Subscribe to The Podcast by KevinMD. Watch on YouTube. Catch up on old episodes!

We sit down with Dike Drummond, a family practice physician, to explore how AI-powered note-writing software is revolutionizing medical documentation. While these tools offer short-term relief to overworked physicians by drastically reducing charting time, could they ultimately lead to increased patient loads and a new form of burnout?

Dike Drummond is a Mayo-trained family practice physician, burnout survivor, executive coach, consultant, and founder of TheHappyMD.com.

He discusses the KevinMD article, “Stop physician burnout: the hidden danger of AI note-writing software.”

Our presenting sponsor is DAX Copilot by Microsoft.

Do you spend more time on administrative tasks like clinical documentation than you do with patients? You’re not alone. Clinicians report spending up to two hours on administrative tasks for each hour of patient care. Microsoft is committed to helping clinicians restore the balance with DAX Copilot, an AI-powered, voice-enabled solution that automates clinical documentation and workflows.

70 percent of physicians who use DAX Copilot say it improves their work-life balance while reducing feelings of burnout and fatigue. Patients love it too! 93 percent of patients say their physician is more personable and conversational, and 75 percent of physicians say it improves patient experiences.

Help restore your work-life balance with DAX Copilot, your AI assistant for automated clinical documentation and workflows.

VISIT SPONSOR → https://aka.ms/kevinmd

SUBSCRIBE TO THE PODCAST → https://kevinmd.com/podcast

RECOMMENDED BY KEVINMD → https://kevinmd.com/recommended

GET CME FOR THIS EPISODE → https://kevinmd.com/cme

I’m partnering with Learner+ to offer clinicians access to an AI-powered reflective portfolio that rewards CME/CE credits from meaningful reflections. Find out more: https://kevinmd.com/learnerplus

Transcript

Kevin Pho: Hi, and welcome to the show. Subscribe at KevinMD.com/podcast. Today we welcome back Dike Drummond, family practice physician, executive coach, founder of TheHappyMD.com. Today we’ll talk about the hidden danger of AI note-writing software. Dike, how are you?

Dike Drummond: Good, really good. How are you?

Kevin Pho: I’m doing well. So we’ve had that intersection between AI and how it can improve physicians’ lives with these ambient scribes, and I wanted to of course hear another perspective on that. So tell us what led you to write this article, and tell us what it’s about.

Dike Drummond: Well, I think there’s a whole bunch of buzz around AI in all sorts of different industries. Everybody’s on the bandwagon, all the stock pundits are getting you to buy the AI stocks and everything. And we’re talking about ChatGPT style large language models of text generation, and what they have the ability to do is mimic human language. And they do it a word at a time, based on, if that was the first word, what’s the most likely second word based upon all the conversations that we’ve analyzed. And if you train it to be a doctor, it can act like a doctor.

First of all, I don’t think a lot of people understand how the software is trained, so let me tell you how the software is trained. A lot of people don’t know what it’s like to have a workflow of seeing a patient in a visit while you’re being recorded by the AI, how you have to change your behavior.

And then I will tell you right now, this is an absolutely unbelievable breakthrough in the ability for you to get home sooner without working harder and have the AI write 98 percent of your note. However, the benefit for that will be only temporary. It’ll be much shorter than you might think, and I am very concerned that what it will do is end up making everything much worse. And the way it will be worse has nothing to do with documentation. It will take off the documentation damper on productivity and expose us to other types of burnout we can’t anticipate until this is widely adopted.

Kevin Pho: So tell us, what do you foresee can happen because of this?

Dike Drummond: Well, it’ll give us a different reason to burn out. Right now, keystrokes, pajama time, the thing that stops you from seeing an extra patient is because of the documentation that’s associated with it. But if I take away that damper, what’s going to happen next?

Let me just ask you this. If I can show you a way to cut your time in the chart by 50 percent, and if your employer, and I’m talking to employee physicians right now, which is 74 percent of American doctors, if I can take the time in the chart down to 50 percent of what it is now, how long do you think it will take your employer to insist that you see more patients?

Kevin Pho: 24 hours? No, I’m just kidding.

Dike Drummond: Actually, when I’ve asked this question to big audiences of doctors, the average answer, and it usually comes in unison, is five minutes. Five minutes. And if they cut your time in the chart 50 percent, how many more patients are they going to ask you to see?

