Welcome to Real Talk on AI and I care with your host, Dr. Scott Morris. And I'm co host, Dr. Rehan Ahmed.
In each episode, we discuss current news, adding a little AI education, and debate innovative AI tools and topics that will change the industry. Everybody in the Scott. And this is that Ray, huh? So today we are going to have kind of a fun little show for everybody really enjoys.
The first one, we always do the knowledge bite and that's in the podcast where we kind of break down. Different terms and tech that kind of have the ability to reshape your clinic and we try to break those down into a little bite size couple minute. Actionable insight kind of stuff. And so today we're going to tackle maybe a topic that you've heard about, maybe just never thought about.
So we're going to break it down. We're going to talk about assistive AI versus autonomous AI. So, you know, and this is one of those topics and it's going to be a term that you need to become acutely aware of and familiar with because it's going to affect what you do in terms of your workflow and your bottom line. But more importantly, it might affect your liability.
And we talked about that a couple weeks ago in one of the episodes about where's your liability going to do? Where's your liability going to be if AI takes over? And so let's kind of cut through that marketing stuff and get down to the basics. So first I was talking about assistive AI.
Think of it almost like your co pilot and you know it's this as we've mentioned many of our podcasts. This is kind of ultimate really hyper super intelligent tech sitting right next to you who knows all of the literature that's out there. And but it doesn't make the final call. It's your research assistant as well.
It acts kind of like as a second set of ears, eyes, sensory mobility, all that kind of stuff and flags. Any abnormalities anomalies that we might miss on some busy Tuesday afternoon when we're on our 50th patient to the day or whatever it is. You know, for example, like let's use example OCT because we do that lot. You know, an OCT scans does a scan.
Well, an assistive algorithm might overlay a heat map or pop up a prompt saying, hey, look right here where we might have just blown over and just a little red spot where it might look and go, hey, yeah, but that red spot wasn't there last time. And it may say here's an 80% or 85% probability of early stage geographic atrophy right here. But that's where it stops. It doesn't make any sense.
This is what we just something to look at. Something to pay attention to. And then it kind of hands the proverbial steering wheel back and decision making back to you. So that the diagnosis, the billing code, the ultimate treatment plan, that's always up to you as the provider.
And it makes you faster and it probably makes you a little smarter, but you're still in the driver's seat. Whereas we get to autonomous AI and unfortunately, and I say unfortunately, because I think a lot of people think of AI as being autonomous. And there are forms that are autonomous and there are things places that's going to be great where autonomy isn't really a copilot at all. And it's that driverless car as we've talked about in the past where it takes the data, analyze it and makes a definitive diagnostic decision.
And you have no part of the decision. And we already see this in some things of clinics today, particularly with like some of the FDA systems we talked about for diabetic or at nopathy and we're starting to see some other forms of aculomics where, you know, we care clinic or high volume practice, the tech or snaps a picture or snaps no CT and then loads in the autonomous software and system spits out some diagnosis. And then even might tell you what to do like hey, this is diabetic right nopathy refer to a rat and a specialist right or whatever you wanted to say, boom done. I never as a provider never even looked at the image and you know, or an administrator, especially in the days private equity, that sounds pretty good because that's a massive win for scalability and catching disease early in the process.
But for providers, that's kind of a big mindset mindset shift for all of us, trust you machine to make all the right decisions. So what's better? The answer is, I don't know, it's kind of strategy, you know, assistive AI is going to help us elevate our experience. It's like you said it's like having the best doctor ever on your side coaching you go and look here, think about this.
Where autonomous AI may screen for something that we may not see, it just make the decisions, but as we talked about many of our podcasts, that's going to be very dependent on data input, how it learns, how it trains for it to totally take over the driver's wheel. So, you know, as we've talked before question, isn't whether AI is going to come to your office, it's it's here, it's coming, if it's not, I just didn't say it's coming, it's here. The question is, is how do we use it? Do we use it as an assistant or do we let it drive?
And I think the answer is, it depends on the workflow that we need of which one we're going to use. I think we're going to have a combination of stuff. Ram, what do you think about that? Yeah, I think that's a great overview of assistive versus autonomous, you know, mental model to think about AI.
