gossips to get the word out. Because no one lives more than 50 miles from one of the five regional centers in Armenia, the Armenian EyeCare Project estimates that it should be able to make a first pass at screening and treating the diabetic population of the entire 3 million person country in about a year.
This novel treatment approach has been about 1 year in operation so far and is being examined by epidemiologists and nonprofits around the world. Its success has prompted support from both donors and drug/device companies that share in the interest of delivering safe, streamlined health care. If necessity is the mother of invention, Armenia has both in abundance, and its needy population coupled with a clever volunteer effort may well show the world the future of diabetic eye care.
Disclosure: Hovanesian reports no relevant financial disclosures. He has been a volunteer with the Armenian EyeCare Project since 1995.