Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes
This article is sponsored by the Dragon Ambient eXperience at Nuance. The exam of the future has arrived with clinical documentation that writes itself.
Recently, I participated in a virtual webinar, “How AI technologies improve the patient experience,” along with Dr. Jesse Affonso, orthopedic surgeon at Cape Cod Orthopedics. We discussed the difference between patient experience and patient satisfaction, the correlation between the two, and the role technology plays.
In a nutshell, I think it’s this: A poor patient experience can lead to low patient satisfaction. A good patient experience can lead to high patient satisfaction. According to the Agency of Healthcare Research and Quality, the patient experience encompasses the range of interactions patients have with the health care system and includes several aspects of health care delivery such as timely appointments, easy access to information, and good communication with health care providers. At the same time, patient satisfaction is about whether a patient’s expectations about a health encounter were met.
This all begs the question: How can one improve the patient experience and satisfaction?
I think we can all agree that a good patient experience includes obtaining a timely appointment, friendly and professional interactions with staff, and good communication and trust in their provider. But we know that this isn’t always the case. It takes too long to get an appointment, the staff is hurried, and providers are looking at their computers more than their patients.
According to the American Medical Association, more than half of physicians are burned out and spend two hours on administrative work for every hour of direct patient care. This directly impacts the patient experience as physicians are scrambling to get to their next patient, document care in between patients, or document after hours. Patients are sitting in waiting rooms, and then when they are seen, the physician is often talking and typing at the same time, unable to focus on them and hindering communication and trust which is not ideal for patients or providers.
With the right technologies, physicians can deliver high-quality health care experiences, and patients can get the health care experience they need and deserve.
That’s what Dr. Affonso did when he started using the Nuance Dragon Ambient eXperience (DAX) over a year ago. With Nuance DAX, he’s able to see 3 to 4 more patients in a day, spend more time with his family, and improve the quality of his documentation. He doesn’t even bring a computer to the exam room. He brings up the DAX mobile app on his cell phone, taps the record button, and automatically captures the conversation.
I have noticed a huge improvement both in the quality of my life and in the patient’s experience. I no longer must bring a computer into the room with me, and I am no longer scrambling to get out of the room to start frantically typing to document the note. I believe I am more effectively meeting the patient’s expectations regarding their overall experience and satisfaction levels.
That conversation is then converted into a clinical note and placed in the EHR for his review and signature. AI-powered DAX is smart and knows what information to include in a note and what not to include. It understands English spoken with any accent and is highly trained in all ambulatory specialties, including primary care and urgent care, and can be used for both in-office and telehealth visits.
The art of medicine is such a vital part of medicine. The science is the clerical part. I think this technology really allows you to focus more on the art of medicine. And the art of medicine is to develop strong patient relationships and to tease out challenging or difficult issues. DAX has eliminated the clerical burden that had been competing with my time with the patient and allowed me to be present and engaged during the visit.
Patients want the best care possible. And physicians want to provide the best care possible. The fusion of art and medicine can deliver a better patient experience with AI-powered, voice-enabled documentation that’s revolutionizing health care.
Jared Pelo is an emergency physician and chief medical information officer, Dragon Ambient eXperience at Nuance. Jesse Affonso is an orthopedic surgeon.
Image credit: Nuance
Transcript
Kevin Pho: Hi, and welcome to the show, where we share the stories of the many who intersect with our health care system but are rarely heard from. My name is Kevin Pho, founder and editor of KevinMD. Today we have a special sponsored episode. Jared Pelo is an emergency medicine physician and innovator. Dr. Pelo helps lead the ambient clinical intelligence team at Nuance, where he is chief medical information officer. He’s here to talk about the Dragon Ambient eXperience at Nuance. The exam of the future has arrived, with clinical documentation that writes itself. Jared, welcome to the show.
Jared Pelo: Thanks so much, Kevin. I’m excited to be here. I’m a big fan; I’ve been reading the blog forever, so I’m excited to talk with you.
