Welcome to Real Talk on AI and I care with your host Dr. Scott Morris. And I'm Carlos 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 Dr.
Scott Morris. And today we have a special guest, Easy Añama, who's a really getting to be a really good friend of mine, one of the young leaders of this industry, and we're so much looking forward to having him on the show. As you guys know, every week we try to break down some complex tech term into bite-sized insights. And we call this the knowledge bite.
And so today we're going to talk about the fuel of the 21st century data. And I know my co-host Rayhan always says, well, data is the new oil. And so we're going to use the fuel kind of symbolism here. Added simply data is just information.
You know, it's your heart rate. The date of your last flu shot or maybe a blood pressure reader and IOP reading. But when we talk about big data, we aren't just talking about a lot of information. We're talking about the three V's.
Volume. So we have massive amounts of data. I mean, think about millions and billions of pieces of information about every single human being on the planet. And then there's velocity that data generated in real time.
Because obviously knowing what your pressure is today is a lot more important in terms of treatment than it was three months ago. And then last but not least, the third V is variety. So that's everything from your doctor, you know, from your handwritten notes to live feeds, to wearable information, to purchasing patterns, to feedback from a history. All of that is just a variety of different kinds of data.
And you can think of a data as like a single drop of water or maybe as we'll call it a single drop of oil. And then big data is well, all the petroleum reserves, well, everywhere in the world. On its own, a drop tells you very little, but in that ocean of oil, it has currents and patterns and a whole lot of power and the ability to create power. So why does this matter for us?
Well, it's all about efficacy. Which is a fancy way of saying, well, how well can we make things actually work? In the old days, we relied on generalized clinical trials, but with big data, we can move towards what's called precision medicine. And we talked about that a couple of our different podcasts by analyzing, let's say genetic data and lifestyle habits of millions of people.
AI can help us predict which specific medication will work for that specific patient based on that specific unique biology rather than use a one size fits all approach. We've talked about before is that big data will build to help us treat on genotype phenotype rather than just, hey, this is what we do for everybody. It also is going to help all of our practices run like a well oil machine by analyzing years of different data. You know, we can predict whether there's going to be a surge of E.K.C.
even before it hits and allow us to kind of prepare for those kinds of things. But more importantly, it's probably going to have a germostromatic effect on what's called the standard of care. And this is about, let's say, the quality of experiences of the patient. Big data is going to allow for what's called predictive analytics.
And imagine wearing a wearable device that monitors your IOP, for example, or your blood sugar through a contact. Instead of waiting for glaucoma to happen, big data algorithms can spot tiny little invisible changes within pressure or optic nerve or ocular pulse amplitude. And alert the provider days, weeks, months, and advance of having physiological changes. So what we're going to do with big data is we're going to move from a reactive system where we fix things once they break, i.e., our current system, to a more proactive system where we prevent the break from happening in the first place.
That's the true magic of big data. It's going to turn numbers into, let's say, more birthdays and more images that we get to see on every one of those birthdays, more milestones and more time with the people we love and care about. So that's your now's my for today. Data isn't just code.
It is fuel to a healthier future. Now, let's get on with the rest of the show. Let's welcome easy to the show. My friend, it's good to have you.
I can't wait to see where this conversation is going to go. For the audience, easy and I talk all the time. Usually it's late at night early in the morning and you know 45 minutes later, like, okay, we saw most of the world's problems. As I said in the introduction, easy is one of those guys who's just very thought-provoking, sees it from a little bit different viewpoint, maybe then I as a seasoned experienced warps veteran does.
He always brings a fresh perspective to the way I see the world. Easy welcome to welcome to real talk. Thanks, Scott. It's great to be here.
We have to get one of these on recording for other people can chime in and see what they think about our conversations. And seriously, obviously, we do easy and I challenge each other all the time and he makes me rethink a lot of stuff that I think. So easy. Today we kind of you and I even I go through this process every time we have a conversation, like, let's just pick a subject and go from there.
And I kind of in our pre meeting, we're talking a little bit about kind of some of the, the, I don't want to say myths of what we do, but kind of some of the oversight of what we do. You know, and I feel like you and I feel the same way and some things is that a lot of what we do are a series of systems we call that workflow. And sometimes we have human in the loop and sometimes we have human should be in the loop and some says we're going to end up with humans not in the loop. Tell me where you see, tell me what you see as some of the biggest myths and challenges of AI.
