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. Well, hello everybody. Welcome to Real Talk. This is your host for today, Dr.
Scott Morris. We have a guest on Real Talk, Ruponsra, Dr. Ruponsra, who is the chief, I say vice president of professional affairs for TopCon. And all of you have been listening to our podcast for a while.
We talk about TopCon a lot. It seems like it's in, in our news conversation or during our podcast every day. So now we get to hear from the guy who's actually got his fingers in it and deal with it every day. So I'm really looking forward to Rupon.
I just want to say welcome to the show. Thanks Scott. I appreciate it. I'm a huge fan of the show.
And yeah, thanks again for the invite. So for all of our normal listeners, you know, we always kind of try to do something we call the knowledge bite where we try to educate. All the audience about some different topic is a late stay AI. And today I was, I was having a colleague of mine yesterday kind of go, Scott, you just published that article on black box, gray box.
And I really understand why the black box is such a problem. So I thought that may be a good knowledge bite. So let's think about the black box problem kind of like this. And so think about imagine walk it.
We'll put it in a little bit of context. Imagine walking to a bank and kind of saying, I'm going to apply for mortgage and you know, you have a good job and you got to get same as a count your history is all clean and you need to hand over your paperwork and a loan officer puts in the computer and goes hits entering five seconds. Go by and he goes, I'm sorry. Your application is denied and you go, well, wait, I mean, why was there specific reason.
And then the opposite of the screen says, well, honestly, I don't know the algorithm just says it's too risky. Now that's not just bad customer service. I mean, that's the essence of the black box problem in artificial intelligence. And in simple terms, black box describes any AI system where we know the input, which in this case was your financial data.
And we see the output decision you're denied. But what happens between them, the internal process is just a complete invisible mystery. I want you to think about how that might apply in medicine when we put input of this is the history, these are observations. And the black box says, well, this is your diagnosis, but we don't get to see the transparency of what's happening in between.
And so that's a big concern in medicine. And so let's think a little bit about what the mechanics of that might be. So if you want to like. If we program a computer, how can we not know how it works?
And that's a really fair question, an traditional software, you know, programmers, but they put in very explicit rules and say if X happens, then do why it's very transparent. But in modern AI, and you could even look at in terms of what generative AI does, but if we think of old school modern AI, and it sounds contrary to this discussion, but modern AI, which is very much you have an input, you have an output. And the deep learning that happens in between that works a little differently. It mimics the human brain and we feed it these massive amounts of data and it process that information through just layers and layers of artificial neurons.
And it teaches itself. Instead of being program, it teaches itself to recognize these patterns and you might look at, for example, what that financial example used it might look at millions of loan applications and create its own very complex mathematical rules for what a good borrower and more responsibly a good pay or back looks like. And these rules aren't written in English or a code that we can see it's just lots of numerical weights and connections and we've talked about that before about how algorithms are based. And there's a box that's incredibly accurate, but very non transparent or what I usually call very opaque.
Well, so think about why does this matter and why should you not just trust the machine if it's usually right. Well, that's a big sentence right there, right? If a machine is usually right. If you want it to be 100% right and medicine, but then you can let's say Marguerman about the gray box.
Do we just trust every physician because their gray box is usually right. So there's two reasons that we struggle with this, you know, one is accountability, any other one's bias. So if an AI diagnosis a patient with the disease, the doctor needs to know why so they can verify it. We can't have a life or death decision made by a system that says trust me.
I just know. Well, we struggle with that in our gray box system and a black box system is even more distrust. And furthermore, I mean these black boxes, they can't hide their biases. If an AI is trained on historical hiring data, for example, the favorite men over women, the black box is going to learn that bias.
So we can't start rejecting female candidates or without transparency. We can't see the eyes being sexist or financially it says, well, hey, you know, you're a guy, you're a better alone candidate than a woman, a lot sexist, that's bias. But we just can't always see it. So what we're going to do is we're going to use this researchers and that's research work on these solutions called X AI.
