October 6, 2024 • 26 mins
Dr. Triola is the associate dean for educational informatics and a professor of medicine at NYU Langone Health, where he is also the founding director of the Institute for Innovations in Medical Education, known as IIME. Dr. Triola received his BA from Johns Hopkins University and MD at NYU School of Medicine. He completed residency training in internal medicine at NYU School of Medicine and subsequently served as chief resident for the internal medicine residency program. He completed a research fellowship in medical informatics at Mount Sinai School of Medicine.In previous roles at NYU Langone, Dr. Triola served as director of the Division of Educational Informatics within NYU School of Medicine; chief of the Section of Medical Informatics within the Division of General Internal Medicine, Department of Medicine; associate director of the Center for Health Information Preparedness, Department of Medicine; and director of research for advanced educational systems within the School of Medicine.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
The following is a paid podcast. iHeartRadio's hosting of this
podcast constitutes neither an endorsement of the products offered or
the ideas expressed.
Speaker 2 (00:09):
The following program is brought to you by NYU Land
Going Health. It's CATS's Corner with doctor Aaron Katz. You're
trusted expert in men's health, providing straight talk on a
wide range of men's health topics and advice on how
to live your healthiest life. Now on seventy ten WOOR.
It's the Chairman of Urology at NYU Land Going Hospital,
(00:32):
Long Island. Here is doctor Aaron Katz.
Speaker 3 (00:36):
Well, good morning everyone, and welcome again to Kats's Corner
here on wr iHeartRadio. So glad you could join us.
We really have a fascinating show for you today. And
oftentimes we talk about a specific entity, maybe it's heart health,
prostate health, bone health, but today it really is innovations
in health and where we are going in healthcare and
(00:59):
in the area specifically of artificial intelligence affectionately known as AI.
And to help us with this discussion, we have a
wonderful guest on I've asked doctor Mark Triola, who is
professor of medicine Associate Dean for Educational Informatics and the
director of the Institute for Innovations in Medical Education, and
(01:22):
doctor Triola really has been an innovator in this new
area of translational information and teaches medical students and he
has a lab that is working on new technologies and
AI driven educational interventions, and I really it should be
a fascinating show. You know, we certainly have heard a
(01:44):
lot over the last couple of years now about AI
and where it's play of its role in medicine. So
I'm really looking forward to having you and this discussion. Mark,
thanks so much for coming on the show.
Speaker 4 (01:56):
Thanks so much for having me. It's a pleasure to
be here.
Speaker 3 (01:58):
So maybe you can start by just telling our listeners,
you know, just a very general question, what is artificial
intelligence and how does that, you know, apply to medicine.
Speaker 4 (02:10):
So, artificial intelligence is the way that computers can think
and reason through large amounts of data in a way
that can sometimes seem and mimic the way humans make
decisions or come to conclusions, or even create new content,
new text or now new images. AI is not new
(02:33):
in medicine. It's been around for decades and we've had
many AI innovations. People wearing an Apple watch listening to
this have AI on their wrist that that can warn
them if their heart is beating too fast or too slow.
But a couple of years ago we got this new
type of AI chat GPT is the most popular kind.
(02:53):
It's this new type of generative AI, which has, particularly
in medicine, unlocked a whole bunch of new opportunities about
what's possible on the science side and on the caring
for patient side.
Speaker 3 (03:05):
And you know, basically, is this something that you know
clearly patients need to know about it? What do they
need to know on the patient side?
Speaker 4 (03:18):
So I think not just patients, but doctors need to
know about it too, because it's going to change everything
in medicine. It's going to change some things a little
in subtle ways, and it's going to change some things
in really big ways. I think my personal opinion is
is that AI, particularly this new type of AI, is
going to solve a lot of the problems and challenges,
(03:38):
inefficiencies and burdensome aspects of medicine and healthcare. It's going
to help us with all of the paperwork, it's going
to help us with coordinating the care that we're receiving.
It's going to help us communicate between doctors and healthcare
teams and patients. But most importantly, it's going to improve
(03:58):
the quality of care that we deliver. It's going to
improve our ability to diagnose, to come up with the
correct answer for that given patient and what care and
what treatments and what tests are best for them. And
it's also going to help us discover new cures and
do new types of research on massive amounts of data,
(04:19):
things that we just couldn't do in the past. So
I think it's really going to unlock a lot of
positive aspects of medicine and really help us move forward.
Speaker 3 (04:32):
Yeah, it really is. It really is an exciting time.
And you say, all of this data that we have,
We certainly have tremendous amounts of data here at NYU.
You know, we have this large medical record, the EPIC system,
and I wonder how that integrates looking at all the
data from patients, does it, and then taking it down
to the individual patient, right, I mean, that's really what
(04:56):
people care about themselves. Of course, at the end of
the day, they want to know, how is all this
data that you have in this system? How can it
help me?
Speaker 2 (05:05):
Right?
Speaker 3 (05:05):
So, yeah, I guess that's a direction that we're going.
Is that is that fair to say?
Speaker 4 (05:11):
That's absolutely fair to say. First, you made an excellent point.
