Moving Medicine
AI's role in health care: supporting, not replacing, physicians
Mar 20, 2024
Featured topic and speakers
In this episode of Moving Medicine, Margaret Lozovatsky, MD, FAMI, vice president of Digital Health Innovations at the American Medical Association, explores the partnership between physicians and artificial intelligence in health care, focusing on AI's role in clinical decision support. This episode was created in collaboration with the PermanenteDocs Chat podcast and is hosted by family physician Alex McDonald, MD.
Speaker
- Margaret Lozovatsky, MD, FAMI, vice president, Digital Health Innovations, American Medical Association
Host
- Alex McDonald, MD, family medicine physician and host of PermanenteDocs Chat, Kaiser Permanente
Listen to the episode on the go on Apple Podcasts, Spotify or anywhere podcasts are available.
Transcript
Unger: Welcome to Moving Medicine—a podcast by the American Medical Association. Today’s episode was produced in collaboration with the PermanenteDocs Chat podcast. It features Dr. Margaret Lozovatsky, vice president of Digital Health Innovations at the AMA. She along with the host of PermanenteDocs Chat—family medicine physician, Dr. Alex McDonald—will discuss the role of AI in health care to support, not replace physicians. Here’s Dr. McDonald.
Dr. McDonald: Good morning or afternoon, wherever you may be watching this from. I'm Alex McDonald and welcome to PermanenteDocs Chat. Today is going to be another great chat. We are talking about AI today. Just a small topic, nothing too exciting. And we are joined by Doctor ... excuse me, Lozovatsky, who is the VP of Digital Health Innovations at the American Medical Association. So, Dr. Lozovatsky, thank you so much for joining us.
Dr. Lozovatsky: Thank you for having me. I'm thrilled to be here.
Dr. McDonald: So, if you're watching live and if you have questions, please drop them in the Q&A box. We're going to ... This is going to be short and high-yield, so make sure you get those questions in early, and we'll try to address as many of them as we can. So, we're going to just jump right in here. So, tell us who you are and what you do.
Dr. Lozovatsky: Absolutely. So, my name is Margaret Lozovatsky. I am a general pediatrician. I practice as a pediatric hospitalist, and I am currently the vice president for Digital Health Innovation for the American Medical Association. I am double-boarded in pediatrics and clinical informatics and have served in multiple CMIO and CHIO roles across various health care organizations.
Dr. McDonald: Wow. So, lots of hats and lots of aspects you can look at this. So, I'm excited for this conversation. So, as I kind of alluded already—AI, augmented intelligence, artificial intelligence, however you want to, whatever term you like using is really all over the news these days and not just in health care. It's poised to be a disruptor in many different industries, including health care obviously. So how much of this is real? How much of this is hype? How fast is AI really being deployed in health care and the health care setting? Are we talking about months? Are we talking about years?
Dr. Lozovatsky: Wonderful question, and of course, top of mind for all of us in clinical medicine. We here at the AMA really think about AI as augmented intelligence, and we use that term very intentionally because there's a lot of technology that's out there in the clinical spaces. We're all aware of it, and we recognize that it's there to augment the physician and the physician's role. And so, we think of it as technology that's going to be helping our physicians with their day-to-day work to be able to care for patients.
In terms of timeline, in the last, I would say, year or so, AI has become very top of mind for everyone, largely because of generative AI that has come out. But when I really think about clinical technology, I think of AI as having been in the clinical spaces for many years, from clinical decision support to natural language processing to a lot of the predictive analytics that is live and being used in the clinical spaces today. So, is it moving fast and will it move faster? I think so, and I think it's already integrated into our clinical environments.
Dr. McDonald: Interesting. And I just want to reinforce the concept of augmented intelligence, and this is designed to help make our jobs as physicians easier and better and improve patient care. I know lots of electronic health systems have reminders that sort of pop up, "Hey, your patient is due for their tetanus. Hey, your patient is due for their flu vaccine."
And I guess even those are some ways, some augmented intelligence that's already being deployed, which we're all very comfortable with, I think. Is that accurate or is that maybe not? I don't want to put words in your mouth or reframe the issue. You're the expert here, so you tell me.
Dr. Lozovatsky: Well, that is a really great example of what we call decision support that's already out there. And often on the backend of that decision support, so the reminders that you mentioned, may be some augmented intelligence that is helping to generate those reminders. We also are all very used to predictive models. So you sometimes see a score that says, "Your patient has this percentage chance of becoming septic," or whatever the clinical scenario may be, and those also have some of the AI models running in the background to help give you additional information. So, it's gathering that data in the back, and it's showing it to you at the time that you're providing care to help you make decisions based on some of that augmented intelligence that's doing the calculations.
