CERA

Science and Research

Ophthalmic epidemiology research

CERA’s population health research is focused on understanding blinding eye disease in the community and developing innovative strategies to reduce its impact, using advanced technology and data science.

Overview

Big data and advanced artificial intelligence technologies are a key focus of our ophthalmic epidemiology research. In one major project, our world-leading scientists have developed deep learning systems to screen for common blinding eye diseases including diabetic retinopathy, age-related macular degeneration and glaucoma, as well cardiovascular disease risk and other conditions associated with ageing. These technologies have been trialled in clinical environments such as GP clinics, endocrinology clinics, cardiology clinics and primary health facilities in remote Aboriginal communities. In another major project, working with an industry partner we are combining a low-cost, portable and fully automatic retinal camera with artificial intelligence screening software to detect 'red flags' in people experiencing potentially life-threatening headaches. Our team used deep learning to develop prediction models for diabetic retinopathy, age-related macular degeneration and glaucoma. We have also established a national data linkage project linking ocular imaging data to systemic health diagnostic codes for the purposes of exploring retinal biomarkers of systemic disease. Other key population health projects at CERA include the first National Eye Health Survey – a study into the prevalence and causes of vision impairment in non-Indigenous and Indigenous Australians – and initiatives to increase participation in diabetic eye disease screening, such as KeepSight. CERA is proud to be designated a WHO Collaborating Centre for the Prevention of Blindness, the only such centre in Australia.

Why this research is important

Our research is developing efficient, accurate and convenient ways of delivering screening services for common blinding eye diseases, cardiovascular disease risk, and ageing risk. This will improve patient outcomes and reduce the burden and cost of these diseases to the community.

Our artificial intelligence-powered smart camera screening for “red flags” at the point-of-care for people experiencing severe headaches will reduce the risk of missed and delayed diagnosis of life-threatening conditions by supporting decisions being made in the emergency department.

Our national data linkage project will result in new databases and collaborative networks that will benefit both deep learning and clinical research.

Key research questions

  • How can we increase participation in screening for common blinding eye diseases such as diabetic retinopathy?
  • How can we increase participation in screening for cardiovascular disease?
  • How can artificial intelligence tools improve the diagnosis and treatment of eye disease?
  • How can these screening systems be applied in different care models in non-ophthalmology settings, like GP clinics, endocrinology clinics, cardiology clinics, emergency departments and Indigenous health services?
  • What is the accuracy, cost-efficiency and acceptancy of these care models in real-world practice?
  • How can we integrate existing and further-evolved deep learning technology to develop and validate a clinical decision system that is able to predict disease outcomes and prognosis?