The practice of medicine is the practice of dealing with
huge amounts of information. We know more about our patients,
and patients know more about themselves than ever before. In fact,
it's estimated that twenty five percent of all of the
information generated on a daily, weekly, monthly basis in this
(05:32):
country is in healthcare and that's you know, some of
that is lab data and blood pressures, some of that
is is insurance things, genetic codes. The amount of data
that we have in medicine is like never before, and
it's a bewildering amount of data. It's overwhelming for patients
and it can be overwhelming for doctors and physicians. And
(05:55):
I think that AI can help us make sense of
that and make it definitely less overwhelming as we navigate
through all of this. You mentioned the individual patient, and
we've been really focused on figuring out Historically, we've been
focused on figuring out cures for diseases and caring for
everybody in the country, and now with this increased amount
(06:19):
of data, and these new technologies, we're moving towards this
model that's called precision medicine, where we recognize that everybody
is different. Everybody's unique, they have different genetic code, they
have different habits in their life, they may respond differently
to different types of treatments for the same disease and
AI and certainly the ability to understand the genetic code
(06:43):
of each individual patient is now enabling us to enter
this new era of precision medicine where we can come
up with a plan that matches your particular needs, not
just for the generic condition or syndrome or disease that
you might have. It's very exciting.
Speaker 3 (07:02):
Yeah, no, I agree. You know, in our specific field
in neurology and prostate cancer, and many of our listeners
are you know, have either been battling prostate cancer, have
had it, or concerned about it. And what we're seeing
in our own particular field is taking all these lumps
of data that you say, the PSA, their biopsy score,
their genomics score, their cat scan, their MRI, and putting
(07:25):
it all together and coming up with predictions and perhaps
even strategies for the best optimal treatment, including as you said,
genetics for that patient to say, look, this is using
our AI and our software programs, this is the treatment
that you should have or that, And I wonder if
(07:46):
you're seeing that. And then I guess the concern there
is is it going to take doctors out of the decision?
I mean, you know, is it going to remove us
as physicians to say you don't need to see you
don't need to see the doctor. You could just use
an AI you know, chat, GDP platform.
Speaker 4 (08:05):
Or something like that, right, right, And that's a big question,
and it is a big concern because we you know,
we'll talk about it. But these AI systems, as amazing
as they are and as optimistic as I am, they
are not without their flaws, flaws that don't quite make
them ready for primetime. Although they are improving very very rapidly.
(08:27):
And I do not foresee a future where the doctor
is out of the loop, and nor would I want
that future. I think that unlike many other industries or
other areas, patients in particular still want and will still
want to talk to their doctor. They want to have
a human connection to really uh get care from a
(08:50):
from a person or from a team of people. That
being said, if I were a patient I would want
my doctor to use AI. I still want to talk
to them. I still want to talk to him or
her and and and have that rapport, but I want
to make sure they're they're able to take advantage of
all of the latest cutting edge evidence information about me
(09:12):
as an individual. Just as you said, and your example
about prostate cancer is a great one where the ability
for you to come up with a personalized plan you
plus AI equals better care than maybe you by yourself.
And and I think that that's the era that we
are looking towards moving into. Maybe the AI can, but
(09:35):
maybe the AI can help with some things that currently
require physicians to spend time on. Maybe there are some
simple things like medication refills or other types of things
where the AI could make things convenient and faster and
more efficient. But the fundamental relationship between the doctor and
the patient is one that I think is important to
(09:55):
be preserved.
Speaker 3 (09:56):
Yeah, one of the things that we've we've heard about
in our field as well as in radiologies, AI may
be able to read X rays better than the human eye,
or read biopsy samples better than the human eye, and
maybe more consistent, you know, in our field of again
going back to the prostate cancer model. But I assume
(10:17):
it's for many cancers you have a pathologist who's reading it.
Oftentimes patients want to get a second opinion, they want
to go to you know, they come to NYU for
a second opinion or read their biopsy because it was
done on the outside in the community that maybe the
pathologist doesn't have the expertise. Let's say that we have
here at n YU. You wonder will there be a
consistent platform out there that could be used and running
(10:40):
a biopsy through an AI type of a program so
that you know, it takes some of that guesswork out
of it and some of that you know that that
human nature, that not human nature, but that human ability,
that limitation of the human I, whereas some of these
computers may maybe better than and humans.
Speaker 4 (11:01):
And there have been many studies in the past couple
of years that have highlighted exactly that that the AI,
under controlled circumstances is able to do sometimes a better
job at coming up with a diagnosis or a list
of possible diagnoses, at finding things. On imaging studies, people
(11:22):
you know, make the point that the AI doesn't need
to eat, It doesn't need to sleep, it can it
doesn't get get tired like us humans. At the same time,
it does not yet have judgment in the same way,
it doesn't necessarily understand the context of that particular patient
(11:42):
and what they need. But I will agree with you
that these improvements are happening very quickly. If you think
just two years ago, CHATCHBT didn't exist, and now it
can create It can create textbooks with images, it can
create videos. You've seen all of the examples of deep
sakes and things like that online and the quality is remarkable.
(12:06):
So the pace at which these AI systems are becoming
accurate and strong and powerful is much much faster than
anything we've really seen in the history of medicine.
Speaker 3 (12:18):
Quite frankly, yeah, I mean, you know, in all in
across our society. I mean, you think about electric cars.
You know, who would have thought we would have that,
you know, ten years ago, or even cars that are
now driving themselves. I mean, I have an associate of
mine who he gets in the car and he puts
on the autopilot takes him home. I mean you think, well,
(12:39):
people want to drive their cars, well, you know, people
want to see their doctors. Well, you know, in the
future you know, it's changing rather rapidly, isn't it. You know,
And when they're not, people aren't even some people aren't
even driving their own car anymore. So really changing rather quickly.
