“And it is coming quickly to a care setting near you,” said Cris Ross at Health 2.0 on Tuesday, touting “small AI and big AI” tools that can help revamp IT systems to improve the experience of clinicians and patients alike.
Mayo Clinic Chief Information Officer Cris Ross put it plainly during his keynote speech at Health 2.0 this week: “Our systems are not adequately supporting our doctors, in lots and lots of ways.”
And he counts his own world-class health system as one of them. Mayo Clinic completed a landmark four-year, 90-hospital, $1.5 billion Epic implementation in 2018. But while it was “an enormous project and by all objective measures we did just fine,” said Ross, “we’re also still at place where our doctors are frustrated and our patients are not seeing a particular difference by us doing that.”
Providers want to know that they have meaningful work, where they are operating in an efficient and effective way and that they’re delivering the best treatment that’s appropriate, he explained.
Want to publish your own articles on DistilNFO Publications?
Send us an email, we will get in touch with you.
“But they’re also looking for joy in practice,” said Ross. “Being a provider is hard. And we make the bar even harder by layering on unbelievable levels of complexity and regulation, which makes their work incredibly hard. We have to help them with that and try to find a way to bring some joy back to their work.”
At Mayo Clinic, he said, “part of what we are trying to do is to pursue the next generation of care.” And to do that, the health system is embracing a wide array of future-looking initiatives such as its just-announced 10-year partnership with Google Cloud, which will offer security and agility – and will enable Google’s AI scientists to work shoulder to shoulder with Mayo’s own researchers, developing new models of care.
‘Small machine learning’
At a more basic level, that transformation will depend on rethinking existing IT infrastructure and workflows. EHRs are a critical tool for maintaining essential data, but they leave much to be desired when it comes to the fast-paced, complex and demanding day-to-day of clinical practice.
“I would make the argument that the systems that we have in hospitals today are about as good as corporate systems you’d see on another side,” said Ross – acknowledging that that comment could make “some of my doctor friends want to tar and feather me: ‘What in the world are you talking about?’
But it’s true: “They all look pretty much the same,” he said. “They are all performing basically the same function of displaying information. So why is it then, if this proposition may be true, that providers are so completely frustrated with the kinds of systems we put in front of them?”
Well, it’s because “the complexity of healthcare is at least an order of magnitude more complex than other kinds of information settings,” said Ross.
“We have to remember that a lot of the systems we’re using in healthcare are serving a purpose not of the physician or the patient but of a broader organization. These corporate systems bring with them every government regulation, all the requirements of all payers, requirements of all quality organizations, all the policies of that organization and all the required procedures.
“No wonder doctors hate electronic health record systems,” he said.
“As a patient, I experienced the fact that our systems are not adequately supporting our doctors in lots and lots of ways – primarily because we’re still struggling with this basic system of automated healthcare and not advancing the program,” he explained.
“So where are we headed? Generation next ought to be systems of intention, insight and cognition – systems that know what we want, help us see what we otherwise wouldn’t see and help us think,” said Ross.
That can be achieved perhaps more easily than one might expect he said, with help from what he called “small artificial intelligence or small machine learning.”
Whether it’s voice recognition, geographic information service, automated reminders or any other of a host of nifty AI tools hidden away in smartphones and mobile devices, there’s “a significant number systems that impact my life all the time – little AI – that keep my life functioning well,” said Ross. “And there are all kinds of ways that we can bring those little tools to healthcare in exactly the same kind of way.”
The tools are advancing all the time. And as healthcare gets better at squirreling them away in unnoticed corners of clinical workflow, they’ll only help improve the clinician experience – and, by extension, the patient’s.
“To know more about me, these systems of learning, systems of cognition depend on intimacy with the people with whom they interrelate,” said Ross. “I have to share personal data about myself If these systems are going to function. But these next-generation systems are what’s necessary for us to get to a new place,” he said.
“Then there’s big AI, the kinds of things that will help us do things like drive diagnosis,” said Ross. “People used to ask me, ‘Is this AI stuff for real?’ Even three years ago, a lot of people said, ‘No, this isn’t coming to healthcare it’s all science fiction.’ But things have changed enormously.
Many of Mayo Clinic’s most exciting recent innovations, from deep learning algorithms that can predict atrial fibrillation to its work with radiogenomics, which combines aspects of AI, imaging and precision medicine, bear this out.
“Doctors aren’t turning the keys over to the machines to make decisions for them,” he said.
Instead, they’re using them to inform their care and help them practice at the top of their license – applying leading-edge techniques that once seemed so “science fiction” to the patients who need them most.
“I hope I am convincing you: this artificial intelligence stuff is real,” said Ross. “Most of this stuff is on a spectrum somewhere from discovery to translation to application.”
Date: September 23, 2019
Source: Healthcare IT News