A new approach to retinal screening using artificial intelligence (AI) is being hailed as a breakthrough for rural and remote communities. LYNN LOGANATHAN from Rural Health West reports.
In Australia’s vast and isolated regions, access to healthcare often hinges not on the severity of a person’s condition, but on how far they live from the nearest specialist. This distance can mean delays in diagnosis, missed appointments, and preventable loss of sight.
Lions Outback Vision has been conducting real-world implementation of AI-enabled retinal screening that’s been hailed as a breakthrough in rural healthcare. Validated through research at Derbarl Yerrigan Health Services and in the Pilbara to detect diabetic retinopathy (DR), the tool is proving its value not only in streamlining access to eye care, but in redefining the role AI may have in delivering equitable healthcare for rural and remote communities.
“We’re not just talking about screening eyes,” said Lions Outback Vision McCusker Director, Professor Angus Turner. “We’re talking about screening for signs of systemic disease, earlier and more accurately, in communities that have traditionally been left behind.”
Why diabetic retinopathy matters
Diabetic retinopathy is one of the leading causes of preventable blindness in Australia. Yet despite how treatable it is, many cases go undetected until significant and irreversible damage has occurred – particularly among Aboriginal and Torres Strait Islander peoples.
“About 98 per cent of blindness caused by diabetic retinopathy is preventable,” Angus notes. “But too few people are being screened, especially in rural and remote communities.”
Unlike countries such as the United Kingdom, Australia has no national screening program for diabetic retinopathy. While there are Medicare item numbers for general practitioners to conduct diabetic retinopathy screening (Item 12325 for Aboriginal and Torres Strait Islander patients and Item 12326 for non-Aboriginal patients), uptake has been minimal – with only around 8,000 claims made since its introduction over nine years ago. The majority of those claims have occurred in Western Australia, thanks to targeted promotion efforts by the Lions Outback Vision team.
A smarter, faster path to diagnosis
The AI-powered screening tool developed by Lions Outback Vision addresses two critical barriers: time and continuity of care. Traditionally, diabetic retinopathy screening involves capturing retinal images and sending them to an ophthalmologist for review – a process that can take weeks. In transient populations, or where follow-up is difficult, this delay often results in patients missing out on vital next steps.
With the new system, the AI algorithm analyses retinal images on the spot, providing a near-instant result back to the clinician. The clinician can then engage a specialist via telehealth to speak directly with the patient and guide further care – all in one appointment. “This is really about collapsing the care pathway,” said Angus. “It’s practical, scalable and culturally safer – and in remote areas, that makes all the difference.”
Training AI to work for everyone
Selecting the right AI platform was critical. Some of the AI systems researched in clinics by Lions Outback Vision showed poor accuracy when used with Aboriginal patients – a clear example of how ethnic bias can creep into machine learning tools if datasets lack diversity.
The team ultimately selected a platform developed by Google, which used self-supervised learning to significantly improve its specificity and accuracy – achieving 97 to 99 per cent accuracy when benchmarked against human ophthalmologists. Thousands of scans were manually reviewed and compared to AI-generated results, ensuring high standards of safety and reliability.
“Bias in AI is a real issue, especially in healthcare,” Angus explained. “We needed a system that works just as well for Aboriginal patients as anyone else. That’s non-negotiable.”
The eye as a window to health
What makes this innovation particularly exciting is its broader potential, with retinal imaging able to detect early signs of other systemic diseases, such as hypertension, stroke and neurodegenerative conditions. “There’s a saying that the eye is a window to the body – and that’s quite literal. We can see blood vessels, nerve layers and other structures that reflect what’s happening in the brain, heart and vascular system.”
Lions Outback Vision’s research underway is already helping reframe eye health as part of a more holistic model of care – one that links vision with chronic disease management, health promotion, and digital health infrastructure.
The idea is deceptively simple. Trained non-specialists, such as Aboriginal Health Workers or nurses in remote clinics, use the AI tool to scan a patient’s eye. The images are instantly analysed by the software, which flags any concerns and recommends further investigation or treatment. No specialist needs to be on-site. No days off work or long trips to the city are required. In minutes, the patient has access to a non-invasive and immediate screening that may help catch disease early – often before symptoms appear.
AI as an equaliser
The success of Lions Outback Vision highlights a broader promise; AI may serve as a crucial workforce multiplier, enhancing the reach and capability of health professionals in the regions. For Aboriginal communities in particular, who face higher rates of chronic disease and poorer health outcomes, the ability to provide early intervention in a culturally safe, local setting could be life changing.
“Ultimately, we see this as a tool that strengthens local primary care,” said Angus. “It enables more proactive, preventative healthcare, and it empowers local providers with real-time decision-making support.”
Imagine AI-supported triage in emergency departments where staffing is thin. Picture smart diagnostic tools assisting midwives or community nurses to make more confident calls in isolated towns. Think of wearable tech that flags early signs of health deterioration and alerts clinicians before an adverse event occurs. These aren’t futuristic dreams – they’re real possibilities, some already in development.
Crucially, AI can also support health equity by standardising care. “Whether you live in Karratha or Kalgoorlie, in Fitzroy Crossing or Fremantle, a well-trained AI system offers the same standard of diagnostic support. That’s a powerful tool for overcoming health disparities.”
Integrating AI
Of course, technology alone can’t solve everything. The real test lies ahead; embedding the model sustainably into rural and remote healthcare systems. Firstly, the regulatory steps for implementation of new software as a medical device need to be undertaken.
Feedback from primary care has been clear that the use of the AI needs to integrate with existing electronic medical record workflow, as well as needing a sustainable business case for the capital purchase and operation of the camera. In addition, building trust in new tools – especially in communities that have experienced decades of systemic health inequities – requires deep cultural engagement, local partnerships and ongoing education.
The team continues to work closely with Aboriginal Medical Services, regional GPs and outreach clinics to support adoption, training and integration. They also hope to see broader awareness among GPs about the MBS items for diabetic retinopathy screening, and long-term support for wider rollout of AI-enabled tools. This is not just about a single piece of technology – it’s about making high-quality, preventative healthcare accessible no matter where you live.
As Angus puts it: “If we can diagnose vision loss early, prevent blindness and spot other risks while we’re at it – then we’re not just improving outcomes, we’re transforming the model of care.”
Credit: Rural Health West
Sign up to Keepsight to protect your eyes
KeepSight is a free diabetes eye check reminder program that makes it easier to take care of your eye health. Register at keepsight.org.au and you’ll receive a reminder to make an eye appointment and help with finding an optometrist if you need one.
