Sniffing Out Hope vs Hype in Digital Health
In this episode, the “Godfather of Digital Health” tells us what he thinks of where we’re going in healthcare.
Dr. Eric Topol is a practicing cardiologist and the Founder and Director of the Scripps Research Translational Institute. He has published over 1,300 peer-reviewed articles, and is one of the top 10 most cited researchers in medicine. He is an original thought leader on digital health, genomics, and artificial intelligence with an enormous following on social media.
Prior to Scripps, he led the Cleveland Clinic to become the #1 center for heart care and was the founder of their medical school. He has also published 3 bestseller books on the future of medicine: The Creative Destruction of Medicine, The Patient Will See You Now and Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again.
Topics covered:
The biggest barriers to innovation in healthcare - and why we haven’t yet realized the potential of technology
Open source healthcare
Evaluating new innovations (and how he sniffed out Theranos as an early skeptic)
The potential of AI in medicine
Listen
Transcript
Halle: Welcome to the Heart of Healthcare podcast. I'm your host, Halle Tecco, and today my guest is Dr. Eric Topol. Eric is a practicing cardiologist and the founder and director of the Scripps Research Translational Institute. He has published over 1300 peer reviewed articles. And is one of the top 10 most cited researchers in medicine.
He is an original thought leader in digital health, genomics and AI with an enormous following on social media. Prior to Scripps, he led the Cleveland Clinic to become the number one center for heart care and was the founder of their medical school. He also has published three best selling books on the future of medicine.
The Creative Destruction of Medicine. The patient will see you now and deep medicine how artificial intelligence can make healthcare human again. But how I really wanna introduce Dr. Topol is that he was a very early supporter of our work at Rock Health when most of the healthcare world was very skeptical of our vision for digital health.
Dr. Topol believed in us, and for that, I will be forever grateful.
Dr. Eric Topol, welcome to the Heart of Healthcare.
Eric: Oh hell, thanks so much. It's great to be with you. And I do remember the early days of Rock Health and how you were rocking it. And I was really so thrilled to see you and all the young folks, all electricity, just lighten it up in, in digital medicine.
Halle: Thank you. Yeah, I will always remember those who supported us because not everyone did. But looking back to when we first met in 2011, would you say that today we are further ahead or further behind in digital medicine than you thought we would be by 2023?
Eric: Well, we're way further behind. There has been some progress, of course.
Yeah. But it takes things like a pandemic to bring in telemedicine and you know, it does, medicine moves so slowly and, uh, you know, at that time so many years ago, we had envisioned the more wide scale adoption of wearable sensors. And potential of remote monitoring and so many things that haven't really been actualized.
So there's been some progress for sure. We're, we're seeing a, a small amount of what is inevitable with respect to consumer patient empowerment, uh, with digital tools. But that has a long ways to go to be where it can be. And hopefully the excitement right now regarding the large language models in ai, uh, will help push things along to some degree.
Halle: Yeah. Do you feel like it's a coordination issue? Like the innovation is there, the patients want these tools, the physicians know these tools can be helpful, but the payers are dragging their feet? Or is there a certain party or stakeholder that's, that has made this more difficult?
Eric: Yeah. Well there's lots of finger pointing, uh, and, and blaming.
But perhaps the biggest problem is there haven't been enough of the pivotal clinical trials to provide compelling evidence. That the, you know, the change it to the new state-of-the-art care of patients. And so that's really on the medical research community and the funding of that and the commitment to, to have that compelling evidence.
So until you have it, it's much harder to get the clinical community to change practice, to get. Payers and reimbursement, uh, all established. So often the rate limiting step, the bottleneck is in that type of evidence base, which we just don't have enough. We have some randomized trials, some things that have a clear path towards that, but yeah.
Uh, so many things that just have been left, uh, in a state of not adequately backed up.
Halle: Yeah. So today, approximately a third of the world's data volume is being generated by the healthcare industry, but we're still using fax machines. My medical records are still, you know, locked up in one physician's office.
