Innovative AI System Addresses Eye Health Disparities in Remote Australia

An advanced artificial intelligence (AI) system designed to capture images of the retina is being implemented in remote areas of Western Australia. This initiative aims to mitigate the risk of blindness among rural and Indigenous Australians caused by untreated diabetes-related conditions.

The mobile health service focuses on screening for diabetic retinopathy, a condition that poses a significant threat to vision and is particularly prevalent in outback regions. It is recognized as the primary cause of blindness among working-age individuals globally.

This innovative service is spearheaded by the non-profit organization Lions Outback Vision in collaboration with technological partners, including Google and Topcon, along with support from the UCL Institute of Ophthalmology and the INSIGHT Health Data Research Hub at Moorfields Eye Hospital NHS Foundation Trust.

After a successful one-year pilot program, funding has been secured to expand the deployment of this mobile AI service in the Pilbara region, which is characterized by its sparse population and limited access to healthcare services. The Pilbara, home to approximately 45,000 residents, spans an extensive area of 507,896 square kilometers.

Under the current model, patients are screened, and their retinal images are sent to specialists for evaluation. If necessary, a consultant reviews the results. However, the new AI-integrated model offers immediate diagnosis at the point of care by analyzing the retina scan and providing instant telehealth consultations with ophthalmologists for high-risk patients. This mobile unit is designed to reach populations in areas where traditional screening services are unavailable.

The next phase of the project will utilize RETFound, the first AI foundation model specifically for eye care, to refine the diabetic retinopathy algorithm tailored for Indigenous populations. This model was developed in 2023 using a dataset of 1.6 million anonymized eye images from Moorfields patients, alongside other data sources. RETFound aims to enhance the accuracy of the AI system and expand its capabilities to identify cardiovascular diseases as well.

This technological advancement is poised to significantly benefit remote communities in Western Australia. Statistics indicate that Indigenous Australians face a blindness rate nearly three times higher than their non-Indigenous counterparts. Furthermore, cardiovascular disease is a major contributor to the disparity in mortality rates between these populations, as reported by the Australian Government.

Key figures in the project, including Australian ophthalmologist Dr. Mark Chia, have highlighted the collaboration between leading data scientists and medical AI experts from the UK and skilled eyecare professionals in Australia. This partnership has led to the development of a unique approach aimed at preserving vision and saving lives in remote regions.

Professor Angus Turner, a leading figure in the Lions Outback Vision team, noted that the rollout of the mobile retinal camera with immediate diagnosis capabilities across the Pilbara has already begun to yield significant positive outcomes. By enhancing access to eye screenings, this initiative ensures that individuals living in isolated areas receive timely diagnoses that can facilitate early treatment and prevent blindness.

Overall, this innovative AI solution has the potential to dramatically improve healthcare delivery in the Pilbara region, addressing critical health inequalities faced by its residents.