And then maybe we'll get into what are some of the opportunities. So tell me, first of all, I mean, you're a young super intelligent guy. I think you probably know more about it and I know. And so tell me, I mean, when you're saying they're talking to people, what are the myths you hear and then what makes you just laugh and go, that's not the way that is.
I think the big myth is that there is some type of human replacement element to AI. I mean, you use it enough, you see all of the weird outcomes and weird outputs you can get from AI. And yet they call us hallucinations. I just call them mistakes.
And yet somehow we then have some people have this fear that it will replace them or somehow disrupt them. And it makes it's challenging for me because it's so interesting to think about like you can use one of these tools and say yes, a human is definitely better. But then when you hire a human, you would never let it just free for all not have protocols, not have guardrails, not have like a action plan for how you would like them to work. And yet you if people implement AI and hope that it just solves all the problems and has all the answers.
And so that's like a huge thing. I think it's interesting. The second thing I think is interesting is that work flows. I feel like the framing of the problem within a practice is more important than whatever the AI will do for you.
And so so many people are going to find that they enjoy AI simply because they met the right company to help them to understand the flow of their practice. Or they're not or they're going to find they don't know what their problem is. And they're going to choose a company that solves a problem they don't have. Right optimistic take.
Of course, truly more likely, especially early on, people are just solving problems. Correct. I knew that. And but the hope is that eventually, you know, the bottom line you realize, whatever you did didn't work, you go back and you and you get these workflows and you understand your business a lot better.
Or just you mean the scientific method, try until you fail them and trying. Yeah, that's crazy. I don't know how much it works though with the business. I hope.
I hope people do implement something like that because then once the workflows down, oh my god, you can implement anytime solution and AI, especially since we talk about what I see is just information technology. So it's like, okay, this is a tool that can improve my understanding of my use of information. Well, when it comes to your business, how much information do you know about your business? Most people know shockingly little actually and I sometimes think I knew more before each are because I had to sit there and folks on it all day and now I'm like, oh, each are tracks that we realize each are going to track a damn thing.
You know, so which is kind of it's very true. Like you know more it's like I read a. I think was a headline that said people who use AI for education, they have better results, but when they stop using it, they have worse results. We're recall for sure.
Well, I mean, that is essentially every tool ever in my opinion, when you use your phone, you save images, you don't remember the events you took photos of. It's like that is the purpose of technology to augment our capacity and our ability. So I don't think that's as much of a bad thing, but it is something to be aware of because. If you're getting this technology, you should know that yes, I'm going to operate a little less in the know that I was before I better make sure it's actually doing that thing I better make sure like the HR is logging or tracking whatnot.
Well, let's break this down, you know, you and I in two weeks for I don't know if this is going to publish this week or next week, so it could be for the odd to could be this this coming week easy and I are going to be in seco at seco in Atlanta talking about reimagining how I care works. And we're spent a lot of time talking about workflow and so let's kind of drill down there a little bit easier. I think that for the audience, you know, first of all understand, you know, we call it SWOT analysis and that's as old as time is what's your strength weakness opportunities threats. And I think that what happens is the place you start when it comes to figure out is AI going to be for me and is it going to be a augmentative tool.
I love that word easy. You know, that's my favorite term augmentative tool. Is that you know, is this going to be an augmentative tool to help me well before you invest a single cent in this or a single second sent or second into this process. First thing you need to do is figure out what's my workflow number one and that number two is what's broke.
You know, what could what what tool do I need to make this part of my patient experience workflow better. Talk to me a little bit about where you see the weaknesses in our current workflow and some of the tools that are out there for the audience that might build a fix some of those workflow problems. Let's start with that. I think the I'm going to take this two part workflow and tools and make it for right before the workflow.
I think there's a huge aspect of like understanding that it's more than just a static internal workflow. There's external vendors external partners external software tools external people, including the patient and the payer and all of this things lots of variables. Lots of variables that go into understanding your workflow and so if you fully map it out it doesn't look like patient comes in intake see the patient. It doesn't look like that there's a lot of eros in yeah a lot of eros in there's computers there's vendors and there's so so when you map it out I think being very specific and almost bulky is useful understanding your workflow because then when it comes to adding tools you realize there's probably a step before that which is removing things like what doesn't need to be here.
