This episode of Real Talk from AI and I Care is brought to you by Barty Software. Welcome to Real Talk on AI and I Care with your host Dr. Scott Morris. And I'm co-host Dr.
Ray Han 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. So Ray Han tell us about some of the news today or one of some of the things you're thinking about. Yeah.
So one article I read recently just the other day in the Wall Street Journal, which is probably among the top papers that I read. It's on my sort of everyday list. I was talking about how AI is going to usher in a new always on icon and always on workforce. And this is something that we've talked about before and it sort of makes sense.
And the idea basically is at least as it revolves to healthcare and AI care specifically, you know, as several years ago, pockets five o'clock or five thirty last patient scene, you're heading out the door and you may have your soft phone or paper on and that's what you meant by always on. And I think healthcare we've always sort of had an always on mentality. But this is this is different. It's as much more always on for the customer or the patient and it's always on for the doctor, but it may not be as negative.
It's not like you're always working. You may just have your bots or your agents who are working for you. So imagine a future where you your your phone calls that your clinic can always answer questions of your patients or there's an avatar of you that's always available to your patients who you do who you have operated on. Patients can always schedule things and we could already store to do that.
But maybe they can even get diagnostics that finance and treatment. Treatments that are always available for patients. This is happening in other parts of healthcare. If patients have a stroke, for instance, they come into the emergency room and it's a maybe a rural hospital that doesn't have the same type of diagnostic ability.
I mean, there are companies, there's a company called vis.ai vis.ai that's able to analyze images to make sure the patients are appropriately triaged. Financial markets are used to this. This is something that's going to permeate not just healthcare and it's really not just eye care, but but the entire society. I think financial markets and health care is going to have already sort of fault it.
But definitely, you know, you think about security systems or other consumer approaches. It's just going to be so ubiquitous that we're not going to notice it. And it's going to have multiple implications. I mean, one, we're just going to have such economic efficient machines.
I think it's just it's going to usperen a new era of what's possible. But the downside is we're going to have global competition. Scott, we're just talking about you. We were talking about a couple of weeks ago.
I mean, yeah, there's no reasons someone in a different time zone can't be competing with us. But we could also be competing for their customers and their patients. I don't think there will be any boundaries anymore. Not that time, not geography.
financially. Yeah, I mean, if you're the best at doing what you do and who knows what that measure will look like, you might be always on for the people who want to pay. And I like what you said, you know, it's not like we're going to I think that we're not going to work nine to five. I think we're going to go, Hey, on Fridays, you know, I work from eight to 10.
But at 10 o'clock, I may be having an evaluation with somebody in Dubai or somebody in, you know, Florida, even though I live in California or Colorado. You know, I mean, it's just. Oh, yeah, I can work patterns are going to change. I think that's great for a lot of us to sort of grew up in a gig economy, traditional nine to five, never really worked for us.
We want to work when it's available and patients will come in when they're available. That means that resource distribution will flatten, right? So what are your busiest days and planets? Oh, I just so we're actually doing this at night time.
My Tuesdays from 4 30 till about seven o'clock are just absolutely crazy. You know, and if I could do that, I mean, I think if I could work from four to seven four nights a week, I wouldn't have to work the rest of the time. I can do enough volume in that time period that I wouldn't need to work the other hours. That's right.
I think I'd be okay with that. Yeah. And you're all good. And there's always economy.
I think we'll flatten. But we'll flatten some more peak demand time because everyone's trying to squeeze in in those times. So I know for a lot of logic clinics, which are not open on Saturdays or Sundays. We're demanding patients come in taking time off work to come in and see us.
But it's a lot of other clinics. You know, Saturdays are super busy days. And because because customers are available, but what if you just had you're just always open? There's someone who's always willing to see you.
There's diagnostic equipment that's always available. And there are clinicians or agents or always able to interpret that. So I read this article and this is from from Wall Street Journal and it was quoting this is a new investment piece from Sequoia Capital, which is of course a large venture capital firm. And so this is and we've talked about a lot, but it was nice to read it in such a in a well thought out way and happy to share the link on the AI and I could website.
Yeah, I think and this is a good leading. I mean, I want to cover a little bit of all kind of some other stuff right now are normal like let's get educated, but it's an late at night. We're going to talk about the patient experience and you know, it's not just about always on being good for providers. It might be really great for lots of patients.
Like you said, so we'll we'll say that we'll cover that here in a few minutes. Yeah. I was going to kind of cover. We always try to talk a little bit about, you know, some new things right?