If you're just tuning up in the morning. Here, if
you're just waking up, we're talking with doctor Mark who
(13:00):
is a professor in the Department of Medicine at the
NYU lango On Health System. He's Associate Dean for Educational
Informatics and the director of the Institute for Innovations in
Medical Education. Let's get into the medical education a little
bit morek and maybe you can tell us. I know
that you're leading this institute, tell us a little bit
(13:21):
about your curriculum and how is that changing for the
new doctors that are out there training now?
Speaker 4 (13:29):
Sure? Sure, And so I'll go back to your self
driving car example, which I think is a great example
because that's another life and death situation about driving a car.
And you know, if you have a fourteen, fifteen or
sixteen year old child, you wonder are they going to
need to learn how to drive a car in the
same way, because how quickly this world is changing. And
(13:52):
that's exactly how we're thinking about our medical students and
our residents. It takes about ten years from when you
start medical school to when you're on your own in practice,
really truly independently taking care of patients. We're only two
years into this AI journey. Imagine what it's going to
look like ten years from now, and so we want
to make sure that are the future doctors that we're
(14:14):
producing out of NYU Grossman School of Medicine are really
ready for this world, and not not just ready in
that they understand how the AI works. They understand what
where the AI can get it right and where it
can get it wrong, and that's very important for them
to learn about. But going back to our earlier conversation,
when to use the AI, what is the what is
(14:36):
the appropriate way to use this? How does it change
the way we communicate with patients, and how do we
really maintain the integrity of being a physician, of the
of delivering care, of protecting the information of our patients
in this world that is being rapidly transformed. And when
(14:56):
the AI can help them do a better job and
can help them make accurate decisions and diagnoses and come
up with cutting edge treatment plans, how to make sure
that they know to use that tool and to learn
about all of the new AI tools that are being
made available to it. So we have trained changed our curriculum.
(15:17):
We're introducing AI early, from the very first day of
medical school. We have our students run through scenarios and
simulations using current AI. We teach them how to do
it in a safe way that protects our patients' data
and protects our patient's identity first and foremost, and we
explicitly give them examples where the AI doesn't get it correct,
(15:41):
so that they can see the pitfalls and the tremendous
opportunities of using these systems.
Speaker 3 (15:49):
It's a really different way of treating and thinking, isn't it.
I mean, and how you're going to critically think about
diagnostic dilemmas and problems when they approach. And it may,
you know, reduce the you know a lot of the
unnecessary things that you know, maybe that we had thought
(16:09):
were so critical in medical school training. Maybe it's going
to just completely change the way that we train our
doctors now and maybe the things that we thought were
so important back then may not be in the future.
I don't know if you've you've thought about that, but
certainly the AAR advances are are making strides in the
(16:31):
training and in the curriculum. And there's only so many
amount of time that we have in medical school, right,
there's only four years, let's say, and we're not going
to extend that. And actually our medical school here on
Long Island is three years now for people that are
really interested in primary care. But you've got to incorporate
all of this and change the way that you know,
(16:51):
our young students think critically about problems, and so you know,
it's certainly I got a hand to you. That's a
it's a big task to do that, I would think,
and it's congratulations on bringing it in very early in
the forefront of their medical school training.
Speaker 4 (17:07):
Oh thank you. And to your point you talked about
how much the care of patients with prostate cancer has evolved.
Think of how much has changed since you graduated from
medical school, since I graduated from medical school, and the
pace of change is ever increasing. So for a medical
student today, the amount of scientific knowledge, the amount of
(17:31):
changing research information, whole new categories of treatments for patients.
By the way, this is all great for patients because
the more complex and more difficult it gets to the doctors,
the more options there are to actually care for patients,
that are more things that are available to take care
of them. But that pace has just gotten faster and
faster and faster, and it's great. It means that we
(17:52):
as a society are moving forward and we're improving our
ability to care. But just as you said, the days
of a metaydical student memorizing the whole textbook, those days
are long gone. That textbook is now exponentially more dense
and information and it's the AI that's going to help
them integrate and retrieve the information they need at that
(18:14):
time that they need it.
Speaker 3 (18:17):
Yeah, I think that's key, you know, because there's so
much information that we learned in medical school that we
never really come across ever again, and never needed again.
It was just for taking a test purposes at that point.
It seemed, you know, all of a pharmacology. And I'm
not saying that that's not important, of course it is
with the basics of it, but so much of it
was it just unnecessary, I think, and now with AI,
(18:39):
maybe it can help to reduce that almost and maybe
almost have a greater focus of the students on the
things that are critically important for patient management and their disease. Well,
what do you think?
Speaker 4 (18:54):
I absolutely agree. I think that the critical skills that
we need to teach our students are continuous learning as
new things come out, and them integrating the latest treatments,
integrating new ways of thinking about things, clinical problem solving.
If we go back to talk about precision medicine, that
really customizing what you're going to do for that particular
(19:16):
patient in front of you. It requires a new mindset,
and it requires using data and tools in a different
way to come up with that correct plan. And so
these are going to become the important skills. What does
not go away, what stays exactly the same in medical school,
it's connecting with your patients, treating them with respect, gathering
(19:37):
a history, and communicating well with patients, Empathy, respect, all
of the things that are of critical importance for being
a physician and navigating those in the era of technology,
whether it's it's AI, or it's the patient portal or
a million other technologies that have changed things are the
(19:59):
kind of things that we really want to focus on
because we're not going to produce We don't want to
produce people who are just driving the electronic health record.