So that's a term you're going to hear where to spend more about that probably in our next podcast, talk about X AI, which is called explainable AI. This is the turning of the black box into a transparent glass box. The goal is to build systems that can use kind of a scorecard, along with the decision telling us I reject this phone, specifically because the debt to income ratio was, I don't know, 5% too high. We're moving towards a future where trust is just as important and maybe more important as accuracy in the AI world, because an AI that's 99% right, 90% right 99% of the time, but can't tell you why is a tool that us as humans are never fully an at trust.
So as we look forward to X AI and bringing transparency what we do, then all of a sudden we have to question, well, are we ever going to have that transparency in the gray box and the answer is probably not. So there's your breakdown of black box problem I'm looking forward to the next one. We'll talk about more about X AI. Remember in the age of AI, asking doesn't work is not really enough.
We have to ask, do we understand how it works. That's our knowledge bite for the day. And now we're going to get to our guest Dr. Rupha, Dr.
Ruphaandra. We're going to talk a little bit. This is, we were talking to our pre-meeting, trying to figure out Rupha, what do we talk about today? It's super interesting.
Rupha, you came up with this great concept about why don't we talk about the why of innovation like why are we innovating why are we doing it. Rupha, start us off like why do we do this? I mean, why do we spend these hours building, developing, thinking about promoting, making these new technologies work? What's the why?
Fantastic question. And thank you again for bringing this topic up. And part of why I wanted to speak about this is, look, you're the bites of knowledge that you provide week after week. I learned so much from those.
And what I feel like is that if I'm able to learn these in bite-sized chunks, learn about these black boxes, when gray boxes, for example, the technology then doesn't become as scary. And now I can focus on what does the technology actually do and how is it going to benefit me as a practitioner and or benefit our patients and the entire healthcare system. And in terms of my why, I want to help you understand my background in my roles, I was in strategy and operations, it's some of the large corporate optometry retailers and moved into sort of a joint venture had grown my joint venture to about 20 practices. And I gave it all up and I gave it all up because I think what is coming down the pike is really going to put optometry in the center of this entire healthcare ecosystem.
And why is that important is it important because I'm an optometrist kind of, but I think what's more important is the understanding that we actually have a broken healthcare system right now. The entire ecosystem is brought up. I actually hugely, I agree, group, just a mess. And politics aside, I mean, when you start thinking about, you know, back when Bill Clinton was president or when Obama was president and even now, you know, what is the right healthcare model?
Well, quite frankly, they're all pretty terrible. We spend, we spend more money on healthcare than any other developed country yet we still continue to have the worst outcomes. And if I could throw in some statistics, I was just doing a paper on this couple of weeks ago, you know, we spend more per capita here in the United States than the next five countries combined and more than the bottom 55 combined. Wow.
And yet we finished number 99 out of 102 in terms of outcomes. Oh, my gosh, like any other business, if you did that, you'd be called out of business. Well, you know, I pride myself on not being 99 for anything. It's good.
You wouldn't be where you are if that was the case group. Yeah, so we've got a lot of work to do. And guess what, I care practitioners. We can be at the center of all this and really transfer healthcare and we can do so through data.
I agree. I mean, we talk about this all the time. My cohort, my co host on real talk, Ray on, you know, he says all the time he goes, date is the new goal. I mean, it's, it's the currency of the future.
It's not even money. It's data. And, you know, what, what you guys are doing at top con and you know, anybody hasn't met root before you got to go check him on a LinkedIn. I mean, his profile of what he's done is super impressive.
I felt like I'm totally inferior and insignificant after I look at his profile. And like, maybe he's done it all, but you know, you're one of those guys that you've got first and experience of that. If you don't have the data, it's really tough to make good strategic decisions. Yeah, and what's, and thank you for the kind words.
What we've all learned from the very first day in optometry school is the words evidence based medicine, whether an optometrist an ophthalmologist, your neurologist. I can tell you that those words evidence based medicine. What does the literature say? What does the evidence say?
Every time I take a CE or I talk to one of these PhD MDs, that's exactly what they're, what they're quoting on us. So what I like to do is I like to simplify it and just think of the data is evidence based medicine and how can we accelerate the use of that data to drive further insights and make us better practitioners and drive better health care outcomes. That's all we're really doing. You know, what I wonder about that route.
