Speaker 01: 0:01

Welcome to the Healthy Matters podcast with Dr. David Hilden, primary care physician and acute care hospitalist at Hennepin Healthcare in downtown Minneapolis, where we cover the latest in health, healthcare, and what matters to you. And now here's our host, Dr. David Hilden.

Speaker 02: 0:19

Hey, everybody, and welcome to episode 17 of the podcast. I am your host, David Hilden. And, you know, we all talk about AI in our lives, but did you know that it's also present in healthcare? So today, we are going to talk about artificial intelligence and how it's showing up in exam rooms, not as a robot doctor, but as a silent partner, taking notes, reducing paper, maybe even saving a little sanity, but we'll find out about that. Joining to help me out today to make sense of the technology and how it is impacting healthcare is Brian Imdek. He is an adult nurse practitioner and a published expert on the topic. Brian, welcome to the podcast. It's a pleasure to be here. Thank you for having me. So set the scene for us. What is a clinical informaticist? Because you deal with those kinds of folks. Yeah. And tell me about your job here at

Speaker 03: 1:04

Hennepin Healthcare. For sure, for sure. I like to compare a clinical informaticist to that tech-savvy family member. Every family has a tech-savvy person, somebody who's looking up the best travel website, who's looking up the best roadmaps and how to get from point A to point B, maybe the budgeting app. We as clinical informaticists do something similar, but in the healthcare environment. And the role is to help healthcare teams and patients use technology better to improve care. Clinical informaticists focus on ensuring digital tools like the electronic health record and different applications are well-designed and they support better, safer, and more efficient care.

Speaker 02: 1:43

So you talked about the electronic health record. We use that here for 20-some years. And most big systems around the country use electronic health records. For our listeners, describe, first of all, what is the electronic health record? What are you talking about there?

Speaker 03: 1:56

Oh, boy. The electronic health record is basically the repository for all of the clinical care that is given to patients. The electronic health record is used to communicate between providers and others in the care team. It's also used to track the care progression and what kind of interventions are done. We use the electronic health record to communicate diagnostics. And it is a huge repository of so much information, and it follows the patient's continuum over a course of time. There's thousands upon thousands of data points for each individual patient. And so the electronic health record is how we convey the care that is given to patients.

Speaker 02: 2:37

So I grew up in the age of the paper charts. We used to have to go to the basement of this hospital, this very hospital we're sitting in, and get a grocery cart. It was a grocery cart. And we would have a stack of manila folders full of paper records, and I would have to pour through those. So if you wonder why your doctor or your nurse practitioner didn't know that you had a CT scan 18 years ago, it's because it was buried in this paper chart. So the electronic health record helps a great deal of that. That's why there's computers in the exam room. But talk about where artificial intelligence now, that's the main topic of today, that's different from the electronic health record, right?

Speaker 03: 3:13

Right, right. So artificial intelligence is basically helping in the circumstance and what we're going to be talking about today is helping providers put their thoughts into the electronic health record to take the information between a patient and a provider and put that into the health record in a way that is clinically meaningful to not only the individual provider, but the patient as well.

Speaker 02: 3:36

So that might be a little bit scary to some people. You're going into your visit and now we're talking about artificial intelligence. How does

Speaker 03: 3:43

it work in health care? So right now there are a lot of different use cases where it's basically taking artificial intelligence, takes existing information and constructs it and puts it into a way that is digestible, that's meaningful for providers and patients. It can help predict disease trajectory. It can help predict whether or not a patient is going to be admitted to the hospital or if they are at risk of deterioration. There's different models in how AI is used in healthcare. The one way that we are using it predominantly is to basically be present during a clinician and patient visit. And during that visit, a recording is basically taken. And the AI takes that recording, leans on algorithms and large language models to basically help generate a note, a clinical note for that visit. And what that leads to is substantial reduction in the effort that it takes for a provider to put together the note for that

Speaker 02: 4:42

visit. So are the clinicians, doctors, nurse practitioners, physician assistants, are they embracing the technology in general? Absolutely.