How do you explain this juxtaposition?
Eric: Yeah, I was trying to remember what is a fax machine, right? I mean, this is incredible. I. I mean, I know how being an old dog that things move slow and uh, it's kinda ritualistic and ossified or sclerotic medical community. But I hope we can start to get things into higher gear just because if we're gonna make the improvements, particularly in the patient, Doctor, patient, clinician relationship, which has eroded steadily.
We, we've got to adopt the technologies that will enable that. Mm-hmm. Uh, but there's, there's lots of issues. I mean, you know, for example, at the fda, they've cleared over 500 AI apps, AI tools in medicine. A lot of those are in radiology and cardiology and some other disciplines. The problem is almost all that data is unpublished.
All proprietary, so there's no transparency. So the health systems and the physicians can say, well, I can't see the data. Uh, it's just five 10 k cleared and you know, I'm not doing anything until this gets published. So this is part of the difficulty of implementing things when you have lack of transparency, lack of, you know, really strong evidence.
Mm-hmm.
Halle: You talk a lot about open source healthcare. What do you mean by that?
Eric: Well, it's just the opposite of this opaqueness. Yeah. Um, that is, um, you have all the data code, everything that's, uh, giving full transparency to the potential users, whether that's health systems, uh, physician practices, you know, whatever that is.
So they can be assured that they understand how it works, what are the nuances, what are the surveillance issues, um, after implementation. So we're missing that to a large degree and that. Just gives another excuse to add onto the list of not implementing the tools that we have before us. What sort
Halle: of impact do you think it could have to move to an open source model of healthcare?
Eric: Oh, wow. Uh, it would accelerate it. Okay. Yeah. But you see, it's kind of against the company interests. They want everything to be proprietary, and as long as that remains, as long as there's unwillingness to share or even publish their data, you know, um, You can't even find a preprint for a lot of this. It's just in the bowels of the FDA that on the public or the medical community will never see.
So if we had that type of O open science open source, then the buy-in would be much easier to obtain. And, and, and, you know, downstream, we would get it into patient care. A lot more depth and speed.
Halle: Yeah. We've seen a lot of big tech companies come and go in healthcare. I don't know how many iterations of like Google Health we've seen.
Um, and we've also seen a lot of startups come and go, right? Some going out with pretty huge implosions between kind of the, the big tech companies and the small startups. Where are you most bullish about innovation coming from and who do you think is most well positioned to scale something meaningful?
Eric: Yeah, well, one thing you can say is Rock Health is still standing and doing well, uh, thanks to your founding efforts. Um, you know, I think the companies have gone through lots of different, uh, phases. Google Health, as you alluded to, You know, has gone through a lot of turbulence and now the Google Health that was at one point envisioned to combine lots of different parts of Google was radically altered.
So Microsoft at the moment, because of their big investment in open AI and G P T four is in a pretty powerful position, uh, even though none of the large language models are truly trained. Medically, just what's in their vast inputs, uh, of, uh, their knowledge base. Uh, data inputs. They can perform pretty darn well for lots of tasks, uh, in the medical world for both patients and for clinicians.
Mm-hmm. So I think Microsoft is at the moment, uh, Google, of course, is chasing them. Uh, and then there's the other big hens like, uh, Amazon and Apple and, and, um, meta and Salesforce and many others, of course, Oracle. So, um, you know, we'll see. It's, it's clear that they all, all of the big tech companies are making, uh, a, a major commitment.
And as you. Touched on Halle, is that they previously have made commitments. Yeah. And then, and then they kind of stalled out because they realized, oh, you gotta get things through the fda. Like, what's that? Yeah. Um, but now I think, uh, it's getting serious because the, the frontier here is so vast. The opportunities are so extraordinary that, uh, without giving this a priority, then they're missing, uh, enormous opportunity.