Yeah, I think there's a combination of both. I mean, at the end of the day, we're taking our humans and you need that human touch, but there are so much that it can do in terms of pattern recognition and data mining and looking at a tons of images, especially as diagnostic technology just multiplies. I mean, we're looking at OCTs, visual fields and there's a lot of data and components here segments. Yeah, I mean, autonomous slow lamp is I'm it's it's going to be there in a couple of months, I just saw one of the prototypes and it's spectacular.
So fantastic. Yeah, they're really amazing devices. And so I do see a future where like I you talk about the office of the future where they, the I care providers sitting in the office and they have like the fire roast, you know, going in the back and they talk into the patient like a normal human being. Like, you know, like a radio, you know, we are basically we are not meant to just be data collectors.
We are data interpreters and synthesizers and communicators and I see the cash. That's the thing that's going to be the biggest part of the new renaissance of what it means to be an I care provider. I'm super excited about that, but I think that that's a good mental model to sort of think about when you evaluate technology, what is assisted and what is what is entirely autonomous and I think. I think about that in the different risk profile those represent is a great way to start.
I think that's a good lead into what we want to talk about fireside today, which was, you know, what does the office of the future look like and I think he used the term unbundled exam. So explain to the audience what you mean by that. I'm looking forward to this conversation because I think it's a fascinating new mindset that many of our audience. So what do I mean by that and bundle the exam and so this is something I just wrote at AI and I care.com I wrote a recent column about.
And Scott, you know, if you think about it, you know, when we were in training, we sort of had this litany of vision, motility, pressure, slipland dilated exam, you know, 24 points and 20s. And then that is sort of encompassed into our building our CPT goes intermediate comprehensive 920 blah blah, you know, all these different things. 99213 versus 99214. Yeah, and for the longest time, the exam was consciously or not was a bundled unit, right?
And this is what you imagine what a intermediate exam is and this is what a comprehensive exam is. But you know, technology and now, especially AI is sort of pulling at that scene, no doubt, and unbundling really important parts of it. And in many cases, and we'll talk about the second is actually providing in my mind more valid. So for example, we talk a lot about oculems, right?
Where is ocule, where is a written a photograph in a detainment? It would be naive to think the I care provider is going to be the place where these fundus photos are. That is not where these companies one, their valuations are not built on the 60,000 I care providers who are in the US. And that is the idea is it will be in retail locations and pharmacies at labs and primary care doctors.
Selfies on your on your phone, I'm holding my phone. It will be everywhere. Now that's a part that's the beginning of the unbundled I exam. That's a part of the ice and it's a written a photograph.
People a lot of information. And now with AI, you could a lot of information that even an I care provider wouldn't necessarily be able to pick up in terms of cardiovascular risk, Alzheimer's risk, all the other things we've talked about a lot. But that's just one part Scott. There are a lot of different parts of the ice and so the next example, you know, the bug of who may be the fourth rail is perfection.
I can't see. There's a hot topic right there. There's a hot topic right. It's being unbundled and I know there's a lot of viewpoints on this, but you know, there are companies now I thought it's one of them and there are the refractive technologies in the past and you know, some were done online.
Almost like a decade ago and that's where it started this concept. But that's you know, again, it would be. I think that's that. Say it for the audience, you know, and this is coming from somebody's been doing this is 35 years of refractions of test 921, you know, it's a test.
It's not different. Any different than a photograph or an OCT or Gonioscopy, which is a procedure or servers test or fluorescing it's just a test. Right. It's a test that really goes only, but only I can do it.
Well, no, I'm sorry you can't. It's not just you and technology can do it. It was harder for it to be sort of pulled at the scene and unbundled so easily, but now with AI and technology and different diagnostics that can be placed again, you know, we're talking about where are patients, where are customers. It's not only in the eye in the eye doctor's office, right, we're talking about retail patients labs at home on your phone, etc.
And doesn't even end there. I mean, those are two examples, but then you have BR headsets now. Well, yeah, but I think even more than that is that you're talking about what's happening for the clinic exam. But think about we've had a couple conversations about this.
What about all of the history, all of the stuff that happens before we as providers ever say hello. All of that doesn't have to happen in the office. We can completely unbundle that and make that happen somewhere else. And here's the other thing this ought to get a few people that'll be really not happy with me for saying this, but retail is already moved out of the office and it went online.
That marketplace of how retail solutions are sold is going to change. You know, I think about you look at surgery, right? That's unbundling. You do the exam in one place.