Speaker 03: 4:49

Absolutely. As you had talked about, we used to do diagnostics. documentation and conductor care through paper, right? And now so much of that has translated into the electronic record. And what that did, it shifted an immense amount of burden from just writing in paper to handing off that administrative burden to clinicians, to providers, to nurse practitioners, to doctors. And on average, for a visit prior to AI, it would take on average about 10 to 12 minutes per visit for a doctor or nurse practitioner to type up that note. Because they literally And they type it. Patients probably see it when they're visiting with the provider. Oftentimes, the provider is there typing away, looking into the computer. Not at you. Not at you. You know, missing that really important dialogue, that face-to-face dialogue. I hear it in my friend networks. It seems like I went to the doctor and they didn't pay attention to me. And with AI, the provider, the clinician, can refocus their attention to the patient, knowing that in the background, all of that really fine, crucial clinical detail is being captured and put into the electronic health record. Ultimately, it's my responsibility. It's our responsibility as clinicians to authenticate and make sure that information is accurate. But the reduced burden and the reduced cognition and effort to remember that information is paramount. And there has not been any other technological solution in my experience as a clinical informaticist that there has been with this AI scribe.

Speaker 02: 6:20

Yeah, you know, and with apologies to all those who developed the electronic health record, I spend an extraordinary amount of my time typing. You know, I went to medical school. I passed a couple classes in science, but the most important class I took in high school was typing. Yeah. You know, I'm a good typer, and it's helped me out a And what a kind of a silly thing to have to say, but I spent a good portion of last Saturday morning typing from my dining room table, my clinic notes. And it took a great long time just to type them even as a good typer. And I don't think that that is probably good for your healthcare. So your clinician has two options. They can sit and type on the computer while you're in with them and then they're not looking at you, or they can do what I actually do choose to do. I tell myself, put your hands off that keyboard, take your hands off that keyboard, take your hands away from that mouse and look and talk to your patient. But the price I pay for that, It's Saturday mornings I'm typing my notes because I didn't have time during the day. So it seems to have clear benefits. But what do you say to patients about your being recorded when you're at one of your most vulnerable moments?

Speaker 03: 7:23

Yeah, it's a great question. And transparency and being forthright is really important. And we do ask permission and we are forthright with informing patients that we are doing a recording. We're using AI to generate the note so I can spend more time focusing on you. And so we keep that level of transparency. And I wouldn't even say selling point, but putting this back into kind of going back to the earlier point that I think there's been a large contingent of patients who have felt like their visits have become a little bit less ingenuous because of the distractions because of the electronic health record. And so patients have been very willing to accept this technology. by and large. Sure, there are some examples in some cases that there's a level of trust or a reason that they'd rather not partake, and that's okay. We can always fall back to typing or doing traditional ways of doing our documentation. But we are forthright, again, with this technology and that it is present. And to be honest, AI solutions are only going to be more and more prevalent in healthcare because of the substantial benefits that it does promise, especially within a complex and highly regulated industry. Back

Speaker 02: 8:35

to that exam room.

Speaker 03: 9:03

consent process that I just spoke to. But basically at that point, I press a button on my phone and during the entire visit, that phone is recording. You just set it there on the desk. Just set it right there in between us. And we'll see if that technology changes in the future. It likely will. But for right now, it's just using your own personal device and app on the phone. Really slick, really easy to use. And that's the thing that's really enticing about it is because so many solutions and technology are hard to embrace or hard to deploy or hard to incorporate into clinical workflows. This is really, really easy. One, it's really easy to adopt and the impact is so substantial.

Speaker 02: 9:40

So this might come as a shock to you and others, but I'm a very verbose guy. In fact, I barely can, I never shut up. What does, how does this thing know? How does this software know what I want put in Jane Doe's medical record and how much I'm just shooting the breeze with it? Yeah, yeah. I shoot the breeze. I'm talking to them about their trip to Norway for the first 10 minutes. And does that have to all go in the chart?

Speaker 03: 10:03

It doesn't. And And that's the really impressive component of this is that it relies on past medical information and data to filter out some of that conversation. It's going to filter out some of where was that recent trip. But it may include something along the lines of patient is having food insecurity or there is a difficulty with something at home. It will incorporate that. And at the same time, it's crucial for the clinician to review one of those notes because maybe it did miss that for you. You wanted to include those details on that trip to Norway. You might have to just type that little bit in. Or maybe it includes a little bit of something that, you know, that didn't need So you might take that out of the note. So that's a balance. But to answer your question, it does a really good job of filtering out some of that chit-chat that is really important for that human connection between a patient and a provider, and also including the really important clinical information. Yeah,

Speaker 02: 10:59

and I'll be honest with you, Brian. I struggled with that when I first heard of this stuff because to me that knowing that my patient took a trip to Norway, I'm just continuing on that example, matters. It really does. I know what they like to do. I know that they like to go hiking or I don't know. I know that they like to go camping in the jungle where they pick up tick-borne diseases. So it is kind of relevant to getting to know people. And so I don't want to lose all of it. But then again, I don't need to write a Herman Melville novel every time in their note.

