Nolan Alexander (Host): Artificial intelligence is becoming a bigger part of healthcare, but what does that actually mean for patients and providers? Today, we'll explore how AI is changing the exam room and helping providers document visits more efficiently to improving communication and creating more time for meaningful patient interactions.

Let's dive in and learn how these tools support care teams, what safeguards are in place, and what patients can expect as technology continues to shape the future of healthcare. This is Health Dose. I'm Nolan Alexander, and we're here with MyMichigan Family Medicine physicians, Dr. Melissa Weckesser and Dr. Jenifer Balawender, discussing AI in the exam room. Doctors, how are you each today?

Dr. Jenifer Balawender: Doing well. Thank you.

Dr. Melissa Weckesser: I'm doing well as well.

Host: Great. Well, we're so excited to have you on here, and thank you so much for your time with us today. So getting to it, during a typical primary care visit, what AI tools might be quietly working in the background, and how do they actually help with diagnosis and treatment decisions?

Dr. Jenifer Balawender: I'll start with one that's quiet in the background, because there are some that are a little bit more in the forefront. So, one of the background tools that we've been using is a tool that helps to just generate a draft response to patient portal messages. And if I take it a step back, you know, there's so much work that goes on behind the scenes outside of the face-to-face patient care room.

And this is a tool that, you know, when we get these messages and, you know, sometimes they can be quite complex and sometimes they can be quite lengthy, it just helps to draft a response that then we review in a detailed way. But it's really helpful to help just bring in even just the structure of the response, you know, the "Dear so and so," and the, you know, "Please let me know if you have any questions."

And then, it can even search the chart. If, for example, I had a follow-up question from a patient, and it referenced the next step that I had said in our office visit last week, and it pulled that in for me. And so, it was just really helpful to bring the information in one space as a draft that then I review and edit, and then send on to the patient.

And so, one of the things for me is to really emphasize that even though I am using this tool that helps automate and kind of cut out some of the busy work of that process, you know, it really is the patient's information and my medical expertise that are going into it, and I'm reviewing it before I send it on to the patient.

So for me, that's a message that I really try to emphasize to my patients so that they're aware that it's not just this chatbot responding to them, that it really is Dr. Balawender responding to their message.

Dr. Melissa Weckesser: Yeah. And to kind of work off of what Jeny said, we use ambient listening in the actual exam room. And I want to make sure that patients know that when we say silent AI, it doesn't mean that it's used without consent. Obviously, we get verbal consent from all our patients before we use any devices, and it usually just sits on the keyboarding area, and it will record the discussion and the actual visit that we will have that day. And then, with that, at the end of the visit, the provider will review the note, make sure they correct any mistakes that might be noted, and then the note's ready by the end of the visit. It's phenomenal. Patients get their wrap-up, and they get their assessment and plan as soon as they walk out the door. So, it really does lead to a lot of patient satisfaction. But always, we get verbal consent before we use any devices like that.

We also in our DynaMedex have an AI component to it now that we use sometimes. But it's built off of base knowledge. It's not just spoon-fed AI information. So, I think it works really well in that aspect.

Dr. Jenifer Balawender: I'll just translate. DynaMedex is like a search engine tool that we use to search medical literature. So for example, as opposed to having to go to every particular source, you know, in a library kind of form that might take a very long time, there's an AI component within this that we can just use to ask our medical question. And then, we know it's coming from trusted sources. And then, those sources are even cited so that we can dive in deeper if we're interested.

I also just want to expand on the ambient listening, we generally call it. So, that's when we're asking patients, you know, if we can record the conversation to help draft the note for us. Melissa, I don't know about you. But for me, it was quite a task to try to scribe for myself and take notes while I'm really trying to focus on the patient, gather information, formulate a plan, put in orders, you know, referrals, all of these things, and then scribing for myself at the same time.

So, this tool just does that scribing component for me, which really frees me up to be able to just know that the details and the information are being documented, and I can really focus on the patient. I don't know about for you, but that's so refreshing for me. And I feel like it just takes one of those tasks off my plate so that I can really focus on the others that I really care about.

Dr. Melissa Weckesser: Oh, I absolutely agree. I think sometimes when I walk out of a patient room, if I get distracted by, you know, a nurse or an MA that wants to ask me another question, I may forget a simple Flonase prescription. But when I go back into my room and I open up the transcription and we have, you know, everything that we had a discussion about. So, it says, you know, "Allergic rhinitis, send in Flonase," and I do it. And it's amazing. So, it's less tasking for my brain to remember every single word that I said. So, it really works well.