We've had this discussion, you know, some of our other guests and Ray, I'm going to discuss all the time is that. I'm not trying to be negative Nellie or what, you know, so negative bias on this and anyway, but I mean, I always think about. Yeah, we look at studies and, you know, we can say many studies are, are cooked before they ever start and that's a whole nother discussion all together another day. You know, but there's some great studies out there, but still we have N values of 200 500 1000 because you know, and then can we really look at genotype phenotype of those.
And so I agree with you. I think the future of medicine is based on better evidence based medicine, but to do that, we need to collect better, more diverse, more unbiased data. We don't really have systems to do that quite yet. And I always put an astrocon that because I know those are coming.
You know, but, but so when you say evidence based medicine, I wonder like, OK, so there's, you know, I mean the statistics vary, but somebody says it's there's somewhere between 16 80,000. And so we have a lot of. And so I think that's the thing, and so many people. And so this is really a way of making care providers between optometry and ophthalmology and the North American continent.
And if you figure, OK, half of them are seeing patients. So there's 30,000 times 20 patients a day. I mean, there's 600,000 patients every day. And so we have a lot of evidence based medicine.
And what's the outcomes when we do when we get back into your why question is, you know, I think the why is because we need to do better. We need to have precision medicine until we can get better, more universal data. We're just stuck on studies and not anything bad against studies. I mean, they're great and they give us a leg up, but we need to move the cog a lot faster than what we're moving it.
And so I think it's better myself. I mean, and you take a look at the number of the number of patients that are in some of these approved normative data sets that are in the instrumentation. It's not even 500. It's 394, 300 and they're FDA approved and all of those are sick eyes.
So, you know, I think what big data allows us to do then it allows us to start to establish these real world data sets. You know, why am I going to like you and I are similar in a lot of ways, but we're also different. And so our demographics are different. So when I'm comparing your eyes, let's say into an OCT normative data set and let's just say, for example, you know, there's 400 eyes.
Now we take a look, you know, to your point earlier. But I was a minus 10 before late, right? So right. So exactly.
So we take your your RX, your body mass index, your age, your blood pressure and all the other factors, genetic factors that you have. And then we got to compare it to 500 eyes. We'd be lucky if there's one eye in there that sort of matches your demographic and I'm different. And perfect.
And so now awesome there. No, thank you. And now if we have what if we had a million people in that data set or 10 million people in the day, or 100 million or 100 million, one billion, right? I mean, that's one eighth of the world population.
It's not even great. Now what if what have Scott's eyes, there's 2000 people with Scott's eyes in a real world data setting. So now we said, you do get to precision personalized medicine. So it's not just the imaging data.
It's the EO Mar data. It's some other longitudinal data that we could be able to track over time, the types of medications. But look, I'm not one of those ODs that has all the letters after their last name. I'm not one of those ODs that isn't as a PhD as well.
So what if, you know, to your point earlier, what if we could compute all of that stuff and even the best in the world can't take all this multimodal input and drive the types of insights that we might actually be missing and save not only site, but also lives from this. Yeah. And quality of life when you think about Alzheimer's, for example, I mean, you look at acologics and you know, I mean, you know, going off a little bit of attention for the audience. And we've had this talk before some other guests on the innovators, you know, is that you look what's happened in acologics five years ago.
I didn't even know if that word existed. I don't even think that was a topic. Now we're talking about diagnosing neurogygenitive disease from a retinal scan with an OCT. And that's insane, right?
I mean, you brought it up in the beginning. You said, maybe we have a future of time to play as such or I care in general plays such a bigger role than it ever has in the before because we're building detect something that not just saves vision AMD. But maybe all of a sudden we can change the structure of neurogygenitive disease. And I mean, we're just figuring out how to diagnosis, but once we start diagnosing more people route now all of a sudden, the PhDs and the scientists are going, well, hey, look at the genotype is for all of these different people.
There's the leal on, you know, chromosome 21, I don't know, making that stuff up. But you know, that all these people share. Well, maybe now we can figure out a nutritional protein that can be used. And fix that.
