This article appears courtesy of mivision.
Centre for Eye Research Australia (CERA), in conjunction with the University of Melbourne is conducting research to gain a better understanding of how artificial intelligence (AI) influences clinical decision-making by a range of health professionals engaged in diabetic retinopathy (DR) screening. Applications that use AI to interpret medical images are increasingly emerging, yet little is known about how the support provided by AI affects clinical decision-making. For example, it is not clear whether AI improves confidence and decision-making speed or whether it might slow some clinicians and reduce their performance by making them second-guess their decisions.
The aim of this study is to investigate the impact of an AI platform for DR grading on decision-making by general practitioners, orthoptists, optometrists, Indigenous eye health workers, as well as trainee clinicians. If AI programs prove to be effective in assisting eye health professionals interpret retinal photographs, they could be particularly useful in areas that have limited access to eye specialists.
Led by medical doctor and Master’s research degree candidate Dr Philip Rothschild, under the supervision of Associate Professor Peter van Wijngaarden and biostatistician Dr Myra McGuinness, the study will determine the impact of an AI platform on speed, confidence, and accuracy of grading.
“AI decision support systems are important as they interpret data autonomously, providing real-time assistance to care providers. The question that we are trying to answer is whether this assistance is ultimately helpful, ” explained Dr Rothschild.
Potential to be Embraced
Earlier research that Dr Rothschild completed as part of his Master’s program, indicated that health professionals are interested in adopting AI for diabetic eye disease detection and grading.
“My initial study assessed ophthalmologists’ perceptions of AI,” he explained. “Participants were a mix of trainees and ophthalmologists, and while both groups expected AI to have a positive impact on eye care, they did have some reservations about the technology.”
Dr Rothschild reported that 71% of respondents held overall positive views of the potential for AI to improve their field of medicine.
“Improved access to disease screening was one of the most commonly reported potential advantages of AI, which helped to inform the design of the follow-up study we’re now undertaking. Ophthalmologists indicated that AI programs could help to remove barriers to DR screening,” he said.
A Team Effort
As a clinician and research student who hopes to eventually work as an ophthalmologist, Dr Rothschild said the project is an exciting one to be working on.
“I reached out to Professor Peter van Wijngaarden in 2018, when I was a final year medical student at University of Tasmania, to express interest in undertaking research at CERA, to further my education in ophthalmology and contribute to the field. He was happy to take me on, and we discussed potential areas of research I could be involved in.
“AI was emerging within eye care and other researchers at CERA were developing a deep learning algorithm to detect diseases, including DR, from retinal photos. We identified that little was known about how eye care providers might interact with the technology and so that was the direction that I chose for my research.”
“I feel quite fortunate to have very supportive supervisors and to be working at the cutting edge of eye health. Sometimes I find it hard to believe that I’m here, helping to address important questions and being part of an organisation that is a world leader in eye research. I’m really appreciative of the opportunity to be involved.”
The survey is available at redcap.link/ipnlckus.