Hello and welcome to our program today.
We're discussing clinical excellence
and cardiovascular imaging. I'm
your host, Jerry. Big naughty.
Today's topic is about the importance of
the user experience in the adoption
of structured reporting, as well as in addressing
the challenges of turning fragmented
data captured in the cardiovascular
service line
to drive outcomes and insights.
Our guests on this program include Dr
Jennifer Hall. Dr. Hall is
a PhD and chief of data
science and the co director of the
Institute for Precision Cardiovascular
Medicine at the American Heart Association.
She leads the association's groundbreaking
initiatives in cardiovascular medicine.
Welcome, Dr Hall.
Thank you, Jerry.
Also joining us is Dr
Scott Flam, an M D. Who's head
of the section of cardiovascular imaging
at the Cleveland Clinic Imaging
Institute.
He's certified in cardiovascular
magnetic residents and in protecting
human research subjects in biomedical
and genetic research.
His specialty interests include cardiovascular
MRI, cardiovascular CT,
ischemic heart disease and congenital
heart disease.
Thank you for being with us today, Dr Flam.
Pleasure to be here. Cherry. Thank you,
Dr Flam. Obviously, a person with your
background and expertise has a keen
interest in cardiovascular imaging.
Let's talk about the importance of the
user experience in C V
imaging. Well, gosh, that's I
think, a pretty easy one in that,
uh, let's go back to I'm going
to date myself here, Jerry, and say that I'm
one of those people who used to have a Palm Pilot
many years ago. And it was a way we
could store information in our portable
phone roll, a Dex sort of thing,
but just absolutely paled
compared to either an
iPhone or an android phone that we have
today that allows us to do so much more
because the interfaces are so much
more intuitive and simple and efficient
to use. And the same thing is true with
the tools that we use in cardiovascular
imaging. Uh, I'm
particularly a dept and
have a focus on cardiovascular CT
and MRI,
and we've got these great scanners that
put out great information. But when we get
to the point of processing the information,
integrating that, converting that into
quantitative information, it
can sometimes be a little bit complex
and a little bit more difficult
to make that happen. So we really want to make
this as efficient as possible
and simple as possible. and so often,
really, the mantra should be, you know,
truly keep it simple. It's that
the simplest interface possible
to allow us to put out all
this information is really the most useful
because people will use it that well, what about
things like auto population of data?
How important is that? Let
me give an example of an
area that I think is really terrific.
And I think our echocardiography furs
have really done a great job with this
and that, uh, the
the analysts, the technologists who
the stenographers who acquire the information
are able to put the information
together and
really almost auto populate.
It sends the data from
their input devices directly to
a report. And so the
report is partially auto populated
by the time the physician
gets to it and starts getting around to it,
and we have only a very little bit
of that In other parts of cardiovascular
imaging, CT and MRI are just really
not like that at all. We're
working on it, and we've got a number of
vendors that are trying hard to make that
more integrated. It's just that there are a lot
of complicated steps along the way, So I think
we need to follow the pathway that
are echocardiography. Folks have done
and expanded into all of the
other imaging CT and MRI
and including the Cath lab and the
electrophysiology lab.
So it's clear from what you've said that workflow
is important and
and structure reporting,
uh, in cardiovascular imaging.
Dr Hall, what are the main elements that promote
clinical workflow excellence
and through that could help drive adoption
of these systems? I think
Scott said it best, and that is just
keeping things as simple
as possible and sometimes
keeping things simple
sounds easy, but understanding
just simple things about
where to put the button that they click. How many
clicks it's going to take? Is the
data auto populated so
they it just automatically
happens,
and we make the clinicians job
easier. Technology is meant to make
things simpler
that we have to remember. It's meant to fade
into the background and simply allow the
clinician to do their job well.
