AI, when and how

OT has spoken to practitioners in the multiple and independent sectors to find out their views on how they believe AI will change optometry.

Positive potential

Optometrist and director of professional services at Boots Opticians, Claire Slade, is watching research into AI develop with interest. She believes that the technology could enhance the clinical care that is offered to patients in the testing room.

“The potential for AI in the testing room is fast advancing and it is very exciting. I think it will transform optometry by enabling the detection of certain eye conditions earlier, increasing patient safety and improving the long-term outcomes of those with eye and general health conditions,” Ms Slade explained.

We should not shy away from new technology that helps us detect eye problems , faster, with more precision resulting in better patient care

For independent practice owner and clinical director at the AOP, Dr Peter Hampson, while AI has not properly arrived into optometry just yet, it is “on the brink” of happening, he acknowledges.

“Though AI is not commonplace in practice currently, research being published by the likes of Moorfields and DeepMind certainly has potential. I see AI as a tool that sits on the shoulder of the practitioner and is able to act as a second opinion or expert consult. It would help optometrists in decision-making and would let them double-check things,” he explained.

For both optometrists, the introduction of AI into optometry is a positive. “If used properly and if used by optometrists, AI could really change how we deliver care. The potential is huge,” Dr Hampson told OT.

However, the optometrist highlights that while this technology is on its way, “there are still large challenges to tackle in making sure that practitioners are comfortable working in this way, as well as its connectivity in practice, which both need to be overcome first,” he said.

I see AI as a tool that sits on the shoulder of the practitioner and is able to act as a second opinion or expert consult

What is AI to optometry?

Discussing what AI means for optometry, Ms Slade explained: “Where I see the investment being placed is into how this technology can work in the testing room to diagnose a patient with particular eye conditions or screen for general health conditions through retinal imaging.”

“Currently, the main conditions that research is looking at are diabetic retinopathy, age-related macular degeneration and glaucoma,” she said.

While AI is currently being developed to support with the diagnosis of three or four global sight loss conditions, Ms Slade predicts that this pool will expand slightly in the future.

“I do see it being used for the detection of an increased number of conditions such as cardiovascular disease,” she shared.

As a result, Ms Slade sees the role of the optometrist expanding in the future. “The eye is the conduit to many conditions. Therefore, the optometrist could play a broader role in general health condition screening as a result of the support of AI, which is going to be really interesting. Optometrists already assess cardiovascular health by viewing the retina, but in the future this could be expanded to give a cardiovascular event risk score by simply taking a non-invasive retinal image. Patients understand that optometrists can already assess for general health conditions by viewing the retina. With machine learning, in the future this could be improved and expanded upon simply by taking a retinal image,” she explained.

The role that the optometrist could play in general health condition detection as a result of the support of AI is going to be really interesting. It is a role that we don’t currently play, but there is definitely a conversation to be had about it

Testing support

So, what does AI look like in the testing room? AI will be introduced into the testing room through imaging technology, the duo agreed.

Both practitioners see the technology being built into devices such as OCT and topographers, “maybe even in visual field screeners,” Dr Hampson said.

“Within these devices will be algorithms that can detect the eye conditions,” Ms Slade said.

Dr Hampson predicts that in as little as five years’ time versions of these devices with a built in AI function will be available on the market. However, he envisages a slower emergence of the technology into practice as a result of the way the profession purchases technology.

“Investment into technology is done in cycles,” Dr Hampson explained. “You have early adopters, such as those who invested in OCT some years ago, then, the larger wave comes and it begins to filter into more practices. A lot of practices are just putting their toes into the water with OCT and I can’t see this large group, who have only had OCT for around 18 months wanting to invest in the next level with AI just yet,” he added.

Enhancing eye care

For both optometrists, AI entering the testing room will have an impact on the level of care and the access to care that optometrists can offer.

“With the needs of the ageing population, paired with the problems that we have with access to care, AI could enable optometry practices, if connected in the right way, to deliver current medical retina consultant levels of care on the High Street and on the doorsteps of patients,” Dr Hampson said.

“Global sight loss is on the increase exponentially,” Ms Slade concurred. “So it is absolutely right that we should be using technology to help us detect and diagnose these conditions quicker,” she said.

Acknowledging that within the profession there may be a nervousness by some about being replaced by machines, Ms Slade responded: “Humans are naturally sceptical of machines doing anything automated – we have all seen the Hollywood films where humans are replaced.”

“Firstly, we must decide how much trust we can put into these systems, but there is no doubt that AI is a form of technology that can really aid optometrists in our objective to help people see better for longer, and having good sight enables a better quality of life,” she added.

“Practitioners may be scared that AI will replace them, but actually most studies show that if you use AI to supplement rather than replace, the outcomes are much better – AI is good, but still tends to be quite literal,” Dr Hampson emphasised.

The retail model that has sustained optics for the last 50 years cannot and will not last, and AI will enable us to further explore the clinical delivery model of care

To complement rather than replace

While Ms Slade believes that AI will “transform optometry” by enabling the detection of certain conditions earlier, she does not feel that machines will be able to successfully detect all conditions and therefore sees this technology as a tool that complements the skills of optometrists.

“Experts who I speak to in AI research are unified in that there are thousands and thousands of eye conditions out there and the machines will not be able to learn them all. AI will be there to assist not replace optometrists in the testing room,” she emphasised.

However, Ms Slade acknowledges and is excited that there are things that AI-driven machines will be able to do that practitioners cannot. “Machines can tell if a patient is male or female just by looking at an image of the eye,” she highlighted.

Take notice

For Dr Hampson, the profession should take notice of AI because “it will allow us to practise in the way that many of us clinicians want to.”

“The retail model that has sustained optics for the last 50 years cannot and will not last, and AI will enable us to further explore the clinical delivery model of care,” he said.

Ms Slade encourages the profession to embrace the potential AI holds for optometry. “We should not shy away from new technology that helps us detect eye problems faster, with more precision resulting in better patient care. We must not forget that the ophthalmoscope, the fundus camera and OCT were all new once too,” she said.

When asked to predict how AI will be used in the future, Ms Slade said honestly: “We are just on the precipice of understanding AI so who knows where it will go in the future. There is not one area where I think that this technology definitely could not help or support.”

Dr Hampson concluded: “I think when someone believes that something cannot be improved upon, that is when it becomes dangerous and at risk of becoming redundant.

“I think most things could be improved through AI.”

Image credit: Shutterstock