← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Dr. Maulik Purohit, Chief Health Innovation Officer at datosX Digital Health Labs
00:12:54
Loading summary
Transcript12 lines
- [00:00]
A
This is Scott Becker with the Becker's Healthcare Podcast. Thrilled today to be joined with a brilliant regular guest. We're joined today by Dr. Malik Purwahit. And Dr. Peretz is going to talk to us about AI, what he's seeing, what's going on in digital, and a lot more. Malik, can you take a moment and introduce yourself and then let's talk about what you're seeing out there.
- [00:22]
B
Hey, Scott, thanks again for having me. It's always a pleasure and honor to be part of the Becker podcast. Yeah. Happy to discuss what's out there. Just as a brief background, my name is Molly Prath, a physician in brain injury, neurotrauma, and then having served as chief health information officer for several health systems, overseeing the digital infrastructure for health systems and everything that entails, which, as you know, in today's world, that's a lot of AI and automation that goes on. And then recently I've stepped out and I'm with consulting company called Datos X, where we actually help validate AI or other digital health technologies. It's been an interesting perspective, having been on the health system side and now on the side of evaluating these technologies to see do they really deliver on what the promises are and helping identify that it's been an interesting run. And I'll tell you, having been a CTIO before, it would have been nice to have had that option or a company like that before because, you know, I think this works for, I can speak for a lot of the CTIO, CMIO, CNIOs, others in the healthcare industry, where you always get pinged online by different various companies and they come and say they've got the best product in the world, it'll solve all your problems, reduce your length of stay, reduce your costs, reduce the FTEs you need, improve quality and solve world hunger at the same time. But when you try to implement them into your system, then there's always a challenge and they work. Sometimes they work, sometimes they don't. But having that validation tool is always great to have ahead of time, which was not an option before. And so it's been an interesting perspective to be on both sides of the. Of the fence, so to speak.
- [02:07]
A
Yes. No. Malik, what's working well in digital health? What's working well in AI? What do you see that's going well that people are thrilled with that? They're like, oh, my goodness, I can't believe we were without this before. What is going well?
- [02:22]
B
Yeah, absolutely. That's a great question. And I'll tell you, the way I used to talk about AI before is that it's definitely artificial and occasionally intelligent. And I think in many ways, on the clinical side, I'd say that statement is still true. But on the operations and other side, I think there's been a huge, huge advantage in so many ways. And I think I'll tell you, one space that I'm super excited about is an area that's been plagued doctors for so long, which is documentation. And this was countless hours in a day and again away from patient care. And as a practicing physician myself, I see this all the time where my interaction with the patient is affected in negative way because I have this pressure of documentation behind me and either I have to do it in real time in front of them, which means my eyes are on the keyboard and on the screen rather than them during the visit, which is not satisfying to me, nor the patient. But I think to me this is a huge improvement in quality because now if I've got an AI tool that can record that conversation in the right format, in the right ways, with the right context, while I can focus on seeing the patient, evaluating them, understanding them, building that rapport, that's a huge add to the system. And I think there's many companies out there that are doing this really well. I think, you know, all the names I'm not going to mention here, but I think there's some, several really good companies that are doing this in an exciting way that allows that entire physician patient interaction to be re evolved into what it was meant to be, which is really caring for the patient and having that human touch. And so I think that's one area that's a huge development. Other areas I see are operations and in terms of automating, backend processes that were manual and tedious and required lots of people to do them. And, and again, it was not satisfying to anybody because it took, you know, hours and days as you were online. But now those are automated so they happen almost instantly. So things like scheduling can often be much better in terms of validation of insurance and those types of things. And so I'm seeing a lot of great progress in those areas. Again, nothing is perfect. There's, you know, just like anything else, there's pros and cons and there's ups and downs in that area, but at the same time, the progress and the innovation and the development has been huge.
- [04:39]
A
Thank you. Are there things that you think you don't have to. Please don't mention brands or names that aren't living up to the promise and aren't Working well that aren't anywhere near where people expect them to be.
- [04:48]
B
Yeah, absolutely. I won't name any names, but I think certainly what I'll say is it's not that they're not living up in the sense that they're bad tools, but I'll say that I think we're still need a lot of evolution there. And so products that get into diagnoses, for example, products that write patient notes, for example, I'm not notes in terms of documentation, but responses in terms of like a FAQ or those types of things, or symptom checkers and those kinds of things, I think those are still needing a lot more work. And I think part of the reason is not so much that the technology is bad, but I think this is going back to an area we've talked about many times, is improving the data quality that we put into these tools. And the analogy I use oftentimes is if you're raising a child right in AI, in many ways you're raising a child in that sense is if you give proper nutrition, if you give proper exercise, if you give proper education, you're going to develop a really good child that's going to become a great adult, or the chances are higher. At least if you have a child where you give them poor nutrition, no exercise, bad education, then the chance of that individual not becoming a productive member of society is much higher. And I think AI is very much the same way is that if you give give AI as a tool the right data sets, the right information, nurture it, give it the right backend processes, then you have a much better probability of becoming a good tool that's usable for the most part. And I think that's what we're seeing now is that the tools that are being developed, like the algorithm itself, may be fantastic, but they need the right training data sets to make sure that the development of that child or AI, so to speak, is heading in the right direction. And again, what we're seeing is that as we're having these tools come out, we're needing to revamp our entire infrastructure around it to help that AI tool become productive and useful in a meaningful way. And so that's, I think what's going on as well in concurrence. And I think it's a great development because as we develop not only the tool itself, but the infrastructure around it, it's going to get better and better. Now the analogy I'll use is sort of like the. Is Tesla, right? This is a, a car that was an Evolution in EV cars. Not the first one, but really a dramatic improvement. But then the concern was the infrastructure of charging around it. But as soon as that infrastructure caught up and now you've got a vehicle and a charging station that helps not only that company, but other EV companies, It's a fantastic development. And I think it's the same sort of concept analogy with AI is that you're seeing the infrastructure around AI get better, and as that happens, the AI itself and the tool itself will get better and better.