Trying to make it simple sometimes
makes us call upon the latest
technology and recently developed
tools,
and I'm particularly
interested in your take on it as a data scientist,
How you see artificial intelligence
driven tools and capabilities,
improving the clinical workflow efficiency
as well as the impact on the adoption
of structured reporting. It's exciting.
The field is certainly exciting. It moves
so rapidly that you know
just what was happening last week is
outdated. Today,
the point to remember about
AI and machine learning is it's there to
again help the clinician.
It's not meant to be the final be all
answer,
Um, but
use that technology and then
put it in the background and only
put out there what the clinician needs
to see.
So what I mean by that is data Scientists
may have
very challenging algorithms
and a lot of math behind
the ML, um, and AI.
But what the clinician needs to see
is simply the output, you know, where
do I click
to get X? And if they want more information,
the documentation is there and
transparent. But it is not
the first thing that they see. We need to be able
to make it simple to utilize,
you know, ai and machine learning to make their lives
simpler and to improve patient lives
as well. You mentioned the clinician perspective.
Let's turn to Dr Flam for that.
Well, I I would just add on
to what Dr Hall was saying
and that this is really a matter of
pulling more information in
to the clinician with the
imaging information or whatever
else what other kind of diagnostic
information was pulled in but be
able to pull in the relevant clinical
information about that patient's so
often we may interpret
an image or piece of diagnostic
information and a relative vacuum.
Uh, and we really want to have that
be a whole to
be able to have the relevant information
regarding that particular patient
that is recent and relevant
and will have an impact on
the diagnostic information that we've
we've been presented. So
I think this is where AI and ML can
really be valuable as well,
by integrating closely with the electronic
medical record and the background information
about the patient.
So when it comes to optimizing the user
experience and and thereby
maximizing the adoption of
these systems, Dr Flam
return to Europe for that.
Some advice, I guess. What should
imaging it vendors be doing
to optimize the user experience and
maximize the adoption of their system?
I think The thing that I worry about
sometimes when vendors
put together products is that
they may focus on
what is the minimum viable
products. You know what is the minimum
thing that can be put out there, too?
Bring this information. And
while that's a practical way to go
and sometimes can allow us to get
the post processing tools
out into the
market into the hands of the clinicians
as quickly as possible, I
think that it's also really important not
to settle there but to be able
to put in the best
interface possible so that
it is simple and easy
to use, pulls in all this relevant information
about the patient and
really makes it so that
the user wants to use it.
Uh, and not simply,
just say, boy, this is something
I need to use to put this
report out to get to this conclusion
about this patient,
make it not there, make it so that they want
to use it, that it really drives their
adoption and drives the desirability
of using that particular product
or tool to help on the patient's
care pathway.
Dr. Hall, did you have anything to add to
that I would simply say
you know, the the analogy. We
use a lot of times as it's like building a house.
If you're going to build a house and
you have the window
person doing the plumbing,
that's never gonna work, and
it's not going to work efficiently. And so what
needs to happen is,
let's say the scientists don't always know
all the expertise that the clinician knows.
So these need things need to be built together.
You need the window person. You need the plumber,
but you need to focus in your areas. Then come
together and go
Wait a minute.
Test this for me. Tell me what we need.
Let's work together on building
the best product.
Cardiovascular departments generate a wealth
of data, and a significant part
of that data is structured on the service.
It looks like an analytics heaven,
but
often the data is stored in multiple places.
Makes it a challenge to derive meaningful,
outcome driven insights.
Dr. Hall, you're the data scientist.
What's your take on this?
We've seen a lot of great
progression in the last few years, and
I couldn't be happier about that because
in not just the clinical world but the academic
world People have their data stored in their
laptops, and they don't share it. In today's
environment, sharing is
the new, you know, catch
all. It is what everybody is trying to
do, because by doing that
you bring data discovery
and discovery for the
patient so much faster.
It's all accelerated. If you can share data,
build a statistical power
to really have the evidence behind things
that we're doing, and we see it all
the time. Data spread all over the place.