Halle: Yeah, well even you mentioned Amazon, even that huge Amazon, Berkshire Hathaway, JP Morgan, cheese, that, that enormous effort, they started that venture maybe four years ago. They disbanded it after three years.
Eric: Yeah. Haven went to hell. Yeah. Haven. Yeah. It was, it didn't last. It sounded really good. Yeah. I mean, they attracted a, a real superstar with a tool Gandi to lead it and, uh, had a lot of potential, but it, again, it reflects the difficulties of change.
In the medical mm-hmm. Sphere, uh, and just cause you have the muscle and resources of a few huge companies, that doesn't mean you can, you can do it. It's a, it's a real challenge and a lot of it is because of the US healthcare system that's fairly unique, lacking universal healthcare as opposed to all other industrialized nations in the world.
That it's a right of each citizen to have healthcare. So we have lots of, Inherent pervasive conflicts and uh, and, and difficulties that make our health system all the more challenging.
Halle: Yeah. But it seems like these companies might not have the patience required.
Eric: I. Yeah, that's part of it. Yeah. And, uh, you know, they have to go through all kinds of fits and starts to finally realize that it's worth, uh, the persistence.
Uh, yeah. Because ultimately, since there isn't any dominant player of the big tech titans, and there's no shortage of startups that have, as you said, some of them have folded already, but there there's many that. That, that don't make it all the more that actually come on board. And so it's a Darwinian process.
We'll see how it plays out, but, uh, I think in the next few years we'll see the most serious priority, uh, that has ever been seen with respect to all the technology companies towards their entries and their major commitments to changing healthcare as we know it today.
Halle: I hope so. So on one of your Silicon Valley visits, when you came to see us at Rock Health, after you visited a then darling of Biomedicine Theranos, and at the time the media was positioning them as the most promising startup of our lifetime, Elizabeth Holmes was time person of the year.
But I remember you coming in to meet with us and you're the first person to ever say, Hey, I'm a little suspicious. I think they're hiding something. And you really had questions about the validity of their technology. Uh, and I always think of you when I think of the Theranos story, because this was before John Caru came out with his findings.
Tell me how, we don't talk about Theranos specifically, but when you go about evaluating new innovations, which you do all the time, how do you determine and kind of get that gut sense of if this is. Hope verse hype verse hypocrisy.
Eric: Right. Well, I do remember that meeting really well with you and the crew and, uh, it was the second time I had met with Elizabeth Holmes.
The first time was when I did the very first video interview with her. I think it was, um, at least 10 or 12 years ago. And, uh, at that time, I, I. I had the impression that she was, you know, really there's something strange about her, but on the other hand, mm-hmm. Uh, as a young person who had recruited her, her own Stanford professor, her mentor, to help her with this aspiration of, You know, creative destruction of lab medicine.
I mean, the cognitive bias was, oh, I hope she's successful. But by the second time I met her, which was a well over a year, year and a half later, when we had spoken about doing the studies you need. Head-to-head studies of Theranos technology versus existing lab technology to show that there was equivalence of results.
She hadn't done anything. Zero and uh, it was starting to become clear. She didn't plan on doing anything and infor, in fact, when it all blew up after John Carreyrou who had to spend 18 months. On this as a bulldog to finally expose her through a couple of e e employee employees that, that finally, um, you know, disclosed what was going on near, behind the scenes.
But the problem I saw with her, and I also expressed, uh, in a New Yorker profile of her is her, uh, unwillingness. We, we talked previously here about transparency. There was none there. Mm-hmm. And, and not she she'd. She wouldn't even let me the first time I, I had come to visit, you know, see the labs and there, that's because there was nothing there apparently.
So, yeah, I got increasingly suspicious and I, you know, clearly, you know, it was thanks to John Carou that really undressed the whole thing and, and blew it apart, but, You know, in, in, it was a pretty unique situation because her idea was very good. And ultimately, while you won't be able to do unlimited number of lab tests through droplets of blood, you'll be able to do quite a few.
