49. Transcript - The future is here_ AI and personalized healthcare with Eric Topol
[00:00:00] Jonathan Wolf: Welcome to ZOE Science and Nutrition, where world-leading scientists explain how their research can improve your health.
Hey Siri, I have a headache and I'm a bit bunged up. What's your diagnosis?
[00:00:19] Siri: Hi, Jonathan. You could be having a heart attack. Please call 999 as you need immediate treatment in the hospital.
[00:00:27] Jonathan Wolf: Erm, okay.
Right now asking Siri, Alexa, or chat GPT for medical advice is a terrible idea. But with recent developments in technology, this looks set to change. AI has become more intelligent, wearable devices are more accurate, and personalized medicine is increasingly mainstream. But hold on a second. Is any of this safe? Should we really trust machines with our health? And will any of this actually happen?
In today's episode, we speak with Eric Topol to find some answers. Eric is one of the top 10 most cited medical researchers in the world. He's the author of three best-selling books on the future of medicine and a practicing cardiologist. He joins us to explore how artificial intelligence may transform your next trip to the doctor.
Eric, thank you for joining me today. I know you are a very busy man and it's a huge pleasure to be able to do this and to talk not only about COVID, which I know is a topic that you've been talking a lot about, but I think to go to one of your big loves, which is really talking about the role of artificial intelligence and what it can mean in medicine.
[00:01:51] Eric Topol: Great to be with you, Jonathan.
[00:01:53] Jonathan Wolf: So one of the things we always do at ZOE is we start with a quickfire round of questions from our listeners. And our scientists always find it really difficult because we have one rule, Eric, which is, you can say yes or no, or you can give a one-sentence answer, but you're not allowed more than one sentence.
And we know that it will always make you a bit uncomfortable, but are you willing to give it a go?
[00:02:15] Eric Topol: I'll try.
[00:02:17] Jonathan Wolf: All right, so the first question that we had from one of our listeners was, in the future when we feel sick, will we contact the artificial intelligence doctor on our phone as our first step?
[00:02:29] Eric Topol: Not on your phone necessarily, but sometimes yes.
[00:02:33] Jonathan Wolf: Alright. Second question. Do you think that regular home testing will become a normal path of our health, like visiting the dentist?
[00:02:42] Eric Topol: More and more, yes.
[00:02:44] Jonathan Wolf: Do you think that wearables like the Apple Watch will play an important role in healthcare in the future?
[00:02:50] Eric Topol: They will undoubtedly increase well beyond just the watch in the years ahead, yes.
[00:02:56] Jonathan Wolf: Should doctors prescribe diet and fitness interventions in the way that they currently prescribe drugs?
[00:03:03] Eric Topol: Well, prescribing it may not be the best, but certainly advocating diet and exercise as an essential part of one's lifestyle on a health plan is critical.
[00:03:15] Jonathan Wolf: And last question. Can artificial intelligence ever replace healthcare professionals?
[00:03:20] Eric Topol: We will go increasingly to a world that is somewhat autonomous and doctorless, but it will never replace. It will just supplant. The role of the clinician.
[00:03:34] Jonathan Wolf: And I think that's something I'd love to get into because I think there's always an enormous amount of fear whenever you talk about artificial intelligence. And my view of this is almost always when you see these new technologies, actually, they allow us to become more productive. You know, no one got replaced with all the technology we've brought into healthcare over the last hundred years. In fact, it's much harder to get to see the doctor now than it was a hundred years ago. So I assume that we will see the same thing, but maybe when we see the doctor, they'll have all of this great knowledge and you know, there's not a lot of that great insight from doctors.
[00:04:06] Eric Topol: Right? I mean, we have problems for sure, of being able to have time with patients and there are solutions, in sight if we work towards those.
[00:04:17] Jonathan Wolf: It's funny, my wife is a doctor and so through that I know a lot of doctors and I think often if you're not a doctor, you see this through the lens of being a patient and it can be very frustrating.
I think it also could be quite scary. You know, most times when I see the doctor, I think there's something wrong and I, I dunno what it is, but talking to my wife and her friends, you know, the number one complaint they always talk about is feeling they haven't got enough time with the patient and that it's sort of consumed. Also in trying to get the information that they really need in order to then do the bit that is really value added. And you know, my wife is sitting here in the Uk. I mean, one of the things she feels is the time has got much worse in terms of availability. Is that a global position you're sitting in, in California? Is that the same, you know, around the world?
[00:05:04] Eric Topol: I think that's unquestionable. There's been really, a degradation of the patient-doctor relationship over the years. We have to turn that around. We can. That was the whole premise of the deep medicine book that I wrote about what AI can do to bring it back. It's the most important thing I think we can do with this technology, in the years ahead, but it's going to take an active, aggressive effort. It's not going to happen by default because we can use these new tools to make things worse. And that's what we've been doing with technology, like electronic health records and many other things, for the past decade. So we, we've really got to turn that around.
[00:05:49] Jonathan Wolf: I think Eric, you are very rare. In general, what we see in health is this divide between people who are, you know, world-class physicians and scientists, and then this completely different world that talks about technology and AI. And in general, there's sort of no overlap at all.
So in fact, you may be on your own in this particular intersection. So I think it's really fun to be able to have you here and talk with us. And actually, you know, I want to share a story just with our listeners for a minute because right at the very beginning of the journey with ZOE, actually, you were one of the first people that I got to go and meet.
And so I was with Tim Spector and we were visiting California and San Diego where you are. Which, you know, for those of you who aren't familiar with San Diego, it has basically the world's best weather. Is that fair, Eric?
[00:06:37] Eric Topol: It's up there. It's one of the best for sure, yeah.
[00:06:40] Jonathan Wolf: And Eric has this fantastic office with this great big glass window that looks out onto like the sea, and you can see the beach and you know, you've flown in from like rainy London and you're like, I've definitely not made the right life choices.
And Eric has clearly made the right life choices. Tim had explained to me, you know, you are, I don't need to say, like an incredibly well-known and respected physician, but we had this fascinating conversation where I came away with a real boost because I think I'd been meeting a lot of scientists and doctors who'd been very skeptical about the idea of applying machine learning and artificial intelligence into healthcare, and also skeptical about doing the sort of enormous scale of studies that we were talking about. And you were one of the very few people who were very excited about the idea and believed there was a lot of opportunity. So, I'm not sorry I've said that to you, but it was a real boost. And I think, you know, I came away thinking that although maybe this whole idea of ZOE was a little bit crazy, it wasn't completely crazy. So thank you.
[00:07:44] Eric Topol: Well, I remember the meeting well and I don't think it's crazy at all. I think it's the future. We'll get there. It won't happen, you know, in the next few weeks or months, but we will get there.
[00:07:56] Jonathan Wolf: I'd love to start with that and I know you've written quite a bit about it and you know, for our listeners though, I think this will be really new. Could you maybe start with how have you seen medicine changing maybe over the last decade or two and what you expect in the future?
Where do you see this going?
[00:08:14] Eric Topol: Well, I'm an old dog, Jonathan. I got out of medical school in 1979, so I've, you know, I started practicing as a cardiologist in 1985, so I'm heading towards 40 years as a cardiologist in the next few. So I've had a lot of experience and I still, of course, practice and see patients, and it's the best part of my week, every week.
And I've watched medicine with steady erosion, particularly in more recent years. The difference in the eighties when I came out and started practicing medicine where the bond between patients and their physicians was precious. It was an intimate, tight relationship. There was trust, there was a presence. There was a really good history taking and physical exam, and you cared for patients. And they felt cared for. Now it's a rush job. Patients have little sense of a real bond, so the humanity in medicine has been compromised, and as a result of that, we're in a desperate situation because of that loss of trust, that loss of that remarkably important relationship, and the inability to listen to patients, we interrupt them within seconds. Don't listen to their story. So we aren't really caring anymore to a large degree. Obviously, there are exceptions, Jonathan, and there are great doctors out there and health systems where you have time. But the biggest insult is that we don't have time and we have shunted our efforts to being data clerks. Pecking away keyboards to enter data or put in orders or tests or whatnot. So we have the tools, I believe, to completely radically improve where healthcare is today, and not only revert back to where it was, in the eighties, but to even go beyond better than that. So that's I think, really exciting.
