Australian AI screening trial looks to improve eye health inequality
Article sourced from Insight.
Led by the Centre for Eye Research Australia’s (CERA) Professor Mingguang He and Research Fellow Dr Stuart Keel, the trial will see 10 outlying primary care services in Western Australia introduce the AI screening system into their practices. It will be testing for signs of diabetic retinopathy, glaucoma, age-related macular degeneration (AMD) and cataracts.
Within seconds, the system analyses retinal images and generates a report indicating whether the patient needs to be referred for further assessment and treatment. The study will test the clinical effectiveness of the new technology, and evaluate how it is accepted by patients and health professionals compared to standard care.
The three-year project – conducted in partnership with the Brien Holden Vision Institute (BHVI) – aims to provide greater access to eye examinations for remote and regional communities, including Indigenous Australians.
Keel said the AI system – which CERA developed with technical partner Healgoo Interactive Med-Tech – had been extensively evaluated using images from Australia, Singapore and China, achieving an accurate diagnosis more than 90% of with time.
“By running a pragmatic clinical trial in remote and Indigenous health centres, the project will assess the real-world accuracy and cost effectiveness of integrating retinal photography and artificial intelligence to build opportunistic screening in rural and remote primary care settings. It will also help identify potential barriers and solutions to the uptake of this service in the local context,” Keel said.
“Our new technology would enable non-eyecare health professionals in regional and remote primary care settings to conduct screening quickly and accurately.”
According to the researchers, up to 500,000 Australian adults suffer from vision impairment or blindness. Indigenous Australians are most at risk, experiencing vision impairment and blindness at three times the rate of non-Indigenous Australians. Late-stage diabetic retinopathy is also disproportionately prevalent in Indigenous communities.
Professor He said: “Artificial intelligence could close the significant gap in eye care services and considerably increase early diagnosis of the four most common blinding eye diseases and reduce the burden of vision loss in the Australian communities that need it the most.’’
The trial is being funded with $1 million from the National Health and Medical Research Council (NHMRC) Partnership Projects Scheme, and matching funds from partners Venture Store and support from BHVI.