Oh my gosh all I just okay because you can't see us i'm just like banging on my head of the wienst wall right now because I was just having this conversation with student doctors yesterday about so much of the stuff that were taught in school were taught as a huge data set of stuff we need to know but we do so many things that are just totally unnecessary they don't contribute to the goal of that particular step so why are we doing it. Like it's not solving the goal of this particular workflow step and get rid of it right so I agree with you you know the fastest way to improve your workflow efficiency is get rid of the crap you don't need to do right and then figure out what tool do I need to do to make the more efficient process you just made better and more efficient correct because almost always you have a really a freebie in my opinion which is like I want to remove that but it'll be something small it'll be like. Small it'll be like that but it's probably the one of the first things you can use technology in this case AI or these AI tools to improve your workflow that but so it's just like it's to me you can start there you know it's like well I don't want to really get rid of this tool because it'll be something so small. Well why why don't you want to get rid of that tool correct is it emotional is it rational is financial is it you know you got to ask yourself those tough questions and too many times like oh let's keep doing what we're doing it's working.
Yeah but it's not working real well yeah so and the working working well and those are not the same process not at all and to your point working well is the goal the goal in my opinion the goal of implementation of some of these tools is to get people to work at their best across time consistently that that is what you get you get to show up as your best self because of the help you have and your staff does too so I think that's another thing to consider. Okay so step one just for paraphrase so step one identify what your workflow is and get rid of the inefficiency steps that you have in the way we're doing it now step two determine is there a tool out there that can help make it better AI embedded or not right so what does that process look like easy. I think that process looks like what isn't going to blow up in my mind. After that I mean I love experimentation probably a little too much so I definitely you should consider what you just don't like or what people your practice don't like to do like the emotion aside what isn't going to have a huge irreversible outcome and three what's going to move the needle for your goals or your key metrics and your brain.
I think like usually those are all kind of the same thing like I think I can't really wear those things. Yeah I agree in fact actually when we're at Seaco I'm going to break this down he's having showed you the slide yet but you know my friend Dr. John Repacus and I do used to do a presentation a couple years ago called what difference is a minute make and we broke it down and you know when we think about what could be saved in terms of money. Time is the most important form of money and so you know we're going to show in this that I'll break it all down is that every minute costs about seven dollars that's what a minute cost so if you can break it down and delete things like step one or number two final process to make what you're new and it reviewed version is making a minute faster that's a seven dollar revenue stream right that's that's a lot of money and so.
You know I agree with you is find out what what's the overhead that's a pain in the butt that nobody wants to do that's critical to mission success. But there might be a better tool out there and you about it up early you know there's lots of vendors and all the vendors of industry probably guys listening you're making tools to solve problems if it's not solving a problem why are you making the tool you know but so you're you're you are solving a problem whether be phone calls or whether it be recall or whether it be. Coverage and benefit eligibility determination or whether it be you know differential diagnosis or pattern recognition and OCT you're making tools to help make everybody's job easier and so it's just step one. So make your pure process more efficient by cutting out stuff step two identify what processes your biggest headache and fix that process with some other tool right right and almost your.
I imagine a lot of these days I come in so early that you just tell me your problems you think solve it. So I'm supposed to be sold the tech and being sold some problem that might be nice just if you really understood the problem you're trying to solve and you know that it'll help go for that in my opinion. I agree so for the audience easy just told you the million dollar secret right there he just just laid it out is. Why would you need to make your job better faster easier don't buy something that doesn't do those three things right and with air more towards a service type of technology because it can be personalized and and it should be useful to you in a very specific way that's the magic of it right so be very specific understand your problem and I think don't change I know I'm being brought we go more details no yeah I think it's good I mean sometimes you need to start with the big picture easy and I think that's great.
So I'm going to go back to kind of another question I asked you earlier so think about so what and I'll be interested to see your response first is what my response is so what do you think is the biggest problem in our current workflow. Data silos. Speaking of my language there my friends go ahead and explain what you mean by data silos because for the audience you've heard Ray on and I talk about this at nausea so I'm going to get off my band of my my lectern here and let easy explain what the challenge with data silos are the challenge with dance also I see it is that just like you have all of these tools you have all this equipment you have all this technology in practice. There's no.