[00:10:26] Jonathan Wolf: And Eric, just before you talk about where we're going because our listeners will be saying, you know, how comes it's worse than 40 years ago?
And I think many of them will have experienced this. And it does seem as though this is a common pattern across many different countries. And there are people listening to this in the States, in Canada, in the UK, in Australia, and all over. And it seems like you get a similar picture everywhere.
We're richer than we were 40 years ago. Like we have more technology. Why is this worse and not better?
[00:10:55] Eric Topol: Well, it's largely because medicine turned into a business. It's especially worse in the US because it's truly a pay-for-service model, but having spent time in the UK in the review of the NHS and other countries, it's also a business. That is the number of minutes. The time scheduling with patients has diminished over the years and the expectations that doctors and nurses, and clinicians across the board are now that they have the added toll of having to enter all the data. Typing it. And spending a lot of time on screens, not even with the patient, not even seeing the patient when the patient is in the room.
So the electronic health record may have been the biggest insult, but this decay of priority over time. And empathy and communication, are all things that have suffered vastly. So again, as I wrote about in deep medicine, when I started, the average time for a new patient was an hour or more to a new consult, and a return visit was 30 minutes. At least.
[00:12:17] Jonathan Wolf: Wow. Sounds amazing.
[00:12:20] Eric Topol: Now we're talking about 12 minutes for a new patient, which is the actual time with the patient, and a lot of that isn't even looking at the patient or seven minutes with a return visit in the United States.
[00:12:32] Jonathan Wolf: I have to say that in the UK right at the moment. So I'm here, you know, it's in the winter.
Things seem under so much pressure that the idea that you would even get a doctor to speak to you on the phone for two minutes sounds like it would be a brilliant outcome. So I think, you know, we're really feeling under intense pressure at the moment.
[00:12:48] Eric Topol: Well, in Asia, the average visit is, in many countries, two minutes.
The visit, not just a phone call. Yeah. So this has been a trend about the business of medicine and, you know, getting in patients like a factory, you know, in and out. The squeeze, if you will. And the problem around the world is that doctors aren't in control. This is run by administrators, the overlord. And their charge is to make things efficient and get those patients through and care, the word care, which is a humanistic thing, that's not a top priority. Do you know what I mean? So part of this is really a profound flaw, losing humanity. The whole reason why we went into healthcare.
I mean, this was the mission. This is what is alluring to people to go into the profession, and we can't even execute as we had hoped because of all these drawbacks that are really serious.
[00:13:54] Jonathan Wolf: Now, Eric, I always think about you as an optimist. So now you've brought me really low. I'm hoping you're going now to flip this around and tell me that we're on a one-way road to a ten-second second appointment with a doctor. By the time I'm 75.
[00:14:10] Eric Topol: No, I am an optimist and I have a solution. I've proposed a solution that I think will take. I hope to see it fully become the norm, the bane of medical practice. The main thing is to describe the unmet need and now take action to get us to where the exciting new place that we can be in the years ahead. If we didn't have a way to get there, that would be depressing, but we have the eminent possibilities to fix this.
[00:14:40] Jonathan Wolf: So you've teased us. Tell us about it.
[00:14:42] Eric Topol: Well, then, is that all these different paths to get the gift of time. So that when you see a doctor, whether it's an outpatient visit or in the hospital at the bedside, that you know, there's no constraints of time.
That there's a chance to really, that there's the presence and that there's really a chance to communicate everything that needs to be expressed. And that's particularly listening to the patient's story. Doing a really good exam and that sort of thing. Now, to get that time, I mentioned there are several paths to get that.
One is, as you've already touched on, in the questions, Jonathan, to give patients more charge. That they can do self-diagnosis screening. Maybe not the final diagnosis, but a good screening. They can have their data and have algorithmic support to interpret it. They can have coaching done virtually on their conditions or conditions that they would otherwise like to prevent.
And of course, things like improving their lifestyle so they have better health in general, which is preventing across all conditions.
On the clinician side. We're talking about liberation from the keyboard. That they wouldn't even exist. We would do this through natural language processing and machine learning.
We would have all the records and all the data. Teed up for the clinician so they don't have to go through page after page, screen after screen to try to aggregate that information. And we would have, you know, so much of this automated so that the residual was, wow, we have time to be together. We have less need for in-person visits and they're going to be for the important stuff, like to discuss a new diagnosis or, you know, to understand what our patient's, you know, deep concerns are, whatever it is. But there'll be a different look at why you see a doctor in the future. We already have many ways to make diagnoses through AI for the patient side.
We have, you know, things like the diagnosis of heart rhythm abnormalities through a watch, diagnosis of urinary tract infections through an AI kit you can get at a pharmacy in many countries. We have skin diagnoses of lesions, rashes, cancers, ear infections, and children. And the list will just. Keep growing, so that most of the routine things that are not life-threatening will be capable of diagnosing through our phones and our wearables, and our data. That's really important. That's what's missed a lot, is that the patient side of this is going to be empowered like never before. Not every patient will use these tools, but yes.
[00:17:46] Jonathan Wolf: I think it's a really fascinating picture, right? All of these pieces sound amazing and, and I guess my first question is which parts of those are actually a relatively near term, Eric, where you feel actually, you know, these could be real things that we're doing within the next, you know, three to five years.
[00:18:03] Eric Topol: Well, the list I just mentioned is here today, this doctor list screening for diagnoses, yes. Yes.
And I think that list is just gonna expand quickly from now on. On the clinician side, for doctors, we are already seeing synthetic notes to replace all the work that's being done to enter notes and data by doctors and nurses, so that synthetic notes are - and I've already seen, they're so far better than the ones that we have accustomed to - based on the conversation, extracted from the conversation between a doctor and patient.
[00:18:40] Jonathan Wolf: So they just listen to the conversation, a bit like when I'm talking to Siri or Alexa and then they condense it down to like something that looks like some doctor's notes about, you know, Jonathan has a sore throat and you know, his knee hurts and you know, we are going to do these tests. Like it can actually pull all of that together?
[00:18:58] Eric Topol: Not only that, but it also says, well, you know, Jonathan, we're going to prescribe you this medicine. And so it already puts that into your pharmacy. We're going to have these tests ordered. They all get ordered. We're going to have your next appointment, you know, in six months. That all gets arranged.
You don't have to touch a keyboard. And that's phenomenal because that's a liberation. And when I did the NHS review with the credible team, I was so privileged to work with, that was the one that everyone was cheering for. That was the one. And that's already started, of course, in many parts of the world.
And it will become the norm in the years ahead. It's not far. We have the tools to do that.
[00:19:44] Jonathan Wolf: And is that live anywhere in the world today, Eric? Is that actually, you know, being experienced by patients?
[00:19:51] Eric Topol: Absolutely. Oh yeah, in fact, it was already two years ago. It was live in Leads, as I understood it, in their emergency department of all places, interestingly.
But no, it's starting to crop up in many parts of the US. Alaska, I learned about it just recently. In China. I mean, it's the real deal, but, you know, it takes time because getting across all specialties, across all of medicine. But it's on the way. And it makes the current way we chart look anemic because of its efficiency.
You know, it's like, why do you have to do this all over again if you could have a machine do it? And by the way, it's improved more because the patient gets to edit the notes. And, you know, access all the notes. And that helps because now there's a work product that's done jointly. So the physician edits, but after the first 25, 30, and 40 notes, the AI has the doc down, but each patient also can weigh in.
So, you know, this is exciting because if we liberate from keyboards, man, that's big stuff.
[00:21:09] Jonathan Wolf: That's really interesting. I'm interested in how much you emphasize this idea that the doctor is now distracted by the computer and the screen and is not sort of looking directly at you. And it's not something I'd ever thought about before, but you're absolutely right of course, that when you are talking now, they are interacting a lot with the keyboard and, you know, that's very different from, for example, the sort of interaction that the pediatrician has with like my child, you know, my son is now older, but, you know, with my daughter, because of course you have got to really give a three-year-old your full attention if you want to get anything out. And of course, you see that's a very powerful sort of interaction that I'd never thought about.
But you're right, you're sort of losing when suddenly, you know, the doctor is actually, you know, largely having to interact with this computer to put in all of this data.
