Vision loss is a global issue with vast impacts across all populations. The startling fact is, much of that vision loss is preventable. Yet increases in common causes of vision loss have been reported in recent years due to worldwide population growth and aging.¹ There is an urgent need to address and reverse such a trend, as it leads to lasting negative impacts to personal autonomy, and individual and family financial security if not resolved. A recent partnership between two vision health champions may be exactly what is needed to address the global issue of preventable vision loss attributed to diabetic retinopathy and diabetic macular edema.
Vision loss solutions on a global scale
Orbis International is a nonprofit organization dedicated to reducing global blindness and vision impairment. Four decades of their focused work include the world’s first fully functioning airborne teaching eye hospital.
Orbis’ hospital-based training programs and permanent offices are located throughout the world to improve quality and accessibility of eye care, and to focus on treatment and prevention of blindness and disease. Orbis and Digital Diagnostics recently partnered with the goal to continue bringing those improvements and accessibility to even more people in need.
Artificial intelligence (AI) impact to preventable vision loss
Dr. Hunter Cherwek, VP Clinical Services at Orbis International, recently interviewed Michael Abramoff, MD, PhD, founder, and executive chairman of Digital Diagnostics, on the acclaimed Orbis Sightlines video and podcast series. During the segment, they take a deep dive into the profound partnership and Digital Diagnostics global initiative to eliminate unnecessary patient suffering of treatable diseases using autonomous AI technology. The episode, “AI & Simulation: A New Frontier” turned a lens on adding the advantage of autonomous AI to fight global blindness, and the implications of such technology on the future, as well as potential outcomes of the partnership between Orbis and Digital Diagnostics. Such a partnership could further “research and future solutions to bring this technology to healthcare professionals around the world.”
Digital Diagnostics’ IDx-DR is an autonomous AI diagnostic system that diagnoses patients for diabetic retinopathy and macular edema at the point of care. The system does not require specialist overread and can be operated by existing clinical staff, making it a simple and effective, FDA De Novo cleared tool to bring specialty care to the people that need it most in a primary care setting.
“The exam is done right then and there within minutes, and there is no human oversight of the medical decision,” Abramoff explained. “[Results are] still discussed with the patient, typically by a doctor, so the doctor is still in the picture. But it is not an ophthalmologist anymore. It means the computer makes the medical decision and makes the call whether this is diabetic retinopathy or [macular edema].”
Dr. Abramoff, a neuroscientist, ophthalmologist, and computer engineer had a dream to use autonomous AI to prevent avoidable blindness. He knew he could make a difference in so many lives, especially when it came to accessibility. “The reason I started this company was to save healthcare costs, improve care quality, and make it accessible everywhere around the world,” he went on to say.
An important component of delivering quality care is the influence of bias. Autonomous AI can help to mitigate the human element to deliver access and an unbiased diagnosis based on science and research, if done the right way. “I started worrying early on about racial bias, and about bias in general in AI,” Dr. Abramoff said. “[One example is that] we built detectors machine learning-, deep learning-based to identify biomarkers, such as hemorrhages and exudates. That builds in an invariance to race and ethnicity in the background color of the retina because it does not matter what background color the retina is. If you have a hemorrhage and an exudate it’s likely diabetic retinopathy.”
There are many aspects to mitigating bias in autonomous AI in addition to the design elements mentioned, including addressing potential bias in the indications for use, validation, and deployment of AI systems. Their conversation continued through the ethical foundation for AI Abramoff has developed, the future of such tools, and the potential long-term impact. “Obviously, we need to think about this as humans, as well as technicians, and clinicians,” said Dr. Hunter Cherwek.
The future of global vision care
The right focus and resources can turn the tide for many millions across the globe who face a wholly preventable loss of vision outcome. Orbis International and Digital Diagnostics together can extend unbiased, accessible, and preventative care to those who need it. That means improved outcomes for so many, who can go on to live lives with the benefit of healthy vision.
Watch Orbis Sightlines episode 8, “AI & Simulation: A New Frontier.”
Learn more about IDx-DR, the first autonomous AI diagnostic system for use in primary care that analyzes images of the retina for signs of diabetic retinopathy and macular edema.
- Flaxman, S R et al (2017). Global causes of blindness and distance vision impairment 1990–2020: A systematic review and meta-analysis. The Lancet, Global Health. Volume 5, Issue 12, December 2017, Pages e1221-e1234. Accessed February 28, 2021. https://www.sciencedirect.com/science/article/pii/S2214109X17303935