I mean, ultimately, the technology to do what she had aspire to will, will probably get there in the next couple years. Uh, not fully, but. You know, to some degree hope. Yeah. Yeah. Yeah. I mean, the last thing it's a great idea. Yeah, exactly. Uh, and you know, I think, you know, obviously she was a criminal and I don't know when, along the way, her good intentions, which I think they were there initially changed to willingness to let patients be harmed.
To have the success of her company. That was, you know, extraordinary. Yeah. But, um, I think when you evaluate, uh, technologies and it's always about the people you know, do they appear to be trustworthy and open, have high integrity, which is it turned out of course, wasn't the case there? And is there technology really exciting?
Is it something that's just. You know, kind of a, a small increment or is it something that could really, uh, be considered hyper innovative? So there she had the hyper innovative side. Yeah. It, it, and uh, unfortunately a lot of the other things were very drastic shortcomings.
Halle: Yeah. A lot of times things have huge potential.
Team seems great and. The technology might be validated, but how do you tell when something has the potential to actually scale within healthcare to be adopted by the right people? Paid for by the right people?
Eric: Yeah. I mean, I think the. The challenge here is, as you know, well, since you evaluated, you know, many of these companies in this space is identifying a, a, a major unmet need.
Mm-hmm. Uh, and then how well almost pa passively without a lot of effort, could it actually be integrated? Yeah. So, I mean, how, how. Much friction and difficulty. Do you foresee, are there significant issues with reimbursement that are gonna have to be confronted? What it, what if is the level, extent of evidence needed to provide to either patients regulatory health systems that they will adopt it?
So there's lots of questions that have to be entertained and it's kind of the synthesis of, of all those, to give you a sense. This could really make a difference. Or it just, is it worth it? And there's a lot of times when it looks great, but then there's unanticipated bumps along the way. You know, whether it's the technology, whether it's, you know, the re the, the regulatory, the reimbursement.
There's so many points of the obstacle course. Of course. And you, it's very hard to predict if, if you're gonna see that kind of obstacle arise or whether it will become relatively easily surmountable.
Halle: Yeah, we, um, we would, in the early days of Rock Health, obviously all the companies we invested in, we had a lot of conviction.
But that, for some companies, starts to fall apart. As you see the challenges in the market, and it could be the greatest idea, but without, you know, market appetite and fit and the. Policy tailwinds, for instance. It's really hard to kind of commercialize some of these great ideas. And you know, it was one of the things that we always tried to do at Rock Health was we knew we had to bring in the healthcare people and people inside the system.
We weren't trying to disrupt healthcare from the outside. It was like, you know, we are bringing new people in, but we need the stakeholders on the inside who understand how this complicated system works to help validate the ideas and bring them to
Eric: life. No question. Yeah, I, I, I, it brings to mind Holly, about, I believe one of the companies was the one with the smartphone to do, to check for children with ear infections.
Yeah. Yeah. CellScope and yeah, I, I, what did they finally make it CellScope?
Halle: So that's a, that's a great example. CellScope unfortunately did not make it. Yeah. See that's, um, great team came outta lab, you know, out of a lab in Berkeley, and I think ultimately for them to distribute, and I don't have the official like postmortem on them, but for them to distribute through pediatricians was just a really, you know, it's a.
Difficult distribution model. Um, I mean, it's
Eric: incredible cause it was, it was ingenious. So simple. I know. I mean, I even used it to, uh, to examine Stephen Colbert's year and his ruptured, his ruptured eardrum on tv, you know, but it was so simple. It worked so well. And you would've predicted I would've, and I knew the CEO and the, the team.
They were amazing. Yeah. Great team. Great team. Just, you know, phenomenal group. Why if that didn't succeed, that shows you how, how hard it is because I know, you know, and why do you have to go through the medical side? Why don't you just go right to parents Yeah. Who don't want to have to deal with the difficulties of urgent care and emergency, uh, rooms and well, maybe
Halle: today what you're doing.