[00:22:00] Eric Topol: Right, right. No, and, and in that is a big deal when you don't even make eye contact. It's awful, both sides. I mean, it's unacceptable and it's been the norm for many years now. And I also want to add Jonathan, it isn't just the time where the computer, in the midst of seeing patients, it's all the added time. Off hours, you know, where you have to still do all your charting and stuff that should have been done in real-time. So it's a burden not just in the midst of seeing a patient, but also well beyond.
[00:22:41] Jonathan Wolf: And it's interesting, I see that with my wife because again, as a patient, it's not something I'm ever really aware of, but there's an enormous amount of bureaucracy dealing with patients outside of, of the room. And I think as a patient, I never really thought about it. You thought the only time anyone's dealing anything with you is when they're there. And then you see there are all sorts of other things they have to do. Falling up on the notes, making sure they're right, sending them, dealing with tests, looking at results, all these sorts of things are a big part. So that's very exciting. I would like to follow up on the other bit that you talked about, about the sort of patient self-diagnosis, because that's something that has obviously been really interesting for ZOE. We had this big experience through the ZOE COVID study of suddenly people logging their symptoms every day. Being able to use that to figure out whether people had covid, you know, what the risk factors were or the rest of it. What I'm struck by is that we had this amazing tool and experience and we still have hundreds of thousands of people who are continuing to use the ZOE Health study to do this. And, it's in no way connected to anything else that is going on in healthcare. And my experience is that there's, you know, quite limited linkage of this idea of sort of self-reporting your data through to the health system. So you are being very optimistic, which I love. Where do you see this really working well?
And what are the gaps to get to the point where we really could be? I always like to talk to people about this idea of preventative medicine. Like surely you ought to be able to figure out things are not quite right a long time before you are really sick. And the reality is that our health systems are not really able to do this in my studies today, right?
[00:24:34] Eric Topol: One of the problems we have right now is the level of depth of data we have for each person. What you've done with your nutrition work is kind of futuristic because you not only need to have the electronic health record and of course, in some countries like the US it's highly fragmented, so you need many different EHR health systems, but then you'll go, of course, you need to give things like the biologic layers, that could be the gut microbiome or the patient's genome, or at least their array, SNIP data. You'd like to have their physiome, which is sensors that you would be able to pick up key metrics from that. Their anatome, which is the scans of their relevant anatomy. Their exposome, the environmental factors, and the social determinants. Their immunome, and on and on. Okay?
[00:25:43] Jonathan Wolf: So it's a lot of stuff.
[00:25:45] Eric Topol: Yeah, a lot of stuff. And by the way, no human being could process that now. Even some particularly confident doctors could never process all that data. All right? So we have to lean on machines, but what we're doing and you're doing, as well as pulling together all this data, multimodal AI and processing it for the individual.
And once you do that, you not only can manage a condition like diabetes or hypertension or depression, common conditions, or obesity, but you can help manage a person essentially in real-time by having that data processed and getting back to the person with an algorithm support, but you also, in the future will be preventing conditions that people are at risk for, which is really exciting because then now we fulfill the fantasy.
The fantasy of medicine is to prevent serious conditions from ever occurring, and when you have all that data at the individual level, you can use it to prevent an illness, a serious illness, whether it be a heart condition or neurologic, you know, cancer, asthma, whatever. This is really where eventually we will be.
[00:27:05] Jonathan Wolf: And Eric, just as you say that, you've been voted the number one most influential physician leader in the United States, which is very impressive. So I want to take this opportunity to say that we like to say that this is possible, but you know, you are much more credible than I am.
Do you truly believe this idea of prevention in healthcare is possible?
[00:27:32] Eric Topol: Yes. I mean, it hasn't been actualized and proven, but theoretically, once you have a person's data, the first step is to say each of us is unique. Accept that each of us is truly unique. Even, biological twins, and identical twins have significant differences.
So once you get past that and you say, okay, well now we can get all the domains of data. And we can process that data, that's one of the bottlenecks right now, by the way. And when we do that, not just at a personal level, we do that in the hundreds of thousands and millions of people, ultimately billions of people.
Then we have a way path to not only treat people better, help them manage their conditions better, but also prevent illnesses. So yes. I mean, we're are ways from that. Not right now, but you can see we will get there.
[00:28:33] Jonathan Wolf: So it's an ambitious goal. You believe we can get there, but we need to be realistic.
When you're talking about the stuff that's near and far, like the prevention in the sense of really being able to figure out that you are at very high risk of a stroke in 15 years. This is a long way off. Perhaps prevention in the way that we think a lot about it, which is improving, you know, diet and things like this, we can see is, you know, I think we feel very good about impacting long-term health, but it doesn't mean we're actually figuring out, you know, the way in advance what's going to go wrong?
You know, the analogy I often like to talk about is just with our cars where, you know, I take the car to be serviced once a year. I still have a car. And they plug a computer into it and they run some diagnostics and they say, oh, you know, the fan belt is about to break and we're going to replace the fan belt, and I drive it off.
And as a result, the car never breaks for, you know, 12, 13, 14 years. And we're used to that in all these other aspects of our life. And I think the experience that, you know, all of us who aren't doctors have is, you know, you go to the doctor and if you're not like really sick with something, there isn't that equivalent or really of being able to do this diagnostic and say, you know what, Jonathan, I know you're worrying about Alzheimer's and heart disease and all the rest of it. Actually, you know what, it's diabetes you've really have got to worry about and therefore this is the stuff that you really need to think hard about, you know, We aren't, as a health system in most cases, right, able to do that. Maybe if I have particularly strong family risk or I've got a very rare gene, or I'm getting really sick, but in most cases, you know, that's a long way on and it's much harder to make the change when you are you're already sick than it is to get you much earlier. Do you recognize this picture I'm painting?
[00:30:30] Eric Topol: Yes. I mean, I wish we were as simple as cars.
[00:30:33] Jonathan Wolf: I accept that.
[00:30:36] Eric Topol: You can digitize a car. And what I've outlined is that we're going to be digitizing human beings. That sounds awful, right?
Oh, we're gonna digitize you. But no, it's for the singular purpose of what you just described, of being anticipatory for potential illnesses that you, specifically are at risk for and knowing someday we're going to have digital twins.
[00:31:03] Jonathan Wolf: And what does that mean, Eric?
[00:31:05] Eric Topol: Yeah. We found these five people around the world that match you at every level as closely as possible.
And by the way, you know, some of them are much older than you, and this is what we learned, how to prevent this condition or that condition. This is what really worked well to treat cancer that you have or going to get potentially. So I mean, we have never done this before. Learn from each other.
We do these clinical trials and let's say we have 10 out of hundred people. That's a big success story in a clinical trial in that we found a treatment that helps 10 out of a hundred. What about the other 90 out of a hundred that take the medicine every day for the rest of their lives? Well, what about if we start getting incredibly precise?
Through digital twins when we have, you know, hundreds of millions, ideally billions of people where we have all these levels of data and we can help share and, you know, make matches so that people will know the best treatments, the best preventions. That's someday. That's further off. That's further off, Jonathan, than what we've described.
[00:32:19] Jonathan Wolf: I mean, I love that because you are sort of articulating in a different language. I think a lot of the ideas that we have at ZOE, that Tim and Sarah and George and I talk about quite a lot. And I think it's wonderful.
I'd love, to pull up a little bit, you've talked about personalization quite a bit here. And I think that's really interesting because, you know, I know that Tim and others talk about the way that when he was taught as a doctor, everything is that, it's a one size fits all solution.
You might diagnose somebody with a particular disease, but then you don't further break this down into lots of different categories. You've got sort of this treatment, but a lot of what you've been talking about is how different we all are, I think, and I'd love for you to talk about how important you think that is, and I guess for people listening is that really going to be able to affect the healthcare they get in, you know, in a decade?
[00:33:19] Eric Topol: Well, sure. I mean, I think, you know, what you've seen, in your work, with Tim and others is exactly what we've seen, which is it's just as simple as, if you have a glucose sensor on and you eat something, and some people have spikes of glucose too, you know, 180, 200 and some people totally flat and the exact same food with the exact same portion.
It was really, Aaron Siegel and his colleagues at the Weizmann Institute in Israel that made that first observation several years ago, which I thought was stunning, but that basically told the story and it's the same across everything in health. Everything.