But it's not such
a big pill to swallow.
If you take it one step at a time and
start building pieces
within each system, each clinical
organizations, you have the right data
for the right treatment to the
clinician at the right time.
Dr. Flame, can you give us the clinical side
of that perspective?
Well, this is an area where I would hope
that AI and machine learning could be particularly
valuable because I certainly recognized
and agree with Dr Hall's comments
that we've got data in a variety
of different databases and repositories
and depositories throughout an
institution and across
institutions, and sharing that data
is difficult
and you know for many years
now, we've tried to do it with a
pretty manual and somewhat Clukey
kind of process in, uh, in
trying to match this data point to
another data point and using
our eyes and and refined
fear defined field names,
et cetera. But I would hope
that this is where AI and machine language
could come in and do all of that matching
for us provide all of the
cross references and the crosstalk
between and among those databases
so that we could have a common
database and repository of information
to help drive that clinical decision making.
When it comes to
trying to to quantify the value
proposition for a system like this, I'm kind
of curious what your suggestion
might be in terms of demonstrating the time
saved or the clinical
outcomes or data storage.
What would your tape beyond it, Dr Flam,
in terms of
being able to demonstrate the value
proposition. But we do have to
develop our demonstrate
to return on investment, develop
and show that value proposition
for all the things that we do. And I
think that there are many examples we could
toss out and give some broad uh,
breast stroke kind of
examples or comments, But
let me just give you a
finite example. I mean, if we're
looking at an electro physiologist
who is doing a procedure in
the e p lab, we
really want to collect all of
the information that has to
do with diagnostic and operations
and throughput in that
particular room and even collecting
all of the,
um, information
regarding resources that go into
that room, whether it be
pieces, whether it be devices
or the electricity and the water
that goes into that particular room. So we
can quantify all of this information
and an easy and reproducible
way, uh, in ways that
we haven't been able to do now we've
taken pieces of
that operation in a
in an E P lab, for example, and
we focused on, Well, how many
of these devices or catheters
do we use? And do we
have a just in time process to
get that in and kind of quantify that?
But we need to look at the totality
of the process and not just
those pieces, but how many people are in
the room, and what does that cost
us and how efficient are they?
And are are we making? Are we
optimizing all of these pieces
for the patient to be able to
get to the proper diagnosis fastest
to get to the proper therapeutic endpoint?
Uh, fastest, most efficient
and
best Most accurately, you
know, these are not easy pieces. We tend
to come at it from pieces and parts,
and we need to do it from a more holistic
process. But all of that information
together do you see a challenge
in articulating that value proposition?
Maybe there's a
bit of a language difference between
the I T department and cardiology.
Absolutely. We've
got our I T folks who
speak their language and they try
to understand the clinical language
and clinical processes. Um,
but in truth, they've
got their specialties, and
the physicians have their specialties.
We certainly have more physicians who
are becoming adept at
more of the administrative
and operational aspects of medicine.
But there's still a minority, and not
all of them understand those sorts
of things, and that's perfectly fine. We really
want many clinicians to absolutely
be at the top of their game, simply caring
for patients. But this
may be another place where AI and machine
language come in and help bridge
that gap so that even though
both sides of this equation that
we're talking about have incomplete
knowledge of the other,
AI and ML can perhaps help
bridge that and allow there to
be a better totality
between the two of
Dr Hall in the two groups.
Excuse me, Dr Hall, um, interested
in your insights as well as
it pertains to expressing the value
proposition.
Our experience has always been
that, in this case, the data
science programs that are very successful.
The I T programs that are successful
are the ones that keep the focus
on the value proposition
that is best clinical care
at the least amount of dollars
spent. Can you perhaps see
more patients? Can you diagnose them
sooner? Can you prevent things
from happening? And so it's again
getting the data in the right place
at the right time.