Speaker 03: 11:28

Absolutely. And that's the balance. It learns based off of past experiences and a huge amount of data. But it can't quite get to the point where it's going to be really specific and individual to that individual provider because we all practice medicine. We all practice advanced practice nursing a little bit differently. And so there are the needs to make those final curations to the note that meet your personal needs. So here's a big

Speaker 02: 11:54

question. Who benefits from AI in healthcare? The hospitals and clinics, the doctors and nurse practitioners and physician assistants, your clinicians, patients, who benefits? All of these. above in my opinion. So first

Speaker 03: 12:07

and foremost, this is driven towards providers. You know, when I look at our healthcare system, our clinicians, advanced practice providers, physicians, nursing staff, therapists, they are the bedrock to a functioning healthcare system. And right now, that's at risk. You know, in a recent study from the American Medical Association, I think this was a couple years ago, 20% of physicians are considering leaving practice in the next two years. Right? That's a scary situation. And in primary care, there's a

Speaker 02: 12:36

shortage of tens of thousands predicted for the next decade.

Speaker 03: 12:39

Right. And that also translates into some providers are needing to reduce their time spent in clinic because the complexities have just grown so immense. Right? So there's the provider benefit to this from a wellness perspective and being able to retain our clinical workforce. Right? And taking these administrative burdens off clinicians have already shown. We have some statistics and some of the work that we've done here at Hennepin on that space. From a patient perspective, we have already alluded to already that patients stand to benefit to have that more human, authentic connection between themselves and their clinician because the clinicians aren't feeling like they're struggling for survival, needing to type along that entire visit. And so patients have shown that they benefit from the eye-to-eye contact, that human listening, that full attentiveness through the visit. Studies have shown that. Many of my colleagues are reporting that as well. All of this translates into the organizations themselves because, again, organizations need to retain their staff right now. They need to retain their clinicians to be able to continue providing care to patients and the communities. There's a substantial benefit to the organization in investing in a tool that has this much impact straight away because there are so many things that we have tried along the lines of improving efficiencies and things for clinician wellness. Nothing really has compared to what we've seen with the AI Scribe solution. Before the break.

Speaker 02: 14:04

Last question is, are outcomes better? What portions of healthcare have improved in the few years it's been

Speaker 03: 14:10

being accepted? Yeah. And a lot of the data, a lot of the information really centers around that provider wellness experience. We have around 140 or 150 providers here at Hennepin using this solution right now. And we did a study not too long ago of the first 60 days, okay? Symptoms of burnout and feeling overwhelmed went down by 30%. Job satisfaction went up by 30%. Stress at work went down by 13% as well. Those are pretty substantial, staggering statistics. Within that as well, we were also able to pull information out of the electronic record to see the amount of time that providers are spending doing their clinical documentation. There was a 40% reduction in the time spent that providers are typing their notes. So each note is taking about five to six minutes less on average. The other component that we're looking at is over the course of the day, we've had some providers that are spending an hour less per day on documentation. That's five hours in a week. That's a lot of time that that provider is getting back to focus on other areas of clinical care are also on themselves so they can get home and not doing their documents. Maybe go see their kids. Go see their kids, not completing results on the sideline of a soccer game or spending time being able to attend to their parents because their parents are aging. All of those things. Because what you're talking about here is pajama time, right? Pajama time. What's pajama time? Pajama time is that time spent in the electronic health record in those overnight hours, typically starting at 7 p.m. and ending at 7 a.m. So we can measure

Speaker 02: 15:43

pajama time by that clinician that you're sitting with, that advanced practice provider, that doctor. We can measure how much time that person is spending off hours in the electronic health record. So that was a great discussion of what AI and what AI scribes can do. But now we're going to take a quick break. And when we return, Brian's going to cover the safety and privacy aspects of the technology. And I hope you'll address the million-dollar question, Brian, am I going to be obsolete? Is AI going to replace people in the future? Don't answer it yet, but that's where we're going to talk about after a short break. Stay with us. We'll be right back.