I mean, it's just, it's insane to think when we have more data, how much we could change the, the trajectory of health of humanity. And I love what you said. There was a number of great nuggets with what you just said there. Let's, let's start with how are we actually even using OCTs.
And then I want to get into, let's get back to the population health of in the United States. I have seen some research whereby and think about how optometry is using imaging. And yeah, I want to credit one of our competitors optos with the ultra wide field. And what they had done, you know, back in the early 2000s was go get an eye exam.
And guess what, you can, you can actually get a screening. Let's say it was $39 for that screening. Now, when we start to translate some of that stuff to OCTs and a lot of the ODS are using OCTs for wellness screenings. What we're finding, let's just take glaucoma as a disease in and of itself, that we're finding that 25 to 50% of the wellness screenings that we're doing with OCT for glaucoma are false positives.
So we're over, we're over referring and putting additional burden on on glaucoma already over and yes. And you can see the articles about the the the delay to get in. I was just that one of the one of the Asco meetings and I was just checking in with someone who had a retinal issue. She can't get into the doctor until February.
I mean, that's the kind of burden. And so we're we're burdening overburdening them with lower level cases. And then on top of it, you know, if you talk to someone like a Mike Shaglasy and glaucoma specials, he'll tell you that glaucoma is a multi factorial very complex disease. And so what are we doing?
We're missing 25% of the patients that actually do have glaucoma. And then over diagnosing people who don't have glaucoma. Exactly. And so, you know, to take all these data inputs, these biomarkers, these multimodal inputs, and then to be able to in your head be able to make a decision, you're probably a much better doctor than I am.
I take a look at it and you have an expert like Mike Shaglasy, we may come out with three different ideas and how to treat this patient. And so, and none of us know that we're right. Exactly. And none of us know that we're wrong.
Other than her ego say, well, of course, I'm right. But to your point, the more data we have, the better the more predictable, the detection of it, the prognosis, and then the outcomes are going to be, but we can't do that without evidence based medicine. And we have to look at evidence in a completely new way in a way that we just personally can't compute within our own minds. Well, it's so funny you said I was having this discussion with one of my student doctors other day.
And he said, how can you just look at this person and just know this answer? I said, well, because I've seen this exact setup. I've been in practice for 30 years. You know, I don't even know how many patients that is over, you know, 25 or 30 patients every day for 35 years or 30 years, whatever it is.
There's a lot of patients. I mean, there's just something here. But I said, that's innately biased because I hear something and I automatically think I'm right. And I don't always do the observation and that is just as dangerous as not knowing is when you think you know.
And you overlook the other observations, right? So there's why I said I told this young man I said you know the future is I believe, which is why you and I are having this conversation and we you and I are on the podiums all the time having these conversations is that this is why I think that we need to develop a model where we look at. Thousands or tens of thousands or hundreds of thousands of patients to find out. I go down a path because I thought I was right.
I go down a path because I was wrong and didn't even know I was wrong. Too very different past. But until we get to, as you say, the data driven model, you know, it's going to keep happening. We all think we're invaluable.
But we're wrong way more than we're right. I don't know about you, but I know I am. So, but I don't want to be 99. I prefer to be in the top three.
And you know what, the problem is not getting any better. It's actually getting worse. So six and 10 Americans. They have some type of chronic disease in this country.
One in five Americans has not seen their physician in over five years. So we're not even going and seeing the doctor. But what we do know is that, and here's the role that I care practitioners can play. We do know that patients or Americans are much, much more likely to see their ECP than they are to their physician.
So then the question becomes, so what if, you know, at the point of care, we're already doing these retinal images anyway. And what if we're then able to run these AI models through these retinal images and then come out with insights, both systemic as well as ocular. Now we're getting the patients earlier. So that's, you think about detection, prognosis, treatment, the outcomes that would happen and then removing the physician burden on, on sub specialty.
And then you think about the cost savings overall. I don't even know if that's calculable. Like it's just, that's a huge number. It's a huge number.
And if you just think about diabetes and cardiovascular disease alone, the numbers are just mind blowing, let alone everything else that's out there. I mean, those are just the big two, right? I mean, just, you know, and I was wondering, I don't think we're really having these conversations yet. But I mean, we're looking at what's happening in terms of proteins and the coroid and all that kind of stuff.