If it's as simple as what you've eaten, then you can imagine, that the fact that we treat people like cattle and we don't recognize their uniqueness insofar as our approach, a general approach, that's also part of the problem. So we have a remedy, we have a fix for that. And it's across everything. And nutrition is one that's, I think, I'm particularly enamored by and wrote a lot about. Deep diet in the book. And the AI diet, if you will, that the New York Times called it.
But, you know, I think this is an exciting area because we spent a lot of our time with what we eat and our sleep and, you know, our exercise, ideally, and a lot of that, a lot of those things can help us preserve health.
[00:34:53] Jonathan Wolf: As we think about the future, I think you've got this picture I think that we know we want to end up with where almost everything has been monitored. There's all of this AI to figure out all of it. It's these incredible tools that support the doctor to figure all of this out. And then, you know, I'm getting this completely personalized solution if I have particular cancer or any of the rest of it. Clearly, we're going to go through some stages from, you know, here to there. And we will also of course have some missteps and we'll think there are things that are really important that we should monitor and they turn out not to matter. I think one thing I'd be really interested in, is I guess, your view about what's the sort of things that we will want to be monitoring sort of overtime on ourselves that you think have value and maybe those that don't.
Because I am struck. There's also like a risk of a lot of companies selling, you know, a particular thing that you should monitor all the time. And I've seen a few of these where, you know, I'm a bit skeptical based upon what I've seen from our data and other scientists. You know, not everything that you can monitor is necessarily important, right? We all know that actually if we think about our own children, there's all of this stuff going on, and actually, you're pretty good at figuring out, actually, they're not quite right, right? You're picking that up in some way. What do you think are the things that, either already we know, have value or are going to have value?
Just, you know, based upon either what you, you've seen yourself or from the research.
[00:36:25] Eric Topol: Well, I don't think it's, they measure all things in all people. I mean, that in itself is particularized, based on the different layers of data. I do want to go back to your assertion that, you know, everybody's going to do this.
I don't think that's ever gonna happen. There's a lot of mistrust in data and privacy of data. There are solutions in sight. In AI, if you talk to AI, computer scientists, every problem in AI, whether it's biased and whether it's privacy, you name it, there's a solution with AI, right?
I mean, it's classic, but most of those are not true. But the one that does resonate is that we will do better with privacy and security, whether it's edge computing or swarm, federated or whatever it is, homomorphic encryption. So the reason I bring that up is that's a barrier as is, you know, bias and as is, you know, other factors that play in this whole area.
But no matter what, we will have many people that are reluctant to use technology in healthcare. I mean, look what we've seen with vaccines. Probably the most compelling data ever assembled. In the history of medicine and look at how much resistance there is. So I don't want anyone to think that this is going to be, you know, a vast majority, but I think it's going to be substantial enough to be very, very important.
So just make sure that that's clear because we tend to get carried away that the way we think is the way everybody thinks. And that couldn't be further from the truth.
[00:38:23] Jonathan Wolf: Totally agree, Eric. I think trust is an enormous part of this. And, the other part of the course is feeling that there's real value, right?
So we always feel better about recording something or sharing our data with somebody like our doctor because we know we're going to get something valuable back.
For the people then who are saying, you know, I want to do this. I'm happy to go and, you know, wear a particular device or whatever. What do you think are the sorts of things that are being monitored today or you can see being monitored in the next couple of years that you think might have real medical value?
[00:39:07] Eric Topol: Well, the first biosensor, which is now used in millions, of course, is glucose, continuous glucose. And that's changed the world for lots of type one diabetics.
[00:39:20] Jonathan Wolf: That's right. And for everybody, that's blood sugar for anyone who's not familiar with it. So this has been able to monitor your blood sugar, you know, in real-time with some little device you stick on your arm for example.
[00:39:31] Eric Topol: Yeah, and this is a big deal because, you know, obviously people taking insulin are at risk for having very low blood sugars, or not having good control of their blood sugars.
So this is a whole lot better than getting a blood test, you know, every three months, like a hemoglobin A1C which is used so, that's the first sensor that got into wide-scale use. And then of course, with the fitness bands like Fitbit and smart watches like Apple and many others now, we are seeing heart rate, heart rhythm, but there'll be others, of course, you know, oximetry, it became important during the pandemic with pneumonia monitoring.
[00:40:17] Jonathan Wolf: This is the amount of oxygen in your blood, Eric, to make sure. That's oximetry, is it?
[00:40:23] Eric Topol: Exactly, so, which can be monitored through a finger or the wrist very easily, but there's no limit.
I mean, eventually, we'll have high-frequency blood pressure, non-invasive passive monitoring, so you won't have to take out a device and measure it yourself. It'll be done automatically, but we already have things. To me, the most extraordinary thing is self-imaging. So using a smartphone with a high-resolution ultrasound probe that you put in the base of the smartphone, you can image any part of your body except your brain.
[00:41:01] Jonathan Wolf: So this is like basically the same as when you know you are pregnant and you do the ultrasound to find the baby, but you can plug it into your phone?
[00:41:08] Eric Topol: Yes. Now, it wouldn't probably be good for a mother, expecting mother to image their baby every few minutes. That's not what I'm getting at. But what I am suggesting, like we're already seeing in heart failure, is that people can image their heart and send a loop to a doctor and you don't have to have any training. The AI tells you how to get the image. All you have to do is be able to put the probe on the chest.
[00:41:38] Jonathan Wolf: That's amazing. So it's another one of these ways where you can push the technology right out to you at home and you can give like really valuable information. Back to the doctor.
[00:41:50] Eric Topol: Exactly.
[00:41:51] Jonathan Wolf: I would like to just do a quick summary, which is what we always do at the end here. And I think, you know, this has been particularly sort of visionary, which is really exciting.
I think that you know, my key takeaway is, you know, from your perspective actually healthcare's been sort of getting worse over the last 40 years. And you paint a picture where it's really true that the relationship between the doctor and the patient is worse. And that actually the technology has sort of been getting in the way.
But the good news is, you do see this sort of sunny uplit future where AI can really change that. And there's a lot of these technologies are sort of there. They're just not in the sort of mainstream deployment, but therefore they could be there in the next few years. You're not talking about decades away.
I think that some of these we will just see when we visit our physician, our doctor, because suddenly, hopefully, they're gonna be able to look at us and engage with us. And there are all of these things in the background. But the other element that you see, I think becoming really significant is people tracking their own data. Now, what exactly that data will be that turns out to be high value is still being figured out. And you talked about blood sugar as an example. You talked about some of the things like blood pressure, but we're going to learn a lot more. And I guess the final thing is, for you, personalization is central to all of this.
That actually, in a way, medicine right now is this thing, you've got a drug that's going to work for us, you said 10 and a hundred people, you know, how do you solve the other 90? How do you figure that out? And so again, this data, hopefully, and this AI allows us to move to a world where not only can you think about prevention, but also you can really think about personalization and I'm slightly putting words in your mouth, but I think you're therefore suggesting there could be a real transformation in the success of medicine if you can really understand that personalization.
[00:43:44] Eric Topol: That's a great summary.
[00:43:45] Jonathan Wolf: Eric. It was such a pleasure. Thank you for taking the time. I look forward to following up and hopefully keeping you abreast of what we do at ZOE and hopefully getting you back on again in the future.
[00:43:57] Eric Topol: That's great.
[00:43:58] Jonathan Wolf: It's a real pleasure. Thank you, Eric. Bye-bye.
Thank you, Eric, for joining me on ZOE's Science and Nutrition today. If based on today's conversation you'd like to understand the difference personalization could mean for your diet and your health, then you may want to try ZOE's personalized nutrition program.
Your ZOE membership comes with our app, which uses artificial intelligence to score millions of foods and meals for you. You also get a personalized meal and recipe recommendations and can chat directly with our nutrition coaches. As a member, you start with an at-home test to understand your biology. We create your individual ZOE scores and a personalized program for you based on your test results and our scientific research.
If you're interested in learning more about ZOE, you can head to joinZOE.com/podcast and get 10% off your purchase. If you enjoy today's episode, please be sure to subscribe and leave us a review as we love reading your feedback. If this episode left you with questions, please send them in on Instagram or Facebook, and we'll try to answer them in a future episode.
As always, I'm your host, Jonathan Wolf. ZOE Science and Nutrition is produced by Fascinate Productions with support from Sharon Feder, Yella Hewings-Martin, and Alex Jones here at ZOE. See you next time.
