Join us for a conversation with Karley Yoder, the GM and Chief Digital Officer of GE Healthcare's ultrasound business. Karley has an extensive background in artificial intelligence and shares her insights on AI in healthcare and how the industry can be using AI more efficiently in the future. She also helps dispel some potential fears about the future of AI and also touches on the work GE Healthcare is doing to advance maternal health.
The Medical Alley Podcast is brought to you by MentorMate. Custom software needs vary significantly, whether you're powering a medical device, overhauling your backend architecture, or re-imagine your patient experience MentorMate can help. Harnessing the technical excellence of Bulgaria, MentorMate provides end to end software services and all sectors of healthcare. With deep expertise in design development, cloud, and software support MentorMate helps healthcare clients administer world-class care through technology. Learn more at mentormate.com.
Gabriela Spence (00:41):
Welcome back for another Medical Alley Association podcast. I'm your host, Gabriela Spence, federal policy and advocacy manager at the Association. Today I'm joined by Karley Yoder, general manager and chief digital officer of GE healthcare's ultrasound business. Karley is responsible for digital product innovation, commercial growth, customer collaboration, AI development, and third-party partnerships to accelerate the impact of digital solutions that drive clinical efficiency and enable better patient outcomes. She previously held the title of vice president and general manager of artificial intelligence, having spearhead GE Healthcare's AI strategy, and leading the team that developed their healthcare specific AI platform, Edison AI. Karley, thanks so much for joining us. Great to have you with us.
Karley Yoder (01:34):
Thanks so much, Gabriela. Absolutely. My pleasure. And I appreciate the opportunity to connect with you.
Gabriela Spence (01:41):
I'd like to start by setting the table a bit. The world of AI is vast even with our more "narrow" — and it's a podcast, so you can't see me, but I'm doing air quotes around narrow discussion of AI in healthcare. So I wonder if you can begin with just a few definitions and background around the subject, let's set some parameters for our conversation.
Karley Yoder (02:07):
Sure, Gabriela. I think that's important because AI has taken on such a hype in terms of what it is and what it does in the past several years, especially. So it helps to take a little bit of a step back. So, so let's demystify it a little bit. The practice of artificial intelligence is actually nothing new. It's been around for 50 years, with data scientists trying to push the boundaries of science forward. But what happened in the last five years was we had a massive breakthrough when it comes to a specific science around deep learning, which is a subset of AI. That is absolutely incredible at pattern recognition. And the reason why we saw this explosion in deep learning is compute power exploded. So thank you, Moore's Law. And the availability of digital data exploded, and that's like kindling to a fire when you bring those two pieces together.
Karley Yoder (03:06):
And when we begin to think about healthcare, it's hard to imagine an industry that's more primed for the impact that this technology advanced can have and the impact it can have on patients and providers all over the globe. But one thing I always like to remind folks, when I jump into a conversation around artificial intelligence is, AI, It's just the tool — albeit a very powerful tool, a very industry changing tool, but it's just the tool. It's not a product in and of itself. And so when we think about a discussion around AI, you can't get lost in the technology or the science. We have to constantly be obsessed with what good does this do for the world for patients, for providers? And then how do we harness this technology to move the ball forward? So it's a little bit of a background and how I'd frame, the science and power of artificial intelligence when it starts to come towards healthcare.
Gabriela Spence (04:06):
That was a really helpful overview. Thank you. Thank you for that. And I'm curious, actually, I'm going to go off a little bit. I'm curious. What makes you passionate about AI?
Karley Yoder (04:17):
Absolutely, Gabriela. So personally I will spend my entire career harnessing technology to drive healthcare forward. My background is biomedical engineering. I've had the joy in my career to work from telemedicine to with state governments here in the U.S., with Apple Health. And all of this, is about how do we leverage the best that technology has to offer to drive the most good, the best outcomes for patients at a global scale. And like I mentioned, with five years ago, with the explosion of deep learning, it became a technology that you just couldn't ignore. Andrew Ng, who's one of the founding fathers of deep learning artificial intelligence, says it best when he looks to the future and says, AI will become like electricity in the future. It will run through everything and become something we take for granted. And so when I fast forward in my head 15 years, I like to think that AI will be what we think of software development today — the basis for every product, for every solution, that we learn how to harness as a tool to drive healthcare forward. And I'm extremely passionate and feel very fortunate to get to be in a place where I get to work with this technology on a daily basis and not just for technology sake, but for the sake of our customers all over the world.