But I think when we can start looking at true blood vessel dynamic blood vessel flow and find out, hey, this person's, you know, got a smaller amount of cholesterol and their central retinal artery than or a larger amount. And maybe they're at risk of that and their central retinal artery. Maybe they're having it everywhere else. I mean, how many strokes and heart attacks are we going to prevent by getting that doctor that patient who hasn't seen a doctor in five years going, hey, this is your increased risk.
I think we need to get you there. You know, and, I don't know if we'll ever be able to calculate those financial numbers or what more importantly is the, we sustained and prevented death numbers. But I think that's the role we all took on when we decide to sign up to be in healthcare. Yeah, and what what I shared at the New England College of Automatries, and I know you had Howard Priscilla previously at their industry collaborative.
And I just talked a little bit about a diabetic screening pilot that we're doing and why are we doing is. So we know that there's 38 million Americans that have diabetes. About 28 or 29 million Americans actually know that they have diabetes, but here's the staggering number that really blew me away. And this is approximate.
I'm going to round up just a little bit here, but about 98 million Americans have prediabetes. And then when you think about those that actually have diabetes about one and four of those have diabetic eye disease and of those total population 1.84 million have vision threatening diabetic retinopathy. Now, this is the part that is the why I want to hit this one here 90% of all of this is present is preventable completely preventable through early detection and through better adherence, but the problem is they don't get follow up care. They don't get eye exams.
It's complicated for them. I don't really feel anything. It's doesn't really bother me. My doctors on the other side of town.
I don't really understand that I need to go see my eye doctor to get screened about 80% of patients were unaware that they actually needed a follow up eye exam. And you know too many medical appointments. We know that all of these comorbidities that go along with diabetes may be able to push the eye exam to the bottom. And then on top of it, the entire referral system between practitioners is completely broken 68% of eye doctors received no communication from the PCP.
They so they got this patient sitting in their chair. They kind of really don't know why they're then they kind of uncover their diabetic, but then how do you circle back and create this connected care model back to the PCP. And I actually have a slide which I showed at your wonderful event vision expo west, by the way, where it's a it's a picture of a doctor, a patient and exam chair, a chart trial lenses care, and they're actually in St. Alma, St.
Lam. That picture was taken in the year 1933 courtesy of the Wisconsin Historical Society, wow. And look, I am not saying nothing has changed in I care in the last 100 years. You and I both know.
I mean, you just mentioned OCTO, CTA, breakthroughs in surgical techniques and pharmacological techniques. But the barrier that we all face in our entire healthcare system is that we all work in a silo. And so go ahead. You just like literally I am writing this article like right over your done is I root thank you for bringing that up because I call it the island right here all.
We were trained and have somehow philosophically believed that the best way to provide care is to be on an island with only one inhabitant and that's our gray box. And tying all that gray box black box stuff back together from our knowledge bite, you know, but it's like I wonder, like how do we do that? I mean, it gets into your why like I'm going to get into the how like how do we fix the healthcare system to be more collaborative and less about what's more about what's right than who's right. I don't have the answer.
I'm always looking for opinions on that one. Well, nobody can do it alone. Nobody can solve this alone. I think that's what we're learning.
And whether it is partnering with some of the university medical systems, some of the large health systems, some of the payers, some of the folks that are in government agencies and the policy makers. And it's also big tech, you know, it is the Microsoft's of the world and you know that we've had a partnership with them since since 2024 and they've opened up a lot of doors for us. But what we realize at top con healthcare is this cannot be done in a silo and top con. We cannot do this alone.
So let's just start with building these data sets. So we if you probably saw and I know you mentioned it last week on the other podcast about the Institute for digital health idea that we have sort of relaunched and we have data sharing partners across the world now that are building in this data set. Well, who has the images? The images on a real world data set is not coming from ophthalmology.
Right. Those are sick eyes that they have from a real world screening population data set, which is extremely important if we're going to build these these I care models or these oculomics models that comes from optometry. It's very, very important that optometry stays connected. They understand the importance of it.