Speaker 00: 16:19

When Hennepin Healthcare says, we're here for life, they mean here for you, your life, and all that it brings. Hennepin Healthcare has a hospital, HCMC, a network of clinics in the metro area, and an integrative health clinic in downtown Minneapolis. They provide all of the primary and specialty care you'd expect to find, as well as services like acupuncture and chiropractic care. Learn more at hennepinhealthcare.org. Hennepin Healthcare is here for you and here for life. Thank you.

Speaker 02: 16:56

And we're back talking with Brian Imdicka about artificial intelligence and AI scribes in healthcare. Okay, so Brian, you've talked us through what it is, but it's natural to think, oh my goodness, somebody's listening to my personal stuff. What is your first response to people when they bring up questions like, oh, I don't trust this?

Speaker 03: 17:15

I validate that question and that concern because in this day and age of interconnectivity and how data is shared, whether it's within healthcare or outside of healthcare, we have to be very cognizant of that. Patient health information, protected health information is the cornerstone of something that we hold very highly dearly in healthcare and here at Hennepin. What we call PHI. PHI, right? The PHI, it's almost in the initial, anytime that we are working within the organization or outside the organization, it is fundamentally crucial that we protect that. And so when we were going down this path of considering an AI-scribed solution, that was paramount. We worked very closely with our legal team. We worked very closely with with our patient privacy team. We worked very closely with our compliance team to make sure that our relationship with our vendor made it sure that we respected PHI to the utmost degree. And those values were shared very much by our vendor as well. How information is shared to generate the clinical note is protected and encrypted at a very high level. It exceeds any regulatory or certifying bodies. And again, it comes down to transparency as well with patients that we conveyed that this is something that we are using. And that PHI, again, is something that we hold very, very dearly and is of the utmost importance.

Speaker 02: 18:33

But Brian, I'm coming to you. I'm telling you all about my personal problems and you've got your phone recording me. How do I know you're not going out to happy hour with your clinician friends and you're listening or you're talking about, let me play this back for you. Okay, I'm being cynical.

Speaker 03: 18:48

No, no, that's okay. That's the first time I've heard that exact question. And to be honest, I don't have immediate access to that recording. And So to answer that specific question, we as healthcare providers are held to the same degree of privacy security. We can only use a patient's health information or components of a visit for clinical purposes only. There is not to be used for sharing stories or is not to be used for personal gain. All of that is strictly prohibited and the same concept applies with the AI Scribe

Speaker 02: 19:22

solution. It really does. Your information is in there in the electronic health record as well, and that is protected 14 ways from Tuesday as well.

Speaker 03: 19:30

Yeah, this is ultimately not a new concept. It's just highlighted because of the recording aspect and that recording is ultimately shared. That's what maybe sets it apart from other solutions that we currently have.

Speaker 02: 19:43

So the million dollar question is, is this going to replace you and I? Is this going to replace the clinician? Is this going to replace all of us? Are people just going to get all their health care from a machine?

Speaker 03: 19:54

Short answer to that, in my personal opinion, is no, right? And I feel, and I think that is really established not only by myself, but others who are very knowledgeable in this field, is what this solution has promoted is that reconnection that human-to-human connection that has been lost by a lot of the distractions and the burdens and the obstacles in the healthcare delivery system. In this particular circumstance, clinical documentation takes a lot of time. Reducing that and that time is reinvested into that human-to-human connection that is backed by that clinical expertise. And so I don't ever see AI Yeah,

Speaker 02: 20:32

and that I think is reassuring and actually does make some sense if we could have AI replace some of these documentation tasks. I have also heard, though, that when you run diagnosis through AI, it's better than I am. And so there must be more applications. In other words, you put in your 12 symptoms and ask a whole bunch of guys like me and then ask AI, AI might be right. I mean, a computer beat the world's best chess player. It's probably true that it's going to beat the world's best diagnostician too. So there must be other roles in which it is helpful. So to round out this episode, could you sort of talk to us about the future? I get it that you can't predict it any better than any one of us, but what do you see? What are some of the cool advances You