Gabriela Spence (05:51):
So Karley paid me a really nice compliment before we got started and called me a really great hype girl, which I take great pride in, but I think you just hype me up about AI. I was fascinated with it before our conversation, during our conversations before this podcast. And now I'm just, I am brimming with excitement over the possibility that AI presents. Before I get lost in the possibilities and those tangents, I have to address, the, I think, scarier side of AI. When I hear or even think about just AI, artificial intelligence, my mind can't help, but conjure images of that movie "Smart House" that many of us grew up with. I think I've been conditioned to be fearful of technology that can think for itself. And I, I hope I can also imagine. I'm not the only one, many people are concerned with AI's accuracy, others fear its replacement of people. And I think we are all concerned about the discriminatory possibilities around AI. How do we manage all these fears?
Karley Yoder (07:08):
Oh, Gabriela, it's such an important question. And let me tell a little bit of a story from my own professional experience before I address it directly. So there is a radiology conference. Most folks listening to this probably know healthcare, but radiologists are the ones who look at medical images, day in, day out and drive a lot of the care that happens in healthcare with the care team. There is a conference every year in Chicago that I refer to as the Super Bowl for radiologists, where 60,000 folks from all over the world, descend in Chicago, for the biggest radiology conference of the year. Let me tell you Gabriela, four years ago at that conference, this was the single most asked question and biggest point of conversation that I addressed with our customers.
Karley Yoder (08:03):
And it was the fear of the unknown, the fear of being replaced, the fear of trust, the fear of things that you can't touch and feel. The next year, the conversation was, "Okay, we're not as afraid of the technology, but how do I adopt it and use it? How do I use it in my workflow?" A year later, the conversation had moved from not how do I adopt it, but how do I pick between the multiple vendors who are bringing me technologies that leverage AI? And I start with this story, because part of the way we need to address fear is we need to take the conversation around AI beyond hype and into the tangible. And we need to show folks that AI can help them with their current jobs and make them better, faster, more efficient at what they were already doing. Think about Siri for us.
Karley Yoder (08:56):
Think about Netflix's next show prediction. Think about when your arms are full of laundry and you realize you're out of detergent and you can yell out to Alexa to order you your next load of detergent. This is all AI, but it's not fear-inducing it's helpful. And so I think that in healthcare, part of what we need to do is as industry and as practitioners is leverage the technology in a way that matters and helps, because you only get one first impression and that first impression matters. And so I think, I think that's a big part of how we manage the fear is we actually put real examples on the table that make a difference for providers and patients. But another thing I'll say is you have to start from the very beginning of AI products' creation with a thought towards ethics and eliminating bias.
Karley Yoder (10:02):
At GE healthcare, four years ago, we created the GE AI ethics principles. For this reason, we knew that this was a powerful technology. And if we didn't sit down and write guard rails on how we were going to harness this, not just for the best impact, but for the most ethical and unbiased impact, we were setting ourselves up to make mistakes. So our four ethical principles are be designed for the benefit, safety and privacy of the patients. So from the very beginning, the patient has to be at the center of what you're doing with artificial intelligence and not just benefit, but safety and privacy matter. Number two, be a trusted steward of data and insights. We all know, and we've all heard that the engine of artificial intelligence is data, but garbage in garbage out. And so we have to be a trusted steward and anyone who creates artificial intelligence has to have a robust, secure way to manage data, manage the life blood of AI.
Karley Yoder (11:11):
Number three is be transparent and deliver robust and reproducible results. This one's so important. You hear a lot in this fear around AI, around the black box concept. Well, I think we can lighten that black box if you will, if we create transparent and reproducible results, just like any scientific approach that demystifies and begins to normalize or make it more tangible, how this AI is created. And number four guard against creating or reinforcing bias. And on this one, I have to say, you can't do this once a product exists. You can't do this in the middle of product creation. You have to do this at the very beginning of an AI project as you think about the data that you collect and how you want to collect it to represent a global population. If you build a model off of just data collected here in the Bay Area, and you expect it to work in China, you're going to have brutal AI that doesn't generalize. And so this guarding against bias has to be something you do at the very beginning of a project. And I really believe if any company takes the time to put these type of thinking these guardrails in place at the beginning of their project creation, we will demystify and move past the fear of this technology and move towards embracing it just like we've seen in so many of our consumer products.
