Skip to main content
Sponsored Video

Generative AI in Healthcare: eKare’s Innovative Approach with CEO Patrick Cheng and Dr. Paul Kim

© 2024 HMP Global. All Rights Reserved.
Any views and opinions expressed are those of the author(s) and/or participants and do not necessarily reflect the views, policy, or position of the Wound Care Learning Network or HMP Global, their employees, and affiliates.

 

eKareJoin Patrick Cheng, CEO of eKare Inc., and Dr. Paul Kim in an insightful discussion on the barriers and opportunities of Generative AI in the healthcare industry.

Discover how eKare is leading innovation by incorporating AI solutions into wound care with their advanced inSight and healBot technologies. Learn about the impact of digital wound measurement and big data analytics on clinical practice and research, and how these advancements are improving efficiencies and addressing challenges across wound care.

Transcript:

Dr. Paul Kim:
Welcome everyone to Tech Spotlight at SAWC. I'm really pleased to be here with my longtime friend and a person I've known in the wound care space for a long time. He's the CEO of a company called eKare. And Patrick and I go way back. I remember when I was working in Georgetown, in Washington DC, and Patrick didn't even have a company at that point. Now they're a major player, especially in the generative AI space, but I want to talk before that. Tell us a little bit about your company and the growth and what your expectations were when you started.

Patrick Cheng:
Thank you for inviting me, and it's always a pleasure speaking to Dr. Kim, and we always learn a lot from you and get a lot of inspiration as well. Yeah, I still remember when we first started. And we go to your office asking for advice. I think we really follow this through and try to build a system to help physicians, more than just imaging, to do a lot more. I think over the past 7, 8 years, we laid the foundation to have a platform in place that enables us to, I guess, take you to the next step, which is all the, our namesake, inside, right? Get the inside from the data, apply AI to it. 

Dr. Kim:
So eKare, how did you come up with that name for your company? 

Cheng:
We have been quite ambitious initially. We started with wound care, but we really wanted to build a little bit more, bigger digital house companies. So we wanted to have a name that's probably a little more universal, not specific to wound care. 
Dr. Kim:
Well, let's talk a little bit about the development and evolution of your company. I think you started in the wound imaging space. So, tell us about that technology and how it's used today.

Cheng:
Yes. Initially we are trying to target the challenge in terms of how to get a measurement right. I think the rule of thumb, the gold standard in terms of measuring healing is still the size change, right? And also some of the tissue types. That's initially what we did, applying computer vision and AI, trying to analyze those things. We started with a specialized high-end sensor, but now we moved away just using the Apple iPhones, because it's getting very powerful. So we get the measurement piece right, and we have been very successful in terms of, in the clinical research space where the measurement is quite important, and then we have also been expanding beyond just imaging to get clinical documentation. So we're getting a lot of also usage in the clinical side of things. 

Dr. Kim:
One of the things, Patrick, you kind of alluded to is obviously the image capture is important. But there's another dimension that people maybe don't think too much about, but that's really the power of your system which is the analytics. Can you talk a little bit more about that? 

Cheng:
Sure. So initially the kind of imaging was kind of the selling piece for us to attract some attention, but really does not create value itself, right? So, as we go into the clinical practice side of things, we understand we have to add more value. We actually were able to, for instance, we were able to implement a system in kind of bigger settings. From the data we saw, we were able to for instance demonstrate significant savings in terms of better reach to patients. So, for some of the specialists, they can manage bigger populations with their limited time and bandwidth, which is an issue across the globe. 

We have this issue here, staffing shortage, we have the same staffing shortage in UK. This is a specific UK study, and we were able to show, for instance, 60% of the people using antimicrobial dressings, they don't have any signs or symptoms of infection, so there's a lot of potential savings in terms of product utilization, and they were able to see that, for instance, some of the patients, they were utilizing the health care system, a lot of resources, but they don't necessarily need to. They're visiting them every day. They don't need to; they can use this technology to monitor them remotely, just checking in once a while. So, all in all, there's like 5 times in terms of savings 

Dr. Kim:
Yeah, I think that's really important, and just for clarity because I know that not everybody as smart as you are, so I'm going to dumb it down for myself and others. So what you're talking about is, it's not just about image capture, doing the accuracy, the billing, and all that stuff, but it's about big databases. And that's kind of the next thing we want to talk about. You're kind of diving into this already. Is analyzing-- and this is the part that's been very elusive for many people-- is analyzing things that work, that don't work, that are overused, underutilized? And you can only do that like really machine learning and heavy computation. And that's where we start talking about treading that line of generative AI. And so let's talk about that. 

I know that's been a dream of yours, that's something you guys think about. And you guys, I remember having conversations in the mid-2015s, 2016, where you were talking about generative AI, and I looked at you with a blank stare, like I don't know what you're talking about. But now it's 2024, and it is everywhere. It's happening in lots of sectors. Health care is probably the most underutilized sector of generative AI. It's in filmmaking, it's in music, it's in social media and everything else, but in the most important space, which is our daily health, it's not been fully utilized. So tell me, one, why that is, why you think that barrier exists, and secondly, how your company is addressing that. 

