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. All right, starting this week on news of the week. Welcome, everyone to another episode of AI and I care Scott. I know you just came back from AOA, and I'm really excited to hear about that.
But before we jump into that, Scott, I wanted to touch briefly on an article that are very capable editors sent out just today on on a news report from it's actually a publication from jamma ophthalmology. It's a trial out of the Chong Shan Optomics Center. It was regarding a smartphone app in China that screened basically a quarter million people. And it caught 20 minute emergencies and 19 of those people had no idea they had it.
But here's a good thing and super interesting thing Scott and I think it ties into something we're going to be talking about it has really nothing to do with AI. That's not the key story here. So they built this app called Capture Tumor. It's something you open on your phone.
It walks you through photographing a spot on your eye or eyelid, you know, like a brown spot, maybe a basil cell or maybe a carot acantoma or something unusual that patients will come off and ask us about. And it tells you right then and there whether the spot looks benign or whether you get seen. Now again, this is in China, not the US. So, you know, different regulatory environment.
And it's just interesting. They use media to for outreach. So they actually. Got almost I had I said 256,000 people to to hear about this app.
And then a 600 or so actually screen themselves at home. Interestingly, 20 cancers, 14 beta cells, six malinomas were found and confirmed by. 100% yeah, yeah. So and that that you see area under the curve, 0.977.
So those are the real numbers, but here's what I want you to say with this is the model itself. The AI is not the breakthrough. We've had slept my AI with actually good accuracy for a while now for these tumors that technology existed. It's around it was like in research papers.
And it was just, you know, it was the workflow. It was how do you put this into systems to actually change management. So what do they do? They built everything around the model.
So the first thing they had public education to get people to look at their own eyes. They had the people that was, you know, was actually being used. And the next day human review every image was in 24 hours. And then there's a proper referral pathway that actually got people into the clinic who could buy up see the leation.
So I guess the the take home is that we have the advanced technology. When we have the algorithms, although we talk about that all the time, we have all the sort of the sexy artificial neural networks. But really it's the workflow. It's the referral loop.
And it's the boring sort of unglamorous part that everybody skips when we're chasing some type of AUC, some type of accuracy levels. And there's something, you know, I talked about with Rumsana China over at University of Wisconsin, Madison. It's often that last mile that throws us off. So this closed the loop in China because it has things that we don't have they have an app.
They actually used to we chat for this, by the way. Maybe it's easier in China, give the regulatory environment, but this is something we could do in America, an American health care system. And so I just wanted to bring that that up to this article again is a jam, up the mall. The link is is set was sent to everyone this morning for those who are subscribed to the AI and I care newsletter.
Please check it out Scott. So tell us about AOA and maybe how this sort of fits in. Yeah, I mean, I actually I'm going to put that in a whole little bit. I want to do a little our knowledge by and then I'm going to tie that into where what some of the things I learned today are.
And then maybe some of the things maybe continue a little bit of our YAI fails and the guard, you know, the hype cycle and all that kind of stuff because I got this. I felt this at vision expo East. I kind of felt this again in a way like there's there's there's a shift happening in the perception of AI. And I want to talk about that before I get in that there was a lot of discussion anyway.
It's been a lot of discussion more about biometrics and I want to talk about what that is because I think sometimes there's confusion about what biometrics are and everybody assumes that's like a watch or ring. You know, all that kind of stuff. And so let's talk exactly what are biometrics. Well, simply anytime you have an automated measurement of some biological or maybe even behavioral care.
And behavioral characteristic like let's say your fingerprint or your voice or your heart rate or your IOP. You know, today the friction in what we do in I care is that we were completely dependent upon static single point clinic measurements like IOP to track some dynamic disease like glaucoma we measure. You know, IOP once every six to 12 weeks at a certain time of the day and it may not be consistent through visits for visits over the course of the year to track some crazy dynamic disease like glaucoma that can be affected by, you know, by heartbeat by respiration by what did you eat that day by, you know, I mean, did you take did you take some medication or some food that increases your coccyte. You know, I mean, there's it's crazy that we still think we can manage diseases that way.
And you know, biometrics, whether you're talking about cardiovascular risk, you're talking about respiratory disease, you know, or even ocular disease biometrics solve or have the potential to solve this by capturing real world. But you know, I think it's not. Can biological data that's continuous and you know, I think it's going to draw my we've talked to a couple other episodes shifting our delivery from these reactive point one single point of care one single data point checkups to really proactive predictive live time continuous care. For all of you who are listening, whether you be a doctor, the clinical staff or administrator, I mean, this has the potential, even though I think we're still in the infancy of this, especially in the eye care world, to transform what we do day to day.
I mean, imagine, and we use this analogy before, imagine a workflow where we no longer have to guess about what IOP and blood pressure are at 4 a.m. and if it's spiking or, you know, depressing based on medication compliance, I think that we're going to have the ability in the near future, whether it be biometrics, do you have an watch or a contact lens or an inter-ocular device, where we can truly have 24-7 gathering of live-time biometric data. And, you know, whether it be like I said, whether it be a smart contact lens, embedded with some micro-sensors, and I've heard that going around for a couple of years now, and I mean, you know, it has a hard time getting out of the FDA, but boy, the technology's pretty impressive. We're not just connecting vision.