It's it's really
that simple, and it takes
a lot of work to get there. But if
the data science team focuses
on that specifically and both
ends the clinician side, the administrators,
the data science side keep that all
in the back of their mind is the main goal.
Then you'll have a better chance of success
for our viewers in a moment. We're going to open
the program up for your questions
before we do. I want to remind
viewers of our guests, Dr Jennifer
Hall of the American Heart Association, as
part of this program, and we thank you for
being here today.
Thank you, Jerry. It's always fun
to be on calls with you as well, Scott.
So always appreciate your insight. And
Dr Scott Flam, head of the section
of cardiovascular imaging at the Cleveland
Clinic's Imaging Institute. Thank
you for being here today to
thank you, Jerry, really a pleasure to be here
and to be here with Dr Call.
And now we're going to open it up for
live questions from our audience.
And just by way of reminder,
you can use the
the box on the right side of your screen
to enter a question
so you can type the question in there. It'll be
received by our background
crew, and they will pass
it on to me for
our Q and a session. We're going to add
to the panel of Dr Hall and Dr Flam
joining us from change. Healthcare
is I tie Galilee. I
tie is the director of cardiovascular product
management at change Healthcare. So I
tie. I'd like you to just say hi so we can
see your face on the screen.
Everyone, thanks Very
happy to be here.
It is joining us from Tel Aviv. So
this is truly an international
event that we're having here.
Let's start the questioning with, um,
the cloud. Let's let's
bring it to a question of the top
drivers in moving cardiovascular
imaging into the cloud. And
I'd like to start with you. Dr Hall,
What are the top drivers and moving, uh,
cardiovascular imaging to the cloud.
Dr. Hall, You may have
neglected to
another mute your microphone
and you need to roll your mouse down
to the bottom of your screen
and you'll see where you can turn your
camera back on and,
um, meet your microphone
while we wait for you to do that. I will pose
that same question
to Dr Flam Doctor Flam.
What do you consider the top drivers moving in
moving cardiovascular imaging to the cloud?
Well, gosh, there.
We do need to move
all this information to the cloud because
it's become so voluminous
and we need access to
it as quickly and
efficiently as possible and doing
so, uh, in our own institutions,
our own repositories, which
we talked about before. Often
being in disparate places and
having different access
routes, et cetera, is just
not as efficient as possible. But
at the same time, I think we have to recognize
that plenty of well, at
least the largest institutions still
have a desire to, for
the most part, hold on to the data themselves
and not put it out to the cloud for
cyber security reasons and
and others that just take a little bit of time
to get over. Um, you know, it's
been a number of years so far, Uh,
and we will get there.
But I think that it may take a little bit more
time just for
the the general comfort
to seep into all of
these institutions and our I t department's
and folks so that we all feel
comfortable. We're all on the same page.
I ty. I know you keep an eye on that market
as well, being in product management,
and so I'd like to pose that same question
from your perspective. What are the top drivers
and moving cardiovascular into the cloud?
Yeah. Thanks, Jerry. So I would like
to maybe go back to two
things that most of them
and Dr Hall talked about
a few minutes ago. The first
thing is, and I think Dr Flynn was just
mentioning that the challenge
of, uh keep deriving insights
from data that's being stored in multiple
separate places. And
I think the cloud again, it's just the technology.
It's an enabler if you use it correctly.
And if you are able to build
cloud native applications in
the right way and and gain the trust
of different users, you're
actually able to unlock that challenge
and and get all this data in a single
place. Of course, you need to have
the right boundaries
and check checks in place to
make sure that yeah, you keep the data
safe that you anonymized the data
that needs to be anonymized. So
definitely, you know, you need the expert
to handle those things, but the cloud
is a great technology
to enable to unlock this data
in this in separate places.
The second point, I would say when you talked about
and think. Both Dr Flynn and Dr Hall
talked about the user experience
in getting users excited
to use the application.
I think again, technology could
be a great in a bar in that domain as well.