You do the surgery in another place. You do the post stop of another place or unbundling a surgical procedure. The exam can be unbundled. This is a history piece that's done somewhere.
There's the clinical diagnostic testing that may come from multiple different places. Having a conversation with an RPHD other excuse me the other day. Who's talking about a device and this doesn't violate an NDA so I can say this is they're working on a device that will actually measure pressure intraoperatively where they'll actually put it in the eye and leave it in the eye. Almost like a glucose monitor slash glucose pump that you know I look at that and go, oh my gosh, you know, now you're getting 20 as we talked about this before again 24 7 365.
You know, I mean, you do information right about what popular pressure is and you know, I mean, and then you take retail and you put it somewhere else and maybe you're taking finance and putting it so that day of. And maybe this is where you were going with unbundled. I don't know, but you know that day of you do everything from the patient walks in the door to the patient walks out or I think that's inefficient ineffective. Traditional, but way updated yeah yeah and I think oftentimes the question was maybe the reflective the reflective person or just start or maybe our reflexor like hey this is not a true item.
It's not a real item. It's not a full eye exam. We're like the patients missing something and you know I think what we're really. The wrong question is is X a full eye exam the real question I think is well who's taking charge of that patient because there's still the issue.
Who's quarterbacking all this right patient goes in gets a gets a great retina photo at their PCP or at you know lab core and gets in some information on their cardiovascular risk or. Or some of you know but who's monitoring their enlarged cup to this or their glaucoma suspect or their gruesome or their patient goes and gets a refraction but you know. Maybe they haven't had an exam in a while and they have hypertensive retina. And you're kind of missing the other side so who's owning all these pieces and that's where I think you know there's a lot of times we think I go out you know are we are we in a replacement what's going on but I think it's up to the eye care industry it's up to us.
To have ownership over it and to really lead the way because as a friend keys are often says you know. And you know we are if we're not able to what is it we're on the menu basically if you're not at the table you are on the menu and. Yeah we either have I agree with you completely Rayhan our industry needs to wake the bleep up in the right now. We're on the menu and we've got to change that and become at the table and I think that I agree with you taking that.
Taking the press at the opinion or the outlook or the perspective that we and you said it earlier perfectly we need to dump the theory of where data collectors and where people who effectively communicate in an empathetic way. And interpret results interpret treatment plans and that's what that's a mindset shift data collection to data interpreters and communicators and that that's that's at every level that's going to happen at academic level it's going to happen in a clinical perspective it's going to happen about how we deal it tell of tell of tell of medicine tell of technology technology surgical robotics I mean it's it's. We at for the audience if you're listening I think that you know Rayhands idea today about the unbundled exam is it this is why I think we're going to look back five years ago now and go. Can you believe we actually had a person committed to everything in one visit.
Not the best way to do it. And so you know the conference for exam taught us you know maybe to put the pieces together right now that the pieces are all sort of leaving the room in different ways it's up to us to sort of put it put it that together right. I didn't they they're still a patient there there still a patient there and so I i'm really excited I think there's a role in that and those who can communicate and be able to have that patient. Is this in relationship or the ones who are ultimately going to be the most successful while incorporating all these new and great AI technologies and new technologies that we're going to have AI or not so.
Yeah I mean I'm really excited about it but there's there's going to be some turbulence there's no question about it. Great great discussion rayon that was a great topic to cover and I really appreciate your time for that so audience you can always reach Scott and. Are you guys me reach me and rayon minds sot is our EHA a our EHA and at AI. A letter a letter i in I care.com i'm sorry dot AI dot AI of hall days about that.
And we're always open for open to new suggestions about topics you guys want to hear about we have a couple ones coming up here that people have suggested I think they're going to be really interesting. Please stay tuned to our website which is AI and I care dot com where you can see all kinds of articles podcasts whether it be this podcast or innovators podcast. Coming up and I'm really looking forward to will kind of give you a little teaser for this rayon I've been working really hard on this for the last probably four or five months very soon we'll have out in a free implementation guide of how to integrate implement. And strategize about putting AI in your practice can't wait that's coming out soon plays pay attention watch your emails on that because we read a lot said I think this is going to be an invaluable resource for all of you who are going how do I do.
We're going to map it out for you one step at a time. We're looking forward to it Scott and thank you again and I can reach just reach out to us we're excited here from our audience and we will talk to you next week.