HIMSS17: Artificial intelligence focus asks the industry – too soon?
While a large portion of the conference focused on artificial intelligence and machine learning, it’s yet to be seen how the implementation in the hospitals settings will play out
HIMSS is kind of like that one old college friend that drops in for the weekend. You get excited leading up to their arrival, have a great time while they’re there and want to shut off your phone and take a nap for a day or two after they leave. To that end, HIMSS17 did not disappoint. It was a whirlwind of meetings, education sessions, vendors in their Sunday best and we’re getting tired just trying to remember it all.
For those of you that couldn’t make it or were busy pursuing lead generations, here’s one thing we learned: Artificial intelligence is hot.
Artificial intelligence was a thing this year
IBM President & CEO Ginni Rometty took the HIMSS mainstage to address the audience Monday morning alongside a string of company announcements.
For those in need of a refresher, IBM Watson came on the scene with an oft-referenced appearance on Jeopardy in 2011. Since then, IBM Watson has been trying to find it’s place in the healthcare industry at large. Rometty seemed eager and happy to talk about the potential and promise of artificial intelligence in the healthcare industry. She stated she believes the industry could be ushering in a golden age of healthcare technology with AI, precision medicine but she noted the effort has to be shaped correctly. Many providers/systems will make technology decisions in the next two years that may impact the industry for the next two decades, she stated, outlining five guidelines for applying AI to healthcare:
- It must provide a range of services.
- It has to provide data transparency.
- It must be domain specific.
- It should be cloud-based to make it ubiquitous but a hybrid cloud securely built to connect across ecosystem
- It must be an open-platform.
IBM Watson Health wasn’t the only vendor touting AI and its potential. Nant Health CEO Patrick Soon-Shiong revealed Nant AI, an augmented intelligence platform as well as Nant Cloud. This follows this month’s announcement that the University of Pittsburgh Medical Center’s (UPMC) partnered with Microsoft, to launch the first project under the new Healthcare NExt initiative aimed at improving clinicians’ workflow, signaling AI investment interest for healthcare.
Google even returned to the HIMSS showroom floor to showcase its progress in smart data, digital health, APIs and machine learning. Still, depending on your size and organization, the flurry of news surrounding AI and increased digital tools use can lead to the question: Is this too much too soon?
But what’s the cost?
The healthcare industry has never been renowned for its speed of adoption, despite the big tech conferences showing off all their latest futuristic capabilities other industries adopted years ago. On one hand, it still seems like the industry hasn’t fully achieved the theme of last year’s conference: Interoperability. Many businesses have become profitable by providing add-ons to legacy EHR systems which can’t function the way many providers want them to. Some physicians have still yet to adopt using email to communicate with their patients. The use of AI and the like seems like a far leap in the face of the industry’s present technological infrastructure.
"While exciting, artificial intelligence in healthcare is still very new, providing great opportunity to explore use cases, data sources, and applications."
Dr. Joshua Newman
On the other hand, AI and machine learning tech doesn’t have to be regulated to the EHR or other legacy systems. One simple example could be to use AI tech to automatically reschedule appointments for patients to a date or time they have checked as preferences. This type of feature could likely assist large systems to expand access to their services (access that at times is being eaten up by primary care offices that offer same or next-day appointments).
Dr. Josh Newman, Salesforce CMO, told Healthcare Dive: "While exciting, artificial intelligence in healthcare is still very new, providing great opportunity to explore use cases, data sources, and applications. What the healthcare industry needs is experimentation, discovery and flexible platforms that enable us to fully understand the power and impact AI will have on healthcare overall."
If you’re a small rural provider, for example, and you look at the cost of some of the big systems, you may have some questions. About 200 individuals were turned away from a packed roundtable discussion during HIMSS on the emerging impacts of artificial intelligence on health IT. We caught up with James Golden, managing director of PwC Health Advisory, after the panel to discuss what surprised him about the discussion and whether or not it actually takes $15-20 million to get into the space.
“It’s not about technology, it’s about building a capability,” Golden said. “How do we think very thoughtfully about what problems can be well defined and matched against classes of AI solutions.” For example, how can a radiology group review 100 images a day compared to 15 images a day?
Golden says he sees clinical trial protocol matching as one low-hanging fruit for groups to enter the space. He noted many individuals during a question-and-answer portion of the panel were from small hospitals and noted they don’t have $10 million to invest in such tech but used open source tools with post-doctorates from nearby universities to implement solutions.
“[Hospitals] are very excited about [artificial intelligence] and are actually very bold about it which is surprising because hospital systems don’t tend to be usually bold."
Managing Director, PwC Health Advisory
Golden says he believes the providers are recognizing that they need to think about how to build high value services that benefit both patients and doctors instead of just building tools for the sake of having tools.
“[Hospitals] are very excited about [artificial intelligence] and are actually very bold about it which is surprising because hospital systems don’t tend to be usually bold. But they’re making investments,” Golden says, adding physicians are really eager to adapt these things not as decision support replacements but as means to spend more time with patients.
“This stuff is coming. It’s coming fast. It’s being viewed as a research project. In the next few years, it is not going to be a research project,” Golden said. “You don’t have the money not to invest in this stuff because it’s going to change everything and it doesn’t have to be a $15 million [investment].”
Follow Jeff Byers on Twitter