And what cloud can bring to the
table here is just the
ability to very frequently update
the user experience, to listen to customers
and actually
close the loop on feedback from the
field very quickly. So you don't need to
get a fix out there and her and wait for
each and every site to deploy it when
they have the time depends on the version. Depends
on you know, the it. You deploy
it in one place very quickly and everybody
can enjoy it. So I think again,
Cloud is not a magic thing. Consult
everything. It's a great technology
that we need to use in the right way
to to bring value. So I think that's
that's what we've seen, you know, talking
to, you know, people in the fielding
and key opinion leaders like Dr
Lam and Dr Hall.
And as a matter of fact, Dr Hall joins
us once again. Thank you, Dr Hall, for
getting past some of the technical issues that
we just experienced. I want to talk
about moving cardiovascular
to be part of a more unified
enterprise imaging
approach rather than just a standalone department.
I wonder if you could talk about the advantages of
that enterprise imaging approach, Dr Hall
And what might be the drivers in
doing that?
Yeah, thanks, Jerry. I think any time
that you can combine data
and people across disciplines
together,
you will have a better result.
I think people thinking, um,
you know, there's been a lot of evidence behind
that across different disciplines and bringing
their excellence to bear in those sorts
of ways helps drive our field
forward. So bringing those experts and imaging
with experts in public health or
are different types of health care together
is really going to bring our fields together.
And so I think, exactly as you're
saying, that will be something
that allows us to solve
the dark matter or unsolved problems
in our field.
I see you smiling, Dr Flam.
Did you want to add something to that? I just
didn't think that I would ever hear the phrase
dark matter in this conversation
today,
I always surprise you, Scott.
Delightful, Delightful. Here
is a question submitted from Menace to
Ari how important this
is for Dr Flam. By the way,
how important is it to have multi modality
imaging in one place? That
is from a storage standpoint and
view, like an image viewer that
is reviewing ultrasound images
along with C T and M R M
R images side by side
in the same viewer, in the same software,
along with reporting capabilities?
And how much time frame
should we expect to finish such cases
reviewing and reporting. But
that's a big unpack that in the order in which
you see fit, that's the That's
a big question. Uh, certainly,
in an ideal world, we would have
tools that allow us to visualize.
See, uh, analyze,
do some post processing, uh,
in in one place. For these various
modalities spread across C T
m r I ultrasound, nuclear,
uh,
the cath lab images and geography.
That would be terrific. I'm
unaware of any tools. Single tools
that do that today and part
of the issue with it is that there are
some different approaches and perspectives
that come to the analysis
of each of those kinds of image
sets that may not have a perfect
overlap. Now, let's not say that I
don't think that there could be tools
developed that would have the flexibility
to highlight certain things
while we analyze certain kinds
of images and then de
emphasize and highlight others when
we move to other sets of images. Right?
I think that's certainly possible. We're just not quite
there yet. And they certainly look forward
to that happening,
you know, as to how important it is
to have those tools together. I think it depends
on what we're looking at. There are some
diagnoses or sets of
images that I have, where I don't need
anything else. I may be looking at Justin
M. R I or just just
an echo or just a
CT scan, and I don't happen.
I don't have a an absolute requirement
to have any other image with me.
And so in in that setting,
one tool is fine.
But others I'd like
to have a full breath of tools
and imaging
openings available so
that I can cross, compare
and understand and correlate what's going
on with other image modalities with the one
I'm looking at, um, today,
I would say that it's a minority of the
time, but I may be biased
in that I don't have that tool available
with all those things available. And if
I did, maybe I
would want to use it more frequently. So
that's that's where we get to that
answer as far as how long
it takes reporting to be. Gosh, we'd certainly
like it to be as as
quick as possible. Uh,
depending on what goes into it. Some
studies are reasonably fast,
whether they're normal or have only minor
abnormalities. Uh, may
not take very long to get reporting
done, but then, in other
situations, we may have things that
are incredibly complex, such as
complex congenital heart disease. Um,
I'm not sure we're gonna get to rapid
reporting for those kinds of cases anytime
soon. Uh, will
simply take longer because there
is such a pain. Ah, pally of information
available that needs to be integrated
and come to a
multiplicity of conclusions. Not
just a single occam's razor kind
of conclusion.