Yeah. Yeah. Maybe today they could, right. Like today, telemedicine adoption is, is far higher on both sides. On the clinician side. Mm-hmm. And on the patient side. So maybe if someone were to kind of resurface the technology today, the telehealth side of it is, is actually played out. Right. We're seeing, um, Much greater adoption.
So maybe it was just a timing thing. They were too early. Maybe.
Eric: I, I, I hope so. Yeah. I, I really think it was a, a first rate, uh, technology. Simple. Yeah. And inexpensive. It could have made a big difference. Maybe it'll resurface. Yeah. Yeah.
Halle: I hope so. I hope so. We've made great strides in technology, but the fact is we're adding costs and we're not living longer.
Like when you actually look at healthcare spending and that is increasing, and then you look at. Our life expectancy and that is flat, if not going down, how do, what do we do? How do, how do we keep doing this and hope for change when nothing is actually changing? We're just making healthcare more
Eric: expensive.
Yeah. The problem is you're referring to this country, uh, the United States where it's unique. It's the only country of the 36, uh, O E C D, industrialized nations that have had such a substantial, almost three year reduction in our life expectancy. And at the same time, expends the most per person.
Healthcare well, well beyond any other country. So we are in a very bad predicament here, and we're going from kind of bad to worse in terms of the gradient of our, uh, expenditures and outcomes compared to all peer countries. So, uh, the, the simple thing, of course, is as we discussed, that, you know, every citizen should have, um, Healthcare, uh, right as a citizen.
And that is a, a a whether when and whether that'll ever change here is, is a real concern. But then if you say, well, okay, you have resources, and how come you don't see improvement in quality of life or life expectancy? You know, why is all this, uh, spending for healthcare not associated with, uh, improved outcomes?
And, you know, one of the interesting things there is, Because our system is designed to do things, you know, do more procedures, do more tests. We have a rabbit hole problem where the wealthy. The Lon, uh, can go and get workup on anything and just get more tests and they'll find things, incidental things, and then all of a sudden there's a complication of one of these procedures.
So we have this paradox of people that are the worried well that are overcooked with medical Yeah. Stuff, because that's our system. It, it basically is propelled. By doing more rather than doing what's needed. And we have no really good prevention strategies, which I think that's gonna change, you know, with the ability to, with these large language models and integrate all of a person's data.
Over time will we'll see that improve. But right now, prevention, primary prevention of illnesses, which is the big step towards improving health span, we, we just don't even show up yet.
Halle: Yeah, it's disappointing cuz we wanna think that our work is, the work that we're doing is, um, we know in at a micro level, we hear from the customers, we hear from the people who are benefiting from these new technologies.
But when you look at it at a population level, we're just not seeing a difference.
Eric: That's hard. No, it, it is. It's, yeah. And I'm sure everybody listening has experiences. Yeah. Who, when they, when they, they had some kind of query medical question and they, the amount of testing that they went through to get an answer seemed to be far more extensive than was necessary.
Mm-hmm. Uh, and you don't see that in many other parts of the world.
Halle: We will be right back after the break.
So your most recent book, deep Medicine, explores the potential of AI to transform healthcare. What do you see as the most promising applications of AI with healthcare? And do you have any concerns about the adoption?
Eric: Well, yeah, you always want to have concerns cuz there's no thing that is kind of foolproof that doesn't carry the potential for harm.
But, uh, I've never been more excited, uh, for our healthcare future. And that's because there's different phases of where ai, and particularly, you know, this large language model, generative AI that's kicked in in recent months to take hold. Uh, the first of course is improving accuracy. Which is a big deal, whether that's from a scan or uh, making a diagnosis, we have.