They share their data in a responsible way with an organization that then can democratize that data and provide that to whether it's research institutions, pharma companies to run run their trials. We don't have to wait 10 years and spend, you know, 10 million dollars on these trials. We can we can expedite these trials by finding the right patients the right time. And then we can get them to be in the right way and then the developers because you know, and I heard you talk a lot about generative AI.
And when did I see I don't think that's the future room. I don't think generative AI is the future. I think it's old school AI. And I guess my point is what generative AI is done for us.
It has accelerated the science of oculomics the last five years or so. Right. And so all of a sudden, what would we're just kind of refining the diabetic models. But to your point, we're starting to find, you know, whether it's Parkinson's Alzheimer's kidney disease preclampsia.
I mean, are you kidding me? Now what we're finding is that there may be PTSD and some some other things that we can actually detect through the eyes, whether it's, it's your people size movements, all of that stuff. So there is what again, what I'm getting at is this connected care. We need to start by sharing our data and then building this, you know, establishing to your point.
The trust and the responsibility that goes along with the AI models and then building these connect could care networks. So we're not living in the silo. But I care practitioners. Patients love going to the eye doctor.
And if they don't love going to the eye doctor, they need to go to the eye doctor because my glasses are scratched. They're broken. I'm seeing blurry amount of contacts. Right.
Or I'm 40 now. So we can play that central role and we cannot live in a silo anymore. And we got to develop these communication protocols with with other providers. Hope that makes sense.
I was going to go like, I totally agree with you. I, you know, if I was to design outline, I think we just clicked off most of the boxes there, you know, about what we wanted to talk about. It was great. I would just kind of wet, where it went.
But I think it's a universal truth is we just, you know, it's that saying we need more data. Data is the new gold or new oil is rayon says, you know, it's, we need more data. When we have more data, we can make better decisions. We need to do it transparently transparently as we talked about in the black box gray box, clear box, you know, philosophy.
But I just think that we, the last, you know, let's say five or six thousand years of medicine have all been done in an island. And I think we're finding that spot where technology is going to let us gather data from a great variety of sources, multimodal data, and be able to have the cognitive electronic cognitive horsepower to truly be able to process it by phenotype genotype. And, you know, I look for, I, I, I look at my young student doctors and go, you're going to see a world of medicine. That I can't even conceptualize like I, my brain's not there yet.
And it's going to be an exciting piece, right? I mean, there's going to be diseases that you and I'm a little older than you, by, well, a lot. But, you know, I mean, there's diseases that when I was there age, we didn't even know they were, we didn't know they existed and now we have treatments. I mean, I think like anti-VegF therapy, right.
I mean, used to be if you had wet AMD, we were just going to it was just a matter of time we're going to wait and watch you go blind and not be able to do anything about it. And now it's treatable disease, right? I mean, I think that we're going to see that with AMD. I really hope we're going to see it in glaucoma.
And we're all talking post year, say, I mean, you look at some of the technologies come out in the end to your site world. I think diseases like dry eyes and allergic conjunctivitis and I'll even venture a guess and say, I think it targeted, it's a tactic disease like care to conus. Some of these young student doctors, you guys who are just coming out of school or a couple years out. That's going to be a disease that we treat.
We don't manage. It's not going to be something we just manage, it's going to be something we solve the problem for in your lifetime. Maybe not my lifetime, but you were lifetime. I mean, we're, you and I get to see some really cool technology and get to talk about some really cool stuff and it excites me.
I mean, I get up at the morning. I'm like, what have you learned today? And so yeah, and I want to, I want to continue that conversation. I do want to circle back to something you said that I think may be important to your listeners and why do you not see gender of a eyes the future?
I think the gender of a eye, well, I mean, I think it is the future, but I think it's a lot further down the road future because generative AI is completely 100% dependent upon data, not data it creates, but raw data. And right now, you know, I'm not, I always get myself in trouble, Roops, the things are sent me up on this one, because I'm just going to get myself in trouble again. I get yelled at every after every one of these podcasts published somebody from some EHR calls me nails at me. But you know, the reality is until we open up the vault and we have access to standardized process, non bias data.