Dr. McDonald: Right. So that's really helpful to know. So how do you think the sort of next generation of generative AI is going to change medicine in the next five to 10 years? And I know it's a bit like reading the tea leaves, but from your perspective, do you think that there's going to be major changes or this going to be a slow kind of iterative process?
Dr. Lozovatsky: I think that there's going to be major changes at the end of this, in terms of how we're able to utilize our technology to be an enabler. So, I think of all the years that I've been in medicine, and I like to talk about how I started with a clipboard, and now the residents and the medical students can't even imagine a world where we don't have technology in every aspect of care. I do think that those changes are iterative day to day, but at the end of the day, when I look at the power of augmented intelligence, I really see it as a tool that can help gather information and feed it to the clinicians.
We think about this cognitive burden that we experience on a daily basis. I'm a pediatric hospitalist, and so when I come into the hospital, there's all of these pieces of information that are being thrown at me, and you're trying to manage so many different patients at the same time. Wouldn't it be great if the system could provide the information that you need for the particular clinical scenario of that patient at the right time and the right place? That's where I think the power of AI is and where I hope to see it moving to.
Dr. McDonald: Yeah, no, that is certainly helpful. And I know just the vast knowledge base in medicine is continuing to expand and continuing to grow, and there's no way that any physician can kind of be ... know everything all at once. And so, I do—just for my own clinical practice—really appreciate some of those reminders and some of those pieces of information being like, "Hey, did you look at this?" Or "Hey, think about this over here."
So, I know that's very, very helpful for me personally as a physician. Now, there are some doctors out there who are hesitant about AI technology, particularly some of this new ambient listening technology for charting, and they're worried about concerns about privacy and automating their work. How would you address these concerns, or how you ... how can we help clinicians feel safe in exploring AI in their clinical context?
Dr. Lozovatsky: It's interesting that AMA did a survey of physicians, and what they found is that a large majority of physicians said that they are excited about this technology, and the things that they brought up are similar to what you had said, which are the fears with—
Dr. McDonald: Right.
Dr. Lozovatsky: —the technology entering the clinical spaces. The biggest concern was actually the patient-physician relationship, and data privacy was the second biggest, which I think are really important. I would say that, like any new technology, we have to be very thoughtful as to how we implement it in the clinical spaces.
And to me, the key to this is number one, making sure that physicians are engaged in these implementations and are thoughtfully helping the technology teams analyze the new technology, the workflow, how it's going to be used. And the second piece that is really critical is it is there as an assistant, a co-pilot, if you will, to the physician. At the end of the day, we're trained to make those clinical decisions, and we will continue to make the clinical decisions, and this will be another tool in our tool belt, if you will, to enable us to care for patients.
Dr. McDonald: Yeah. No, I think your point about physicians sort of being on the ground floor and really making sure we are involved in the early implementation here is so critical. I think some physicians have been burned maybe by the implementation of the electronic health record where, let's be honest, a lot of physicians were not involved, and we feel like that's not very intuitive.
This is probably coming whether we like it or not as physicians. And being involved and being on the ground floor, we can at least help frame it and make sure it's done in a way that's safe and that works for both patients and clinicians. And so that's my own 2 cents.
We have a question, our first question in the chat here, which actually is tied very closely to that. How do you see augmented intelligence really helping to alleviate physician burnout and improve physician well-being both now as well as in the future?
Dr. Lozovatsky: I think one of the most critical ways that it can be incredibly useful is to help alleviate some of the administrative burden. Again, we talk about the cognitive burden that we experience. Part of that is the management of the clinical information, but a lot of it is really administrative burden, things that the computer can take off our plate. So prior auth being a perfect example of a situation where you really can use some of this technology to help take some of that burden off.
I think that the chart review that we do, and the processing of information can be summarized. Some of those new tools are just up and coming, and I see a lot of potential in those. The tools that are coming out that are used to help with the in-basket that are writing those responses that we can review, we still have the review, and we make sure that they're accurate and that they're clinically appropriate, but again, takes away some of the administrative burden that we experience.
Dr. McDonald: Yeah, yeah. And just I guess from my own experience, I've been able to use some of these tools that kind of can listen, ambient listening tools, to the patient encounter and then help create documentation, clinical documentation for that. And honestly, I asked the patient, obviously I put my phone down. I have a conversation with the patient. I don't look at the computer once, and then I walk out of the room, and much of the note is already created for me using this technology. And just for me personally, that's what I find much more fulfilling about that patient-physician relationship.