A very serious problem with, uh, major medical errors in this country. Mm-hmm. Over 20 million a year. Uh, and we can fix a lot of those by using AI as a support, as an assist tool by better interpretation of scans, not missing things, not giving false positives, uh, but also, The data, you know, not missing things that's in a patient's, uh, electronic records, which are oftentimes are multiple records and not keeping up with the latest in the medical literature, the corpus of medical knowledge.
So that should improve on a longer term beyond the fact that we're gonna see the automation. Uh, hopefully rescue the so-called what I call keyboard liberation and all that movement. We can talk about that. Yeah. But the longer term view, the, the, the thesis of the Book of Deep Medicine is this counterintuitive notion that we can use technology to improve the humanity, to restore the humanity in medicine.
Uh, but that is the gift of time. That it can provide us would then give us the, the what we need to restore that patient doctor relationship. And that means having a presence, a trust, a empathy, the communication, you know, that the patient knows you have their back. We don't have that largely today. We had it in the 1970s and eighties when I was just coming into medicine, but it, it's basically eroded.
And so I do think AI will get us that back if we are, if we go after it. It won't happen by accident, but if we make that a priority over time. The gift of time could get us to a very, uh, I important, uh, restoration of what is the essence of medicine, which is the human component, the interhuman bond. Yeah.
Halle: And are you finding that providers are warm to AI or skeptical? Or both?
Eric: I think both. I think that what they really see the benefit, as you can imagine, Hallie, it could have saved them money. Okay. So yeah, the idea here is, oh, you mean we could get rid of keyboards in the office and the doctors and the nurses, they wouldn't have to type stuff and the synthetic notes would, would then lead to.
Prescriptions and follow-up appointments and the lab, uh, orders and fol and pre-authorization notes and, you know, everything would be done. Wouldn't that be great? And then we can have the, the doctor see more patients and they could read more scans and more slides. So, Uh, the idea that, you know, the, the coding could be more efficient with ai.
These are the kinds of things that health systems say, oh yeah, I could make more. But the reality is, if we understand the power here, which is substantial, uh, particularly with the large language models, when you're pre-trained specifically for these. Important purposes, then we have a kind of a reset. Uh, and it's the problem Holly, is, as you know, well, most health systems, uh, are run, you know, by um, uh, administrative managers, the overlords of American medicine.
Right, and they have different interests than the patients and the clinicians. And so we, we have to come to a better place where this relationship becomes center stage. And I envision someday, Halle, that we'll see health systems instead of now when they would advertise that they have robots to do the surgery or you know, whatever they're advertising, they might advertise and have billboards about.
We give the gift of time to our. Uh, patients and our, our doctors and nurses, because that's what's missing now. And that rush, that lack of time. Yeah. Um, which is what leads to so many mistakes.
Halle: Yeah. Yeah. I feel like, uh, providers are always running late and you get less time than you want, so. Mm-hmm. Um, and you know, the biggest complaint I hear from the PHY on the physician side is really like, I'm spending way too much time in the emr.
Eric: Yes, exactly. And then a lot of that's off hours and weekends and, uh, no less, you know, throughout the day we can't even look, uh, eye to eye with a patient. Yeah. Because of that. So this is what has been a serious compromise of the profession and, um, It's not just felt by the doctors and nurses with profound disenchantment and depression and burnout, uh, and even suicides because of the inability to care for patients, which is why we went into medicine, of course.
So, uh, this has to be remedied quickly. And I, I do think in the next couple of years, this will be one of those very short term radical uptakes. It's already being piloted in many health systems throughout the country with different tools that. That get us started in this automated way, and I hope that that is something that we'll see, uh, an accelerated update because we need it.
Yeah.
Halle: Can you tell us or explain to us what algorithmic fairness is?
Eric: Hmm. Well, there's lots of different ways that algorithms can be biased and unfair. Uh, discriminative, some of it is a good part about the inputs. So, because, for example, let's say you had a, uh, Algorithm that would interpret a skin lesion, like a photo on a smartphone.