Current clean way to get them to all flow together so that you can use them all in utilize them all it's like on that easy and working on that. But it's like if you try to I don't know do anything trying to cook a meal you don't want your eggs upstairs you don't want you know you're seasoning in the basement like that's kind of what it's like and the meat out in the yard and the vegetables are at the grocery store right so right and making works to get them you must use a different. Key is it is large set of keys that you have to look to it's like craziness so the way I see by them with different currency and speaking a different language each person it's and that is the key part the speaking a different language I think that you know with a lot of these tools especially large language models which started out with like Google translated. You know processing the different forms of communicating like I think that all a lot of the devices in tech we use in practice it's siloed it's separate and every tool speaks a different language right and so one of the problem with having 300 different vendors are actually there's almost 580 vendors now and I care I counted the other day on us doing the project for something and unfortunately none of them talk to each other.
Correct and I don't think the HR's make it easier and which is it's so countertuitive to how. People practice because you want to take each part each tool and combine it to make an assessment on a patient or provide the best care so why it's kind of like it's kind of like an algorithm right if you think about input there's processing and there's output and that outputs then used for the next sequence of input processing output except the fact our industry doesn't work that way you know we have an output and then we got to translate to a different language. And put a different term and go through summary it's not raw data output that goes to raw data input there's processing in between and where do you store it all your brain well and you know that has to see where I think the biggest problem is called the gray box right we talk about the black box of AI I think the biggest problem is the gray box or what I often refer to the end of one right is that we're all processing this data or brain we take an output from this device for that device or this history that is. And we process it through that big gray gray box that we have between our ears but that box is so full of bias and air and exhaustion and frustration and you know I mean there's I think I count as 136 kinds of bias and as I was building algorithms for my project you know I encountered every single one of them right it's all bias and so that processing unit has got lots of problems so I was going to kick up that what a one of the things that I'm going to do is I'm going to do that.
I'm going to kick up that what a what providers in healthcare I just I can't provide us are expected to retain regurgitate review you know analyze spit out a new outcome is insane and that information is doubling every three months I mean it's just how we think cognitively that the gray brain the gray box can keep up I think we're a little crazy on that one right is there's no way. I don't know where and I can. I can't. I can't.
So interesting because you have like you're doing math. You have to just recall some specific fact that you have to do maybe some fashion of styling and build it's just like the the mouth of. Behavioral modification psychology management. And so understanding which insurance does what thing and you know what if I do this I'm going to need prior authorization so maybe I should do something else and is there even evidence based about that right.
And it's not even necessarily just a bad thing I think it's just it doesn't really have to be that way because if you think about like autopilot you can fly a plane and not be an autopilot but if you're going to be a commercial pilot chances or you're going to want to want to. That's just how it goes and what is all autopilot do mostly it controls and helps you process all the information is going on in a plane. I don't know this for sure these are just speculative but the point is like what does that look like in a practice and I think how many of us are on autopilot we can't not be in applying anything new. There's nothing well as far as the gray box short but in a practice I think there's a possibility that you could have some of these a lot of these processes.
Well, handled themselves you get to the point where you're the judgment your judgment matters more. And then you're all right we had a podcast of the day and it'll probably go on here before this one does called connectors right I think that our new rule part of our new role is going to be connecting information we're already connectors right where that we take output we connect it together we process and put it out but I think we're going to be connecting a lot of that output to our patients in a way they can understand I think that we're going to go from data gathers to data interpreters to data connectors. That's the role process we're going to undergo and I will say you don't buy it you it doesn't matter what how many stars it has you don't really buy anything off Amazon without reading the reviews you want to see other people think about it like humans love humans and that's the name of the time. Most of the time human love the right so the goals that a lot of these tools can help strengthen that human human connection I know it sounds very out there but it is the truth as far as economic well you know you started the beginning of this and you said you're very first phrase I think your second phrase was AI is never going to replace us and I agree it's never going to replace not my lifetime probably not your lifetime easy year a few decades younger than me it's never going to replace us but it is going to augment us for sure and then the second phrase you is but it will disrupt everything and I completely agree it will disrupt everything that we do because what we do is easy to disrupt because it's not very efficient.