[00:00:00] Jonathan Wolf: Welcome to ZOE Science and Nutrition, where world-leading scientists explain how their research can improve your health.
Hey Siri, I have a headache and I'm a bit bunged up. What's your diagnosis?
[00:00:19] Siri: Hi, Jonathan. You could be having a heart attack. Please call 999 as you need immediate treatment in the hospital.
[00:00:27] Jonathan Wolf: Erm, okay.
Right now asking Siri, Alexa, or chat GPT for medical advice is a terrible idea. But with recent developments in technology, this looks set to change. AI has become more intelligent, wearable devices are more accurate, and personalized medicine is increasingly mainstream. But hold on a second. Is any of this safe? Should we really trust machines with our health? And will any of this actually happen?
In today's episode, we speak with Eric Topol to find some answers. Eric is one of the top 10 most cited medical researchers in the world. He's the author of three best-selling books on the future of medicine and a practicing cardiologist. He joins us to explore how artificial intelligence may transform your next trip to the doctor.
Eric, thank you for joining me today. I know you are a very busy man and it's a huge pleasure to be able to do this and to talk not only about COVID, which I know is a topic that you've been talking a lot about, but I think to go to one of your big loves, which is really talking about the role of artificial intelligence and what it can mean in medicine.
[00:01:51] Eric Topol: Great to be with you, Jonathan.
[00:01:53] Jonathan Wolf: So one of the things we always do at ZOE is we start with a quickfire round of questions from our listeners. And our scientists always find it really difficult because we have one rule, Eric, which is, you can say yes or no, or you can give a one-sentence answer, but you're not allowed more than one sentence.
And we know that it will always make you a bit uncomfortable, but are you willing to give it a go?
[00:02:15] Eric Topol: I'll try.
[00:02:17] Jonathan Wolf: All right, so the first question that we had from one of our listeners was, in the future when we feel sick, will we contact the artificial intelligence doctor on our phone as our first step?
[00:02:29] Eric Topol: Not on your phone necessarily, but sometimes yes.
[00:02:33] Jonathan Wolf: Alright. Second question. Do you think that regular home testing will become a normal path of our health, like visiting the dentist?
[00:02:42] Eric Topol: More and more, yes.
[00:02:44] Jonathan Wolf: Do you think that wearables like the Apple Watch will play an important role in healthcare in the future?
[00:02:50] Eric Topol: They will undoubtedly increase well beyond just the watch in the years ahead, yes.
[00:02:56] Jonathan Wolf: Should doctors prescribe diet and fitness interventions in the way that they currently prescribe drugs?
[00:03:03] Eric Topol: Well, prescribing it may not be the best, but certainly advocating diet and exercise as an essential part of one's lifestyle on a health plan is critical.
[00:03:15] Jonathan Wolf: And last question. Can artificial intelligence ever replace healthcare professionals?
[00:03:20] Eric Topol: We will go increasingly to a world that is somewhat autonomous and doctorless, but it will never replace. It will just supplant. The role of the clinician.
[00:03:34] Jonathan Wolf: And I think that's something I'd love to get into because I think there's always an enormous amount of fear whenever you talk about artificial intelligence. And my view of this is almost always when you see these new technologies, actually, they allow us to become more productive. You know, no one got replaced with all the technology we've brought into healthcare over the last hundred years. In fact, it's much harder to get to see the doctor now than it was a hundred years ago. So I assume that we will see the same thing, but maybe when we see the doctor, they'll have all of this great knowledge and you know, there's not a lot of that great insight from doctors.
[00:04:06] Eric Topol: Right? I mean, we have problems for sure, of being able to have time with patients and there are solutions, in sight if we work towards those.
[00:04:17] Jonathan Wolf: It's funny, my wife is a doctor and so through that I know a lot of doctors and I think often if you're not a doctor, you see this through the lens of being a patient and it can be very frustrating.
I think it also could be quite scary. You know, most times when I see the doctor, I think there's something wrong and I, I dunno what it is, but talking to my wife and her friends, you know, the number one complaint they always talk about is feeling they haven't got enough time with the patient and that it's sort of consumed. Also in trying to get the information that they really need in order to then do the bit that is really value added. And you know, my wife is sitting here in the Uk. I mean, one of the things she feels is the time has got much worse in terms of availability. Is that a global position you're sitting in, in California? Is that the same, you know, around the world?
[00:05:04] Eric Topol: I think that's unquestionable. There's been really, a degradation of the patient-doctor relationship over the years. We have to turn that around. We can. That was the whole premise of the deep medicine book that I wrote about what AI can do to bring it back. It's the most important thing I think we can do with this technology, in the years ahead, but it's going to take an active, aggressive effort. It's not going to happen by default because we can use these new tools to make things worse. And that's what we've been doing with technology, like electronic health records and many other things, for the past decade. So we, we've really got to turn that around.
[00:05:49] Jonathan Wolf: I think Eric, you are very rare. In general, what we see in health is this divide between people who are, you know, world-class physicians and scientists, and then this completely different world that talks about technology and AI. And in general, there's sort of no overlap at all.
So in fact, you may be on your own in this particular intersection. So I think it's really fun to be able to have you here and talk with us. And actually, you know, I want to share a story just with our listeners for a minute because right at the very beginning of the journey with ZOE, actually, you were one of the first people that I got to go and meet.
And so I was with Tim Spector and we were visiting California and San Diego where you are. Which, you know, for those of you who aren't familiar with San Diego, it has basically the world's best weather. Is that fair, Eric?
[00:06:37] Eric Topol: It's up there. It's one of the best for sure, yeah.
[00:06:40] Jonathan Wolf: And Eric has this fantastic office with this great big glass window that looks out onto like the sea, and you can see the beach and you know, you've flown in from like rainy London and you're like, I've definitely not made the right life choices.
And Eric has clearly made the right life choices. Tim had explained to me, you know, you are, I don't need to say, like an incredibly well-known and respected physician, but we had this fascinating conversation where I came away with a real boost because I think I'd been meeting a lot of scientists and doctors who'd been very skeptical about the idea of applying machine learning and artificial intelligence into healthcare, and also skeptical about doing the sort of enormous scale of studies that we were talking about. And you were one of the very few people who were very excited about the idea and believed there was a lot of opportunity. So, I'm not sorry I've said that to you, but it was a real boost. And I think, you know, I came away thinking that although maybe this whole idea of ZOE was a little bit crazy, it wasn't completely crazy. So thank you.
[00:07:44] Eric Topol: Well, I remember the meeting well and I don't think it's crazy at all. I think it's the future. We'll get there. It won't happen, you know, in the next few weeks or months, but we will get there.
[00:07:56] Jonathan Wolf: I'd love to start with that and I know you've written quite a bit about it and you know, for our listeners though, I think this will be really new. Could you maybe start with how have you seen medicine changing maybe over the last decade or two and what you expect in the future?
Where do you see this going?
[00:08:14] Eric Topol: Well, I'm an old dog, Jonathan. I got out of medical school in 1979, so I've, you know, I started practicing as a cardiologist in 1985, so I'm heading towards 40 years as a cardiologist in the next few. So I've had a lot of experience and I still, of course, practice and see patients, and it's the best part of my week, every week.
And I've watched medicine with steady erosion, particularly in more recent years. The difference in the eighties when I came out and started practicing medicine where the bond between patients and their physicians was precious. It was an intimate, tight relationship. There was trust, there was a presence. There was a really good history taking and physical exam, and you cared for patients. And they felt cared for. Now it's a rush job. Patients have little sense of a real bond, so the humanity in medicine has been compromised, and as a result of that, we're in a desperate situation because of that loss of trust, that loss of that remarkably important relationship, and the inability to listen to patients, we interrupt them within seconds. Don't listen to their story. So we aren't really caring anymore to a large degree. Obviously, there are exceptions, Jonathan, and there are great doctors out there and health systems where you have time. But the biggest insult is that we don't have time and we have shunted our efforts to being data clerks. Pecking away keyboards to enter data or put in orders or tests or whatnot. So we have the tools, I believe, to completely radically improve where healthcare is today, and not only revert back to where it was, in the eighties, but to even go beyond better than that. So that's I think, really exciting.
[00:10:26] Jonathan Wolf: And Eric, just before you talk about where we're going because our listeners will be saying, you know, how comes it's worse than 40 years ago?
