Q&A: Why clinician engagement is key for success with healthcare AI

Earlier this month, the newly spun-off GE HealthCare revealed its second planned acquisition as an independent company. It signed an agreement to acquire Caption Health, which developed AI-enabled guidance software for ultrasound imaging. 

Karley Yoder, general manager and chief digital officer for ultrasound at GE HealthCare, said the guidance software allows more clinicians to confidently use ultrasound. Artificial intelligence has become a hot topic in healthcare, but she notes it’s important to focus on what it could do rather than the novelty of the technology.

“No one now talks about software as an important thing to have, right? We all are used to a language like, ‘What does software enable your phone to do that it couldn’t do before?’ When it comes to AI, we still kind of get caught up in the technology, as opposed to the outcomes,” she said. 

Yoder sat down with MobiHealthNews to discuss the deal and the industry’s changing views about AI. 

MobiHealthNews: GE recently signed an agreement to acquire Caption Health. Can you tell me a bit about how this deal works within GE’s larger strategy?

Karley Yoder: My previous role at GE HealthCare for five years was building out our AI practice, and I’ve been in this ultrasound-focused role for almost two years now. GE HealthCare has a strength and a focus when it comes to AI. When we looked at our ultrasound portfolio, and we looked at what Caption Health had created – they’ve created something truly novel for the space, and it’s scan guidance.

So it’s AI that helps you get to the right image, as opposed to understanding what’s in that image. And what this does is it expands who has access to use ultrasound. It takes an incredibly powerful tool, and puts it in the hands of primary care physicians, the emergency room, GPs [general practitioners], folks who typically in the past maybe wouldn’t have been comfortable with ultrasound.

The vision of our ultrasound franchise is for ultrasound to be a tool that can be used by nearly every clinician in nearly every clinical setting in nearly every country in the world. This is an incredibly important piece of the puzzle to expand access.

MHN:  I saw an interview you did several years ago where you were discussing a survey where health systems said that they didn’t see AI as a priority. And that was in 2018. How do you think that viewpoint has changed? 

Yoder: Every year I’m blown away by the advances that happen in the data science space. We are fortunate at GE HealthCare to have some of the brightest minds in the overlap of data science and healthcare. But I think that some of those advances have more to do with how AI is integrated and deployed than even how it’s built.

So even back in 2018 maybe 2019, 2020 you could build really good data science. But if you didn’t make it invisible into the workflow, it was clunky. It was hard to use. I’ve been using this example for maybe five years, but it’s like in Google Maps, right? If you had to go to a different app to see the change of direction — or, God forbid, a different phone — you’d never use it.

There’s just some usability norms that I don’t think we necessarily got right with the first generation of how we thought about AI. But I think one of the big improvements that we’ve done, and they have been a huge focus at GE HealthCare, is [thinking about], how do you build something trustworthy? How do you deploy it invisibly? And then how do you engage users to understand how to use it within their workflow?

Build, deploy and engage. Because if you just focus on the build, you’re never going to see the change management or adoption that you’ll see if you figure out how to integrate it in the right way. 

MHN: It’s interesting that you’re pointing to workflow, because that’s something that I hear a lot with any digital health tools. If it’s not part of a clinician’s workflow, they’re not going to be able to easily use it. Could you elaborate a little bit about how you see that working with AI? 

Yoder: Within existing workflows, that means it has to be integrated within the tools that you already use today. That is step one. You can’t add new screens, you certainly can’t add new steps. It’s got to be invisible and just help you be more effective at what you’re doing.

But I also think AI will introduce new workflows. You know, I think about the EMR. We took a very painful paper process and created a pretty painful digital process. How do we take a moment to reimagine workflows with the advent of AI?

Multi-modality AI is a true north and a future we want to get to. How can we get all of the data that surrounds the patient to come together and to inform clinicians on the right path forward? This is really what GE sees as the future when we talk about precision care, when we talk about our strategy on D3. 

How do we take all of the data that surrounds the patient and run intelligent AI on it to inform the best care pathways for any given patient? That’s not where we are today, but that’s the future we can strive for, as long as we take these incremental steps to use the AI that exists today to drive better access and drive better efficiency.

More regional news

Employer health benefits company Gravie scores $179M and more digital health fundings

Digital therapeutics startup Mindset Health scores $12M in Series A funding

Maven Clinic buys fellow virtual women's health company Naytal

Source: Read Full Article