So, I think, if anything, this will help prevent us from staring at our computer screens the whole patient encounter and really actually be able to connect with the person, which is why I think a lot of us went into medicine. That's my own very brief early experience. So, I have to put that caveat in there, but I know that a lot of patients, excuse me, a lot of physicians feel like they're just tied to the computer, and they can't actually connect with their patients on a personal level. And I think this technology may really help pave the way to get back to that.
Dr. Lozovatsky: Yeah. And that is a really good point, and something that we hear across the board from physicians is that they often will have this experience of being tied to the computer and not being able to pay attention to the patient. So, I absolutely agree that the technology can help to actually enhance the patient-physician experience.
Dr. McDonald: Yeah. So how can physicians really kind of address some of these benefits and challenges head-on? Like I said, this is sort of the wave has started coming, but what recommendations would you give to individual physicians who want to learn more or kind of want to address these concerns, be it at their clinic, their hospital, their health system level, or even the policy level in the state or nationally? How would you recommend physicians to go about addressing some of these concerns or challenges or benefits that might be coming?
Dr. Lozovatsky: Well, I think that, as we discussed, the engagement is really critical. So, I love that you're asking this question because a lot of folks may not know how to get engaged. One of the ways that within institutions people can get engaged is to get to know the physicians that are working on this. So, most organizations will have a clinical informatics team or a clinical informatics leader, whoever that may be.
And I do talk a lot about the fact that clinical informatics as a specialty is fairly young, and as we think about technology becoming so integral in our care for patients and everyday lives, it's really important to have the folks that understand that technology and understand clinical medicine be at the forefront of that.
So, I would encourage all physicians to get to know those teams, to understand what their organizations are doing, and to help provide input to the clinicians and those teams that are working with IT departments to implement the technologies so that they are intuitive so that they are implemented in a way that's most useful in the clinical spaces.
Dr. McDonald: Yeah. No, that makes perfect sense. So, another question here in the chat. This is great. Keep the questions coming. I see lots in the queue here. "How do you think that physicians should inform patients about the use of AI in the treatment setting?"
Dr. Lozovatsky: We talk a lot about transparency and the importance of transparency. Just as it's important for physicians to understand what technologies are being used in their care for patients, I would say that it's also very important to be transparent with patients, and that's why it's so critical for the physician voices to be engaged in these conversations.
I also would add that we at the AMA have put out a document recently with some of the newest trends about AI, so people are looking to learn more. That's another really good resource, and we will continue to put out those materials to help people stay on top of some of the new trends that are happening in this technology.
Dr. McDonald: Yeah. Do you think patients are kind of learning about this and aware of this? I mean, again, this is not just in the health care industry. This seems to be across all industries, this explosion of AI and the generative AI in particular. Do you think patients are aware of this, or they're seeking this information out themselves?
Dr. Lozovatsky: I think so. I think patients are reading just like we are. They're looking at the news, and they have a lot of questions. "How is this going to be impacting my care?" So, I do think it's critical to address those questions to help people understand. And patients are interacting with some of our clinical technologies. They're looking at the patient portals. They're seeing some of that information. So, I do think it's going to be really important for that transparency to exist within those tools as well.
Dr. McDonald: Yeah. What do you think about physicians who are worried that AI will just replace them and replace their job? Is that something we'd be concerned about?
Dr. Lozovatsky: I personally don't think we should be concerned about that. I really do see it as an augmentation tool, and that's why I really like that terminology because it is there to assist physicians in caring for patients, not to do the work of physicians. But it is important because the burden, as we talk about, is clear.
We constantly hear about burnout and all the things that are coming at the physicians. So, the fact that this technology is going to be there to be assistive in that work is really, really critical. I just do not think it's there to replace the work that we do because that is going to require a human to be able to process, to diagnose, to care for patients.
Dr. McDonald: Yeah. No, I agree with you entirely. I'm not worried about being replaced by a computer anytime soon, at least. And I know a lot of my patients, they really seek that face-to-face interaction, that connection. Even though it doesn't always feel that way from our perspective, I think a lot of patients really want that relationship, and computers can't do that despite how much we've tried.
Another great question here in the chat, which is a little bit of a … connected to some of the information we talked about prior, but obviously physicians are maybe not experts in computer science or technology. And yes, getting involved with your chief information officer or whoever's going to be kind of leading this charge within your organizations and important. But where can doctors just learn more about this technology and become sort of more expert in new technologies so they have more knowledge about it and then can then help understand the limits and the benefits and the implementation?