Host: So if I were to come visit each of you, I wouldn't hear all the click-clacking on the keyboard while we're having a conversation?

Dr. Melissa Weckesser: I don't click clack to begin with. I try not to use it in there. But yes, we would ideally not have to hear any more of that.

Dr. Jenifer Balawender: I was the opposite. I would be click-clacking away. Well, you know, really trying to do a good job multitasking. But, you know, when you're seeing multiple patients in a day, it's just so much information. And if I didn't take notes, and even just like Melissa said, then I might be missing something. And so, you know, it's both in terms of just capturing that information in real time, but just making sure because it's so complex and there's so much. So yes, it's so refreshing to—I still have my computer open because I still need to reference maybe labs or information or imaging, although I'm not click-clacking like you mentioned.

Host: So, I've talked with some of my friends, and many patients worry that AI means less time that you're spending with your doctor. I'm curious, from your perspective, does it help or hurt the human connection in the exam room?

Dr. Melissa Weckesser: So, I actually think it gives me more time with my patients, because obviously we have schedules to keep and we're trying to manage making sure we finish the note and send in medications as well as give our patients the appropriate time that they need to discuss their concerns. And I can actually know that when I walk out of the room, that extra five minutes or seven minutes I would've been spending documenting, I can spend in the room with the patient.

So in that aspect, I feel like it gives back to that appointment time. But also when I don't spend so much time at home doing notes, I come to work a lot more refreshed and not so tasked and exhausted. And so, I'm a lot more engaged in the room because, you know, I'm not thinking about the two hours of documenting I have to do after work. I'm going home and enjoying time with my family and coming back and doing it all over again.

Dr. Jenifer Balawender: Yeah, I agree. I just feel like I can be more focused. Like I said, it's one last multitasking thing that I'm trying to do. And so, definitely, my focus can be even more on my patient, which is really awesome. And it's tangible. I think they really feel that when they come to see me too.

Host: That's very important. I mean, all of us as humans, we want to be focused and refreshed, and I think it's incredibly important for our physicians to be that way when they're in the exam room. And along the lines of AI, how do you as physicians maintain your clinical judgment and avoid over-relying on the AI recommendations that you see?

Dr. Melissa Weckesser: I think the resource that it comes from really matters. Like, we were talking about earlier, the DynaMedex is a group of different journals that have been vetted and, you know, are studied. So, it's not like going onto ChatGPT and just texting or typing in something and you have to decipher where this is coming from.

So, the tool that you use is really important. And then, also building off of the knowledge that you already have, right? So, everybody has their specialties or what they're really good at. So if, for instance, family medicine is what I do, so I can tell if something sounds just off, like if the dosing is off or, you know, I can always vet it myself and look up another valid source. But I think number one is know where it's coming from, not just a random AI Google search.

Dr. Jenifer Balawender: Yeah, I totally agree. I'm really aware that MyMichigan has established pretty significant governance around the use of AI tools. And so, because of that, I know that any tool that we're using that is helping with our patient care, either directly or indirectly, has really been well-vetted. Just like any other medical treatment or intervention, it's been looked at in terms of, you know, what are the risks versus benefits and how are we going to deploy this, you know, with our clinicians. I know because I'm aware of that process behind the scenes that any tool that we are given to use really has been well-vetted.

In addition to the extra layer of recognizing that it's just augmenting what we're doing, right? It's not replacing our care. And so, just like Dr. Wekesser said, you know, we have our medical degrees and we have the baseline knowledge. But sure, it's handy to double-check it against something else or make sure that I'm not missing something.

As we're using these tools, part of the governance that we have at MyMichigan is that we're keeping the human in the loop. That means that no medical decisions are going to be made without our clinicians looking that over. And so, again, I think I have a lot of confidence, you know, just knowing that that's the case and that we're not using any tools that haven't gone through this process.

Host: And speaking of confidence, a concern that many people share is privacy. How is patient information protected when AI tools are involved? What sort of confidence can you instill in us?

Dr. Jenifer Balawender: Sure. Well, again, by creating this governance process, we can ensure that that's one of the things that's well-vetted so that our patient information is very secure, just like it would be within our electronic medical record. And so, that assessment for security is on the list of things that are checked for every single AI tool. And it's exactly why the tools that we use have to be approved by this process, so that those things have been checked.