Kevin Pho: Well, thank you so much for your readership and support. First off, I’m just interested in hearing your story and your journey to where you are today.
Jared Pelo: Yeah, so for me, it really started in 2012. I’d been practicing for a few years after residency. I did my residency at the University of Virginia, and I was practicing in Lynchburg, Virginia, in a busy emergency department, 100,000 patients a year. I grabbed my partner during a break, and I said, “Hey, Andy, you seem like the happiest guy in our group. How do you stay so happy?” And he said, “Listen, Jared, every day I have off, I dread coming back to work.” To me, that was a little bit of a punch in the gut, the worst pep talk I’ve ever heard. He followed it up, though, with a couple of nuggets. He said, “You know, really, the people who seem the happiest are those who either go into administration or those who innovate and try to make health care better.”
So I tried exactly what he told me to do. I became a medical director for a year, and I am not the hospital administrator or medical director type. That’s hard. I tip my hat to everybody who does that day to day; that’s hard work. But during that period, I was driving about an hour to go to a satellite hospital where I was a medical director. We had a scribe program at the mothership, but not at that place an hour away. Around the same time, Google Glass was just becoming generally available for innovators. So I saw that, and I thought, “Why can’t I just wear Google Glass? My scribes back in Lynchburg can follow along.” I started grabbing one of my scribes from Lynchburg and taking them with me on these drives to this smaller hospital, and we started talking about this. We started talking about how, really, in the future, if we just capture all this data, someday computers will be able to do this for us.
So in 2014, we started a virtual scribe company, and in 2017, we were acquired by Nuance. We became part of Nuance because their vision was completely aligned with ours. And then, a little bit over a year ago now, the Dragon Ambient eXperience became available, where you do record your visits and then AI writes your note for you. So that’s how I ended up where I’m at.
Kevin Pho: I talk to a lot of physicians who are interested in getting involved in that innovation space. If you were to advise these physicians, what would be your number one actionable tip?
Jared Pelo: It would be to get involved. I mean, when you look at health care and all the problems in health care, and you talk to other providers, we have a lot of the answers. We know what we want. When I first started the first company, I had three other physicians who came on board as co-founders, but they were too scared to cut back their shifts to really commit and do the work that needed to happen. You really have to make a sacrifice and say, “I’m committed to actually trying my hand at this.” There’s no guarantee that you’re going to be successful when you try to innovate in health care, but if you want to make a difference, you really have to dig in. You can do it on the health care system side. I think that’s safer than starting a startup, though I would not say that’s a sure bet at all. But that’s how I decided to do it. I really just think it’s making that decision, “I want to do this,” and jumping in and trying it.
Kevin Pho: Now, did you have a business background prior to your career in medicine?
Jared Pelo: No business background, actually. I went to a startup accelerator. That’s why I was in Lynchburg, Virginia, and I ended up in Durham, North Carolina, because Lynchburg, Virginia, is not the tech mecca of the East. But the Triangle, Research Triangle Park, the Triangle down here in Raleigh-Durham, has a lot of great tech. So I came down here and joined an accelerator that helped me get things off the ground and find my co-founders, all those types of things. So yeah, if you’re going to start a startup, I would definitely say find the right mentors, find the right people to help you get off the ground.
Kevin Pho: Now, the COVID pandemic has really accelerated the involvement of artificial intelligence in health care. In your opinion, how has AI evolved over the last decade?
Jared Pelo: It’s almost a question of how it’s evolved over the last few years. I mean, AI 10 years ago, let’s just take Siri, for example. When we first started dealing with Siri, she was kind of fun, and she had some rules-based jokes she might tell you, but you wouldn’t really call her intelligent as a virtual assistant. I remember when we first became part of Nuance, talking to researchers about what we wanted them to build with our data, and they would say, “Oh, so you want the Harry Potter experience. You want the quill floating in the air next to you, writing a summary for you.” I was like, “Yeah, that’s exactly what I want.” When we said that, they were like, “You’re crazy. That’s impossible,” or they said it’s probably seven or eight years away. And now, three years later, we’re already at the point where we’re writing a majority of the clinical documentation just with AI.