We want to produce physicians who are going to engage
with and care for patients, and so that's really important
to us to maintain.
Speaker 3 (20:16):
That is really well said, Mark. That is really as
well and so important the the empathy, the professionalism, how
to reduce anxiety when coming into a hospital or needing
a procedure, or being faced with a diagnosis of let's
say cancer or another illness that may affect one's life.
You know, that's where the human ability comes in. And
(20:39):
that's something that maybe we can focus more on in
our medical school education that AI won't be able to
replace there. And that was really really well said, Thank
you for that. I also want to bring up an issue,
you know, I remember when web md came out, you know,
and everyone is you know, using WebMD, and how there's Google,
(20:59):
and you know a lot of the elderly people, I
think are concerned about how fast paced things can go.
I am too, I mean, and how fast paced and
are we able to keep up with it? And you know,
you wonder are people that are let's say elderly that
didn't there aren't millennials that didn't grow up with you know,
while they were three years old with a laptop in there,
(21:20):
you know, in their height chair. How are they able
to navigate and to use AI? Is it going to
be just way too difficult for them or is it
not as difficult as one may think.
Speaker 4 (21:32):
It has been a remarkable journey in these past two years.
I would say before two years ago, AI was difficult
to use. It was often very technical. You had to
really understand what was going on. This new type of
AI like chat cheapt and I encourage everyone listening, if
you've never tried it before, just go to chat schepyt
dot com and ask it a question. It's so much
(21:52):
easier to use because it's just like talking to a person.
It's like having an interactive conversation with an expert, with
a with an assistant, with whatever it is you want
you want it to be, and so it's it's natural.
It's it's as natural as as having a conversation. There's
no code, there's no computer code to write, there's there's
(22:13):
nothing really sophisticated about it. You have to ask it
good questions to get good responses. But as we've seen,
everybody is now really starting to use this. The growth
of the use of this type of AI has far
outpaced almost anything, with the exception of some social networking stuff.
But everybody's using it young, old, you know, different different
(22:35):
levels of reading literacy and health literacy, and they're starting
to do really interesting things with it. People are using
it to understand the medications that they're on. I saw
an example on Twitter where somebody took a picture of
their pill bottle. They uploaded the picture to Twitter and said,
tell me the side effects of this, and chat GPT
understood exactly what the medication was and the dose, and
(22:56):
it gave them a whole print out of information and
education about that about that particular antibiotic. So I think
that this is going to really be helpful, easy and
useful to far more patients than the older types of AI.
Speaker 3 (23:14):
Yeah, I mean, and you wonder, I mean, I guess
patients or myself are wondering, you know, how are these
things on chat GDP validated? Like how do we know
that there it's not fake news as I mentioned earlier
or something, you know, but it could it be fake?
Could it be something that it could cause harm to
someone if they just rely on this chat GTP. Is
(23:37):
there's somebody behind the scenes that's checking all of this
to make sure that it's all right, that it's all
correct or and is it continuously as you said, these
machines are learning, is continuously. If it made a mistake,
would it correct itself and so that the next time
it wouldn't make that mistake. I mean, it could can't
be or could it be perfect all the time?
Speaker 4 (23:59):
It is far from perfect and that so that's one
of the biggest, really the biggest concerns. We don't really
know yet how accurate these systems are. In studies that
have been done on medicine looking at the accuracy of
the ability of CHAT, GPT and other similar systems, There's Google,
(24:19):
Gemini and Anthropic Clause. There's many many different ones looking
at their ability to answer medical questions. It gets it
right the vast majority of the time, but it gets
it wrong a significant amount of the time. As these
past two years have gone by, the accuracy of these
systems has improved. These systems sometimes do something called hallucinate,
(24:41):
where they make stuff up. That's kind of what they're
designed to do. They're designed to write wedding speeches and
translate poems and do sort of fun things like that.
But when it comes to medicine. Making things up is
not good. So there have been many examples in healthcare
uses of AI where these systems have fabricated clinical values,
(25:02):
even lab test results and things like that. So understanding
not only how accurate and valid they are, but also
how these systems are trained. They're trained on information from
the Internet, and just as you said, if you feed
it fake news, it will spit out fake news as
the answer. And ensuring that these systems are trustworthy and
(25:26):
trained on accurate information, which is something we're working a
lot on here at Nyulango, and we're actually building our
own large length or our own AI models based on
data we know we can trust. That's going to be
a big part of the future. And until we get there,
it's not quite ready for primetime, but it's moving very
quickly and the accuracy of these systems is improving very quickly.
Speaker 3 (25:51):
Yeah, I can imagine in why you would take the
lead under your direction and then maybe work with other
major academic centers across the United Slates States and globally
to further validate things and to see how our results
compare to their results, and you know that's a validation thing.
Thank you so much, Mark, It really was a pleasure
(26:11):
talking with you. You're terrific and NYU is really fortunate
to have someone like you at the Helm in this
most exciting I think field in innovation in medicine. Well
that's a show everyone. I want to thank you all
for tuning in and tune in next week here on
Kats's Corner, we'll be back again with a great show.
This is doctor Aaron Katz.
Speaker 2 (26:30):
You've been listening to Katzer's Corner. Come back every week
to hear more straight talk on a wide range of
men's health topics and advice on how to live your
healthiest life.