Dr. Lozovatsky: Yeah, absolutely. And I don't think anyone is expecting for physicians to be experts in this field, but I love the question of how can I learn more because we all want to at least have the basic understanding of what's happening. I mentioned some of the tools that the AMA has put out there, and I would encourage you to look at them. I also would encourage you to speak to the physicians in your organization that are experts.
So, depending on who your electronic medical record vendor is, for example, there are classes you can take through them. Your informatics team probably has some materials within your organization where you can learn more about that. There are clinical informatics classes that you can take. There's lots of materials out there, but they, I think, would be most useful to folks when they're tailored to their clinical space and their institution.
Dr. McDonald: Interesting. Yeah. No, I think that's the key is there's a lot of different pieces out there. And I think physicians, by and large, we want to know, we want to understand, we want to learn not just about medicine but everything adjacent to medicine as well and does seem like there's some great tools out there and you've already mentioned some of these pieces.
So, I think we've gone through most of what I had here. So, I think we've covered this really well. I feel like most of our questions have been answered here. There are a few things I think straggling in here, but more importantly, I have a question for you. I guess, in some of your background in informatics, how do you get involved in informatics, and do you encourage other physicians to learn more about medical information technology?
Dr. Lozovatsky: Yeah. Well, of course, I am slightly biased, but I am really interested in this field. So, my background is I do have a computer science degree, but I do not think that's necessary for this field. And as I came up through my training, we were doing a lot of implementations. That's when EHRs were just starting to have a presence in the clinical spaces. And I found very quickly that I was the person that could talk to the technology teams and that was able to get our needs met if you will. And so, I became the go-to person. That's how it all started. And what I really enjoy most about this field is solving problems.
When a clinician is trying to take care of a patient, and they're frustrated because they can't get something in the system to work, that is the best opportunity for us to take away some of the friction, if you will, that exists. So I started doing more and more of that work. I would encourage folks, if you have any interest in this space, to look into these opportunities, and I mentioned a few of them. But one of the early things I did is, I became a physician builder, and so then I could actually make changes in the system. There's lots of these things that you can do within your institutions to make incremental changes.
Dr. McDonald: Yeah, that's great. I've always been fascinated by computers. I'm better at using them than building them or fixing them once they break, but something that we're all... the next generation, having grown up with computers, might be better at it than we are.
So to your point earlier about paper charts, a few years ago, our computer system went down for a couple of hours in the middle of clinic, and all the residents were freaking out. And then some of our older doctors just brought out a pencil and paper, and they just kind of got back to it. So, it'll be interesting. What's old is new again. One last question here. What is your biggest concern about AI in clinical medicine?
Dr. Lozovatsky: My biggest concern is that it will be implemented in places without the appropriate due diligence.
So it's not so much that I'm concerned that these technologies are coming into the clinical spaces because we've had new technologies over the years in all of our clinical medical care. What I think is going to be critical is we talked about having that physician voice, making sure it's aligned with the workflow, making sure it's safely integrated so that there is transparency into where it's used, how it's used, and really looking at bias and making sure that it is providing the output that we're expecting.
Dr. McDonald: Yeah. Yes, computers are designed to do exactly what we tell them to do, right. And whether you implement that technology with bias or not can certainly affect the outcome. So that's an important piece to consider and why it's important that we're all involved and we're all at the table as we deploy these technologies to make sure we are collaboratively working to mitigate any implementation bias, which will then result in the end outcome too. Well, this is wonderful. We could, as always, go on forever and ever, but again, I try to keep these short and high yield. Last question, and perhaps always my favorite question, what makes you most proud to be a physician?
Dr. Lozovatsky: The list is long. I struggle with answering this question with just one thing. I am so proud of all of my colleagues that are out there every day, putting in 150% of effort to care for patients and that truly want to do what's right for the patients in front of them. And so, over the years, I've worked with a lot of physicians, and what I have found is that in our profession, people really, truly are so resilient and are such hard workers. That's what makes me proud to be a physician.
Dr. McDonald: Wonderful. That's great. Well, Dr. Lozovatsky, thank you so much for joining us today. I really appreciate your time and your expertise and all your insights.
Dr. Lozovatsky: Thank you for having me. It's been a pleasure.
Unger: Subscribe to Permamente's Docs Chat podcast to never miss an episode and register to take part in upcoming live chats. Visit permanente.org/amadocschat.
Disclaimer: The viewpoints expressed in this podcast are those of the participants and/or do not necessarily reflect the views and policies of the AMA.