Questions submitted by Ken Martin
is one that I'm gonna pose to Dr
Hall. But I welcome, uh,
either of our other Panelists to chime in if they'd
like to participate in this. But the particular
question from Ken comes in as
a question about tracking patient
consent.
What are the best practices to track
patient consent
to use their imaging as part of a patient?
Population studies where, UM,
artificial intelligence and machine learning
are being used for further insights
and value?
Dr. Hall.
Yeah, thanks, Jerry, that
thanks for the question. That's really an
important question today. There are several,
um, software programs that are beginning
to do this a little bit better.
Um, there's nothing perfect
out there yet today and
and there, you know.
You know, I guess it's just it's a
really difficult question to answer because
it depends on your institution,
um, and the studies that these
individuals are part of usually
as part of the study, I would
suggest going back, Um, if
you know that there are particular
studies research studies
being done at your institution, of which
they would be a part of getting in
touch with that principle investigator
and or study coordinator, and working
through them directly
is the most direct route to go.
Um, and then certainly making
sure you have that consent to do that.
Um, is the right answer,
Dr. Flam, did you have anything that you wanted
to add on that?
I think I can only add
that it is a complex
issue. And
I'm gonna leave it up to the folks
at the American Heart Association
with incredible
resources and digital
experience to lead us through
this issue.
Very well.
Another question we have dealing
with best practices deals with
what imaging leaders can do.
Uh, what kind of best practices they might follow
to maximize the clinical outcomes
and clinical efficiency?
And what role does a top notch user experience
Play in that? We'd like to
start with that one.
I'm looking for hands. Anyone, Bueller.
Well, I can certainly start,
um,
as as far as
maximizing efficiency.
I mean, I think we get
back to some of the issues that
we've already talked about and that is having
image display tools
available,
having image analysis,
having some cross talk, cross
referencing between the different
imaging tools,
as well as integration with
our electronic medical record to pull
the kind of information that we need
that is relevant current,
uh, available and
and potentially impactful uh,
that goes into the imaging diagnoses.
And so
if we have all that information, preferably
cloud available, so it's available
quickly. Uh, we can integrate
that information, uh,
as rapidly as possible and
get to a relevant,
uh, and we hope, accurate diagnosis
and then getting to patient outcomes.
We have a much harder time with that
with imaging, because
imaging doesn't lend itself
as well
to a direct relationship
to outcomes, it's often quite
critical along the pathway.
But it's more difficult to come out
with a direct relationship between
an imaging result and an outcome.
And we tend to associate
that more with a particular drug therapy
or device therapy or surgical procedure.
Uh, and and so that gets more complicated.
However, we have an opportunity talking
about AI and machine learning and deep
learning and pulling all this information
together,
I think we're gonna have a much brighter
future, an opportunity to
be able to highlight and recognize
how important imaging is
in a patient's care pathway and
providing better outcomes for those patients.
Very good
do either of our other Panelists want to
chime in on that one?
I would only add that
are the American
Heart Association has funded a number of
data scientists
that have done great work in the area
of machine learning and
artificial intelligence
with imaging and
using that for predictive modeling.
Um, and I think the field
is is certainly moving
in that direction. And, uh,
we have a long ways to go. But I think
even sharing best practices, which
is something that has grown organically
out of our funded scientists, that they've gotten
together and
share best practices together
on many of the pipelines
that they're building and a lot of the practices
that they're doing, so more of that and open
data sharing. I think,
um is the way to go.
I'd like to bring I Tiger Lily back into
this conversation because I think this next
question
may fall in your bailiwick. I tie. Since
you work in product management.