- [07:26]
A
Yes. And talk about where in AI are you most excited currently? What's most exciting currently?
- [07:34]
B
To me, the most exciting is areas that people probably haven't heard about, which is seeing things that have not been seen before. Let me give you an example. Before we had lab tests and other things, we relied almost exclusively on history, meaning what symptoms the patient had, what they're going through, and then examination, those kind of things. And then, you know, blood work and those kind of things were not really that possible 100 years ago, 200 years ago, whatever time frame it may be. And then, you know, eventually we started developing lab work where you get what's called the CBC or complete blood count and you can see the amount of different blood cells in the, in the blood itself. And then that allows you to help diagnose and adds information to those things. But human eyes could not look at a blood sample and give you that information. Human yet to have another technology, lab machines, others to evaluate that blood, to look at that and identify that in a more meaningful way. And what I'm seeing Now is with AI algorithms, we're seeing that in terms of, for example, EKGs, where even EKG reading and the human eye can see only so much within that ekg. And it's critical what they do see, especially in the acute setting and ruling out heart attack or not. But there's other fundamental things that are within that electric reading that can identify structural things, for example, like a long term heart issue or a new development or genetic issue, those kind of things. There are companies coming out with algorithms with AI that are reading EKGs in a sophisticated, amazing way that gives you insight from an ekg, simple EKG that we could not have before, simply because human eyes cannot see certain things. Same thing with radiology, same thing in other areas. What I feel excited about is extending the ability to diagnose what is going on, somebody which then allows you to have the right treatment for them. And everything in healthcare starts with a diagnosis, because if you don't know what's going on with the patient and what's causing it, then it's really impossible to take care of them in a meaningful way. And so this allowing us to expand that and look at things in a much different way, much more holistic way. I think these are exciting things that are on the horizon.
- [09:47]
A
What specialty of the specialty is likely to have serious negative impacts from AI and sort of income, business, what people do like, I don't see an orthopedics in the short term, but what specialties are there that you think will be seriously impacted by AI and so forth?
- [10:08]
B
Yeah, great question. I think all of us will be impacted by AI. The in general there's going to be a positive impact. But I do see to your question some negative impacts in what are called evaluation management visits, E M visits for short, which are visits where there's an interaction between the patient and the physician and there's a dialogue and then there's a diagnosis and then there's a discussion of what needs to be done. And that can be in primary care, that can be rheumatology, that can be in my core specialty, physical medicine, rehab in brain injury, brain wellness, it can be neurology. And the reason I say that I can see potential negative impacts is because in that interaction, because of our healthcare system and the structure that we have right now, we see that there is, let's just say there's a want of a better interaction and want of better time and those things and from both sides, from both the physician side and the patient side. And these 15 minute visits are not satisfying. Patients are turning to online tools and ChatGPT and other things to help supplement that visit or that knowledge base. And we're seeing, and you've seen this in news recently, in the last week alone, we've seen some of the issues with some of these online tools that give you the wrong advice or very dangerous advice. And so the concern is in the short term, as we're evolving to a much better place, are we going to have incidents where the AI tool gives bad advice or contradicts a physician advice or contradicts known evidence based practice and as a result causes harm to the patient because they follow the wrong advice for better or worse? Oftentimes we look at technology as being better than the human in terms of giving us advice. That danger is ever present as we're evolving towards a better place. To me I can see that as the biggest concern is supplementing your normal visit with an AI tool or technology that actually contradicts known evidence based practice and leads you to a worse place in terms of your own care or your own diagnosis. And that's concern, especially in the younger group, because that's more likely to use these tools in the youth. And so that's the scary part of this is can we mislead somebody to the wrong path simply because the tool is not ready, but it's been used in that way?
- [12:20]
A
No. A fascinating perspective and always curious as to where AI is going to make it a lot easier for specialists versus where it's going to hurt their practice and so forth. The jury is very much still out a lot of uncertainty and certainly some of the specialties that people thought would be disrupted a long time ago by this have not. And so it's always hard to really project it. Dr. Perry, it is always fantastic to talk to you, one of the brightest people I get a chance to talk to. Thank you for joining us today on the Packard Healthcare podcast. What a pleasure. Thank you very, very much.
- [12:50]
B
Thank you for having me. Always a pleasure.