And I think many of them will have experienced this. And it does seem as though this is a common pattern across many different countries. And there are people listening to this in the States, in Canada, in the UK, in Australia, and all over. And it seems like you get a similar picture everywhere.
We're richer than we were 40 years ago. Like we have more technology. Why is this worse and not better?
[00:10:55] Eric Topol: Well, it's largely because medicine turned into a business. It's especially worse in the US because it's truly a pay-for-service model, but having spent time in the UK in the review of the NHS and other countries, it's also a business. That is the number of minutes. The time scheduling with patients has diminished over the years and the expectations that doctors and nurses, and clinicians across the board are now that they have the added toll of having to enter all the data. Typing it. And spending a lot of time on screens, not even with the patient, not even seeing the patient when the patient is in the room.
So the electronic health record may have been the biggest insult, but this decay of priority over time. And empathy and communication, are all things that have suffered vastly. So again, as I wrote about in deep medicine, when I started, the average time for a new patient was an hour or more to a new consult, and a return visit was 30 minutes. At least.
[00:12:17] Jonathan Wolf: Wow. Sounds amazing.
[00:12:20] Eric Topol: Now we're talking about 12 minutes for a new patient, which is the actual time with the patient, and a lot of that isn't even looking at the patient or seven minutes with a return visit in the United States.
[00:12:32] Jonathan Wolf: I have to say that in the UK right at the moment. So I'm here, you know, it's in the winter.
Things seem under so much pressure that the idea that you would even get a doctor to speak to you on the phone for two minutes sounds like it would be a brilliant outcome. So I think, you know, we're really feeling under intense pressure at the moment.
[00:12:48] Eric Topol: Well, in Asia, the average visit is, in many countries, two minutes.
The visit, not just a phone call. Yeah. So this has been a trend about the business of medicine and, you know, getting in patients like a factory, you know, in and out. The squeeze, if you will. And the problem around the world is that doctors aren't in control. This is run by administrators, the overlord. And their charge is to make things efficient and get those patients through and care, the word care, which is a humanistic thing, that's not a top priority. Do you know what I mean? So part of this is really a profound flaw, losing humanity. The whole reason why we went into healthcare.
I mean, this was the mission. This is what is alluring to people to go into the profession, and we can't even execute as we had hoped because of all these drawbacks that are really serious.
[00:13:54] Jonathan Wolf: Now, Eric, I always think about you as an optimist. So now you've brought me really low. I'm hoping you're going now to flip this around and tell me that we're on a one-way road to a ten-second second appointment with a doctor. By the time I'm 75.
[00:14:10] Eric Topol: No, I am an optimist and I have a solution. I've proposed a solution that I think will take. I hope to see it fully become the norm, the bane of medical practice. The main thing is to describe the unmet need and now take action to get us to where the exciting new place that we can be in the years ahead. If we didn't have a way to get there, that would be depressing, but we have the eminent possibilities to fix this.
[00:14:40] Jonathan Wolf: So you've teased us. Tell us about it.
[00:14:42] Eric Topol: Well, then, is that all these different paths to get the gift of time. So that when you see a doctor, whether it's an outpatient visit or in the hospital at the bedside, that you know, there's no constraints of time.
That there's a chance to really, that there's the presence and that there's really a chance to communicate everything that needs to be expressed. And that's particularly listening to the patient's story. Doing a really good exam and that sort of thing. Now, to get that time, I mentioned there are several paths to get that.
One is, as you've already touched on, in the questions, Jonathan, to give patients more charge. That they can do self-diagnosis screening. Maybe not the final diagnosis, but a good screening. They can have their data and have algorithmic support to interpret it. They can have coaching done virtually on their conditions or conditions that they would otherwise like to prevent.
And of course, things like improving their lifestyle so they have better health in general, which is preventing across all conditions.
On the clinician side. We're talking about liberation from the keyboard. That they wouldn't even exist. We would do this through natural language processing and machine learning.
We would have all the records and all the data. Teed up for the clinician so they don't have to go through page after page, screen after screen to try to aggregate that information. And we would have, you know, so much of this automated so that the residual was, wow, we have time to be together. We have less need for in-person visits and they're going to be for the important stuff, like to discuss a new diagnosis or, you know, to understand what our patient's, you know, deep concerns are, whatever it is. But there'll be a different look at why you see a doctor in the future. We already have many ways to make diagnoses through AI for the patient side.
We have, you know, things like the diagnosis of heart rhythm abnormalities through a watch, diagnosis of urinary tract infections through an AI kit you can get at a pharmacy in many countries. We have skin diagnoses of lesions, rashes, cancers, ear infections, and children. And the list will just. Keep growing, so that most of the routine things that are not life-threatening will be capable of diagnosing through our phones and our wearables, and our data. That's really important. That's what's missed a lot, is that the patient side of this is going to be empowered like never before. Not every patient will use these tools, but yes.
[00:17:46] Jonathan Wolf: I think it's a really fascinating picture, right? All of these pieces sound amazing and, and I guess my first question is which parts of those are actually a relatively near term, Eric, where you feel actually, you know, these could be real things that we're doing within the next, you know, three to five years.
[00:18:03] Eric Topol: Well, the list I just mentioned is here today, this doctor list screening for diagnoses, yes. Yes.
And I think that list is just gonna expand quickly from now on. On the clinician side, for doctors, we are already seeing synthetic notes to replace all the work that's being done to enter notes and data by doctors and nurses, so that synthetic notes are - and I've already seen, they're so far better than the ones that we have accustomed to - based on the conversation, extracted from the conversation between a doctor and patient.
[00:18:40] Jonathan Wolf: So they just listen to the conversation, a bit like when I'm talking to Siri or Alexa and then they condense it down to like something that looks like some doctor's notes about, you know, Jonathan has a sore throat and you know, his knee hurts and you know, we are going to do these tests. Like it can actually pull all of that together?
[00:18:58] Eric Topol: Not only that, but it also says, well, you know, Jonathan, we're going to prescribe you this medicine. And so it already puts that into your pharmacy. We're going to have these tests ordered. They all get ordered. We're going to have your next appointment, you know, in six months. That all gets arranged.
You don't have to touch a keyboard. And that's phenomenal because that's a liberation. And when I did the NHS review with the credible team, I was so privileged to work with, that was the one that everyone was cheering for. That was the one. And that's already started, of course, in many parts of the world.
And it will become the norm in the years ahead. It's not far. We have the tools to do that.
[00:19:44] Jonathan Wolf: And is that live anywhere in the world today, Eric? Is that actually, you know, being experienced by patients?
[00:19:51] Eric Topol: Absolutely. Oh yeah, in fact, it was already two years ago. It was live in Leads, as I understood it, in their emergency department of all places, interestingly.
But no, it's starting to crop up in many parts of the US. Alaska, I learned about it just recently. In China. I mean, it's the real deal, but, you know, it takes time because getting across all specialties, across all of medicine. But it's on the way. And it makes the current way we chart look anemic because of its efficiency.
You know, it's like, why do you have to do this all over again if you could have a machine do it? And by the way, it's improved more because the patient gets to edit the notes. And, you know, access all the notes. And that helps because now there's a work product that's done jointly. So the physician edits, but after the first 25, 30, and 40 notes, the AI has the doc down, but each patient also can weigh in.
So, you know, this is exciting because if we liberate from keyboards, man, that's big stuff.
[00:21:09] Jonathan Wolf: That's really interesting. I'm interested in how much you emphasize this idea that the doctor is now distracted by the computer and the screen and is not sort of looking directly at you. And it's not something I'd ever thought about before, but you're absolutely right of course, that when you are talking now, they are interacting a lot with the keyboard and, you know, that's very different from, for example, the sort of interaction that the pediatrician has with like my child, you know, my son is now older, but, you know, with my daughter, because of course you have got to really give a three-year-old your full attention if you want to get anything out. And of course, you see that's a very powerful sort of interaction that I'd never thought about.
But you're right, you're sort of losing when suddenly, you know, the doctor is actually, you know, largely having to interact with this computer to put in all of this data.
