Like many professionals, medical executives would be forgiven for believing that a huge wave of Artificial Intelligence fueled transformation is about to disrupt everything about the way they work, and earn a living.

The truth, as ever,  is a little more sanguine. Of all the professions, healthcare is likely one of the safest. In the end, it’s about care. And care, in all its detailed forms, isn’t easy to replicate on machines.

Still, technology will usher in major changes.

The most immediate impact of AI in medicine, for instance, will be felt in diagnosis. That’s already apparent in areas such as radiology. But as Professor Enrico Coiera, Director of the Centre for Health Informatics at Macquarie University, and leader in the AI health debate says: “Diagnosis is just one day in the life of a patient’s healthcare journey. Everything that follows diagnosis makes up the majority of healthcare, so healthcare professionals haven’t got much to worry about.”

  • Register today: Professor Enrico Coiera is a keynote speaker at the Wild Health Summit on October 16 in Sydney where he will also chair a panel on the rapid introduction of AI into healthcare. Which-50 readers can use the promo code WILDTMR for a 20 per cent discount on their ticket to the summit.

“Healthcare won’t run out of jobs, because there is such a huge, unmet need for care. If we were to cure the top 10 diseases, we would then get to focus on the next 10, and so on.”

Coiera sees the efficiencies inherent in medical AI as helping the problem of this unmet need.

But he is also worried that between the business of technology, the desire for speed to solve efficiency issues in the face of cost pressures in health, and the iterative nature of the introduction of AI into medical workflows, we are missing something vital.

“A lot of decisions in healthcare are going to be shaped by input from algorithms and while that is overall a good thing, we need to understand much better in what circumstances it will also be a problem. We need to be very clear about the framework in which we operate when we’ve made that conscious decision to hand over part of the care to an algorithm,” Coiera said.

Clear evidence

He warns that often we are introducing AI into medical processes without clear evidence of how we should be working with that technology, or without providing proper training to the healthcare professionals.

A simple example he cites is computer-based prescription writing, which is now becoming widespread.

“When you prescribe as an unaided human there are a set of errors you will make, and bringing in a computer can help to significantly reduce those errors. The issue is that using a computer introduces new types of errors, and we aren’t teaching people how to use the technology to manage these new types of error,” Coiera said.

He cites the recent example of a rural GP whose prescribing error led to the death of a patient. The GP’s defence was “the computer didn’t tell me it was wrong.”

“That’s automation bias”, says Coiera, who believes that computers aside, all clinicians need to be trained in the safe use of an electronic medical record, given how pervasive this technology is starting to become.

In a recent blog on the BMJ, Coiera said: “Like any technology, AI must be designed and built to safety standards that ensure it is fit for purpose and operates as intended.”

He also said: “Humans should recognise that their own performance is altered when working with AI … If humans are responsible for an outcome, they should be obliged to remain vigilant, even after they have delegated tasks to AI.”

Coiera remains a very strong AI advocate, maintaining that its potential to increase efficiencies in an over-stretched healthcare system is enormous.

But he is also worried that the evidence for safety around the introduction of AI into medicine remains mostly slim, and that regulators rarely if ever get a look in when something iterative occurs inside existing automated systems.

“In aviation, if you develop a new model of aeroplane you have to meet stringent testing and accreditation guidelines before it can be used for the public. We don’t say, ‘oh, aeroplanes are a great idea overall, so new aeroplanes are OK to go’,” he said.

“There is currently no safety body that can oversee or accredit any of these medical AI technologies, and we generally don’t teach people to be safe users of the technology. We keep on unleashing new genies and we need to keep ahead of it.”

Professor Enrico Coiera is a keynote speaker at the Wild Health Summit on October 16 in Sydney. He will also be on the Q&A panel of the same meeting which will be asked how we meet the challenges of allowing the rapid introduction of AI into medicine while keeping the technology safe for patients. Go to https://wildhealth.net.au/ and check out the other sessions for this important one day summit and use the promo code WILDTMR for a 20% discount on your ticket.

LinkedIn
Previous post

Video: How marketing operations can simplify compliance

Next post

LinkedIn loses bid to stop startup scraping its public data

Join the digital transformation discussion and sign up for the Which-50 Irregular Insights newsletter.