Mm-hmm. And you wanna say, is it's cancer? Is it a rash? You know, what is it? Well, if you only train the algorithm, uh, with people, uh, who are a European ancestry and know people of color, it's gonna be a lemon. It's gonna fail. And so that's one issue of lacking fairness and a profound bias. But then you have, uh, another example is, uh, Optum, where they had tens of millions of people, uh, in their Optum, uh, data resource where they made this, um, conclusion that people of, uh, African ancestry had a different resource.
Use than those of Europe of European ancestry. It was completely flawed because of wrong assumptions. Uh, so one of the things that we've learned about algorithms, most of the problem is not the AI per se, it's the lack of. Of deep interrogation of the inputs, things that are, are embedded in our culture or a lack of, uh, foresight in the people who are writing the codes, who are doing the training, who are doing the testing and validation of the algorithm.
Most of the time, once you get past that and you, you vet it and, and you look carefully at. The input of data is, it is diverse as it needs to be. Then you, you, you're, it, the algorithm will probably not be the, the problem. But unfortunately, AI gets blamed largely when it's really people. Uh, it's really the it, it's the culture and the people that are developing algorithms that need to take on the responsibility.
Yeah,
Halle: well, hopefully we can have AI that can be, uh, self-critical and understand if it contains bias due to the
Eric: inputs. Well, you're bringing up one of the things about AI that's funny, Allie, every time there's a problem identify with ai, the AI scientists say, we'll use AI to fix it. Uh, but unfortunately, That might work to kinda deconstruct a, a, a, a deep neural network to understand what are the features that make it accurate.
But it's much harder as we've learned to use AI to, to find misinformation or to detect bias and lack of fairness. Hopefully it will help, but we hear human judgment and, you know, deep interrogation is gonna be necessary until we have, uh, assurance that AI can help. Eradicate this problem. It is a significant problem, and we've seen it with so many things.
I mean, you remember the, the problems with oximetry through, uh, wrist wristbands and how Pulse oximetry Yeah. If the person of color, you know, didn't work very well. Uh, and this, and, and released at scale without having that data. So we have to do better. I hope we will, I hope we'll have learned from the genomics community because they're a decade or more to figure out, you know what, if you don't study people of all ancestry and ethnicity and have data that's rich, you're, you're not gonna do well.
Halle: Yeah. Do you think that we're gonna see AI helping improve access and equity for underserved populations, or is it gonna be. Kind of like everything else has gone, which is it helps the wealthy first and then hopefully trickles down.
Eric: Yeah. Well, this is something that I'm excited about too because I've seen some great examples of how AI tools have been, uh, adopted much more quickly in low and middle income countries than here.
Hello. So, um, my favorite example, there are many, uh, is a smartphone ultrasound. Because what you're seeing in Africa, in the hinterlands of Africa, Indian, other countries, is that people with no training to acquire an ultrasound can have the AI direct them even to get an echocardiogram, which is complicated cuz it's videos and it's the heart.
It's in motion. But other parts, any part of the body except for the brain, you can get an ultrasound image through the smartphone with a probe. That just, you know, attaches to the base of the smartphone. So the AI will say, you know, as long as you put it like on the abdomen, it'll direct you, you know, how could, how do you get an image of the kidney or, you know, the gallbladder or the liver or whatever.
And, um, then it also can provide auto interpretation. Very cool and it's been used, uh, in Africa for a diagnosis of pneumonia by when it's used for the lungs. And what's really extraordinary is you have the AI not only helping to guide, to acquire the image, and you don't even know as the person that you got the image, cuz it's auto captured.
As soon as the AI detects it's high quality and then you get. Interpretation of the image with the algorithm. So this is a, a great tool, uh, for places that don't have this technology widely available. Uh, no less the expertise in interpretation. Now, of course, you always want to have humans in the loop.
When you're gonna have a management, uh, issue of how do you treat the patient. Yeah. Uh, but this helps to get the initial screening, kind of get a handle on what's going on with patients. I think it's, it's gonna play out to be important. And it's just one of, of several examples like that. Yeah. What's the, what's
Halle: the name of the company or the.