Some will what are different concepts that we need to learn about. And so tonight I want to talk about deep learning and so I know we've mentioned a couple of the other ones and so I want to kind of go into a little bit more detail is when we think of AI, AI is a subject right and then underneath that subject is a couple different breakdowns, but one of them is machine learning and how do the machines actually learn and part of that, a sub-paint, it's piece of that is what's called deep learning and you know, in the reality is deep learning is really just mimicking kind of our own human structure function of the brain. Instead of a neural network that's truly neuronal, we have an artificial neural network and these networks are really just designed to kind of recognize patterns and make decisions kind of much like human thought processes, but just on a lot larger and faster scale. And so people have asked us what's got explained that in English.
And so, you know, I've been kind of over the last couple of months been trying to think about what it is and and I want you to think of it like this is and we're going to use it kind of in a little bit of the concept of machine vision, but let's say you want to teach a computer to recognize a picture of a dog, right? My dog's looking at me right now like you're going to feed me, but let's say you want to teach a machine to look at or recognize a picture of a dog. Well, deep learning is kind of a way to use a system that as we said, it's kind of inspired by the human brain and I want you to think of it as a stack of interconnected layers very much a three-dimensional neural spider web if you want to think like that. Well, when you show a machine a picture, it goes through let's call it the first layer.
Let's think of it as the outside network, the outside layer of this big three-dimensional wall. And then this layer might learn to spot simple things like an edge or a corner or a color. And that output then goes into the next layer down, which then takes the images it got and the things it learned and says, well, now let's combine that with slay, the other maybe more complex shapes like a circle or an oval like a pupil or a long curved thing like a tail. And this information then goes from one section that learns a little bit to the next section of more complex shapes to more complex shapes to more complex shapes.
And as it passes through these layers, let's think as it gets deeper and deeper so deeper learning. Each layer then kind of learns to recognize increasingly complex patterns by combining the patterns it found in the previous layer. And so each subsequent layer recognizes more and more things like and eventually get that what recognizes an eye or an ear or a whisker or a tail or a tongue wagging. And it takes all this different things.
And by the time you get to the last layer, you're really combining all these features and go, well, if it has droopy eyes and upright or floppy ears and you know, it's got a tongue wagging and it's got a snout and it's kind of furry and it's got a tail. Well, is it a cat or is it a dog? And so it makes a guess and says, okay, this is a dog. Well, if you take that and you do that enough times like in humans, we did that as infants growing up and learning and going, no, that's not a ball.
That's a block. And you know, our parents went, no, not ball block. Well, the machine doesn't kind of has that same feedback in terms of different forms of learning. We learned about that a couple weeks ago.
But to get good, the system needs thousands or maybe millions of examples of pictures of dogs. Both dogs and maybe not dogs, maybe cats. And every time it makes this guess, it says, okay, was I right? And if it's wrong, it goes, well, I need to adjust this.
I didn't read it right. And over time, it just gets really very accurate. This is, I'm right every time. So deep learning and kind of thinking about like, it's essentially a machine system with just lots and lots of layers kind of like our brain that each layer learns kind of to identify slightly more complex patterns and puts those together.
Similar to just the way our brains are from experience. And now you can use that sort of, we use the example machine vision, but you could use it for speech or translating languages or let's get really crazy and think like self-driving cars or maybe an autonomous slit lamp or an OCT. Huh. How to recognition is still pattern recognition?
It's just the level of training we have to do. So hopefully we learned a little bit about deep learning. And I've hopefully provided an example of something that makes sense to people. But you can learn about you can look it up.
You can do those things. But if you think about the way the brain works, this is just an electronic non-biological version of the way our brain works. This episode is brought to you by Barty, the AI powered all-in-one EHR built for Icare. With AI scribe, voice over internet phones, websites, payment processing, and Rcm all bundled into one system, Barty is redefining modern Icare.
It's been less time clicking and more time with patience at Barty.com. That's absolutely fantastic. And I love the analogy to a kid growing up. And that's how we were basically the brain is a pattern recognized, recognizing machine.
Hey, let's go back, you know, because we want to talk and you and I have these conversations all the time. But I've really been thinking a lot about how does AI going to affect the patient experience? Like you were talking earlier about some of stuff like maybe it's going to change the way we provide care. But what if we flip and think about what are we all in this for?
We're there to take care of people. If there are no people that take care of we don't we don't even I don't really have much to do as physicians. And so the question is is that you know what do we think the patient experience is going to look like? I mean how is AI going to impact that?