And that's not supposed to be designed right like the inefficient things are eventually supposed to be disrupted by technology so if you choose to hold on to that then you're essentially just committing to being disrupted yourself as well because I imagine some people are going to adopt AI and they will get some some advantage and people aren't and they won't. So I don't agree with that like that that's he's very harsh but it is the truth that doesn't necessarily mean they won't survive but it be the equivalent of never adopting a auto refractor or even a. But we're going to have people we're going to have to evolve easy if we don't evolve we're going to be eradicated I mean it's. We just have to evolve with technology and that you know I think this is where you and I talked about this countless hours and you know if the audience you guys hear this over and over and over again when every single person on the podcast says the same thing maybe we should be listening is you know we have to evolve to become more effective and more efficient and we have tools to do.
It's not going to be the either places going to be the it's going to be the practices the doctors and health care systems that evolve the replace the ones that don't. And honestly this conversation that I was sometimes we're like I don't want to. It feels obvious all those things aren't always obvious but like just go back in time and look at a commentary 50 years ago go back look at us 100 years ago you would never practice like that so clearly the tools are going to involve how we do it changes. I mean to me it's just what it is you know so I agree I you know I agree so we try to keep these somewhat in the 20 to 30 minute range so I'm going to kind of ask you one big question and I think this is real talk easy part one of probably about 10 that we're going to have because easy and I can talk about this for hours.
So when you look at we talked about the myth of AI what's what's the great hope. What's what's that you look at you're you involved with it probably more than anybody in our industry on a daily basis so when you look at this and go what what drives you to go this is this is the great hope of AI. Deep personal understanding of individual specifics and an extent and any crazy awesome and clear way of communicating that to another person. Wow so now what I thought you were going to say you got me off guard on that one so I'm a patient deeply understanding my body and deeply understanding what's going on and how I feel about how I look.
I feel about how I look at the car glasses on or anything and being able to communicate that to my provider my provider to then understand my problems and be able to provide the solutions for me like just it sounds crazy but if you think about it current. current. All right, large language models at least have a deep understanding of internet activity in our digital behavior but it's very broad. So but it's going to get very acute it's going to get very pinpoint specific but I think those two things live it together everyone's specific.
Understanding plus that broad understanding and just being able to dynamic communicate between the two I think that solves a lot of problems. I think that is my like the pool angle because then you can start doing things like oh yeah my house just knew I wanted at 76 degrees right now so just. Change the temperate because my rings and my body temperatures low and I when I get to the house and it's going to increase the temperature in the house. Now granted I think a lot of the stuff and we talk was all the time is.
You know there's the promise of AI and are we at the peak of it oh my gosh we're a decade away from the peak of it that doesn't mean that we don't need to start paying attention and using those tools that are coming out to improve some part we'll go back to that workflow story we need the tools to improve some part of our workflow are all the tools out there they can improve every part of a workflow no not yet but. And every one will be different. And so it's like it's an evolution. I really think you'll get the super specialist who like to practice in a certain way in service certain type of population and community and I will allow them to have that type of practice absolutely absolutely absolutely.
Alright so everybody audience hopefully you've got to learn something out of this I think the big tips for today or easy like hey look at workflow. Trim out the fat of what you currently doing you don't need to do and then when you're done doing all that with your workflow then start looking at hey what tools are out there that can improve my biggest need like which part of my workflows most broke that can save me the most minutes that can make me the most revenue per encounter. And look at those tools and there's lots of those tools being developed you know adopt those tools put them in place yeah I was saying if you don't want to cut something out be sure to write down why what it is that's keeping you can cutting it out that might be something that a tool so we'll solve. Absolutely and then hopefully we gave the audience we gave you guys a little bit of optimism about you know AI as easy said is not here to replace us it's here to augment us.
It will disrupt what we do and we have to evolve. Great words of intelligence they're easy thank you so much my friend for being a show we got to do this again I'm sure we could do this for at least another. At time 20 or 25 hours. Thank you thank you have me it's an honor and a pleasure to be here.
Alright everybody thank you for joining us in real talk if you have questions thoughts concerns please feel free to reach out to me it's Scott SEO tea that's just one tea. At AI in I care the letter A the letter I in I care dot AI happy if you want to be on the show please contact me happy to have more many guesses we can. To get as many different perspectives on how AI and other types of great innovations are going to change the way we practice I care thanks for joining us. 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.