Now let me just do some math here. The AI softwares, and there are two dozen of them, and right now the people who created these software companies called this the land grab phase of this market, everybody’s trying to lock up as many groups as possible with their software. But the softwares are all almost identical, and each of them have an average per user cost per month of about $100. About $100.

And let me ask you this. If you could see one more patient, and I’m a family doc, so my patients don’t kick in a lot of cash to the cash flow, but if I see one more patient, if you’re a primary care doctor and you see one more patient, about how much money do you think that will add to the top-line revenue of the organization?

Kevin Pho: Probably about a hundred bucks.

Dike Drummond: OK, so if they install this software and it cuts my time in the chart by 50 percent, can I see one more patient a month? Can I see one more patient a month? How many are they going to ask me to see?

Kevin Pho: Probably 10 more a day.

Dike Drummond: But the thing is, if I can just see two more patients a month, remember the software costs 100 bucks, the patients give me 200, I’ve got a 200 percent return on my investment from just seeing two patients a month. This is a chief financial officer’s bonanza.

The return on investment for this, if you see three extra patients a day for 20 days in a month, and now we’ve got $3,000 for a $100 investment, they’ve never made such a good investment in their entire life. They’re going to be absolutely head over heels willing to install this as fast as possible, and then they’re going to crank up your patient volumes, because every dollar of this is dollars that they weren’t realizing previously. It’s like bottom line profit, profit on profit.

So if I unleash you from seeing, let’s say, 22 patients a day, and I want you to see 30, and the documentation is not the problem, the documentation isn’t the thing that overloads you, task overloads you, what will be the next barrier to your health and well-being that we run into? It’s going to be a brain limitation, not a typing limitation. It’s going to be some sort of fatigue of some cognitive ability, and we already are aware of several of these, right? We already are aware of decision fatigue, we’re already aware of compassion fatigue.

How many times can you take a complicated history and deal with a complicated patient, bam, bam, if there’s no delay that’s mediated by your need to do the medical records? What will it take? And I have a fear, because I deal mostly, about 80 percent of my work is remedial for folks with burnout or disrupted behavior, the rest is performance coaching, but I know enough about what can break people to know that we don’t know what comes next, because everything’s been choked off by the need to document. It’s never allowed us to go fast enough to hit the cognitive barriers we’re going to run into when this is fully implemented.

Kevin Pho: Now, you’re of course a burnout expert, and you’ve talked to a lot of hospital systems, and physician burnout has been under radar. We talk about this multiple times on this podcast and on KevinMD, and you give seminars and coaching and you talk to hospital systems about burnout all the time. So don’t you think that these systems would have burnout at their forefront, and just by increasing the volume, increasing the profit, that’s going to defeat the whole purpose of these AIs? Do you think they have that type of insight and empathy into physicians, or are they about the bottom line?

Dike Drummond: Oh, well, there’s a huge polarity in the center of the business of medicine. OK, there’s two sides of the house. There’s the physician side, the worker side, the side that does the clinical encounters that produce the income for the organization, and then there’s the business side of the house.

Now the way it works is, imagine a yin and yang symbol. They’re connected like a yin and yang. You can’t separate one from the other. But what you always hear is, no margin, no mission. I tell you what, no mission, no margin. It works both ways. But nowadays, especially in America, especially if your company happens to be owned by private equity, the financial considerations are ruling the considerations of the industry.

So another way to express this dichotomy, this polarity at the heart of it, is suits versus scrubs. And I’m using the word versus. I’m sorry, I apologize, suits and scrubs. It’s not suits or scrubs, it’s not margin or mission, it’s margin and mission. But can you slow down the business side of the house to allow this transformation to occur?

If you think about Friedman economics, the purpose of a business is to give maximum return to the shareholders. And what I want to say is, not in health care. That should not be the purpose of a health care business. A health care business should not treat maximizing the profit as the sole reason that the business exists. A health care business, because we’re in the business of delivering health to a population, a health care business should have profit as a constraint. You have to have a positive margin or the business ceases to exist. With a positive margin, we can still do our life-saving work in the front lines. But maximizing that profit, we know what happens, it’s happening right now, it breaks the doctors and nurses and staff in the front lines. It breaks them.