[00:22:00] Eric Topol: Right, right. No, and, and in that is a big deal when you don't even make eye contact. It's awful, both sides. I mean, it's unacceptable and it's been the norm for many years now. And I also want to add Jonathan, it isn't just the time where the computer, in the midst of seeing patients, it's all the added time. Off hours, you know, where you have to still do all your charting and stuff that should have been done in real-time. So it's a burden not just in the midst of seeing a patient, but also well beyond.
[00:22:41] Jonathan Wolf: And it's interesting, I see that with my wife because again, as a patient, it's not something I'm ever really aware of, but there's an enormous amount of bureaucracy dealing with patients outside of, of the room. And I think as a patient, I never really thought about it. You thought the only time anyone's dealing anything with you is when they're there. And then you see there are all sorts of other things they have to do. Falling up on the notes, making sure they're right, sending them, dealing with tests, looking at results, all these sorts of things are a big part. So that's very exciting. I would like to follow up on the other bit that you talked about, about the sort of patient self-diagnosis, because that's something that has obviously been really interesting for ZOE. We had this big experience through the ZOE COVID study of suddenly people logging their symptoms every day. Being able to use that to figure out whether people had covid, you know, what the risk factors were or the rest of it. What I'm struck by is that we had this amazing tool and experience and we still have hundreds of thousands of people who are continuing to use the ZOE Health study to do this. And, it's in no way connected to anything else that is going on in healthcare. And my experience is that there's, you know, quite limited linkage of this idea of sort of self-reporting your data through to the health system. So you are being very optimistic, which I love. Where do you see this really working well?
And what are the gaps to get to the point where we really could be? I always like to talk to people about this idea of preventative medicine. Like surely you ought to be able to figure out things are not quite right a long time before you are really sick. And the reality is that our health systems are not really able to do this in my studies today, right?
[00:24:34] Eric Topol: One of the problems we have right now is the level of depth of data we have for each person. What you've done with your nutrition work is kind of futuristic because you not only need to have the electronic health record and of course, in some countries like the US it's highly fragmented, so you need many different EHR health systems, but then you'll go, of course, you need to give things like the biologic layers, that could be the gut microbiome or the patient's genome, or at least their array, SNIP data. You'd like to have their physiome, which is sensors that you would be able to pick up key metrics from that. Their anatome, which is the scans of their relevant anatomy. Their exposome, the environmental factors, and the social determinants. Their immunome, and on and on. Okay?
[00:25:43] Jonathan Wolf: So it's a lot of stuff.
[00:25:45] Eric Topol: Yeah, a lot of stuff. And by the way, no human being could process that now. Even some particularly confident doctors could never process all that data. All right? So we have to lean on machines, but what we're doing and you're doing, as well as pulling together all this data, multimodal AI and processing it for the individual.
And once you do that, you not only can manage a condition like diabetes or hypertension or depression, common conditions, or obesity, but you can help manage a person essentially in real-time by having that data processed and getting back to the person with an algorithm support, but you also, in the future will be preventing conditions that people are at risk for, which is really exciting because then now we fulfill the fantasy.
The fantasy of medicine is to prevent serious conditions from ever occurring, and when you have all that data at the individual level, you can use it to prevent an illness, a serious illness, whether it be a heart condition or neurologic, you know, cancer, asthma, whatever. This is really where eventually we will be.
[00:27:05] Jonathan Wolf: And Eric, just as you say that, you've been voted the number one most influential physician leader in the United States, which is very impressive. So I want to take this opportunity to say that we like to say that this is possible, but you know, you are much more credible than I am.
Do you truly believe this idea of prevention in healthcare is possible?
[00:27:32] Eric Topol: Yes. I mean, it hasn't been actualized and proven, but theoretically, once you have a person's data, the first step is to say each of us is unique. Accept that each of us is truly unique. Even, biological twins, and identical twins have significant differences.
So once you get past that and you say, okay, well now we can get all the domains of data. And we can process that data, that's one of the bottlenecks right now, by the way. And when we do that, not just at a personal level, we do that in the hundreds of thousands and millions of people, ultimately billions of people.
Then we have a way path to not only treat people better, help them manage their conditions better, but also prevent illnesses. So yes. I mean, we're are ways from that. Not right now, but you can see we will get there.
[00:28:33] Jonathan Wolf: So it's an ambitious goal. You believe we can get there, but we need to be realistic.
When you're talking about the stuff that's near and far, like the prevention in the sense of really being able to figure out that you are at very high risk of a stroke in 15 years. This is a long way off. Perhaps prevention in the way that we think a lot about it, which is improving, you know, diet and things like this, we can see is, you know, I think we feel very good about impacting long-term health, but it doesn't mean we're actually figuring out, you know, the way in advance what's going to go wrong?
You know, the analogy I often like to talk about is just with our cars where, you know, I take the car to be serviced once a year. I still have a car. And they plug a computer into it and they run some diagnostics and they say, oh, you know, the fan belt is about to break and we're going to replace the fan belt, and I drive it off.
And as a result, the car never breaks for, you know, 12, 13, 14 years. And we're used to that in all these other aspects of our life. And I think the experience that, you know, all of us who aren't doctors have is, you know, you go to the doctor and if you're not like really sick with something, there isn't that equivalent or really of being able to do this diagnostic and say, you know what, Jonathan, I know you're worrying about Alzheimer's and heart disease and all the rest of it. Actually, you know what, it's diabetes you've really have got to worry about and therefore this is the stuff that you really need to think hard about, you know, We aren't, as a health system in most cases, right, able to do that. Maybe if I have particularly strong family risk or I've got a very rare gene, or I'm getting really sick, but in most cases, you know, that's a long way on and it's much harder to make the change when you are you're already sick than it is to get you much earlier. Do you recognize this picture I'm painting?
[00:30:30] Eric Topol: Yes. I mean, I wish we were as simple as cars.
[00:30:33] Jonathan Wolf: I accept that.
[00:30:36] Eric Topol: You can digitize a car. And what I've outlined is that we're going to be digitizing human beings. That sounds awful, right?
Oh, we're gonna digitize you. But no, it's for the singular purpose of what you just described, of being anticipatory for potential illnesses that you, specifically are at risk for and knowing someday we're going to have digital twins.
[00:31:03] Jonathan Wolf: And what does that mean, Eric?
[00:31:05] Eric Topol: Yeah. We found these five people around the world that match you at every level as closely as possible.
And by the way, you know, some of them are much older than you, and this is what we learned, how to prevent this condition or that condition. This is what really worked well to treat cancer that you have or going to get potentially. So I mean, we have never done this before. Learn from each other.
We do these clinical trials and let's say we have 10 out of hundred people. That's a big success story in a clinical trial in that we found a treatment that helps 10 out of a hundred. What about the other 90 out of a hundred that take the medicine every day for the rest of their lives? Well, what about if we start getting incredibly precise?
Through digital twins when we have, you know, hundreds of millions, ideally billions of people where we have all these levels of data and we can help share and, you know, make matches so that people will know the best treatments, the best preventions. That's someday. That's further off. That's further off, Jonathan, than what we've described.
[00:32:19] Jonathan Wolf: I mean, I love that because you are sort of articulating in a different language. I think a lot of the ideas that we have at ZOE, that Tim and Sarah and George and I talk about quite a lot. And I think it's wonderful.
I'd love, to pull up a little bit, you've talked about personalization quite a bit here. And I think that's really interesting because, you know, I know that Tim and others talk about the way that when he was taught as a doctor, everything is that, it's a one size fits all solution.
You might diagnose somebody with a particular disease, but then you don't further break this down into lots of different categories. You've got sort of this treatment, but a lot of what you've been talking about is how different we all are, I think, and I'd love for you to talk about how important you think that is, and I guess for people listening is that really going to be able to affect the healthcare they get in, you know, in a decade?
[00:33:19] Eric Topol: Well, sure. I mean, I think, you know, what you've seen, in your work, with Tim and others is exactly what we've seen, which is it's just as simple as, if you have a glucose sensor on and you eat something, and some people have spikes of glucose too, you know, 180, 200 and some people totally flat and the exact same food with the exact same portion.
It was really, Aaron Siegel and his colleagues at the Weizmann Institute in Israel that made that first observation several years ago, which I thought was stunning, but that basically told the story and it's the same across everything in health. Everything.