What's the gender of AI going to learn from, right? And we need a different model than what we have to gain the data that these gender of AI models can learn on because right now they're only getting, you know, we have certain companies to say, oh, but we're letting the data lose some like, but your data was a click box. It wasn't strategic algorithmic thinking it was I clicked a box. Well, that's biased by itself, right?
We didn't really have a strategy of how we collected the data. We just counted on somebody to go, well, that looks like the right box. Well, I'll click it and then we're teaching a generative AI to learn from it. And that's, that's kind of create problems that I'm, you know, I'm forecasting into crystal ball and saying, you know, I look forward to technology and I hope generative AI can solve a lot of these problems, but we need to give it better data.
Right now we don't have a better data source. Yes. And this is one of the reasons and it's a wonderful point. You may this is one of the reasons why we need to continue with our clinical training.
You know, the educational institutions that stuff is not going away. We need a big topic at the Neco conference was critical thinking skills. We need to continue to build the critical thinking skills develop and understand the differential diagnosis. And, you know, to your point, these AI models are basically clinical decision support tools for us.
It doesn't remove the doctor from the equation. And as you know, over time, you can get drift. Right. And so will it ever really remove the doctor from the equation.
No, it can, it can provide a number of double clicks or insights for us. But at the end of the day, doctors, we have to continue to be doctors and partner with other doctors to ensure that the outcomes are not a set of better. Yeah, AI is an augmentative tool to help us make better decisions. People, I say this all that people treat people.
That's absolutely. Absolutely. And I did want to point out a few, a couple of things that there are a number of AI models that are going through FDA approval. They're in development or going through approval process right now.
And so not only in our lifetime, but in the next 12 to 18 months. And people like you and I are going to start to see things we never thought we're going to be possible. And what may have taken seven to 10 years to develop the rate of acceleration of innovation. It is just, it's just continuing to accelerate.
And you know, what would normally take seven to 10 years is now, you know, on the device side, we're actually thinking 18 months, three years tops, maybe depending on how complicated it is, but more long lines about 18 months. And so we also have to be prepared not to lock ourselves into instrumentation that is not a colomax ready. I think, you know, in the past, it kind of shopped around for furniture and saying, well, we can mix and match and that stuff's good. But I think in the future, and this is where again, I don't want this to be a top-con infomercial.
But when you take a look at what top cons doing, we've evaluated over 650 different companies in the past two years. And then you kind of see that the press releases that seem like they're happening every week, every other week on LinkedIn with top con or with Ali Tafreshi, who's our CEO. But that has been the back story has been several years of work, along with vetting and we have an incubation and acceleration program within our innovation center. And so trying to help some of these companies guide them and push them across the finish line.
And now we're at this inflection point Scott, where we're going to start to see a number of these companies get across the finish line. So we'll start to see a lot of this. And it is going to change the way that we practice like now. And so I would say, a, make sure if you're in the, if you are in the market for replacing some instrumentation.
You have to understand what's coming down in the next 12 to 18 months and you have to be a calomics ready. Number two is what I hear from people is, well, it's just a lot of talk and then it's these data centers and nothing's profitable. But let me, let me share with you a real world scenario and you did a wonderful job you and Rahan did. It vision expo what West with the videos, painting what that picture could look like.
And so imagine a world where Leslie sitting in your chair and you say Leslie, what brings you in today. Well, I'm here for an eye exam and to get screened for heart disease. Well, who recommended you come in the American Heart Association and the AOA did. This is all hypothetical.
But I'm not saying the AOA, the HIV endorsed this. Yeah, this is just a hypothetical scenario I'm putting together. But this is what the future could look like in the role that we're playing in the perceived role that we're playing with Americans. Go get your eye exam and get screened for heart disease.
And guess what? It's not painful. Maybe they don't even have to take your blood pressure going in front of an exam. I'm going to do a wellness screening on you with a retinal image, but I'm going to run an AI model through it and then within 30 to 60 seconds.
I'm going to come up with insights about risk profile. And by the way, as long as you're getting your heart disease check, let's look for neural degenerative disease as well. Right in the same scan. Amazing.