Speaker 1 (26:41):
The proceeding was a paid podcast. iHeartRadio's hosting of this
podcast constitutes neither an endorsement of the products offered or
the ideas expressed.
The following is a paid podcast. iHeartRadio's hosting of this
podcast constitutes neither an endorsement of the products offered or
the ideas expressed.
Speaker 2 (00:09):
The following program is brought to you by NYU Land
Going Health. It's CATS's Corner with doctor Aaron Katz. You're
trusted expert in men's health, providing straight talk on a
wide range of men's health topics and advice on how
to live your healthiest life. Now on seventy ten WOOR.
It's the Chairman of Urology at NYU Land Going Hospital,
(00:32):
Long Island. Here is doctor Aaron Katz.
Speaker 3 (00:36):
Well, good morning everyone, and welcome again to Kats's Corner
here on wr iHeartRadio. So glad you could join us.
We really have a fascinating show for you today. And
oftentimes we talk about a specific entity, maybe it's heart health,
prostate health, bone health, but today it really is innovations
in health and where we are going in healthcare and
(00:59):
in the area specifically of artificial intelligence affectionately known as AI.
And to help us with this discussion, we have a
wonderful guest on I've asked doctor Mark Triola, who is
professor of medicine Associate Dean for Educational Informatics and the
director of the Institute for Innovations in Medical Education, and
(01:22):
doctor Triola really has been an innovator in this new
area of translational information and teaches medical students and he
has a lab that is working on new technologies and
AI driven educational interventions, and I really it should be
a fascinating show. You know, we certainly have heard a
(01:44):
lot over the last couple of years now about AI
and where it's play of its role in medicine. So
I'm really looking forward to having you and this discussion. Mark,
thanks so much for coming on the show.
Speaker 4 (01:56):
Thanks so much for having me. It's a pleasure to
be here.
Speaker 3 (01:58):
So maybe you can start by just telling our listeners,
you know, just a very general question, what is artificial
intelligence and how does that, you know, apply to medicine.
Speaker 4 (02:10):
So, artificial intelligence is the way that computers can think
and reason through large amounts of data in a way
that can sometimes seem and mimic the way humans make
decisions or come to conclusions, or even create new content,
new text or now new images. AI is not new
(02:33):
in medicine. It's been around for decades and we've had
many AI innovations. People wearing an Apple watch listening to
this have AI on their wrist that that can warn
them if their heart is beating too fast or too slow.
But a couple of years ago we got this new
type of AI chat GPT is the most popular kind.
(02:53):
It's this new type of generative AI, which has, particularly
in medicine, unlocked a whole bunch of new opportunities about
what's possible on the science side and on the caring
for patient side.
Speaker 3 (03:05):
And you know, basically, is this something that you know
clearly patients need to know about it? What do they
need to know on the patient side?
Speaker 4 (03:18):
So I think not just patients, but doctors need to
know about it too, because it's going to change everything
in medicine. It's going to change some things a little
in subtle ways, and it's going to change some things
in really big ways. I think my personal opinion is
is that AI, particularly this new type of AI, is
going to solve a lot of the problems and challenges,
(03:38):
inefficiencies and burdensome aspects of medicine and healthcare. It's going
to help us with all of the paperwork, it's going
to help us with coordinating the care that we're receiving.
It's going to help us communicate between doctors and healthcare
teams and patients. But most importantly, it's going to improve
(03:58):
the quality of care that we deliver. It's going to
improve our ability to diagnose, to come up with the
correct answer for that given patient and what care and
what treatments and what tests are best for them. And
it's also going to help us discover new cures and
do new types of research on massive amounts of data,
(04:19):
things that we just couldn't do in the past. So
I think it's really going to unlock a lot of
positive aspects of medicine and really help us move forward.
Speaker 3 (04:32):
Yeah, it really is. It really is an exciting time.
And you say, all of this data that we have,
We certainly have tremendous amounts of data here at NYU.
You know, we have this large medical record, the EPIC system,
and I wonder how that integrates looking at all the
data from patients, does it, and then taking it down
to the individual patient, right, I mean, that's really what
(04:56):
people care about themselves. Of course, at the end of
the day, they want to know, how is all this
data that you have in this system? How can it
help me?
Speaker 2 (05:05):
Right?
Speaker 3 (05:05):
So, yeah, I guess that's a direction that we're going.
Is that is that fair to say?
Speaker 4 (05:11):
That's absolutely fair to say. First, you made an excellent point.
The practice of medicine is the practice of dealing with
huge amounts of information. We know more about our patients,
and patients know more about themselves than ever before. In fact,
it's estimated that twenty five percent of all of the
information generated on a daily, weekly, monthly basis in this
(05:32):
country is in healthcare and that's you know, some of
that is lab data and blood pressures, some of that
is is insurance things, genetic codes. The amount of data
that we have in medicine is like never before, and
it's a bewildering amount of data. It's overwhelming for patients
and it can be overwhelming for doctors and physicians. And
(05:55):
I think that AI can help us make sense of
that and make it definitely less overwhelming as we navigate
through all of this. You mentioned the individual patient, and
we've been really focused on figuring out Historically, we've been
focused on figuring out cures for diseases and caring for
everybody in the country, and now with this increased amount
(06:19):
of data, and these new technologies, we're moving towards this
model that's called precision medicine, where we recognize that everybody
is different. Everybody's unique, they have different genetic code, they
have different habits in their life, they may respond differently
to different types of treatments for the same disease and
AI and certainly the ability to understand the genetic code
(06:43):
of each individual patient is now enabling us to enter
this new era of precision medicine where we can come
up with a plan that matches your particular needs, not
just for the generic condition or syndrome or disease that
you might have. It's very exciting.