Gabriela Spence (12:52):
I have to commend you on your verbiage, Karley. You clearly have talked about AI many, many times. It's reassuring. I love the transparency you've brought and I really have to applaud GE for these principles. In reading through them in preparation for this conversation, I just found them very insightful, very reassuring. I'm also very on board with this AI stuff. Now that we're talking about it a little bit more as I'm sorry if it sounds a little bit sales pitchy. But it's funny that you mentioned that, you know, the real examples. It's not fiction, it's not "Smart House" the movie. It's not, it's not created from Hollywood, we're already living it. I love your examples about show prediction and Alexa and Siri. I mean, there's just so many applications for AI that already exist that I don't think we even are aware of. I mean, even my thermostat can predict, you know, based off of weather patterns outside, does it need to be hotter or colder? I can set things. I mean, it's really interesting. And it's the efficiency aspect of AI that I really love and that I want to focus on for these next couple of questions, but, you know, how do we really harness this advanced technology to change habits and workflows and policy? Maybe you can speak a little to that.
Karley Yoder (14:18):
Absolutely. And you know, just to transition off the last topic, you know, I'll paraphrase one of the smartest radiology meets AI thinkers in our space, Curt Langlotz from Stanford. And he said, AI will replace radiologists, but radiologists who leverage AI will probably replace those who don't. And so I really think, you know, one of the ways to move past the fear and, Gabriela, move towards advancing this technology and changing habits is if we shift the mindset from man versus machine to man plus machine, and what can we do better together?
Gabriela Spence (14:57):
Wow. I'm sorry, I have to comment on that. That is profound. That is so good. And it will really push us into this next evolution of workforce and job types and impact. I think we can make such a more positive impact across many fields, not just healthcare, utilizing some of these. That was, that was stellar. Thank you, please continue.
Karley Yoder (15:25):
Absolutely, Gabriela. And it's, it's really — look, I think the way we advance this fastest is if we make AI invisible, or what do I mean by that is if, if we stop talking about products as valuable, because they have AI and start talking about products that are valuable because they make you more efficient, they make you more impactful. They give you time back in your day. And AI is just a subset of that, right? And so I think when I think about how do we change habits, policy, it's really, especially in healthcare, it's a workflow question, right? So think about the doctor who comes in eight in the morning with 50 items in their inbox that they need to look at. The current practice today is first in, first out. So like many of us ,we'll start at the top of our inbox.
Karley Yoder (16:22): Karley Yoder (17:16): Gabriela Spence (18:30): Karley Yoder (19:29): Karley Yoder (20:41): Karley Yoder (21:43): Karley Yoder (22:46): Gabriela Spence (24:22): Gabriela Spence (25:22): Karley Yoder (26:00): Karley Yoder (27:08): Karley Yoder (28:19): Gabriela Spence (29:02): Karley Yoder (29:50): Gabriela Spence (30:30):
We'll just work from the top to the bottom. But in a healthcare setting, what if the item at the bottom was the one that if you intervene right away would save a patient's life? With AI, we have the chance to flip the whole way of working on its head in a way that makes us smarter, more efficient at the jobs that we were already doing. And I really think this is where that man plus machine becomes so powerful. But again, I'll give another example from the consumer space. We've all used Google Maps. I doubt many of us are using Apple
We never adopt the technology, no matter if it was embedded with artificial intelligence. Now, Gabriela, I'm using air quotes, but what we've done in our consumer products is made AI invisible, used it as a tool to drive technology changes that matter to us as the consumer. This is exactly what we have to do in healthcare and what, honestly, with a lot of technologies, we haven't been good at doing in healthcare in the past. Think about the EMR, which is a phenomenal step wise change and taking us from a paper process to a digital process. There's no doubt there, but in a lot of ways, we took a painful paper process and created a painful digital process where the clinician is still turning away from the patient to typed in their computer. they're not able to spend the time with the patient. We have to do better when it comes to the integration of AI, into our healthcare workflows and healthcare products. We have to think about it, how we can honestly give time back to the doctor to engage with the patient, which is what every doctor got in healthcare to do.