So I just think sometimes we forget about the exponential growth of technology and how much smarter things get and how much faster. I really think we’re at that tipping point where we’re going to see things go from concepts, for so long, to becoming the real thing. I mean, think about streaming video back in 2000. Everybody thought it was a great idea, but there wasn’t bandwidth, there wasn’t technology, it wasn’t going to work. Now almost none of us have cable anymore; we all stream our video. So I think we’re just getting to that tipping point where we’re going to see AI go from concepts and being used minimally to being used widely and really driving a lot of what we do, not just in health care but everywhere.
Kevin Pho: Now, I’m looking forward to seeing how AI can improve the lives of physicians. During the pandemic, one of the themes on my blog and the podcast, of course, is physician burnout. Surveys report almost half of physicians indicate that they are burned out, and of course, that number has only grown during the pandemic. Where should leaders look to leverage automation, and more specifically conversational AI tools, to limit this administrative burden on clinicians?
Jared Pelo: Yeah, I mean, when you look at physician burnout, there are lots of different reasons around it. Of course, administration is a big one. Loss of autonomy is another one. ZDoggMD talks about the moral wound. I think it’s all about how you fix the problems so that you can connect with the people you’re caring for, and what technologies, if you’re going to use AI to fix that. How do you implement an AI program that actually helps providers and patients connect? I think that at the heart of it, physicians, health care providers across the board, chose this profession because we felt called to it in one way or another. Either we love the science, or we love to help people, we love to connect with people. And then we got into it, and the reason we feel burned out is that it’s not fulfilling that need that we entered health care for.
So if we can use technologies like DAX to go ahead and offload a bunch of my clerical work, so I can actually look my patient in the eyes when they’re telling me stuff, and I can look for nonverbal cues because I’m not so busy staring at the computer typing, I can actually be a much better health caregiver. I think that’s huge. And we know that if you’re experiencing burnout, you’re twice as likely to have a medical error in the next three months. So there are so many good reasons to try to fix the burnout problem.
Kevin Pho: Let’s go ahead and talk about the Dragon Ambient eXperience, or DAX. How can ambient technology help bring back the human element of patient care to improve the patient-provider relationship?
Jared Pelo: Yeah, so the way DAX works is really just taking that conversation that you would have with your patient and doing all the documentation and clerical work for you. I like to say it’s the provider-patient relationship of the 1970s with the technology of the 2020s. We want discrete data. We want better data to take care of our patients, absolutely. But we also want to connect with our patients and have a great relationship with our patients. So I think that with DAX, the way we really tackle this problem is: Let’s solve the number one pain point for physicians, which is being stuck looking at a computer instead of my patient, and then start planning toward all those other things you’re going to be able to do with structured data and automated data capturing, all of the downstream things you can think of where AI can be implemented.
Kevin Pho: So I’m an internal medicine primary care physician, and I was wondering if we could talk in detail about a typical workflow. If I were to use DAX and I walk into the exam room, walk me through the step-by-step process of how this product would work.
Jared Pelo: Yeah, so what you do is you pull out your iPhone. You see a schedule of your patients that day, and you pick that patient. If you do pre-charting or chart review before you enter the room, you could say a few things: “Patient’s following up; hemoglobin A1C came back high, so the office called and asked her to come in.” Go ahead and enter the room and have a normal conversation. All we’re going to ask is that you be explicit around certain things. For the physical exam, while you’re doing the physical exam, just call out your findings. If you train DAX that you have templates, DAX is going to go ahead and pull those templates in, and you can just call out the abnormals: “Rales at the bases, bilateral lung fields. Trace pedal edema, bilateral lower extremities.” Whatever you’re finding on your physical exam, call those things out, and DAX is going to switch out the normals for the abnormals. When you get to your assessment and plan, just go through it and explain it to the patient: “For your diabetes, we’re going to increase your Lantus to 15 units. For your hypertension, we’re going to keep you on lisinopril. For your depression, we’re going to increase your Zoloft.” If you go ahead and explain things like that, when you leave the room, DAX is done.
