Dr. Eric Eskioglu
0:00 AI is already in our lives, whether or not we know about it or believe it. You know, when you talk about your cars, those are all sensors that give you information. You use that pretty accurately. Now, do you question, is my beeper if I'm making lane change accurate or not? Maybe the first time you drive the new car, but once you get used to it, you can trust it almost all the time. You're still vigilant, you're still making sure that it's okay. But after a while, once you notice that it's almost 100%, correct. You don't think about anymore, and it becomes almost like an extra sensory organ that you don't have.
Gina DiPietro
0:39 That's Dr. Eric Eskioglu, Novant Health executive vice president and chief medical officer, on how artificial intelligence brings together vast amounts of data to quickly find insights and efficiencies. You're listening to Industry Insights: A healthcare podcast presented by Novant Health. I'm your host, Gina DiPietro. In this episode, we explore how healthcare is using AI to solve real problems like health equity, cost of care, and better patient outcomes. More on that to come. But let's take a step back and start with what is data? And how is it used to get artificial intelligence? Here again is Dr. Eskioglu.
Dr. Eric Eskioglu
1:20 So the data portion is your anything a patient generates throughout their lifetime. Whether it be regular physical checkups, regular care, or unfortunately have to end up in a hospital if they get labs if they had x\X rays, MRIs. So this is all data. But as we started growing as a nation, and you know, our medical knowledge actually scientifically has increased tremendously. 50 years ago, our medical knowledge was doubling every 100 years. And currently, it's doubling every 72 days. And it is expected to double every 30 days in the next two to three years. So you can imagine the amount of medical knowledge and the data. We're going to be overwhelmed with all of this. We need to put it to good use. And from this data, you can get predictive analytics, which is the initial step to get into artificial intelligence. And that predicts, you know, things that could happen to a patient before it happens, which is very helpful. Then you can do what's called machine learning. And that's more getting structured data in an area where you can start not only predicting things, but also expecting things and also working with things that we can prevent. The final stage is artificial intelligence. And this one has two components. It's got recurrent neural networks, which is data that's dynamic that is streaming. So you can think of a heart monitoring for EKG, that is a data that's continuously streaming, you have to make decisions on the spot. And then there's also a convolutional neural network, which is the static data such as if you look at an MRI of a liver. That data statically, you have to turn it into an artificial intelligence. Now, what's exciting to me is artificial intelligence, we already have it around us. Whether or not you drive a Tesla car. Tesla cars are equipped with it. Whether or not you have an Alexa, if you're an Amazon Prime member. They not only when you log in, they can tell you what you have bought in the past, but what you may like from your profile. Artificial intelligence is going to be the fourth industrial revolution. I truly believe in that. In healthcare, if we don't adopt it, with the increasing complexity of patients, increasing amount of data, doubling of the medical knowledge every 72 days, which is going to decrease to every 30 days. We are not going to be effective and efficient and we're not going to have good outcomes. So it is imperative that all of us get into this field.
Gina DiPietro
3:37 Keep in mind here that AI will not replace people in healthcare. Machines will never replace human qualities like empathy, intuition, or compassion. But it's important that people like physicians and nurses adopt AI to better care for patients. Dr. Eskioglu weighs in on where he thinks it will have the most impact.
Dr. Eric Eskioglu
3:57 Physicians and nurses who use AI are sure to replace the physicians and nurse who refused AI. They are going to be much more effective, much more efficient. You know, I'm a vascular neurosurgeon, so people are surprised to hear me say that AI is going to have the biggest impact on primary care. I've said it for many years. You know, when you want to get a heart care, you think about several institutions around the country. When you want to get cancer care, you think about several institutions. But when you think about primary care, you don't hear somebody say, I'm going to get go to Cleveland Clinic and get my primary care or I'm going to go to MD Anderson to get my primary care. The reason it's going to change primary care. Those are the physicians that have the longest tenure with the patients. They have data, maybe q&a with a patient, you know, 20 to 30, sometimes 40 years of data. Because when you're with a primary care, you're you're with that person for a long time. With that data comes a lot of treasure trove of things we can predict. So, you know, we're going to be able to use artificial intelligence to predict trends from that data. Whether it be labs, whether it be MRIs, whether it be from natural language processing, your notes you wrote 20 years ago or 15 years ago. Next time you go to your primary care physician, ask them how many notes they go back to. And how many labs or x rays do they go back to see what went on. I can guarantee it's no more than one or two. So there's a lot of trends that were missing. AI is going to come into that space where it's going to get the physician or the nurses' or nurse practitioners' attention saying, 'We see a trend here from the last 15 years. And you may want to look into that.'
Gina DiPietro
5:31 Here's the bottom line. Physicians and nurses can use that data to precisely diagnose and even predict health outcomes before they happen. Here again is Dr. Eskioglu with more on the potential of AI to reduce the cost of care, and why it could be a great equalizer for health equity.