So right now I’m the only person I’ve heard have any reservations about implementing these softwares. I’m the only person I’ve heard express a fear that we’re going to run into a different cognitive barrier to physicians seeing more patients, and it’s going to be even more destructive than just simply being overloaded by documentation tasks.

Kevin Pho: Now when you talk to physicians and they’re using these AI note-taking tools, and full disclosure, my presenting sponsor is one of these note-taking tools, sponsored by Microsoft, so full disclosure there, are you seeing it make an appreciable difference in terms of their pajama time and documentation time?

Dike Drummond: Yes. No, I have limited exposure, because I’m not associated with any of these companies and I’m not on site when they’re rolling them out. But I have had members of my community trial one of the ones that you can actually just get for free for 20 days, and you plug it into your cell phone and off you jolly well go.

Here’s a workflow for these softwares, and first of all let me tell you how they were trained. So if you’ve ever messed around with ChatGPT, it will have a conversation with you, and if you’ve never messed around with ChatGPT, I encourage you to do so. Just ask it to tell you the six principles of poodle grooming, it’ll have it there in less than 30 seconds.

But what they have the ability to do is take a ChatGPT engine, which can be taught almost anything, and then what you do is you go into a big health care company and you grab their EMR, and you grab a million records out of that EMR that were written by family practice doctors. So you’ve got a million family practice doctors’ notes and you feed it to ChatGPT, and that category is family practice. And now when it listens to what’s going on in the exam room, it can write a passable family practice note, actually a scary passable family practice note.

And so when you log into this software you make two choices. What’s your specialty, because it’s been trained in these channels, family practice, pediatrics, OB/GYN, radiology, whatever, and do you want a SOAP note or an H&P? And it’s on your phone and it’s on your computer, and your phone is the recording device.

And so what you do is, hey Mavis, how’s it going? Hey, let me turn on this virtual scribe and let’s get started. Boink, turn it on, set it off to the side, do your visit. The only adaptation you have to make is you have to speak your physical, you have to say the physical findings. Chest is clear, heart regular in rhythm, no murmurs, no pedal edema, no rashes, and you’re going to put your positive and negative findings.

And then at the end you’re going to do a summation, you’re going to talk to Mavis about what the plan is, when to come back. Mavis may hit you with the list of, oh by the way, and three other things, because if it’s seen a million family practice notes, it’s dealt with that, right? And then you say, thanks Mavis, and you hit the done switch, and by the time you get back to your desktop in your office the note is there. And it’s a note, it’s not a transcription of your conversation. It has digested everything you said without saving it, and then spit it out using the training of the million family practice notes it had seen.

And I tell you, I have had a couple of folks in my community, of folks who are challenged to understand how they’ll make it to retirement because they’re so stressed by their job right now, literally the first note fall to their knees in tears and say, oh my God, I’m going to make it. I had one person who had downloaded the software onto her computer but never used it, and one day her MA didn’t come in, so she didn’t even have an MA, and she decided to go live on the software on the day she didn’t have an MA, and saw 27 patients and was done 20 minutes after the last patient left, with all her notes completely acceptable in the chart.

And there are initial studies of small groups of doctors, and all of these companies are trying to get their studies out into NEJM Catalyst and stuff like that. There are studies that show it can lead to up to 50 percent decrease in total time in the chart and the elimination of pajama time at home. So yes, it can make a huge difference. But it’s being just trickled out these days.

And here’s my recommendation to anybody who’s listening. Be an early adopter. Get really good at this, because it will be mandatory. Some of the software has actually learned from your interaction, so get it on your computer and get started so you can begin to teach it to speak like you, and get on the bandwagon. This is not something you want to resist or fight, and it should make everything much easier temporarily, until the visit volumes go up.

Kevin Pho: So what do you recommend? Do you recommend physicians resist this AI push on the potential long-term risk of being forced to see new patients, or should they take that benefit now short term and hope for the best going future? What do you recommend physicians do when given the choice of using AI note-taking software or not?

Dike Drummond: Certainly neither of those two options. OK, so here’s what I recommend. Be an early adopter, get it on your cell phone, get it on your computer, be one of the power users of this technology in your group, and be right up front. You have to have some sort of representation in the C-suite that says, look, you can’t jack us up 50 percent in a week. You can’t do that.