So I've been really really really thinking hard about that lately about what does the eye care patient of the experience of the future look like? And you know, and you and I've talked about I'm going to do this at Vision Expo meeting in 2025 in Las Vegas. I'm going to do the patient experience. I'm going to show some examples of this.
So it's been really in my brain a lot about, you know, I think it's going to come in phases. I don't think it's all going to happen at once. But I think and it's already starting to happen. But you know, like you said, always aren't.
Well, what does that mean for a patient? I think about my last patient today, he's a radiologist at one of the hospitals. And this is only night. He's like, Scott, this is the only night I get out before eight o'clock at night that I can actually be here to see people.
And I start every day at seven and you don't work on weekends. And so I really like you guys. And this is my night. This is the only night I can do this.
I've been looking forward to this for four and a half months. And you're like, no, not first of all, thank you. Yeah. But was that really the patient experience I want to give that they have to wait four and a half months to see me?
And so let's talk a little bit about that. I mean, where do you think? Well, I think the patient is going to change. Yeah, I think it's a short term, no longer term view.
For the short term, I think about all the pain points that patients, flash customers feel. And that's where I think there's going to be a lot of innovation. For example, a pain point is simply, there's two pain points. One, identifying the doctor, to making the appointment.
So let's just start from the beginning. Identifying the doctor. Okay. So how do you know which I get professionally?
You're going to see, you're going to go online and try it right now. The state of the art is you do some Google searching. Well, in the future, I think we're going to have Scott's going to have an agent. I'm going to have an agent.
My mom's going to have an agent. And we were probably just asked based off who we are, who are priming for doctors and maybe what our problem is, it's going to start interacting with those doctor agents. And so there's going to be this whole sort of underworld that sort of happens. We all answered with Will said a whole.
Yeah, we're dark webbed after we just talked about about. Yeah. Yeah. I think agents are going to talk to agents and say, who's the best match?
I think, you know, we're going to look at our personalities and go, which your personality is a provider and what's that patient's personality? And I want to get something to sing my personality. I think there's going to be a lot of preference and profile. And I mean that in a good way that goes on behind the scenes to find somebody because I don't I don't I'm assuming everybody is listening, you know, we always have that one patient in the day that is like, God, I just don't connect.
Like I just I try everything. I just can't connect with that person the way I'd like to. And we're just not a good match. You know, I'm one of those today and she's just probably needs to see somebody else because we don't we don't connect.
But maybe our agents will help us fair it that out. So we find the people we do connect with. And it's probably more of a longer and by some point I'd say longer term. I mean, maybe a couple of years I have no idea.
But I think the stuff here and now within the next six months, certainly there's already companies that are working on just getting an appointment in a reasonable time. If you're a patient and you're having symptoms that are consistent with, you know, maybe retinal issues. And maybe you already saw and you know, maybe there you probably should see a retinal specialist or maybe not. It could fair it's in those symptoms and say, well, this person may be the best.
Maybe you have some families to retinal skis and maybe you're better off seeing a retinal specialist. And if they're available, that's probably the best person to see. So I think rather than seeing someone like an occhioplastics, right? And that's happened where the patient is totally mismatched from a from a standpoint, you're just not seeing the right specialist.
And that's just through all of health care. So I think those types of things are going to happen very quickly where they're going to manage the right patient with the right doctor and being able to schedule it perfectly without the pain points that we have right now or literally you're on average out of the screen as that statistic. It's like 15 to 20 minutes that a patient will call and maybe multiple phone calls to a doctor's office to try to make an appointment. So I think those type of very sort of low hanging fruit that's going to be the first thing we're going to see.
And that's going to be the first thing that patient sees how AI is improving their overall experience. I agree. And I you said, I think you're I think we're three or four years away from that at max. I mean, that's coming really soon.
But I'm going to challenge you, Rayhan. I want to talk about I think the problem happens earlier than that. I think the problem happens in the awareness stage, right? So I always think of what is the patient flow, patient experience looks like?
And about of the third of the flow happens before they ever even talk to us. Right? First of all, they have this awareness of I have this symptom. Is this something I should be worried about?
And we go back to the agents. And I think that we're going to have agents that go, well, let me take a really thorough comprehensive history and find out if this is something you need to worry about. And if it is, maybe we need to fast track this. And if it's not, maybe it can wait a week or two or three or four or whatever.
You know, I think the awareness is there. I think also the other states to talk about is once you though there's a you're aware of an issue. And then you as you said, you have to look for solution, which doctors best for me. I think this is a little bit more long term.