Speaker 3 (07:02):
Yeah, no, I agree. You know, in our specific field
in neurology and prostate cancer, and many of our listeners
are you know, have either been battling prostate cancer, have
had it, or concerned about it. And what we're seeing
in our own particular field is taking all these lumps
of data that you say, the PSA, their biopsy score,
their genomics score, their cat scan, their MRI, and putting
(07:25):
it all together and coming up with predictions and perhaps
even strategies for the best optimal treatment, including as you said,
genetics for that patient to say, look, this is using
our AI and our software programs, this is the treatment
that you should have or that, And I wonder if
(07:46):
you're seeing that. And then I guess the concern there
is is it going to take doctors out of the decision?
I mean, you know, is it going to remove us
as physicians to say you don't need to see you
don't need to see the doctor. You could just use
an AI you know, chat, GDP platform.
Speaker 4 (08:05):
Or something like that, right, right, And that's a big question,
and it is a big concern because we you know,
we'll talk about it. But these AI systems, as amazing
as they are and as optimistic as I am, they
are not without their flaws, flaws that don't quite make
them ready for primetime. Although they are improving very very rapidly.
(08:27):
And I do not foresee a future where the doctor
is out of the loop, and nor would I want
that future. I think that unlike many other industries or
other areas, patients in particular still want and will still
want to talk to their doctor. They want to have
a human connection to really uh get care from a
(08:50):
from a person or from a team of people. That
being said, if I were a patient I would want
my doctor to use AI. I still want to talk
to them. I still want to talk to him or
her and and and have that rapport, but I want
to make sure they're they're able to take advantage of
all of the latest cutting edge evidence information about me
(09:12):
as an individual. Just as you said, and your example
about prostate cancer is a great one where the ability
for you to come up with a personalized plan you
plus AI equals better care than maybe you by yourself.
And and I think that that's the era that we
are looking towards moving into. Maybe the AI can, but
(09:35):
maybe the AI can help with some things that currently
require physicians to spend time on. Maybe there are some
simple things like medication refills or other types of things
where the AI could make things convenient and faster and
more efficient. But the fundamental relationship between the doctor and
the patient is one that I think is important to
(09:55):
be preserved.
Speaker 3 (09:56):
Yeah, one of the things that we've we've heard about
in our field as well as in radiologies, AI may
be able to read X rays better than the human eye,
or read biopsy samples better than the human eye, and
maybe more consistent, you know, in our field of again
going back to the prostate cancer model. But I assume
(10:17):
it's for many cancers you have a pathologist who's reading it.
Oftentimes patients want to get a second opinion, they want
to go to you know, they come to NYU for
a second opinion or read their biopsy because it was
done on the outside in the community that maybe the
pathologist doesn't have the expertise. Let's say that we have
here at n YU. You wonder will there be a
consistent platform out there that could be used and running
(10:40):
a biopsy through an AI type of a program so
that you know, it takes some of that guesswork out
of it and some of that you know that that
human nature, that not human nature, but that human ability,
that limitation of the human I, whereas some of these
computers may maybe better than and humans.
Speaker 4 (11:01):
And there have been many studies in the past couple
of years that have highlighted exactly that that the AI,
under controlled circumstances is able to do sometimes a better
job at coming up with a diagnosis or a list
of possible diagnoses, at finding things. On imaging studies, people
(11:22):
you know, make the point that the AI doesn't need
to eat, It doesn't need to sleep, it can it
doesn't get get tired like us humans. At the same time,
it does not yet have judgment in the same way,
it doesn't necessarily understand the context of that particular patient
(11:42):
and what they need. But I will agree with you
that these improvements are happening very quickly. If you think
just two years ago, CHATCHBT didn't exist, and now it
can create It can create textbooks with images, it can
create videos. You've seen all of the examples of deep
sakes and things like that online and the quality is remarkable.
(12:06):
So the pace at which these AI systems are becoming
accurate and strong and powerful is much much faster than
anything we've really seen in the history of medicine.
Speaker 3 (12:18):
Quite frankly, yeah, I mean, you know, in all in
across our society. I mean, you think about electric cars.
You know, who would have thought we would have that,
you know, ten years ago, or even cars that are
now driving themselves. I mean, I have an associate of
mine who he gets in the car and he puts
on the autopilot takes him home. I mean you think, well,
(12:39):
people want to drive their cars, well, you know, people
want to see their doctors. Well, you know, in the
future you know, it's changing rather rapidly, isn't it. You know,
And when they're not, people aren't even some people aren't
even driving their own car anymore. So really changing rather quickly.
If you're just tuning up in the morning. Here, if
you're just waking up, we're talking with doctor Mark who
(13:00):
is a professor in the Department of Medicine at the
NYU lango On Health System. He's Associate Dean for Educational
Informatics and the director of the Institute for Innovations in
Medical Education. Let's get into the medical education a little
bit morek and maybe you can tell us. I know
that you're leading this institute, tell us a little bit
(13:21):
about your curriculum and how is that changing for the
new doctors that are out there training now?
Speaker 4 (13:29):
Sure? Sure, And so I'll go back to your self
driving car example, which I think is a great example
because that's another life and death situation about driving a car.
