• Politics
  • Diversity, equity and inclusion
  • Financial Decision Making
  • Telehealth
  • Patient Experience
  • Leadership
  • Point of Care Tools
  • Product Solutions
  • Management
  • Technology
  • Healthcare Transformation
  • Data + Technology
  • Safer Hospitals
  • Business
  • Providers in Practice
  • Mergers and Acquisitions
  • AI & Data Analytics
  • Cybersecurity
  • Interoperability & EHRs
  • Medical Devices
  • Pop Health Tech
  • Precision Medicine
  • Virtual Care
  • Health equity

IDx Is Ready for AI to Break Through in 2019


Autonomous AI can improve healthcare productivity and lower costs for patients, according to IDx founder, Michael Abramoff, M.D., Ph.D.

artificial intelligence

This past year was a big one for IDx, whose first product, IDx-DR, became the first autonomous artificial intelligence (AI) diagnostic system to gain approval from the U.S. Food and Drug Administration (FDA). Next year is poised to be another big year for AI in general as IDx and many other tech heavyweights strive to continue the push for easier access to autonomous artificial intelligence systems.

We recently spoke with IDx founder Michael Abramoff, M.D., Ph.D., to learn more about his company, his competition and how autonomous AI might affect healthcare and patients in the near future.

“Autonomous AI systems have massive potential to improve healthcare productivity, lower healthcare costs and improve accessibility and quality,” Abramoff said in a recent release.

>> READ: How AI Is Shaking Up Healthcare, Beyond Diagnostics

IDx’s Plans for Autonomous AI

IDx-DR, was designed to detect diabetic retinopathy, the most common cause of vision loss among patients with diabetes and a leading cause of blindness among working-age adults. Its FDA clearance earlier this year was a major milestone that will likely help propel a market with great promise for medicine.

But what is autonomous AI? These kinds of systems do not require a physician to interpret the images or results. This drives down cost and makes it easier for patients to get answers about their health more quickly.

Abramoff said autonomous AI solutions will shift the point-of-care from specialty clinics to primary care. Not only could this increase access for patients, it could lower the cost, depending on a patient’s copay.

For example, patients with diabetes are supposed to get their retina checked once a year for blindness. Patients need to make appointments with their eye care provider three months in advance, and most people forget. Abramoff told Healthcare Analytics News™ that roughly 50 to 70 percent of patients don’t get the exam, even though they need it.

Abramoff suggested that these exams must happen inside the primary care provider’s office. With autonomous AI, the exam takes only a few minutes, and the system can be operated by anyone who has had proper training. The patient sees a physician who analyzes the results and determines what needs to be done. If the diagnostic test yields abnormal results, the patient is then referred to an eye doctor.

The goal for Abramoff is to reach more patients than he did this year with his IDx-DR system. While he didn’t provide the number of patients who had used the technology, he said many received referrals and “thousands” have used it. Multiple health systems have adopted the system and are evaluating how it affects their workflow.

IDx is working on diagnostic autonomous AI for macular degeneration and glaucoma, skin and ears and Alzheimer’s disease. Abramoff is working on a screening test for Alzheimer’s, but he’s facing challenges because autonomous AI is most suitable when a large group is at risk and a good treatment plan already exists.

The Future and Challenges of Autonomous AI

IDx is not the only company working on autonomous AI systems — and Abramoff said his competition should increase next year.

Researchers at Google are working with doctors to develop AI that could automatically identify diabetic retinopathy through examining retinal photos. Google hopes this technology can screen more patients for the condition than doctors can on their own. The technology grew stronger after after human beings judged a small subset of retinal images, which further trained the technology and resulted in fewer errors.

In October, CA Technologies unveiled a new platform that the company claims will allows IT teams to automate or eliminate key tasks and create self-healing applications. The company leverages innovative AI, machine learning and automation capabilities.

Additional companies are looking to expand the space by working on autonomous AI systems for other organs. More clinical trials for new AI systems will likely occur.

Going forward, Abramoff thinks that interoperability between AI technologies and electronic health records (EHRs) will increase. The results of AI tests will be most effective if they enter the EHR, he said. Then the two systems can automatically exchange information, like scheduling, diagnoses and results. If there is no interoperability, physicians are not maximizing the gains they have and are not increasing efficiency in the workplace.

AI experts are facing some challenges with this technology, though.

>> READ: Tech Is Targeting Diabetic Retinopathy. But What Does the Market Look Like?

Abramoff said that there are not good, objective data for everything, which impedes the development of AI systems for certain clinical areas, like pain.

There are also very few regulatory guidelines for how this technology can be used, and Abramoff and other health-tech insiders argue that the approval process needs to be quicker. It took eight years for the FDA to approve the IDx-DR, as the agency was unsure whether it was safe for patients. Federal agencies and Congress are now working on how to implement AI into healthcare safely, and the FDA has also made strides to streamline the process.

Despite the excitement of many, some experts still oppose to the implementation of autonomous AI.

Tom Dietterich, president of the Association for the Advancement of Artificial Intelligence, believes fully autonomous AI systems should never be built.

“By definition, a fully autonomous system is one that we have no control over,” Dietterich said at DARPA’s “Wait, What?” conference in 2015. “And I don’t think we ever want to be in that situation.”

Get the best insights in healthcare analytics directly to your inbox.


Another Start-Up Advances AI's Assault on Diabetes

Who Is Responsible When AI Fails?

Ethical Concerns for Cutting-Edge Neurotechnologies

Recent Videos
John Glaser
Shereef Elnahal, MD
Related Content
© 2024 MJH Life Sciences

All rights reserved.