Dr. Eric Eskioglu
5:49 There's two things to AI that's really going to help us out on. We're going to be able to lower cost, because we're going to lower the covariation channel in the treatment modalities by using AI. What I mean by that is we've been looking at pneumonia, uncomplicated pneumonia admissions. And you know, before that homogeneous patient population, the outcomes are pretty similar, they're good. But we're noticing that some physicians order what we call the kitchen sink or the Cadillac workup. They order labs every day, x rays every day, consults. Some of them order only every other day, or every three days. So with AI, we're going to be able to predict after how many labs you've had normal do you need to stop ordering labs? Is one lab or two labs enough? So that's going to help with clinical variation. What we call economic value enhancement. And by doing that, we're going to be able to lower the healthcare costs, thereby increasing the access to underserved communities and people who have less or no access today. The other part of AI is we're going to help with social determinants of health, but also it's going to provide health equity. And what I mean by that is I'll give an example. In our stroke care, we've been very successful working with a startup company called Viz.ai. We were one of the first in the country to adopt them. when they were just at the beginning stages. It has proven to be a complete hit with us and our patients have been the true winners. This is an artificial intelligence layered on top of our telemedicine and we're able to detect a stroke on a CT scanner before the patient's out of a CT scan. On average, we've been able to save about 10 minutes per stroke patient, which equates about 19 million brain cells. Quality of life has increased. They don't die and this has had a tremendous improvement on the length of stay. But by doing that, we've provided health equity because you know, if you're a patient in Elkin, North Carolina, which is one of our affiliated hospitals that works with us. Or if you're a patient in uptown Charlotte, guess what? You get the same standard great care for stroke. There's no variation. You get the same care. So you don't have to be in a metro area to survive a stroke. And you know, that I'm proud of. So it will be the great equalizer. AI will be the great equalizer for health equity. And I truly believe that. Now, a couple of caution points. One of them is we have to make sure that we don't get bias entered into AI algorithms. And to prevent that, we need to make sure all populations data are entered not just a certain segment of a population. So we need to make sure Caucasian, African American, Asian American, Latino as well as any Native American, these populations have to be entered into the AI databases, because then you can get a true picture of what you need to do, rather than having a bias. So those are areas that we have to be cautious about.
Gina DiPietro
8:42 As Dr Eskioglu put it, healthcare systems should approach AI with three pillars. I'll let him explain.
Dr. Eric Eskioglu
8:48 First of all, from safety and quality. You know, we want to make sure whatever AI project we do is going to improve the safety and quality. So there may be an operation where somebody may be diagnosing the patient, but it has to improve the safety and quality. The second one is it has to lower the cost by unifying the protocols. The third one is it has to provide access to health equity. So those three pillars have kept us really in check. I always said, if you provide the highest quality clinical care with the help of AI, if you lower the cost, you are going to have quite a bit of money coming back in to be able to put for your mission, in not for profits. You will be able to invest in further things. And these things are not cheap. But you know, if you save more lives, if you prevent hospital acquired infections, if you provide serious safety issues in the hospitals - you'll be able to get to a point where you know you're one of the best in the country. And people will seek you out.
Gina DiPietro
9:45 Looking ahead, he's especially excited about a few other things as well. I'll let him wrap up our conversation.
Dr. Eric Eskioglu
9:51 The other things that's going to speed up AI is the quantum computing. It's here. It started last year. Google was the first company to be able to achieve that. So that's going to rapidly increase. The other thing that excites me is the fact that we can provide health care equity, using AI for all the populations we serve. I think that's going to be tremendous. I've always said it, the basic promise of democracy is health care access to everyone. And if you don't have your health, you can't have your economic upward mobility. So health is the most important asset you have. Forget about your house, forget about your car, forget about your bank account. The most important asset in your life that you will have is your health. The other thing I'm really interested excited what we call ambient intelligence. So artificial intelligence is your data, what we do with your data and how we make it work for you. Ambient intelligence is the data surrounding how does that affect your health? That's going to take more and more center stage since the environmental factors. We're going to be able to notice that if you're getting out of a bed, are you more likely to fall? We're going to be able to notice that. And that's what you know, a lot of our lives revolve around right now with our cars. What the sensors they have is AI, but what we call ambient intelligence, it gathers the data around your environment, and feeds it to you. So there's going to be two components -the AI, which is going to be your own personal medical healthcare data, that's going to affect how you get treated, but also the ambient intelligence, which includes social determinants of health. And those two combined with astute physicians who really utilize AI and again, I program in Python, you don't have to be a programmer in Python, you just have to have a basic understanding of AI and how it's going to help your patients. So I'd like to think people to think outside the box when we get to that point.
Gina DiPietro
11:45 Gina DiPietro here and thank you for listening to this episode of Industry Insights: A healthcare podcast presented by Novant Health. You can find more episodes in our Industry Insights library. We've touched on healthcare supply chain, orthopedic trends, and even workforce burnout. There's tons of great content there. So feel free to browse around on Apple, Google, Spotify, or anywhere you listen to podcasts.