I love these examples. You do a really beautiful job of giving us kind of this etherial concept and then breaking it down into a real world application. I'd like to dive into that a little bit more. Give us a few more of these examples, these case studies of how we can get data moving better. And I like to include a specific ask. you know, I do a lot of work on our maternal health initiative. You and your colleagues have been incredibly helpful in providing guidance in this space. And maybe we can start there. You know, the CDC says 60% or more of maternal health related deaths can be prevented. And I really admire GE's approach to minimizing that statistic by a commitment to improve patient safety and outcomes through optimizing technology for healthcare staff. Tell us some more stories.
Absolutely, Gabriela. So first I will say the opportunity to use data to apply better health healthcare outcomes is limitless. So the average hospital creates 50 petabytes of data. 50 petabytes of data might not mean anything to you. So think about if you took the DNA sequence of every person in the U.S. and then multiply that by three, That's how much data we're talking about coming out of one hospital in one year, but less than 3% of that data actually drives anything actionable, right? We have 97% of this digital trove of data that sits there unused, ignored almost, when it could be helping providers address issues like what you just brought up around maternal health. So we have to do better. We're fortunate in healthcare that we've already kind of had a digital revolution, which is the first step before any AI revolution, because you have to lift data into a form where it can be leveraged, but we're terrible at using that data.
I say we've started to become great at creating data, but horrible at leveraging data healthcare. And so let me give you one example in the women's health space,on how we can leverage this data to do better. So, Gabriela, imagine your mom, your friend, yourself going in for a 20-week fetal exam. So this is the midpoint in a pregnancy. This is, you know, a little bit of a longer ultrasound exam. The sonographer or the clinician, depending on where you are, spends a lot of time, walking through this exam, capturing many different views of the baby, explaining what's going on to the mother. And one of the things that we've done here at GE healthcare is we created a tool called Sono CNS. And this tool automates the CNS or brain area exam for a baby in utero.
Now, remember when we're talking about ultrasound in a fetal exam, we're talking about capturing a picture of a little guy or gal who is moving around within a patient who is anxious and trying to understand what's happening. The last thing you want is your sonographer to be clicking buttons over and over again, unable to engage with the patients. So what we've actually done with Sono CNS is we've automated this process and removed 80% of the time and steps out of the scenario of capturing this image so that the clinician can spend more time with the patients and less time fiddling with the device or fiddling with the technology. This is the future. This is where we're going across the board, where we want to tackle the problem where clinicians only spend 27% of their time engaging with patients, Gabriela. My husband's actually a primary care doctor.
I don't know if we talked about this the first time we connected. And I can tell you, he didn't get into healthcare to get really good at data entry into an EMR. He must be the exception, but he got into do something that makes a difference for patients all over the world. And we're trying to do that here at GE, from the work we've done with our critical care suite, which is an x-ray AI use case where we can prioritize critical conditions right to the top of the radiologist's queue. What we've done with our shock ultrasound suite, where critical moments that are in a shock context. It is a terribly detrimental disease if clinicians can't move quickly. And if you're able to use a device that automates your steps and makes a novice operate at the skillset of an expert, the outcomes and impact for patients are what we see change and in a positive manner. And so these types of use cases where we take critical conditions and critical cases and put them at the top of the list where we take a long task and we can cut it in half where we take unique cases that may have been missed because it's either a less experienced position or a really rare case that just isn't something that's normally seen and we use AI to capture and find it, these are the type of applications that AI is really making strides to make feasible, to begin to change in the healthcare setting.
Oh, this is so cool. I just, my enthusiasm keeps building with more of your responses and stories. And I love these real world applications because, you know, not everyone's going to experience pregnancy or childbirth, but I think a lot of us have experienced some sort of traumatic medical event, either ourselves personally or through a loved one and bringing those stories home, I think punctuates the importance of, I'm actually going to quote you some, I love how you phrase this when we chatted beforehand: AI puts doctors back into the patient care business and not the data collection business. That is very profound to me, and I can imagine would resonate really beautifully with a lot of our healthcare staff and practitioners and the doctors that listen to these. So that makes me hopeful and excited and also would ideally give doctors more time to spend with patients.
I think that's often a frustration on the patient side, but we don't often get to look at the complications they're dealing with and just collecting the right information. So, very illuminating, very helpful. Thank you. I want to look a little bit to the future before we wrap up. I know you were involved in and actually led GE's, and I'm blanking on the name. You'll have to help me out here, but the medical startups, the program you guys have. Tell me about what you're seeing in, you know, the up and coming devices innovations. A lot of them are integrating AI, is that correct?