If there’s anything you didn’t want to say in front of the patient, “Patient’s showing malingering behaviors,” or whatever you might not want to say in front of the patient, and in this day and age with open notes, it doesn’t really matter, but if there’s anything that you didn’t want to say in front of the patient, you can say it right when you leave the room. As soon as you hit stop, that complete recording is put through the AI, and we parse out the speakers, so we get clean transcripts of everybody who’s speaking. Then it’s put through our natural language generation model, and it writes the note for you, separates it into sections, pulls out discrete data points, and can do all of the things that you would normally do.
We still don’t think it’s quite ready to go directly to a physician yet, so we do put it through a quality review. We have somebody trained up in internal medicine. They listen to the whole recording, read over that note, make any corrections or edits that you might need, and then they put it in the EHR for you within a couple of hours. Our contracted time is four hours, but usually within a couple of hours, that note is back in your EHR. It’s there; you read it and sign off on it. What we found is that we save most doctors about six minutes per patient, so if you see 20 patients in a day, that would save you two hours of documentation time.
Kevin Pho: So if I’m having a conversation with a patient for the history of present illness, for instance, DAX would transcribe that. What would that look like in the final note?
Jared Pelo: That’s a great question, and I think the best way to really talk about it is to show you. We have these primary care models where we’ve trained DAX from physicians meeting with their patients. So what I want to show you, I’m going to share my screen with you. I’ve pulled it into a Word document, so it’s not in our interface; I can’t show you that. But I’m going to show you a note that was written completely by AI. No human has touched this yet, other than me copying and pasting it over to Word. I’m not going to talk a whole lot; I’m going to let you read through it. But this was, you know, I told somebody, “You just went to the ER, you have these medical problems,” and we had just a totally normal conversation, and this is the output.
You can see where it’s calling out some past medical history in the HPI. It talks about going to the emergency department because you had chest pain while doing yard work. They said it was not a heart attack, no blood clots in their lungs. If you were to listen to the conversation, it’s completely normal. In the conversation, he said, “Well, my wife made me go,” and that doesn’t necessarily make it into the note, because that’s not necessarily medically relevant, although it does bring in the patient’s grandfather, because I asked about family history; he had heart disease. It went ahead and pulled in a review of systems from the conversation. I told it to use my normal physical exam template, which is a big, complete physical exam template, so it pulled in this whole physical exam. And then down at the bottom, it went ahead and wrote my assessment and plan for me.
We’re getting to a point where I’m getting very excited, because of the point I want us to get to, right? Right now we have a quality reviewer, and it takes a couple of hours, and that’s great. But it gets me excited about the point where I get that note within seconds, or I’m talking to my patient and there’s intelligence coming back to me, if I set up certain alerts. I really want to give the control to the provider. I’m kind of tired of administrators and health care organizations, or even technology companies, deciding when to alert a provider. I want a provider to be able to choose: “Hey, if a patient tells me everything that’s needed for a CHA2DS2-VASc score, alert me and tell me the score.” I think it would be really cool when we can deliver that kind of intelligence to a provider and make it automated, or even just at their fingertips. It gets me really, really jazzed.
Kevin Pho: So to be clear, is this the final note that goes to the physician?
Jared Pelo: Oh, thanks for asking that. I do want to be clear: This note would not be seen by a physician today. This note would go to a quality documentation specialist. They would listen to the whole encounter, fix everything in the note that might be an error, and then it would be placed in the EHR for you to sign off on. So yes, this is not the final note.
Kevin Pho: Now tell me how DAX would differ from, say, a medical scribe. There are several options, where we have to actually have a scribe in the room with you. There are now virtual medical scribes, where a live person on the other end would transcribe the note. Now, how would DAX contrast with those other options?