Cheng: 
I think it's just amazing when I started as a graduate student in engineering school, we're doing neural networks already, applying to image analysis. So, that was actually kind of precursor to all these kind of more complicated transformer models. But I think just in the last year, these whole improvements, in terms of computational power and the large data sets we were able to consume, really create a model that could kind of be emergent, so to develop some capabilities we did not anticipate. Initially, it was developing for like finishing the sentence kind of tasks, but now you can see that it's doing some pretty impressive work.

So, we were really excited when we saw these announcements. We jumped right on it. Starting from our company itself, we are really leveraging this. So, for instance, we have some really smart developers, was able to do what usually takes 4 person job, like front end, back end, database, kind of full stack. He was able to use AI himself, prototyping our first generative AI solution within a month.

Dr. Kim:
So, I just want to just repeat that. So, the machine was teaching the machine on how to do the job, is that right? 

Cheng:
No, he's just using the AI for instance. I wanted to create a server. Could you show me how to do it? It's generating the code for him. He just runs it. Of course you have to have some background in the AI space, in the programming space to really using it, right? So we're using it through and through in our company, so the productivity boost is amazing. 

Going back to health care, I think health care is going to be the space getting the most benefit out of this AI, because this AI is very good at consuming a large amount of data and also very good at handling some of the mundane tasks in health care. I think you guys spend 30 % of your time doing administrative work, right? Highly, there's a lot of compliance related stuff, a lot of documentation related stuff. I think those perfect tasks for AI, I think the reason why AI hasn't really been picked up in the health care space. I think there's a lot of excitement already. People started to play with it. 

The reason is there's still some limitation. One, I think the biggest challenge is regulatory approval. So, depending on what your claim is, you want to use AI for decision making, that's a higher bar to cross, right? But I think there's lower hanging fruits for us to do, like for instance, education. You will make medical education a lot more enjoyable. You can do the just checking, double-checking some of the administrative work, it'll be perfect, right? So, I think those are the low-hanging fruit I will see more adoption quickly. 

And the other reason for the slow adoption is how scary, so complex right? People sometimes don't understand the scale of the complexity of the health care system. We have been in this space so long, we are always learning every day, but if you think about it, it's like 4, 5 trillion dollar industry. That's the one was bigger than Japan or Germany's GDP.  So trying to apply it across the board, I think it'll be difficult with this complexity of the system we have. 

I think the last piece is really the variance, for instance every institution practicing medicine differently, especially in the wound care space. There's really lack of standards, based in our experience. So, we were fortunate enough to have a little bit of a global perspective. We work with different markets. So, we always run into these challenges. We cannot build a kind of a decision support system based on the US practice and transfer to other markets.

For instance, no other market is using tissue products, using hyperbaric oxygen therapy, using all these things. So, a lot of this data here probably doesn't really apply to other places, and that also means if you can actually train your AI based on your own data that fits your own need, that will be the future; that's what we're working on.

Dr. Kim:
So, I think that that makes you make a great point, Patrick. One of the things, the technological challenges, is to be on the local level. So, for example, where I work, it's different from where I work versus 1 mile away at a different hospital. They may have the same EMRs, perhaps the patient mixes different, populations are different, their problems are different. And so the customizability that AI is needed for each individual, I think is the biggest challenge. Because people think that, oh, machine learn everything, get the biggest database, and I don't think that's the right approach either. 

Cheng:
Yeah, that's the challenge, right? When you look at, even though the, whatever, Mad Palm developed by Google can pass the medical exam 70%, but that's all generic stuff. When it gets to the nitty gritty, how you practice medicine, there's a lot of nuances, different organizations have their own systems.

Dr. Kim:
Well, people talk about the complexity of generative AI, and when I talk about it in terms of health care, I'm like, well, there's no other more complex system than the human being. That is the most complex system you'll ever encounter. So, as smart as generative AI will be, you're still dealing with another system that's more complex, which is the human person. 

Cheng:
Yeah, that's millions years of evolution. 

Dr. Kim:
Absolutely, and we don't always do everything right, and we're prone to failure and injury, but I think that's one piece that I think people like, oh, it's all data, it's all data-driven, big data, that's all we need. I'm like, no, it's very individualized too. That's the nuance and the trick, I think is the challenge for companies like yourselves.

In the last couple of minutes, Patrick, I just want to talk about where you see the future of your company, specifically, not the tech space, just your company. 

Cheng:
We wanted to build a system that everybody can use, that can add value to people, be an enabler to help people get through their daily job quicker, faster, and in the same time, generating the data that could help the next generation of health care providers, especially in the wound care space, to do a better job in terms of treating the patient, getting a better outcome, right? So that's our specialization. 

So, we are working on, of course, continuing to developing, there's a lot more than technology, we need to do to get to that point. You know, workflow is a big deal. We have to make sure this is easy for people to use, part of the workflow, and we have to make sure we can integrate with a system, existing EMR system, getting the data flow. 

I think last piece is we wanted to just, medicine is a multimodal practice; you're looking at the patient, you're smelling it, you're trying to do a lot of things. I remember the experience at one time, you actually really have to smell the wound to kind of detect some of these potential infections. So, there's more than just text data. There's image data. There's other things. We're working on bringing all those data into one platform. And then I think the next step is really applying the AI to makes sense out of the data. 

Dr. Kim:
Perfect. Well, I want to thank you Patrick for joining us on this Tech Spotlight. I would encourage the viewers to go on their website for further information. Thank you very much.