And you know, if you have a fourteen, fifteen or
sixteen year old child, you wonder are they going to
need to learn how to drive a car in the
same way, because how quickly this world is changing. And
(13:52):
that's exactly how we're thinking about our medical students and
our residents. It takes about ten years from when you
start medical school to when you're on your own in practice,
really truly independently taking care of patients. We're only two
years into this AI journey. Imagine what it's going to
look like ten years from now, and so we want
to make sure that are the future doctors that we're
(14:14):
producing out of NYU Grossman School of Medicine are really
ready for this world, and not not just ready in
that they understand how the AI works. They understand what
where the AI can get it right and where it
can get it wrong, and that's very important for them
to learn about. But going back to our earlier conversation,
when to use the AI, what is the what is
(14:36):
the appropriate way to use this? How does it change
the way we communicate with patients, and how do we
really maintain the integrity of being a physician, of the
of delivering care, of protecting the information of our patients
in this world that is being rapidly transformed. And when
(14:56):
the AI can help them do a better job and
can help them make accurate decisions and diagnoses and come
up with cutting edge treatment plans, how to make sure
that they know to use that tool and to learn
about all of the new AI tools that are being
made available to it. So we have trained changed our curriculum.
(15:17):
We're introducing AI early, from the very first day of
medical school. We have our students run through scenarios and
simulations using current AI. We teach them how to do
it in a safe way that protects our patients' data
and protects our patient's identity first and foremost, and we
explicitly give them examples where the AI doesn't get it correct,
(15:41):
so that they can see the pitfalls and the tremendous
opportunities of using these systems.
Speaker 3 (15:49):
It's a really different way of treating and thinking, isn't it.
I mean, and how you're going to critically think about
diagnostic dilemmas and problems when they approach. And it may,
you know, reduce the you know a lot of the
unnecessary things that you know, maybe that we had thought
(16:09):
were so critical in medical school training. Maybe it's going
to just completely change the way that we train our
doctors now and maybe the things that we thought were
so important back then may not be in the future.
I don't know if you've you've thought about that, but
certainly the AAR advances are are making strides in the
(16:31):
training and in the curriculum. And there's only so many
amount of time that we have in medical school, right,
there's only four years, let's say, and we're not going
to extend that. And actually our medical school here on
Long Island is three years now for people that are
really interested in primary care. But you've got to incorporate
all of this and change the way that you know,
(16:51):
our young students think critically about problems, and so you know,
it's certainly I got a hand to you. That's a
it's a big task to do that, I would think,
and it's congratulations on bringing it in very early in
the forefront of their medical school training.
Speaker 4 (17:07):
Oh thank you. And to your point you talked about
how much the care of patients with prostate cancer has evolved.
Think of how much has changed since you graduated from
medical school, since I graduated from medical school, and the
pace of change is ever increasing. So for a medical
student today, the amount of scientific knowledge, the amount of
(17:31):
changing research information, whole new categories of treatments for patients.
By the way, this is all great for patients because
the more complex and more difficult it gets to the doctors,
the more options there are to actually care for patients,
that are more things that are available to take care
of them. But that pace has just gotten faster and
faster and faster, and it's great. It means that we
(17:52):
as a society are moving forward and we're improving our
ability to care. But just as you said, the days
of a metaydical student memorizing the whole textbook, those days
are long gone. That textbook is now exponentially more dense
and information and it's the AI that's going to help
them integrate and retrieve the information they need at that
(18:14):
time that they need it.
Speaker 3 (18:17):
Yeah, I think that's key, you know, because there's so
much information that we learned in medical school that we
never really come across ever again, and never needed again.
It was just for taking a test purposes at that point.
It seemed, you know, all of a pharmacology. And I'm
not saying that that's not important, of course it is
with the basics of it, but so much of it
was it just unnecessary, I think, and now with AI,
(18:39):
maybe it can help to reduce that almost and maybe
almost have a greater focus of the students on the
things that are critically important for patient management and their disease. Well,
what do you think?
Speaker 4 (18:54):
I absolutely agree. I think that the critical skills that
we need to teach our students are continuous learning as
new things come out, and them integrating the latest treatments,
integrating new ways of thinking about things, clinical problem solving.
If we go back to talk about precision medicine, that
really customizing what you're going to do for that particular
(19:16):
patient in front of you. It requires a new mindset,
and it requires using data and tools in a different
way to come up with that correct plan. And so
these are going to become the important skills. What does
not go away, what stays exactly the same in medical school,
it's connecting with your patients, treating them with respect, gathering
(19:37):
a history, and communicating well with patients, Empathy, respect, all
of the things that are of critical importance for being
a physician and navigating those in the era of technology,
whether it's it's AI, or it's the patient portal or
a million other technologies that have changed things are the
(19:59):
kind of things that we really want to focus on
because we're not going to produce We don't want to
produce people who are just driving the electronic health record.
We want to produce physicians who are going to engage
with and care for patients, and so that's really important
to us to maintain.
Speaker 3 (20:16):
That is really well said, Mark. That is really as
well and so important the the empathy, the professionalism, how
to reduce anxiety when coming into a hospital or needing
a procedure, or being faced with a diagnosis of let's
say cancer or another illness that may affect one's life.
You know, that's where the human ability comes in. And
(20:39):
that's something that maybe we can focus more on in
our medical school education that AI won't be able to
replace there. And that was really really well said, Thank
you for that. I also want to bring up an issue,
you know, I remember when web md came out, you know,
and everyone is you know, using WebMD, and how there's Google,
(20:59):
and you know a lot of the elderly people, I
think are concerned about how fast paced things can go.