The questioner many, She says
that
because of the complexity of
software positions and clinicians
spend a significant amount of their time doing
administrative tasks. And
the question is whether you believe that there's a need
to take care of such things
through software using automation
so that the physicians, the clinicians,
can concentrate more towards patient
care
Your thoughts I chi.
So I think
the answer is definitely yes
again at the end of the day. Like I think both
doctor from Doctor hell mentioned
the idea is for the system to help
the clinician do their job
and and serve that, you know, to the patient
care walkthrough. So I would say
it's also a little bit connected to the previous
questions in my mind, because we're looking
at What can we do to help?
At the end of the day, the physicians to focus
on the patients and forget about other
things so it could be taking all
the high volume, repetitive tasks
that are happening every
day and try to automate them. We use
AI machine learning to do that. One
example would be
a lot of our customers indicate
the challenge with
image quality in measurement
variability between different modalities
between different technicians with different
skill sets and experiences. And that's that's
something that AI has proven
to be able to reduce, variability
significantly
and actually also improve efficiency because
if you don't have to take all the images at the time
of the study on the mortality, But can
look at what the A. I can just fix
the one that you don't agree with. You just save
80% of your measurement work.
So that's an example for both. I think
another thing that we see a lot is the whole,
uh, charge.
Capture inventory management.
It's more on the invasive side
of cardiovascular. But
typical, you know, invasive procedure
can have dozens of charge
codes and inventory items,
and we want to make sure that the physicians are actually
treating the patients.
Opening data is replacing the malfunction
involved and not dealing with
scanning inventory items and making sure
that you reduce stock or capture
the charges emitted it on something. So
definitely, we think software systems
have the role and and and also
the responsibility to do executive
and allow clinicians to focus on patient
care. The last thing that was me, I
would mention, is again. We see today
sometime in his connection between technology
and workflow,
because if sometimes the technology will
get you to do some
of your studies on the ideology side,
some of your studies, for example, cities and MRI,
some sometimes even nuclear medicine,
and in some of the studies on because the other side
but the workflow could spend
across both
the cardiac city. That's happening, maybe
on the theology system would need to
be used in a structural heart procedure.
For a prep, you need to see the classification
of the valve. And if you don't have the software
tools to allow you to connect the dots and
actually be able to run your
workflow,
forget about the different actions
and technologies around the world for in a seamless
way, then I think
your efficiency is automatically
being produced. I think that's
that's an important part of any system that gets
into to serve commissions.
I know that Dr Hall has something
she'd like to add Dr Hall
just quickly. I'll just let
you know that there and several of you already
know this. There's many companies out there
that are already optimizing work floor
workflow in the hospitals and using
machine learning and AI and predictive
modeling to figure out number of beds
and discharge
and many things. So that is coming
and it's coming your way to separate
workflow and use machine
learning and predictive modeling to help
you achieve those goals.
So you as the physician are not
handling those things
very good. And Dr Flame, I saw you
shaking your head affirmatively during many
of those answers. Well,
I just think that there are so
many pieces of the puzzle that can
be or should be automated.
I think of so many reports
that I put together with C T E R
M R I that I'd love
to have more information pulled
from the electronic medical record automatically
inserted into the report
or measurements
that would, um, adjust
some of the measurements that
I've already obtained or technologists
have obtained. We're putting into the report,
uh, so that
all needs to be done. Uh, we should have that
today, and we don't, but I'm looking
forward to it as soon as possible.
Very good. Uh, Peggy from,
uh, H C A. Healthcare says
imaging is typically seen as an independent
support service.
This results in fragmented leadership.
And so she's asking what's the best practice
for incorporating
imaging
within CV services?
Who would like to go first?
No, I think I think that's
a doctor hall question
that's outside of my area.