Jared Pelo: I mean, the biggest thing is you’ve got technology that does the bulk of the work, and that means several things. It means that it’s going to get smarter and smarter. If you’ve ever worked with scribes, they’re great. I trained at the University of Virginia, one of the first sites where scribes were really big. I had scribes as a resident. But I’d say there are different levels of quality per scribe. They come and go. They’re always going to med school, and you have to retrain a scribe, and most of that training is on you. They call in sick, they go on vacation, they go on lunch break. DAX doesn’t do any of that. DAX just gets smarter and smarter and smarter, and it’s just a push of a button, and you go ahead and document any time of day. So to me, that’s it: It’s scalability, it’s technology, it’s more intelligence, it’s the roadmap, the vision. All that stuff just gets me. Scribes are great, and there’s so much pain out there, I would never say don’t get a scribe. But I’d say if you have the opportunity to try out DAX, try it out, because it’s the future.
Kevin Pho: Now, one of the other effects of the pandemic is the rise of virtual care. I was wondering if you could share your opinion in terms of the role of AI when it comes to virtual care encounters.
Jared Pelo: Yeah, I think one interesting thing to me is that when the pandemic first started, many people assumed virtual care was going to be easier for the clinicians. But now you don’t have an MA who’s doing work for you or a nurse who’s doing work for you, some of your documentation. You’re actually doing the intake of the patient; you’re doing all of it, just with the patient in front of you. And then you’ve got to document that whole conversation, and you can either be looking down at your screen or looking at the patient. It can be even worse for a patient to have their provider typing right in front of them than having the provider turning their back to them. So I really think that with telemedicine and telehealth, having a physician clearly focused on you while still using AI to help document is the way to go.
Kevin Pho: And to be clear, DAX has a pretty robust option when it comes to virtual encounters.
Jared Pelo: Yeah, really. It’s able to document on any telehealth or televideo platform. You can go ahead and use DAX, and it’s going to easily capture the audio and process it.
Kevin Pho: Now, what do you see as the future opportunities for automation tools like DAX going forward?
Jared Pelo: I mean, I spoke a little bit about it when I talked about automated calculators coming back to us. I think Nuance has a lot of exciting partnerships with EHRs. We have “Hey Epic,” “Hey Cerner,” and “Hey Meditech.” This last week, we announced integration with Athena’s mobile app, so there’s the ability to use that virtual assistant to pull up old notes, place orders, pull up labs, all that stuff. I think the more we make it so that we save clicks, the more powerful the technology is. So that’s kind of the short term. And then I think longer term, we’ll just see more and more technology enabling the provider with the right data at the right time, because usually it’s not that providers don’t have the data. It’s hard to get to, and there’s too much of it to actually find what’s useful at times. So I think it’s AI needing to help us find the right data at the right time.
Kevin Pho: Now, this of course sounds great to primary care doctors like myself, having an AI tool passively create a note that comes back in a few hours. But when it comes to implementing these tools with the administrators who make those financial decisions, certainly there’s a concern about return on investment. So the next question: When do you think it’s reasonable to expect health care organizations to see a return on investment on these types of AI investments?
Jared Pelo: So there are all sorts of different measures you can look at when you’re looking at ROI. I mean, we know that burnout on its own is a multibillion-dollar problem each year. We know that when health care organizations lose a provider, it’s about $500,000 to replace them. So there are good reasons to make your physicians happy and to keep them working at your institution. Of course, there’s increased productivity, increased throughput, and faster charting, so you can close those charges faster. Really, within a month or two, you should be seeing DAX pay for itself. And there are multiple different ways to look at how it’s paying for itself. One of the cool things that we love to do when we bring somebody on is pick out the measures they want to look at and start measuring them from the beginning. One thing we’ve seen, just talking about burnout: We target providers who are burned out, so our pre-go-live number is 72 percent who are experiencing burnout, but within a month, that’s down to 17 percent, which is a pretty amazing decrease. So yeah, there are lots of good ways to find that return on investment.