Dr. Melissa Weckesser: Nolan, in regards to privacy, I was thinking back, because I was just thinking about all the stuff that's available. And when I first started medicine, we used to record on a phone line, and it would be sent to some scribe sitting in an office somewhere. And, you know, someone's information was being given to transcribe into a note. So in regards to that, the parameters that Jeny's talking about I feel like are actually more structured and, you know, protective of patient information than some random person writing up my note 10 years ago or 12 years ago. So in that aspect, I feel like all the checks and balances are there right now.

Host: Well, let me throw this situation out there. I mean, I don't think anyone's going to necessarily sign up to have a landline phone out there listening. But if someone's just uncomfortable, can a patient opt out of the AI recording during their visit?

Dr. Jenifer Balawender: I think that's really important. You know, we want to be transparent about our use of AI with our patients. And if one of my patients says they don't feel comfortable, no problem. And, you know, in terms of the ambient listening tool that we use, we can turn it on and off. So, we might be going through our conversation and at any time, and I emphasize, "Hey, you want me to pause it? I can pause it." Teaching the patients about the tool for me is part of the process because that way now they feel informed, they're educated, so they can feel comfortable when they see me, and they also can trust that at any time, you know, I could start it and stop it or just, you know, "You know what? Today, I want to talk to you, and I don't want that to be on." Okay, no problem.

Dr. Melissa Weckesser: Yeah. I've had patients, actually just a handful to tell you the truth, but they have said no, and we have said, "Okay, we don't need to use this for this visit." And it's been just fine.

Host: That's great to hear. What questions should patients feel comfortable asking their providers about how AI is being used in their care?

Dr. Melissa Weckesser: I don't think there's any question that they can't ask, because obviously AI is in the forefront in any aspect. So, no different than asking your provider about any health concern that you have. So if you have a concern about AI and its utilization in medicine, bring it up. You have the ability and we can answer to the best of our ability. I can't say I know as much as Jeny does in regards to all the AI and the technology component, but I will answer to what I can.

Dr. Jenifer Balawender: I think it's really important that they feel comfortable and that we encourage them to ask these questions. Personally, what I see, like out on the internet and in social media is a lot of unregulated AI. And I think that creates a lot of mistrust. And so, you know, part of my job, I feel like, as a clinician is to not only help, you know, teach my patients and help educate them and help them with their health, but also help so that they can feel comfortable knowing, you know, about the tools that I'm using and being transparent about that. So, I definitely welcome any questions about the use of AI.

And I think the questions that you've highlighted, Nolan, are really some of the key points that I would want my patients to ask. And like I mentioned with my portal message example, if anything, I really want to go above and beyond, and I try to educate my patients in the office. Look, when you get a message from me, and there is a disclaimer at the bottom that says, "AI was used to help formulate this message," I really want them to understand what that means so they can trust that this message really is coming from Dr. Balawender." And I actually put a lot of thought and intention into that, as opposed to it being kind of an automated message. And so, unless we explain that to our patients, I think that might be unclear.

Host: Well, I'll leave you with this. What excites you most about the future of AI in healthcare?

Dr. Melissa Weckesser: I think time efficiency, accuracy of diagnosis. A patient and provider-centered approach, because I feel like this helps in all components. Not only does it help the patient, but it helps the clinical staff, it helps providers. So, I'm just looking forward to a more fluid, efficient pathway forward

Dr. Jenifer Balawender: I couldn't agree more. I think, just in terms of the opportunities of the kind of care that we can deliver if it augments what we can already do today, in addition to helping to decrease a lot of the more administrative tasks that either take us away from patient care or take up our time before, in between, and after our eight hours of patient care or 12 hours of patient care a day.

There's a really significant problem with clinician burnout. And I think that if AI can be used to help trim away some of those inefficiencies and some of that administrative kind of busy work, that really frees up our clinicians to do the patient care that they really want to do.

Host: This has been such an insightful conversation with you both today. Thank you so much for your time and your wisdom on this matter. We really appreciate it.

Dr. Melissa Weckesser: Absolutely.

Dr. Jenifer Balawender: Really appreciate the opportunity to speak on it. Thank you.

Dr. Melissa Weckesser: Thank you, Nolan.

Host: That was Doctors Melissa Weckesser and Jenifer Balawender. Have questions about AI in healthcare or your care experience? Talk with your provider and visit our website to learn more about the tools and technology supporting patient care. For more information, go to www.mymichigan.org.