If we’re going to increase visit volumes, and we know that they will, we have to have some way of increasing and watching. Pause, increase and watch. Pause, increase and watch. Pause. Because we don’t know what the next barrier to physician well-being will be, but it’s going to be an overload at a cognitive level, and most of the time the documentation overload was only at a keystroke level. It was, you had to type, and typing doesn’t involve as much thinking as seeing more patients. So we could have people go completely on the fritz.

Now I look forward to the time where, you remember The Matrix, right? I would like to have that big plug in the back of my head, me think my notes, it’s like, plug, OK, great. And that’s never going to happen. But this is a major advance. It will be temporary, and what we have to do as physicians, again, money and meaning, profit and purpose, suits and scrubs, we’ve got to reach some balance that says yes, this will increase our ability to see patients, billable encounters, RVUs, and we must keep our finger on the pulse of the well-being of the doctors as we advance the patient load, even though we know that doesn’t mean they have to spend more time dictating. We’re looking for something else.

Kevin Pho: Are you already seeing some hospital systems and practices pressure physicians to see more patients because of AI technology?

Dike Drummond: Not yet, because what’s happening is, I don’t think there’s any place in the country, and if there is, they’re keeping wraps on it, because there’s some companies that want to publish because they have a small market presence and they want to step up. There’s other companies with big market presence, like I would imagine Dragon and their DAX software, or Abridge, or one of those, they want to keep their successes under wraps.

But I’ve not heard of any department or service inside a hospital system where the whole department and service had adopted it, where you could jack up everybody’s productivity quota together. Because you wouldn’t want, I don’t think you would want, and this would be a really strange other issue, you have five doctors take off and love this software and they’re seeing 10 more patients a day than you, making twice as much money, and you’re back here in, you know, AI 101, stuck. What would be divisive among the reputations of the staff. I mean, there’s just ramifications for things that we can’t see yet.

But it used to be that you could be bad at documentation, because you wrote the great American novel late at night after your kids were in bed and you keystroked it all and you weren’t templated and all of that, and you would be humored by your organization because you could see 18, 20 patients a day, which was acceptable. It was a little under half of the MGMA average. Now, when we’re talking about this, your number of patients on the MGMA averages, as this gets adopted, is going to shoot so high that you’re going to be forced to adopt it too. You’re going to have to rise to this new level of charting efficiency.

Kevin Pho: So I heard one of my other guests saying that AI isn’t necessarily going to replace physicians, but the physicians who are comfortable with AI are going to replace those who aren’t.

Dike Drummond: Correct. And then the other thing that’s happening is, I mean, I remember the first time I heard one of my clients who I had helped them get a physical scribe in the room with them, and I checked in with them later on because we had dialed it in and they were working really well, and I said, “How’s that scribe going?” She says, “They’re great, and they’re not enough.” And I said, “What?” And she said, “Yeah, all the notifications that come out of the EMR, all the messages that come from my staff, all the messages that come from the patient portal, I still have to take those.”

And so now what we’re seeing is some of the AI softwares are broadening their footprint to include everything documentation wise that ends up on a doctor’s desk. Some of them are even taking over some of the tasks of a receptionist, with a little tablet in the waiting room that takes your history, right, and reconciles your meds and all that kind of stuff. So there’s things that are now addressing patient portals. Basically they’re chatbots, the kind of chatbot you see on any website. Patient portals, patient questions, EMR streamlining, all that kind of stuff. It’s going to spread out and cover a lot of that.

Kevin Pho: We’re talking with Dike Drummond, family physician, founder of TheHappyMD.com. Today’s KevinMD article is “The hidden danger of AI note-writing software.” Dike, as always, we’ll end with your take-home messages to the KevinMD audience.

Dike Drummond: Be an early adopter. Grab yourself a copy of this software, especially if your institution is not rolling it out, even if your organization’s not rolling out a study. If they roll out a study, be one of the early adopters, become one of the power users. And then if you have any pull with any leader up the chain of command, give them a heads up right now on the concern you have, if you share my concern that there’s going to be a new burnout risk because we’ve taken away the throttle of documentation. We need to be in front of that, so that when the CFO starts checking out the ROI, they don’t just break you.

Kevin Pho: Dike, as always, thank you so much for sharing your perspective and insight, and thanks again for coming back on the show.

Dike Drummond: You bet you.