I am too, I mean, and how fast paced and
are we able to keep up with it? And you know,
you wonder are people that are let's say elderly that
didn't there aren't millennials that didn't grow up with you know,
while they were three years old with a laptop in there,
(21:20):
you know, in their height chair. How are they able
to navigate and to use AI? Is it going to
be just way too difficult for them or is it
not as difficult as one may think.
Speaker 4 (21:32):
It has been a remarkable journey in these past two years.
I would say before two years ago, AI was difficult
to use. It was often very technical. You had to
really understand what was going on. This new type of
AI like chat cheapt and I encourage everyone listening, if
you've never tried it before, just go to chat schepyt
dot com and ask it a question. It's so much
(21:52):
easier to use because it's just like talking to a person.
It's like having an interactive conversation with an expert, with
a with an assistant, with whatever it is you want
you want it to be, and so it's it's natural.
It's it's as natural as as having a conversation. There's
no code, there's no computer code to write, there's there's
(22:13):
nothing really sophisticated about it. You have to ask it
good questions to get good responses. But as we've seen,
everybody is now really starting to use this. The growth
of the use of this type of AI has far
outpaced almost anything, with the exception of some social networking stuff.
But everybody's using it young, old, you know, different different
(22:35):
levels of reading literacy and health literacy, and they're starting
to do really interesting things with it. People are using
it to understand the medications that they're on. I saw
an example on Twitter where somebody took a picture of
their pill bottle. They uploaded the picture to Twitter and said,
tell me the side effects of this, and chat GPT
understood exactly what the medication was and the dose, and
(22:56):
it gave them a whole print out of information and
education about that about that particular antibiotic. So I think
that this is going to really be helpful, easy and
useful to far more patients than the older types of AI.
Speaker 3 (23:14):
Yeah, I mean, and you wonder, I mean, I guess
patients or myself are wondering, you know, how are these
things on chat GDP validated? Like how do we know
that there it's not fake news as I mentioned earlier
or something, you know, but it could it be fake?
Could it be something that it could cause harm to
someone if they just rely on this chat GTP. Is
(23:37):
there's somebody behind the scenes that's checking all of this
to make sure that it's all right, that it's all
correct or and is it continuously as you said, these
machines are learning, is continuously. If it made a mistake,
would it correct itself and so that the next time
it wouldn't make that mistake. I mean, it could can't
be or could it be perfect all the time?
Speaker 4 (23:59):
It is far from perfect and that so that's one
of the biggest, really the biggest concerns. We don't really
know yet how accurate these systems are. In studies that
have been done on medicine looking at the accuracy of
the ability of CHAT, GPT and other similar systems, There's Google,
(24:19):
Gemini and Anthropic Clause. There's many many different ones looking
at their ability to answer medical questions. It gets it
right the vast majority of the time, but it gets
it wrong a significant amount of the time. As these
past two years have gone by, the accuracy of these
systems has improved. These systems sometimes do something called hallucinate,
(24:41):
where they make stuff up. That's kind of what they're
designed to do. They're designed to write wedding speeches and
translate poems and do sort of fun things like that.
But when it comes to medicine. Making things up is
not good. So there have been many examples in healthcare
uses of AI where these systems have fabricated clinical values,
(25:02):
even lab test results and things like that. So understanding
not only how accurate and valid they are, but also
how these systems are trained. They're trained on information from
the Internet, and just as you said, if you feed
it fake news, it will spit out fake news as
the answer. And ensuring that these systems are trustworthy and
(25:26):
trained on accurate information, which is something we're working a
lot on here at Nyulango, and we're actually building our
own large length or our own AI models based on
data we know we can trust. That's going to be
a big part of the future. And until we get there,
it's not quite ready for primetime, but it's moving very
quickly and the accuracy of these systems is improving very quickly.
Speaker 3 (25:51):
Yeah, I can imagine in why you would take the
lead under your direction and then maybe work with other
major academic centers across the United Slates States and globally
to further validate things and to see how our results
compare to their results, and you know that's a validation thing.
Thank you so much, Mark, It really was a pleasure
(26:11):
talking with you. You're terrific and NYU is really fortunate
to have someone like you at the Helm in this
most exciting I think field in innovation in medicine. Well
that's a show everyone. I want to thank you all
for tuning in and tune in next week here on
Kats's Corner, we'll be back again with a great show.
This is doctor Aaron Katz.
Speaker 2 (26:30):
You've been listening to Katzer's Corner. Come back every week
to hear more straight talk on a wide range of
men's health topics and advice on how to live your
healthiest life.
Speaker 1 (26:41):
The proceeding was a paid podcast. iHeartRadio's hosting of this
podcast constitutes neither an endorsement of the products offered or
the ideas expressed.
Popular Podcasts
Las Culturistas with Matt Rogers and Bowen Yang
Ding dong! Join your culture consultants, Matt Rogers and Bowen Yang, on an unforgettable journey into the beating heart of CULTURE. Alongside sizzling special guests, they GET INTO the hottest pop-culture moments of the day and the formative cultural experiences that turned them into Culturistas. Produced by the Big Money Players Network and iHeartRadio.
40s and Free Agents: NFL Draft Season
Daniel Jeremiah of Move the Sticks and Gregg Rosenthal of NFL Daily join forces to break down every team's needs this offseason.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.