That's right, Gabriela. And you know, I think a closing statement and looking for the future is there's so much good that can be done by leveraging AI in healthcare that no one company can solve this alone — and no one company should. We need to create ecosystems of partnerships. And when you think about what that ecosystem of those partnerships look like, we think about three legs of the stool, if you will. So one, you have to work closely with customers. So to the conversation that we had earlier, in this dialogue, if you don't solve the right problem with AI, you're wasting the technology. So you have to walk in the shoes of the folks who are leveraging this technology and make sure you're pointing this technology at problems that really matter. So you have to be deeply embedded with customers and partnering with them on AI creation. Number two, technology partners. The advances in cloud computing, things like GPU's, things like data management are vast.
And GE Healthcare, we partner closely with folks like AWS, Intel, Nvidia, Microsoft, leveraging the best of what they can do, but adding our healthcare domain on top of it. And I think that's essential. And third, what you were really teeing up there, the explosion of startups in healthcare AI is so encouraging to me. At the last count, there's over 400 startups with nearing $4 billion of venture capital and outside funding flowing into this. If you zoom into the medical imaging where deep learning is especially impactful, there's 200 startups with over $1.5 billion of funding. And not all of these startups are going to hit home runs, but I sure hope a lot of them do. And part of our thinking in the Edison developer program and why we've invested in our Edison platform, which is our intelligence platform here at GE Healthcare, is because we need to think now about how you weave all of those different solutions into one tapestry for customers, so that they have a unified experience when they're accessing this technology,
and don't feel like they're working with 25 different pop-ups on their computer. And so as, as we've developed our Edison developer partnership program, we've really invested on building roads and bridges, if you will, that allow us to be a really strong partner to this burgeoning startup community, and help them more quickly bring their technology into the workflows and products of our customers all over the world. And I truly believe that I look to the future in healthcare, but any industry, we have to be thinking about partnerships and ecosystems. If we want to move at the speed that AI will allow us to.
Karley, I am so excited about the future of healthcare, just based off of this brief conversation we've had. I think the future's really bright and grateful for GE's, you know, one of your colleagues called it an agnostic approach, and it really is very holistic. You guys have a great emphasis on the partnerships you try and build. You talk the talk, but you do walk the walk. And it's great to have, companies like yours here in Minnesota's Medical Alley, creating these innovations so we get to really see that impact bringing that future forward. Before we wrap up, is there anything else you want our listeners to know around AI and healthcare, What GE is doing, wide open question. However you want to answer it.
Sure. Hey, I will take the movement to just encourage folks to jump in. if you're passionate about healthcare, if you're passionate about technology, I can't think of a better time to jump in and play a role in bringing our digital future to life. Digital AI is the future and now is the time to make real progress. And I want the smartest minds in the world tackling these problems so that we can move the ball forward. And Gabriela, I really appreciate the time to connect today and I'm grateful to the Medical Alley team and the broader organization. And just thanks for the chance to have the dialogue.
Thank you. It has been a joy talking with you today. Thank you for your passion and your expertise. And I would add to your, to your call to action, you know, the smartest minds, absolutely. But the passionate ones as well. If you've experienced things firsthand, that makes you a fierce advocate in a way that's very unique. So I could not agree with your sentiment more, please dive in. Reach out if you're interested in getting involved, but don't really know where to start. let us know. My colleagues and I love just helping the community thrive and building everybody up. So please reach out and I'd like to give a big thank you to our listeners. Your continued support allows us to bring you amazing conversations like this one. So if you enjoyed it, please consider subscribing to the Medical Alley podcast and giving us a five-star rating. Thanks again. And we'll see you next time.
Karley Yoder (17:16):
Gabriela Spence (18:30):
Karley Yoder (19:29):
Karley Yoder (20:41):
Karley Yoder (21:43):
Karley Yoder (22:46):
Gabriela Spence (24:22):
Gabriela Spence (25:22):
Karley Yoder (26:00):
Karley Yoder (27:08):
Karley Yoder (28:19):
Gabriela Spence (29:02):
Karley Yoder (29:50):
Gabriela Spence (30:30):