If it's as simple as what you've eaten, then you can imagine, that the fact that we treat people like cattle and we don't recognize their uniqueness insofar as our approach, a general approach, that's also part of the problem. So we have a remedy, we have a fix for that. And it's across everything. And nutrition is one that's, I think, I'm particularly enamored by and wrote a lot about. Deep diet in the book. And the AI diet, if you will, that the New York Times called it.
But, you know, I think this is an exciting area because we spent a lot of our time with what we eat and our sleep and, you know, our exercise, ideally, and a lot of that, a lot of those things can help us preserve health.
[00:34:53] Jonathan Wolf: As we think about the future, I think you've got this picture I think that we know we want to end up with where almost everything has been monitored. There's all of this AI to figure out all of it. It's these incredible tools that support the doctor to figure all of this out. And then, you know, I'm getting this completely personalized solution if I have particular cancer or any of the rest of it. Clearly, we're going to go through some stages from, you know, here to there. And we will also of course have some missteps and we'll think there are things that are really important that we should monitor and they turn out not to matter. I think one thing I'd be really interested in, is I guess, your view about what's the sort of things that we will want to be monitoring sort of overtime on ourselves that you think have value and maybe those that don't.
Because I am struck. There's also like a risk of a lot of companies selling, you know, a particular thing that you should monitor all the time. And I've seen a few of these where, you know, I'm a bit skeptical based upon what I've seen from our data and other scientists. You know, not everything that you can monitor is necessarily important, right? We all know that actually if we think about our own children, there's all of this stuff going on, and actually, you're pretty good at figuring out, actually, they're not quite right, right? You're picking that up in some way. What do you think are the things that, either already we know, have value or are going to have value?
Just, you know, based upon either what you, you've seen yourself or from the research.
[00:36:25] Eric Topol: Well, I don't think it's, they measure all things in all people. I mean, that in itself is particularized, based on the different layers of data. I do want to go back to your assertion that, you know, everybody's going to do this.
I don't think that's ever gonna happen. There's a lot of mistrust in data and privacy of data. There are solutions in sight. In AI, if you talk to AI, computer scientists, every problem in AI, whether it's biased and whether it's privacy, you name it, there's a solution with AI, right?
I mean, it's classic, but most of those are not true. But the one that does resonate is that we will do better with privacy and security, whether it's edge computing or swarm, federated or whatever it is, homomorphic encryption. So the reason I bring that up is that's a barrier as is, you know, bias and as is, you know, other factors that play in this whole area.
But no matter what, we will have many people that are reluctant to use technology in healthcare. I mean, look what we've seen with vaccines. Probably the most compelling data ever assembled. In the history of medicine and look at how much resistance there is. So I don't want anyone to think that this is going to be, you know, a vast majority, but I think it's going to be substantial enough to be very, very important.
So just make sure that that's clear because we tend to get carried away that the way we think is the way everybody thinks. And that couldn't be further from the truth.
[00:38:23] Jonathan Wolf: Totally agree, Eric. I think trust is an enormous part of this. And, the other part of the course is feeling that there's real value, right?
So we always feel better about recording something or sharing our data with somebody like our doctor because we know we're going to get something valuable back.
For the people then who are saying, you know, I want to do this. I'm happy to go and, you know, wear a particular device or whatever. What do you think are the sorts of things that are being monitored today or you can see being monitored in the next couple of years that you think might have real medical value?
[00:39:07] Eric Topol: Well, the first biosensor, which is now used in millions, of course, is glucose, continuous glucose. And that's changed the world for lots of type one diabetics.
[00:39:20] Jonathan Wolf: That's right. And for everybody, that's blood sugar for anyone who's not familiar with it. So this has been able to monitor your blood sugar, you know, in real-time with some little device you stick on your arm for example.
[00:39:31] Eric Topol: Yeah, and this is a big deal because, you know, obviously people taking insulin are at risk for having very low blood sugars, or not having good control of their blood sugars.
So this is a whole lot better than getting a blood test, you know, every three months, like a hemoglobin A1C which is used so, that's the first sensor that got into wide-scale use. And then of course, with the fitness bands like Fitbit and smart watches like Apple and many others now, we are seeing heart rate, heart rhythm, but there'll be others, of course, you know, oximetry, it became important during the pandemic with pneumonia monitoring.
[00:40:17] Jonathan Wolf: This is the amount of oxygen in your blood, Eric, to make sure. That's oximetry, is it?
[00:40:23] Eric Topol: Exactly, so, which can be monitored through a finger or the wrist very easily, but there's no limit.
I mean, eventually, we'll have high-frequency blood pressure, non-invasive passive monitoring, so you won't have to take out a device and measure it yourself. It'll be done automatically, but we already have things. To me, the most extraordinary thing is self-imaging. So using a smartphone with a high-resolution ultrasound probe that you put in the base of the smartphone, you can image any part of your body except your brain.
[00:41:01] Jonathan Wolf: So this is like basically the same as when you know you are pregnant and you do the ultrasound to find the baby, but you can plug it into your phone?
[00:41:08] Eric Topol: Yes. Now, it wouldn't probably be good for a mother, expecting mother to image their baby every few minutes. That's not what I'm getting at. But what I am suggesting, like we're already seeing in heart failure, is that people can image their heart and send a loop to a doctor and you don't have to have any training. The AI tells you how to get the image. All you have to do is be able to put the probe on the chest.
[00:41:38] Jonathan Wolf: That's amazing. So it's another one of these ways where you can push the technology right out to you at home and you can give like really valuable information. Back to the doctor.
[00:41:50] Eric Topol: Exactly.
[00:41:51] Jonathan Wolf: I would like to just do a quick summary, which is what we always do at the end here. And I think, you know, this has been particularly sort of visionary, which is really exciting.
I think that you know, my key takeaway is, you know, from your perspective actually healthcare's been sort of getting worse over the last 40 years. And you paint a picture where it's really true that the relationship between the doctor and the patient is worse. And that actually the technology has sort of been getting in the way.
But the good news is, you do see this sort of sunny uplit future where AI can really change that. And there's a lot of these technologies are sort of there. They're just not in the sort of mainstream deployment, but therefore they could be there in the next few years. You're not talking about decades away.
I think that some of these we will just see when we visit our physician, our doctor, because suddenly, hopefully, they're gonna be able to look at us and engage with us. And there are all of these things in the background. But the other element that you see, I think becoming really significant is people tracking their own data. Now, what exactly that data will be that turns out to be high value is still being figured out. And you talked about blood sugar as an example. You talked about some of the things like blood pressure, but we're going to learn a lot more. And I guess the final thing is, for you, personalization is central to all of this.
That actually, in a way, medicine right now is this thing, you've got a drug that's going to work for us, you said 10 and a hundred people, you know, how do you solve the other 90? How do you figure that out? And so again, this data, hopefully, and this AI allows us to move to a world where not only can you think about prevention, but also you can really think about personalization and I'm slightly putting words in your mouth, but I think you're therefore suggesting there could be a real transformation in the success of medicine if you can really understand that personalization.
[00:43:44] Eric Topol: That's a great summary.
[00:43:45] Jonathan Wolf: Eric. It was such a pleasure. Thank you for taking the time. I look forward to following up and hopefully keeping you abreast of what we do at ZOE and hopefully getting you back on again in the future.
[00:43:57] Eric Topol: That's great.
[00:43:58] Jonathan Wolf: It's a real pleasure. Thank you, Eric. Bye-bye.
Thank you, Eric, for joining me on ZOE's Science and Nutrition today. If based on today's conversation you'd like to understand the difference personalization could mean for your diet and your health, then you may want to try ZOE's personalized nutrition program.
Your ZOE membership comes with our app, which uses artificial intelligence to score millions of foods and meals for you. You also get a personalized meal and recipe recommendations and can chat directly with our nutrition coaches. As a member, you start with an at-home test to understand your biology. We create your individual ZOE scores and a personalized program for you based on your test results and our scientific research.
If you're interested in learning more about ZOE, you can head to joinZOE.com/podcast and get 10% off your purchase. If you enjoy today's episode, please be sure to subscribe and leave us a review as we love reading your feedback. If this episode left you with questions, please send them in on Instagram or Facebook, and we'll try to answer them in a future episode.
As always, I'm your host, Jonathan Wolf. ZOE Science and Nutrition is produced by Fascinate Productions with support from Sharon Feder, Yella Hewings-Martin, and Alex Jones here at ZOE. See you next time.