I'm really happy to see that. I'm totally agree. You know, I'm super excited about the possibilities. I think you brought up a great point.
I want the audience to I want to repeat. You know, I think people, and especially in our industry, we're as healthcare providers, we're a little myopic. And we're a little conservative as a general rule. And we're a little slow going.
Let's just wait and see. I mean, even the I consider myself an innovator and a very much an early adopter at the at my slow stuff. And I'm even think I'm conservative and slow to the slow to pull the trigger. But I think we are.
I think people are just waiting to see, but I think we're going to be brought up a great point is the speed of evolution is no longer years or decades. It is months to small number of years. And we have to be willing as an industry to accept the fact that things are going to change. And we'll go back to that phrase.
I always say it. I'm not trying to scare the audience when I say this. But you brought up before it. A.I.
is not going to replace doctors. Never going to happen. Not in my lifetime. But the doctors who embrace A.I.
are going to replace the ones who don't. And that's not not trying to be a fear monger. I'm just state and the basic facts for the audience. You've got to keep abreast of what's happening.
It's going to be very quick. Like group you said, you know, I think if it is obvious, you guys can't see my hands, but you know, the slow punishes very vertical in terms of what we're going to see in the near future. Yeah. And we're finding the same thing.
And so, you know, when you think about A.I. and health care. It's already here. It's been around and who sort of leads this, you know, cardiology as well as radiology leads this space.
Who is going to far surpass both of those disciplines. I care is I care. We have huge amounts of data. Real world data to your point multimodal data.
It's the only organ on the body that can be three dimensionally image of the micron level and imagine low cost non invasive rapid through one simple scan. Where else can you get that done? It's you. You drop the mic, right?
That's that's that was it. Hey, Roop. So I will kind of call it stops here. We tried to do that 45 minutes or less.
Roop. This is a fantastic conversation. We got to do this again. Like I feel like I mean, I have an entire page of notes that I want to save for next session because I got like 10 different tangents.
I want to go off on that. I want to go off on that. I want to go off on that. Things that we covered.
This was super exciting. I really do appreciate your time. Roop. I know you're a busy guy and you're traveling all the place trying to educate.
The I care industry about all the new things that both TopCon has and things to be aware of. So thank you for your commitment to the industry over the last couple decades and really thank you so much for your time today because audience. I'm somebody who's in it all day every day. Some really great insight from Roop about, hey, what you need to the why you need to pay attention.
Yeah, if I could take just thank you again. If I could just celebrate you and Rehan just just for a minute here is that we need. People to get involved. We need leaders in this space and optometry to push it forward.
So we just don't end up spinning dials. That's the last thing I want to do in the last third of my career. And so it's going to take the five to 15% of leaders within the space to push it forward. And that starts by education.
It starts by creating forums for discussion. And we have to be relentless about it. And what you had mentioned earlier in the conversation was hey, oculomics. It's kind of the word of the year.
And three years ago, a lot of people didn't know what oculomics was, but now we kind of do. And just to understand the word and then what it can unleash for us. You know, in the entire healthcare system is the next step. And then the next step is how do I get involved?
How do I build the foundation? The next step is the activation of it. And that's where we can raise that middle 65% of practitioners to then level their game up to the top five to 15% and teard point. There's going to be folks here that just aren't going to do it.
But again, I want to circle this back to having an optometrist and an ophthalmologist hosting a podcast on AI on how we need to better work together to deliver better healthcare outcomes. And it's breakthrough. And I just appreciate what you guys are doing. Thank you very much.
And thanks audience for listening. Hopefully you guys learn something today. And hopefully we made, Rupin, I made you think about something and you're listening to this in the car on the way home or the way to work or in the bathroom or wherever you listen to your podcast. Hopefully we made you go, huh, I never thought of it that way because this group, as you said, the only way we're going to change things is education, open debate, open discussion and getting rid of the politics of what the initials are behind your name or where you work and understand we're all in this together to help change the way people live their lives.
Right. Thank you so much for attending everybody. You've been listening to Real Talk, an AI and I care, your weekly podcast to keep you informed about AI technologies revolutionizing I care.