But I don't think it's going to be on Google reviews. I think it's going to be based on evidence based medicine. I mean, I think that as AI moves further and further down the path, we're going to have scores as providers that say, Hey, you know, I'm really good at treating this. I'm a four out of four I treat this, but I'm a three out of four for treating this.
Right? And maybe part of that matching system is, Hey, I have, I have issue a, I should go see somebody as a four out of four of an issue. If they meet the rest of my logistics like they're within 20 miles or 10 miles or five miles and they're under my insurance coverage. If insurance even exists at that point, you know, and so I think there's the awareness issue.
Well, is this is problem that's serious? A solution who can provide the solution? And then as you kind of said, do I like that person? Am I attracted to that person?
Do we match or are our goals the same? You know, I'm a pretty Eastern Western medicine kind of doc. I think both work depending on the situation. And I had a patient today just lit into me because I even suggested that she took an antihistamine.
I mean, just I don't ever take drugs. Okay. You know, so she probably would have been matched with another doc in the area who is totally a homopathic and a homopathic doc. So, you know, all of that I think getting you through to the walk, you know, this through the scheduling stage and walk in the door.
I think that whole patient experience is really relatively low hanging fruit. I agree. I think that's all coming pretty quick. Yeah.
Certainly there would be challenges. I think that the accuracy and quality of the data, especially in terms of when you're talking about ranking doctors or scoring doctors or even having, is it is good necessarily for patients to be matched with a doctor that they will like and maybe sometimes they should have been matched with some of them. So, it's not the panacea here. It's not the solve.
It's not going to solve all the problems. But definitely low hanging fruit and some of these operational issues are sure. And I totally agree with you. A lot of these issues is even before they pick up the phone or a cell phone or they're looking at their medical asses to schedule for them, it's going to be at their awareness stage and totally agree that that's actually earlier on in the patient journey before they actually even recognize themselves to the patient.
But they know taking to the next stage, you know, and I think maybe this is still a little bit pre-visit, but definitely is intravisit is that so, you know, I think that you're going to make a diagnosis or you're going to develop a treatment plan. And so many the times we go, here's your diagnosis and here's what we're going to do for treatment. And we walk out the room and we cover all that in 30 seconds because that's the 20th patient today. We've had a similar modality for group.
For that patient, that was the one and only time. And if they were at all distracted, they missed it. Right? And I think that what we talk about it being an agent or we talk about it from an educational perspective like, hey, maybe they should have always on access to, well, this was my diagnosis.
Let me watch this video. And then of course they get home and they're spouss ass what they have. And they're like, I don't know something. I had to take drops.
I don't know what exactly. So, you know, I think that next step of how do we communicate 24, seven, they can always watch it. They can always watch what medication you put them on. They can learn a little bit about the drug and maybe what the side effects are.
And, you know, when I got a lady today, she got a horrible allergic conjunctivitis. And you know, when you put predacitate on the first day, they're going to have all that white precipitated steroid in the corner of the eye. And I literally didn't think to tell her about it until she was walking up the door. I said, oh, wait, whoa, whoa, whoa, wait, wait, tomorrow, you're going to have all this kind of white gunk.
That's normal. Don't worry about that. It's going to be okay in a day or two, but it will be there. Now, if I hadn't said that, she'd walked out the door and thought, hey, I was crazy tomorrow.
And B is I got the wrong diagnosis and C is she's out of control and he's going to see another doctor. You know, so maybe that stuff is pre-baked into education that doesn't necessarily happen in the examiner in the office, but something they have 24-7 access to. That changes the experience in that when they have it, then they don't have this, oh, my God, something's not working. And that fear factor that comes post visit, now maybe we can educate them all the way along.
And I guess you can even take one step further and go, now people could also pre-what would be the word for that. I would have to come up with some cool word for this one. But almost like a they can pre-shop what their options are for I care, right? I mean, maybe people go, I want to try multi-focals, what multifocus is going to be best, multifocal contacts and be best for me or what multifocal IOL is going to be best for me.
Instead of it, I'm sure you've had that conversation with patients, hundreds of not thousands of times about well, which multifocus best for you, maybe we can pre-educate. Maybe our agents can pre-educate the patient based on their preferences about what might be important to them. And everybody has a better outcome. Yeah, it's definitely happened.
I think it's an exciting time, especially for both physicians and patients. One of the biggest draws in AI in general is healthcare and this is going to be one-hopes, a really exciting time for us and better for our patients overall. Always a good conversation, my friend. One-hope.
Talk to you next week. 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. We would like to thank Birdie Software for supporting this episode of Real Talk for AI and I care.