Kevin Pho: We’re talking to Jared Pelo. He’s an emergency physician and innovator. Dr. Pelo helps lead the ambient clinical intelligence team at Nuance, where he is chief medical information officer. He’s here to talk about the Dragon Ambient eXperience at Nuance. The exam of the future has arrived, with clinical documentation that writes itself. Jared, I was wondering if you could share a real-life example or a case study of the difference the Dragon Ambient eXperience made for a physician or a practice.
Jared Pelo: It’s hard for me, honestly, to pick which one to share, because I have so many. We have the physician whose wife sent us a valentine saying, “Thank you for saving my marriage.” We have the physician who said, “I was going to retire in six months; now I’m going to practice for five more years.” I love those ones. We have the provider who wrote us a note and said, “Hey, I missed that the patient told me they have a history of breast cancer, but you documented it, and that triggered an MRI, which found a cancer.” So there are lots of those kinds of soft studies for us. But one that’s more of a hard study that I really like is a cardiologist working in a critical access area who was able to increase his capacity by 24 percent. Those patients who were waiting months and months to get in to see him were now only waiting weeks to get in to see him. He was still going home earlier than he’d ever gone home and eating lunch, all of the benefits that we kind of promised. So there are lots and lots of great case studies to share.
Kevin Pho: Now, of course, I have a primary care bias as a primary care physician, but I was wondering if you could give examples of how DAX can be used across specialties. You mentioned cardiology and oncology, but give some examples across specialties, and maybe some of the procedural specialties that may benefit from DAX.
Jared Pelo: Yeah, when you look at ROI in the procedural specialties, it’s pretty easy, because the more patients you can bring in, the more you can take care of, the more you can take to the operating room if they have something that needs to be operative, and the better you can compete against your competition if there are several of you in the area. DAX currently works for every ambulatory specialty other than behavioral health. We currently do 32 different specialties, really practically every ambulatory specialty out there. Next year, we’re going to start going into the hospital, which I think is really exciting. Procedural specialties work really well with DAX.
Kevin Pho: Now tell me about the future of DAX. You mentioned where you want DAX to go, but where are we with the technology itself now, and contrast that to where, ideally, you want it to be going forward.
Jared Pelo: Yeah, so I think my goal is that nine months from now, we’re allowing providers to see those draft notes and giving them the opportunity to choose whether they have a quality review check or whether they just go ahead and check it themselves. We also have a really great ambient device that you can put in your room, and that allows you to use those virtual assistant skills. But I see that ambient device becoming part of the care team. I can imagine a day when your MA brings somebody into that room, introduces you to DAX, and DAX finishes up the medical intake or asks all the questions for the annual wellness visit that haven’t been captured yet. So I really think DAX becomes part of the care team. I think five years from now, physicians will mostly refuse to practice without some sort of ambient intelligence aiding them in their documentation and other tasks.
Kevin Pho: And my final question, Jared: What’s your take-home message that you want to leave with the KevinMD audience?
Jared Pelo: I really want people to understand that the future is now. There are things out there that will solve your major pain points, that will solve your problems. Even if it’s not DAX, it’s worth going out into the market and finding things that make your life better, that make your patients’ lives better, and those technologies are around. So put a little effort into finding what will work for you. If you want to try DAX, definitely reach out. As you can tell, I can talk about DAX all day. I’d love to talk to people, and your listeners are welcome to reach out and contact me.
Kevin Pho: And for those interested in finding out more about DAX, can you share a website or contact information?
Jared Pelo: Yeah, you can really just go to the Nuance website, and we have an ambient clinical intelligence piece in there. Anybody who wants to email me directly is welcome to, and my email is just my name at Nuance, so Jared Pelo at Nuance.
Kevin Pho: Jared, thank you so much for sharing your time and insight, and thanks again for being on the show.
Jared Pelo: Thanks, Kevin.
Listen and followApple PodcastsSpotifyYouTubeMore apps · Browse all episodes
