How Might Chat GPT Influence Practice Management?
Welcome back to Podiatry Today Podcasts, where we bring you the latest in foot and ankle medicine and surgery from leaders in the field. I'm Dr. Jennifer Spector, the assistant editorial director for Podiatry Today, and on this episode we have with us our guest, Dr. Jim McDannald, who's the founder and director of Podiatry Growth, which provides online strategy and services for podiatry clinics. In this episode, he's going to discuss a new and emerging digital tool that may not initially seems like it applies to podiatry practice, but he'll share what we know so far and how this could impact our practice management strategies. Thank you so much for joining us yet again, Dr. McDannald, and we're looking forward to hearing more about this topic. So for those of us like myself, who barely know what it is in general, what is ChatGPT?
Dr. Jim McDannald:
Yeah, I'll try to describe it in terms that maybe you, the audience and then my parents could maybe understand if I were try to explain it to them. So ChatGPT is a company that has created a computer program that's built from what they call a certain type of artificial intelligence called machine learning. So machine learning is a way of teaching a computer program to learn from data rather than having to have specific instructions for every possible scenario. In the case of ChatGPT, it was built using a large amount of text from the internet, books, articles, websites. And basically, this program was trained to analyze all that text and learn to predict what words and phrases might come next in a sentence, so based on what it's seen before. So this allows ChatGPT to understand and generate human-like language. It's not at an academic level where it can write theses and do things like that.
But it can understand how words are parsed together and continue building things off of that. So when you type in a question or a statement as you chat to ChatGPT, the program uses machine learning to basically analyze what you've written and figure out what you're asking or saying. So it then searches a huge database of information to find a relevant response. Not all of it's factual, so be careful, we'll talk about it in a little bit. But to your question or statement and generates a response using human language that has been trained on. It really works on the types of prompts. So we'll get into some of these prompts later. But for example, "What are the best places to visit in Montreal when I'm on vacation?" It'll go out and scour the internet, but it'll also find books, it'll find texts, and then it'll try to put together maybe the top five places and an explainer on those subjects. So that's not just say, I guess, Montreal vacations are podiatry related, but we'll get into some of those more podiatry-based ideas in a little bit.
Dr. Jennifer Spector:
What right now are some of the more traditional implementations of ChatGPT?
Dr. Jim McDannald:
ChatGPT is obviously a new version of AI, in a way. I think if anyone has a Gmail account that's listening to this, you probably will notice that after you start typing certain things into Gmail, there's auto-complete. So auto-complete is one idea like, oh, that was before ChatGPT, right? So this AI utilization of that. I'd also say that there's different ways that machine learning has been utilized, whether it be the way that Google ranks websites, it's crawling the internet of all of this data finding which links to links and makes those Google results on a page. So there's a lot in the tech world that's been utilized with this machine learning and AI, but now it's going to starting to bleed into healthcare and bleed into medicine in ways that we can utilize.
Dr. Jennifer Spector:
Well, that's fascinating, honestly, in your opinion and observation on this then, how might something like ChatGPT apply to podiatric practice?
Dr. Jim McDannald:
Yeah, there's a lot of examples. We'll go through a list here just so people can get some ideas and play around with it. I think it is an interesting thing right now where it is free, at least as of the recording here, maybe that will change. But there is a pro-plan currently for $20 a month where people can get access, because right now it sends to be overused in a way that it's down quite a bit because people are using the system. It runs on the Cloud, on servers. But some ways that podiatrist can test it out for free and some ideas that I've utilized with some of the clients I work with and some ideas that I've gotten from other places that, number one, it's a way to make a response, an email response or a letter response to insurance providers for a claim denials. So it's a way of actually almost personalizing in a way that doesn't seem very intuitive like, "How can I use this thing to make personalized emails?"
But for example, it's really what you tell ChatGPT to do, and these are what we call prompts. So let's say, I have a 10-year-old patient, I send a letter for orthotics to Blue Cross, Blue Shield, not slamming on them, this is an example only, they denied an orthotic for pediatric flatfoot. So I basically would type into ChatGPT, respond to denial of claim from Blue Cross, Blue Shield for a 10-year old with flatfoot deformity who needs orthotics, cite some academic sources. And you would be surprised what you'll see. You'll see a letter, you can even say, maybe you could say the person's name, you could say your name signed by this. And within 10 or 15 seconds, you basically have a form letter with scientific abstract call-outs there within the document.
And obviously you're the editor. I think one of the thing key things I want to be very specific about, is that this is a creator, but you have to be the editor. It's like self-driving cars right now, you don't want to go from New York to LA and just not touch the steering wheel. You have to be aware of what's going on with ChatGPT to make sure that the information is accurate and you're not doing any harm. That's the whole aspect of why we're in medicine in the first place. But it can be a huge time savings for something like that. So that's the first example, I would say
Dr. Jennifer Spector:
Almost like a virtual intern where you're still in charge, but they might be able to help you do some of the legwork.
Dr. Jim McDannald:
Yeah, that's huge. I think there's a lot of podiatrists I've talked to these days who are working with almost like an overseas or an AI transcription of all of their clinic visits. So gathering all that data, all that information and you talking to your patient about the way you treat certain things could lead to huge time savings in the future. But like you said, you have to watch over it with editing, but this denial of claims is one example. I would say another example too is just communications with patients. Over time what might happen is not necessarily through ChatGPT, but through machine learning is that you have your patient record within your EMR, you've seen how you've treated them in the past, and they can basically summarize the entire course of all of their treatments with you in a very succinct way.