Product.
Eric: Well, there's several. Okay. Uh, the one that I've seen used, uh, at least the one that was, uh, in the, in a Big New York Times feature, uh, was the, uh, I'm trying to remember which ultrasound smartphone, but it was with caption, which I think is now part of ge. Okay. So they used one of the startup guidance for Yeah.
Getting the, uh, acquisition of the images.
Halle: Yeah. Very cool. So you write, you treat, you teach and you convene. I'm not sure what else you do, but, um, have you ever thought of starting a company? I.
Eric: Well, I have been involved, uh, with a, uh, on a few startups. Yeah, yeah. Uh, usually they, they failed. Um, but, uh, you know, I, I guess, uh, I wonder, I'm trying to remember if any, have I been successful?
Um, not really. Well, you know, one that's still in suspension, you know, I work with my friend Steve Quake at uh, Stanford, who's done a lot of startups, he's a really hyper innovative guy on a company we call molecular stethoscope to take, uh, RNA circulating in the blood, a tube of blood and be able to, uh, do like a, a test of multiple organs to see if there's any malfunction that that company's still alive.
Hopefully, eventually it will find its place, but outside of that, you know, I think. I don't ever really, I don't know if I want to do that anymore because I'm not, I don't have a good track record. Uh, I've done a few and they haven't really, they survived. They didn't completely, uh, go, go bankrupt, but they didn't really do what they were intended to do, which is disappointing, but mm-hmm.
What's more fun is to see technologies like we discussed. Uh, and help be as an advisor, or at least even informally, to bounce ideas off to help catalyze a success.
Halle: Maybe VC is another hat you could
Eric: wear. See, I don't know about that one. But yeah, I mean, I interact some with VCs. You know, they're oftentimes, they're about a lot of sharp people, but the problem I have is they tend to really crowd out the inventors.
The founders. Mm-hmm. Yeah. And I don't like to see that, you know, they, yeah. They, they, they are often people that don't have resources and before you know it, they don't even have much of the company and their idea. Yeah. Um, so I, I don't know the vulture capital, I mean, some of them are really good, I guess, but I, I often think that, uh, it'd be great if some of these.
Founders with great ideas could, could try to avoid much in the way of ec now that obviously in many areas they're absolutely vital, but in some of the medical startups, it, it's good if, uh, as, as far as I can see if to get some real good momentum before you have to resort to getting a lot of funding through, uh, venture capital.
Oh,
Halle: absolutely. You don't then you don't have to give up as much of your company. Yeah, yeah, exactly. Um, Amazing. Well, for kind of to close things out today, what advice would you give to our audience of aspiring healthcare innovators and leaders?
Eric: Well, I wish I could trade places with them cuz they're had to be younger than me.
Right. And, uh, it's so exciting to see what's going on right now because, um, there's so much room for, uh, the innovative spirit and to actualize these opportunities that seem to be limitless. Uh, and, you know, everything from the, the early. Short-term benefits that we had to gutting hospitals as we know them today, to have patients be able to be treated in, in their homes, to virtual health coaches, to digital twin infrastructures where we, we, a true learning system of the planet, of our species.
There's j it's just hell. It's so exciting and I think, uh, what a space to be in. So your, your crew, uh, they got a lot to look forward to and, uh, I'm sure they'll, uh, they'll have a lot of successes along the way. Well,
Halle: thank you. I say my favorite thing about working in healthcare is the people. It's what keeps me in this otherwise very challenging and slow industry.
And Eric, you have always been one of my favorites. I appreciate everything that you do for this industry and how much support you've given me in my career. So thank you and thank you for being here.
Eric: It's a mutual Halle, keep up the great stuff and I'll be, uh, hoping to have more conversations with you in the Times ahead.
Halle: Yeah. I need to come visit you in San Diego.
Eric: That'd be great.