We're tracking things like IOP or tear-film glucose, and they find that might be a better test than looking at blood glucose. You know, finding out what happens at IOP during sleep or other biomarkers in the tear film, you know, things that can be measured kind of whatever we want without necessarily having to do some point in time device. So, you know, I think that biometrics, and like I said, I think we're very much in the infancy ray hand of this, but I'm very excited to see where biometrics are going to impact not just our overall general well health. Like our watches, we always joke, my wife is like, I only know if I slept well, if my ring tells me I slept well.
And I'm like, well, maybe just did you sleep well? Like, did you get up at sleep well? But I think about a day, I don't think is that far off when we ask a patient, how do you think you've been doing in terms of your blog comment? And they go, well, according to my phone, it says that my pressure spikes at 4am every day.
But if I take this medication at the right time, when if I go to bed, it doesn't spike. And we start involving patients in their own care, because I think as we look at people with watches and phones, they're like, oh, my blood pressure is up. My respiratory rates down, my resting heart rate is different. And they start taking better care of themselves, because they have lifetime feedback.
I think when we can start doing that for eye care, that might change a lot of what we do. That's, I couldn't agree more. I think the idea of smart wearables, becoming part of ultimology, and given the fact that we have a wearable that many of us wear all the time, compared glasses or contact lenses or contact lenses, a window into our own health. It's only a matter of time.
And it's already happening where we're getting really good information, not just about our eye health, but I think of our systemic health. So, well, so tell me about it. That kind of rules in, yeah, I'm going to rule in. For any of you who haven't read Ray Hans' last article, it was called the iExam one bundle.
We did a little bit of that on our last podcast. You need to read that article. And even though that wasn't necessarily brought up during the AOA, that during the AI, the innovative innovation summit, excuse me, the AOA puts on in the first day of AOA, this year I was fortunate enough to leave one of the panels again. And not just that panel, but a bunch of other panels and some other other courses I said in on, you know, the talk was less about AI and technology as a whole, but an overlying impact of, I think we're starting to see a change.
And it's at the beginning, there's no doubt about unbundling what we do. You know, the start be conversations about, well, what about these histories at home? What about diagnostic testing centers or what's the impact of biometrics going to be on when we have more data and now we're moving more towards telehealth? Instead of seeing somebody every six or eight weeks, maybe we're having a telehealth visit every time their compliance changes.
Or hey, we see a trend that we call them and say, hey, here's how it goes. You know, I, but the arguments kept coming up is that all of this is going to require a change in the way we think about what an exam is. It's going to require a change in how we deliver care, where we deliver care, when we deliver care. And those are all interesting debates.
And I think that nobody had answers, right? I mean, it was one of those that I love this phase of an innovation shift is everything. Well, we know there's a problem and we know there's a solution, we still not implement the solution into the problem yet. And you know, that's good, because it gets people to start thinking about, and that's what we're trying to do right now for audiences, start thinking about, well, what happens when you unbundle the exam as Ray Hunn said in his paper and our last podcast?
You know, what does that look like? How does that change delivery? How to biometrics come in? How does our adoption of telehealth come in?
At the end of the day, it's going to be a total disruption to the workflow we currently have in our standard exam. Like, we know, and I think you said this in the paper, you know, the day of they walk in the door and we check their history, we take their history, and that takes 15 minutes and we do the exam, you know, if you do a refraction, we do its program, and we dilate before we look at the eye, which is great, before we look at SLITLAMP, which is crazy, totally inefficient. And then we do, you know, the day of having this complete exam unit, I think we're, as you said in your paper, we're about to have a major shift in that, and we're starting to hear those conversations from some of the forward thought, you know, thought leaders going, maybe it's time to do things different, what would that look like? And, you know, I don't have the answer, I mean, Rand, I'm happy to toss ideas off, but what the answers may be, but I think we're starting to have those conversations, and that's the beginning of an evolutionary shift in the way we provide care.
Yeah, I think you said it best when workflow is going, is the ultimate solution, that's where our solutions are found. We are here to solve problems, AI is a tool, but AI itself, I think this is where a lot of the frustration with AI comes in, because people just come in and say, well, AI is a solution, and AI is AI, that, well, in fact, it's just a tool for a workflow problem. I mean, we are trying to solve problems here. And so AI is definitely going to be part of that.
And I'm really excited when we have an AI-native type of clinics, when we sort of go past this sort of transition period that we're in, but we're in it. I mean, we're in this transition period, not for a long time. We're in the middle of it. Absolutely.
And so we are transitioning the old way of doing old CPT codes, legacy type of billing into a new way where certain threads of that exam are now being teased apart. And so it's obviously disruptive and anxiety inducing for a lot of people, but we think the better care is on the other side. I agree. And you know, we've said this before, is AI is just another technology that if you have a weak workflow, it's going to exploit it, it's going to dump, it's going to show it.
And it's going to say this doesn't work. And the AI is going to fail because the workflow fails. If we fix our workflow, and whether AI or any other technology for that matter is, you know, affects that and positively impacts our workflow, we're going to all think it's a home run. But when we have bad workflow and one of these technologies don't work or slow us down or create new problems, don't blame the AI.
Take a look at your workflow. And fix it. Well, I think it's any better way to end. Thank you everyone for listening to another episode of our podcast.
All our materials available at AI and I care dot com. Please feel free to reach out to us. Rehan R E H N at AI and I care dot AI and Scott SCOT one T at AI and I care dot AI. Great chat Scott.
Look forward to chatting with you all all next week. Have a good one everybody.