You could basically pull all that data or pull all those visits and say, ChatGPT, summarize the entire lifetime of visits I've had with this person in 200 words or 300 words. And they could try to pull out the key findings or the patterns and stuff. So you can have, maybe Mrs. Jones is curious about, what did we do over the last year? Or they want to have specific questions about their treatment for X type of problem. This is something like ChatGPT or AI or machine learning can do to help make really relevant and personalized communications. They don't necessarily require you to go in and read all of her charts, write her personal letter on your computer and send her an email or a message in a way that's HIPAA-compliant, so that's another way.
So another way is correspondence with other providers. Right now, I'd say a lot of these things it's going to overtake is I say generic templates. I think in a lot of practices that are run very well, we have these templates for whether it be patient referrals or doctor referrals or different things where we're going to send them this kind of fill in the blank letter, which isn't horrible. But like I said, with machine learning and ChatGPT and those things, if Dr. Smith sent you a patient that has maybe a wound or something like that and they need to have a specific update as far as what's happening, you might be able to pull that from your EMR, send the last six months of care from this patient to Dr. Miller, and it just does that. Obviously you're going to scan through it, going to read through it, you're going to edit it. But that's something that could be potentially huge. I don't think that's available yet, but that's something to consider.
Thank you notes being a little bit more personal. "Thanks for the referral. You've sent me this many patients." It just pulls that from the data. There's ways different data can get pulled into these communications to make them more personalized in a way. Also, I would say that whether it be website or marketing copy in writing could be a huge thing. I do quite a bit of writing myself, so I'm not trying to put myself or Jennifer or anybody at the staff at Podiatry Today out of business. But it is going to be one of those things where you no longer have to stare at a blank page. It is something where you get some ideas, maybe they're not perfect, but at least you're not staring at that blank page anymore. You say, "Give me topics about heel pain. Give me five different ways to treat heel pain." Or, "Give me some ideas about how a patient can get... What are the best five types of exercise for someone that's having balance issues." Or something.
Obviously check the literature or ask for the literature to be given to you to make sure that it's legitimate. But this could be some way of putting stuff on your website, creating webpages, blog posts, other types of material, come from a marketing or website perspective. And then this also could be used a little bit in social media posts as well. "Write me a Twitter thread about why heel pain is the worst," or something. And have it be relatable to patients. You can't also change the tone of what it sounds like. Obviously if you want it to be more familiar or more formal, these are things you can change within ChatGPT, which is pretty interesting, it's not just like... You can tell it to make a joke about something. There's a lot that it can do. And then we're just skimming the surface on some of those use cases right now.
Dr. Jennifer Spector:
And so now that we know a little bit more about what it is, what one thing do you feel that podiatrists should do now to either learn more about it or test out incorporating it into their practice?
Dr. Jim McDannald:
I'd really say go check it out on the web and just play around with it, to be honest. Hopefully, you've gotten some good information from myself on this podcast and it'll pique your interest a little bit to test out some different things, maybe some processes that involve creation or writing in your clinic that you think might be a way to save time. Because that's the biggest thing to me, is that if you play around with it, you might find some ways that it can save some time in your practice. And like I said, I think things are moving this way in the future. So whether it be, like I said, people are recording their entire conversations, their entire patient encounters, and it's pulling information about how you treat patients. So it's a way to maybe improve the way you practice.
Maybe you notice something that you do, something you want to change or tweak and ChatGPT picked up on that or some machine learning said, "Hey, the patients have been getting better, either with you or somebody else had been doing these things." It's a way to really improve patient care at the end of the day. And I would say that text is the first area that's going to be affected by machine learning and AI and ChatGPT, but this is going to move into video. This is going to move, and it has a little bit already, it's going to move into image creation. There's these different ways to create AI images, like say instead of... One of the things I usually preach is that you need to have authentic photos of you taking care of your ideal patient.
I can see in five or 10 years where you take a selfie of yourself with your phone and then you basically give an AI or a website that image and say, "Show me treating a 80-year-old woman with arthritis. Show me treating like a 12-year-old, she's in the treatment chair, I'm in the doctor chair treating her for this problem." And it creates that image for you and it's completely created from the ether. It's just one of these things where it feels like magic. But I think it's one of those things where you have to... We're a ways away from that, so I'm not trying to say it's happening tomorrow, but just get familiar. Just like the generation before us had to get used to what Google was and understanding it.
And I think people understand Google a lot from a consumer perspective, but from a strategic practice owner perspective, it's a different game or it's a different kind of way of maneuvering. And the same thing will happen with ChatGPT. Get used to it, see if there's some kind of processes within your practice that might be able to become more efficient because of it. Because I think we are moving towards this ChatGPT, AI machine learning in the future and it's not that you shouldn't be left behind, but I think if you're curious and you want to provide great care for patients, it's a great way to move in that direction.
Dr. Jennifer Spector:
Well, we can't wait to see where this technology could take us into the future. Thank you so much for sharing your knowledge today, and thank you to the listeners for joining us. Make sure to check out this and other episodes of Podiatry Today Podcasts on podiatrytoday.com. SoundCloud, Apple Podcasts, and your favorite podcast platforms.