Eye diseases such as diabetic retinopathy, glaucoma, and age-related macular degeneration affect millions of Americans. Diabetic retinopathy is one of the main causes of vision problems and blindness among working-age adults. Over 3 million people in the United States have glaucoma, and this number is likely to increase by 50% in the next 20 years worldwide. Finding and treating these diseases early is very important to stop vision loss. But many people have a hard time seeing eye specialists because there are not enough doctors, especially in rural or less-served areas.
AI-assisted screening and automation can help solve these problems. AI tools can quickly look at medical images, like retinal scans, find early signs of disease, and tell doctors which patients need follow-up. This helps clinics see more patients and may lower the need for many specialist visits.
In eye care, AI is mostly used to automatically screen for diseases like diabetic retinopathy and glaucoma. Dr. Travis Redd from the Casey Eye Institute says AI is mainly used for screening diabetic retinopathy. There are three AI software tools approved by the FDA in the United States for this. These AI systems look at retinal images and can detect disease with about 90% accuracy, similar to or better than human experts.
For glaucoma, AI tools like Optain Health’s Eyetelligence Assure Suite have early detection accuracy of over 95%. They can also cut the time to diagnose by about half. These systems use machine learning trained on many eye scans to find small signs of disease that people might miss. Quicker diagnosis helps patients get treatment earlier, which can save their vision.
Even though AI has strong potential in eye care, there are still problems with data that slow down its wide use in the U.S. One problem is not having enough large and varied datasets to train AI programs well. Eye data, like retinal images, come from many different machines. Many machines do not follow the standard DICOM format. Without the same format, AI results may differ and it is hard to connect AI with other practice systems.
Also, privacy rules like HIPAA in the U.S. limit how patient data can be shared. This makes it hard for different hospitals and clinics to work together to build better AI. This slows down the work needed to make reliable AI systems. Dr. Redd says AI companies need to work closely with eye doctors to make sure the tools answer real medical questions and keep high quality.
A survey by the American Medical Association (AMA) asked over 1,000 U.S. doctors about AI. Almost two-thirds saw benefits in using AI in healthcare. But, only 38% were actually using AI tools in 2023. Many doctors worry about AI causing bias, breaking patient privacy, and new legal risks. Dr. Jesse M. Ehrenfeld, the AMA President, says it is very important that humans stay involved. “Patients need to know there is a human being on the other end helping guide their course of care,” he said.
It is important for AI to be clear and for doctors to stay responsible when using it. Developers must make sure AI tools handle patient permission, protect data, and avoid mistakes—especially for different groups of people where AI might not work as well.
Using AI to automate screenings can make it much easier for patients to get eye care, especially those who find it hard to get appointments. AI can do the first round of checking and decide who needs to see a specialist and who can wait. This helps reduce long waits at specialist clinics and lets care reach places like primary care offices, pharmacies, or mobile programs.
For example, AI retinal screening machines can be used in community clinics. Diabetic patients can get quick testing there without waiting a long time for a specialist. This means they get diagnosed and treated sooner, which helps slow down the disease and protect their eyesight.
AI also helps telemedicine grow. Patients in faraway or poor areas can get eye screenings and follow-up care without traveling. In the future, AI-powered wearable devices might watch eye health constantly and warn early if something changes.
AI helps not only with screening but also with running clinics better. Medical office workers and IT managers can use AI to automate routine front and back office jobs. This makes clinics run more smoothly.
Busy clinics deal with scheduling, phone calls, insurance checks, and paperwork. Simbo AI is a company that uses AI to answer calls and handle appointment reminders and insurance questions. This lets staff spend more time on tasks that need people.
AI can also help with clinical records. AI that understands natural speech can write up notes from doctor-patient talks. This reduces paperwork for eye doctors and lets them spend more time with patients. Research shows 56% of doctors believe AI can help lower paperwork, which is important to stop doctors from getting too tired.
AI is also useful for reading eye images fast and accurately. This cuts mistakes caused by tiredness or different interpretations. It makes diagnosis quicker, reduces the need to take images again, and helps manage patients better.
AI can also sort urgent cases by using special programs. These programs can spot cases that might cause serious vision loss and send them for quick review, making sure patients get care on time.
For those managing eye clinics and thinking about using AI, there are important things to consider:
Using AI in eye care is part of a bigger change in healthcare to use digital tools for better results and efficiency. The AI healthcare market in the U.S. is expected to grow a lot— from $11 billion in 2021 to about $187 billion by 2030. Tech companies like IBM, Apple, and Microsoft are investing in AI healthcare.
Automating office tasks such as scheduling, billing, and answering calls with AI helps clinics lower costs and give better patient service. For example, AI chatbots and virtual assistants can help patients anytime by sending reminders or answering questions outside clinic hours.
AI does not replace eye doctors but helps them. Dr. Eric Topol said AI should be seen as a “co-pilot” that helps doctors by giving data insights, lowering mistakes, and creating care plans fit for each patient.
Medical practice administrators, owners, and IT managers who learn about and use AI in eye care will be able to help more patients, run clinics better, and improve care results. There are challenges, but working together with technology makers, doctors, and policy leaders will guide safe and effective AI use in eye health.
AI has the potential to transform ophthalmology practices by enhancing diagnostics, patient management, and streamlining administrative tasks, ultimately improving accessibility to care.
AI is mostly implemented in autonomous screening for diabetic retinopathy, with only three FDA-authorized AI-enabled software as medical devices for this purpose in the U.S.
Barriers include small data sets for training AI, difficulties in data sharing due to privacy concerns, and non-standardized imaging formats.
AI can automate initial screenings, helping identify patients who need further care while reducing unnecessary appointments, thereby increasing care accessibility.
Large language models could be integrated into electronic health records to automate clinical documentation and reduce administrative burdens of physicians.
Dr. Redd recommends improving data sharing, ensuring DICOM compliance in medical imaging, and developing appropriate reimbursement models for AI usage.
Collaboration ensures that AI models develop clinically relevant solutions and address meaningful questions, enhancing their added value.
Physicians express worries about AI potentially introducing bias, risking patient privacy, and creating new liability issues.
Nearly two-thirds of physicians indicated they see advantages to using AI, but only 38% were actually using it as of 2023.
AMA President Jesse M. Ehrenfeld emphasizes the importance of a human guide in patient care, regardless of AI’s potential advancements.