Speaker 03: 21:17

know, there's a short-term and a long-term answer to that question. I think short-term, there are going to be the things along the lines of how AI will help suggest orders and diagnostics for a particular visit. I talked a little bit earlier about how the EHR is full of thousands of data points for an individual patient. AI can help summarize those, so we're making sure that we're capturing all the clinical information that could be lost otherwise, and so there is the ability to to do chart summarizations. There's also a future in where this is not just for nurse practitioners, physicians, assistants, doctors. I see this being a solution very soon for nursing. I think that's really important. Our nursing colleagues are facing similar burdens and attrition that physicians and advanced practice providers are. So on a short-term basis, I see this broadening. This is going to become more and more ubiquitous in care. It's going to be part of the healthcare delivery. I think in the what is that information and how can we capture patients' experiences that are happening outside of the conversation between the provider and the patient? How are we capturing impact of medications and taking medications? What are the socioeconomic factors that are influencing a patient's care? How is a day-to-day lived experience from a patient being potentially captured with AI in a collaborative and with all the safeguards that we've already talked about? So we have So I really

Speaker 02: 22:52

like what you said there, especially when you talked about nurses and others. Because, folks, you really don't want your nurse sitting at a computer documenting what he, she, or they do all day long. You would rather them be doing the high-level skills that they're trained to do. And so it's all over healthcare. Your therapist, your physical therapist, your OT, all of the aspects of healthcare, it could be more than just you and I in our lives. So I really like what you said about that. So Brian, to close us off. What excites you the most, or maybe what worries you the most, about AI in healthcare in the future?

Speaker 03: 23:26

Yeah. I think what excites me most with AI and what we've experienced in these early stages is that substantial reduction in the burdens that clinicians face and the ability to focus more on patients, to focus more on themselves, to find more fulfillment in practice, and to have that better human-to-human connection with patients. That's why all of us went into medicine. That's why all of us went into is to help people. And again, there's been so many obstacles to that human connection that AI really has an opportunity to help get us back to that human connection that is lost beyond just in healthcare right now in my mind. So that's what really excites me. I think there's so much more potential down the road and this will change very quickly. I think the thing that scares me most is that if we rest on our laurels and say, yep, we have AI scribes and we can There is so much more that we need to do to fix our antiquated healthcare system right now. And just to reduce clinician documentation burden and not focus on the larger picture, we have to keep going. We have to look at new ways that we interact with patients. We can't continue to go down the current ways that we're delivering care. And so how do we leverage technology? How do we leverage the people who are doing the work and the processes to ultimately make this a better experience for everyone, patients and providers? That's brilliant. and Brian, if there's anything AI could do, what would you wish for? Getting more of a holistic view of what is impacting our patients' health and what our patients' lives are, right? Again, so much is focused on those 15 to 20 minutes that we have with patients. We can't crack the surface. How can AI broaden our understanding of patients' lived experiences while also leveraging the improved connection that patients have with their providers? And I think that that is what I hope to see in Brian, brilliant

Speaker 02: 25:24

advice. And listeners, you're going to be hearing, I guarantee you're going to be hearing more about AI throughout your life, but in healthcare as well. And so I hope you got a glimpse of what that technology might look like next time you go into your clinic visit. Brian Imdik is probably the biggest expert that I know in my personal and professional life on the subject. So Brian, thank you for being here. It's my pleasure. Thank you for having me. It's been great hearing your wise advice. on a topic that affects us all. Listeners, I hope you've enjoyed this episode and I hope you'll join us for the next episode when we will talk about chronic obstructive pulmonary disease with pulmonologist Dr. Caroline Davis. It's going to be a great show. In the meantime, be healthy and be well.

Speaker 01: 26:06

Thanks for listening to the Healthy Matters Podcast with Dr. David Hilden. To find out more about the Healthy Matters Podcast or browse the archive, visit healthymatters.org. Got a question or a comment for the show? Email us at healthymatters at hcmed.org or call 612-873-TALK. There's also a link in the show notes. The Healthy Matters podcast is made possible by Hennepin Healthcare in Minneapolis, Minnesota, and engineered and produced by John Lucas at Highball. Executive producers are Jonathan Comito and Christine Hill. Please remember, we can only give general medical advice during this program, and every case is unique. We urge you to consult with your physician if you have a more serious or pressing health concern. Until next time, be healthy and be well.