Chief information officers and other IT leaders at hospitals and health systems are being bombarded with a wide array of AI technologies and are trying to navigate a burgeoning marketplace with more than its share of hype.
In fact, advice on procuring artificial intelligence filled the better part of a morning at last week’s 2024 HIMSS AI in Healthcare Forum, where digital health leaders compared notes on how to separate signal from noise.
Healthcare organizations must weigh the pros and cons of many different platforms and tools – which can’t always deliver on their lofty promises – while identifying strategic partners who can help meet the challenges of integrating new AI technologies within their existing networks and workflows.
Tips for approaching providers
What digital transformation leaders want foremost from AI vendors is for them to be honest about their ingredients, according to Lee Schwamm, chief digital health officer at Yale New Haven Health System.
“Number one, you’re an AI company, and number two, you’re a platform – it is okay to not be either of those,” he said during the Taming the Wild West of AI in Healthcare panel on Friday.
“You can say, I’m a company built on X, and we’re starting to infuse our product with some AI – that’s very reasonable. That’s what most of you probably are.”
“You need to have a really good understanding of how you fit into the workflow, because that is the problem with point solutions that we have today,” added Eve Cunningham, chief of virtual care and digital health at Providence.
“Actually integrating that into the workflow that we have right now with the tech stack and the infrastructure that we have is an extremely complicated process,” she added. “So you can have the best point solution in the world, but if they can’t integrate, and there isn’t a path to integrate, and you don’t know how to speak the language of understanding that, there’s just a very low level of interest in engaging on that front.”
It’s important too to target sales pitches appropriately and find the right decision-maker, noted Dr. David Newman, chief medical officer of virtual care at Sanford Health. “Trying to get your hook in seven people actually backfires.”
He said he was talking with his 15-year-old daughter about his role speaking about vendor relationships “and about how best for them to contact us.”
“She said, ‘It’s like them sliding me with DMs,’ which is right.”
The key point, Newman said, is that technology vendors must know what a provider’s mission is, and what problem it is looking to solve before reaching out.
“That way, instead of skipping over your email, I’m going to answer you,” he said.
“It isn’t just about relationships and knowing somebody,” but rather about having objective evaluation of the technology, Cunningham added.
Many AI enhancements address physician survival and workflow, but testing them out in partnership with vendors is causing providers “pilot fatigue,” she said.
The panelists encouraged vendors to understand from the providers’ perspective – is it a physician productivity enhancement, or something revolutionary that gives health systems something they haven’t had before?
“Have you ever actually sat in a doctor’s office and looked at how many clicks they’re doing? There is no room for one more click,” said Cunningham.
“Sometimes, they’re good enough that we’re willing to break our workflow to adapt to that because it is a superpower,” Schwamm added.
“You have a mature enough product that we’re not going to build your product for you,” said Cunningham. “We’re not your dev shop.”
Provider vision, costs and backlash
There is no simple answer for managing costly product pipelines, Schwamm said: “Maybe I do want someone who’s really going to partner with me and build their product roadmap against what my horizon looks like.
“At the moment that we’re in now, you don’t have the luxury of [saying] ‘This is going to be my system forever,'” and knowing a technology module is “rip and replace,” may be favorable.
There’s “not a lot of appetite to rip and replace, but sometimes we have to do that,” Cunningham acknowledged, such as with ambient listening technologies.
Digital healthcare leaders need to consider, “What is it going to look like three years from now, five years from now, seven years from now?”
In her vision of the technology-driven doctor’s office five to seven years from today, a number of things are happening simultaneously.
“Maybe there’s a big screen in the room, and there’s no keyboard,” she described. “I’m having a conversation with the patient. My note is being drafted. All of the things that we’re talking about, ‘Hey, you have COPD’ and all the data from the patient’s chart that’s relevant to the COPD is surfacing up.
“‘Hey, we need to get an order for more PFTs for you,’ and the order goes in. ‘Hey, here’s a little bit of knowledge about the next best actions,'” she continued.
“All of that’s happening in the room, and I walk out of the room, and all that work is done.”
To gauge what stage providers are currently in with their AI adoption, a recent HIMSS Market Insights survey looked at the uses of AI among healthcare organizations for positive impacts and to uncover the challenges they face integrating AI into their workflows and existing technologies.
Nicole Ramage, HIMSS senior market insights manager, joined Schwamm to explore the questions surrounding AI and insights from the report.
While nearly half of the organizations surveyed this past spring are larger organizations with 7,500 employees or more, the data showed, unsurprisingly, that “the smaller organizations are less likely to be further along in their AI adoption journey,” said Ramage.
“I think your data shows very nicely that it’s a capital-intensive process, and it also involves the ability to think about the workflows you’re going to go after” and the leadership structure that it will require, said Schwamm – noting that hospitals “are severely underwater from the devastating impact of COVID plus changes in our age and population,” and ROI is deteriorating.
“It’s a continuous downward trend, and expense is a continuous upward trend,” he said. “So, that’s not a great formula.”
Ramage asked Schwamm what he sees as the greatest opportunity for AI transformation in patient care and operational efficiency over the next three to five years.
“The simplest things to go after from a transformational perspective are operational workflows or back-office stuff because they don’t touch patients,” he said. “They’re very low risk, and they’re relatively unregulated.”
While the greatest financial opportunity is currently in back-office operations, he said, there is also the increasing employee backlash against the human replacement effects of AI that could be coming, according to Schwamm.
There are four ways to get ROI, he noted.
“You either renegotiate a contract for a lower price or cancel a contract,” he said. “You make it easier to do the same thing and do it at a lower cost, or you reduce your labor.”
AI is going to reshape healthcare and its workforce.
“Whether you agree or not, it is going to be your next member,” said Sunil Dadlani, chief information and digital officer of Atlantic Health System.
Charles Jaffe, CEO of HL7, said that he is concerned about the politicization of the process. “The promise of AI is not a threat to anyone,” he said. “It is a challenge to make their jobs easier.”
Small organizations are, however, in a very vulnerable position, said Schwamm. They cannot afford to fall behind.
Ramage asked Schwamm what approaches he recommends for smaller organizations to effectively promote AI adoption while also maintaining staff commitment.
“If I was in a small organization, I’m capital constrained, the stuff is hitting me,” he said.
Without in-house expertise and funding to hire consultants, he advised smaller organizations to partner with organizations they do not compete with, then divide and conquer, and vet technologies as a collaborative.
“You have a group of five or six health systems, about your size, different parts of the country, no competition between you, and you say, ‘Hey, Jones Regional Hospital. How about you take on this issue of the back-office thing around, and we’ll take on the question of patient navigation?'” Schwamm suggested.
“And then maybe even collective bargaining, right? Think about group purchasing opportunities.”
Data footprints and law of subtraction
From a leadership perspective, the biggest challenge is who owns AI data, Schwamm said.
“Previously, if I worked with you and I gave you data for you to be able to process a transaction for me, at the end of the contract, you destroyed my data or you gave it all back to me,” he said.
When data is used to train an AI model, “giving me back my data doesn’t give me back the intellectual property you’ve extracted from my data.”
When integrating AI into specific domains “most mature healthcare systems are going to end up with really almost like a change management function that is devoted to AI, because it’s been such a big piece of those change management projects and it has such labor implications.”
Healthcare is one-third of the world’s data – and it’s growing and doubling every 46 days, Dadlani noted in the ethics panel.
“So, you need to have more and more technology and interoperability principles in place to really make smarter decisions about patient outcome, patient safety and moving to the next generation of that,” he said.
This is where the laws of addition and subtraction come into play.
“Anytime you are trying to add more and more and more technologies, you need to think about and take a platform approach. Where can we simplify the process?” Dadlani asked.
When an organization keeps adding technologies, it increases complexity, which he said means higher administration costs, more data failures and more data breaches. If it’s a point solution and cannot integrate with other technologies, “it’s of no use,” he said.
Andrea Fox is senior editor of Healthcare IT News.
Email: afox@himss.org
Healthcare IT News is a HIMSS Media publication.