Well, I guess
I guess when it comes to that question I'm sorry,
Peggy, but we
We don't have a Panelist who can help
you with that. So we're gonna have to move on to
Laura's question. Laura
is asking,
uh, regarding the PACs integration.
There are some integrate herbal software
solutions that work as an extension
within hospitals, own packs.
And there are some internal solutions working
as separate desk apps
or web apps, depending on the provider.
What
is your opinion on those two different solutions,
huh?
Well, let me let me toss out
an answer. And then afterward, I would
really like to hear what it says
about this. Uh, but
we certainly have both
options available.
Uh, we have some tools
that are stand alone, some tools
that are integrated with packs and some tools
that are both.
And I
think that
they each have their pluses
and minuses, at least today.
And this we
may really be I
don't want to say stuck with it, but it may
simply be where we are today
that we don't have the full integration. I think
earlier on in in these
discussions, we talked about having
a single tool
or orp anomaly of
tools available that would allow us to see
different imaging modalities and
have different, uh, tools,
measurement options, processing ability
for that all integrated together.
And that might be the very best option for
us. Uh, but we don't have that
today. And I think in the current
environment,
I look at the tools that I have,
and if they're stand alone, I'm sort
of used to that. If it were integrated
into packs,
I'm not necessarily sure
that I would like that better. But I'd
like the option to be able to try that and
see if it worked better for
me if it were a
smoother, more efficient, more,
uh, more obvious and
intuitive to be able to use those tools
if they were integrated.
Well, I tie. He teed you up, so
I'm talking to you. Thanks,
Jerry. So definitely, I think
it's It goes back to the classic
battle between best of breed
and a single vendor.
If you want. You know all of your
applications to be from single providers,
that we have a unified, smooth user
experience. You have to sometimes compromise
on capabilities,
or if you want to choose for each and every
cast that you're performing, you want to choose
the best of breed, uh, cutting
edge solution. You oftentimes find yourself
with many two separate standalone
systems that you need to toggle between to
complete your work. And I think
from what we've seen that that's
how it's used to be a piece from what we
saw, you know, out there in the market.
But I think technology today is
at the point that people should
stop compromising, at least on a lot
of things. You can still have these standalone
systems today, but the the level
of integration
and the ability to embed systems. And
when we move from declines to
Web apps to zero footprint,
this becomes easier and easier to still
use those standards and best of breed solutions
and embed them very seriously into
into your workflow into your system. So
I think pax historically was
the backbone. See packs of imaging
workflows. I think those
integration capabilities should
allow us to maintain that backbone,
but bringing all these new capabilities
that us as a as a you
know, PACs vendors won't have
the resources to always provide
the best
and most cutting edge solutions,
but we should definitely be able to bring that in
and allow,
you know, conditions like Dr Flam you
to use them inside the workflow
in the most in this way.
Alright, folks, it's time for you
to get out your crystal balls
And look five years down
the path. five years in the future.
What emerging technology do you see
most? Disrupting
the cardiovascular imaging market
In the next five years, I'm going
to start with Dr. Hall on this one.
Yeah, I
think. Did you get Did you get your crystal ball
from Amazon? Yeah.
I'm just now pulling it out
of of the drawer here. Jerry, Um,
I think it's really gonna be
driven by a number of
things. Um,
certainly technology and
the speed at which
at which it is moving
and computational power.
So the speed at which
clinicians can pull
up images, bring them all together
and analyze them with,
you know, other data at their
fingertips. Like Scott is saying,
the electronic medical record,
the ability to have
templates at the ready. So these
tools and reports
are
Kind of a thing of the past. They almost
automatically write themselves one
Scott is able to transcribe,
you know, his expertise
down um, in those
very difficult cases,
and I think he will see what we talked
about, which is the separation of workflow
from real clinician,
um, duties, which is really taking
care of the patient. And that
will be,
you know, driven by a lot of machine learning
on that workflow side, which
is something maybe we didn't predict many
years ago.
Dr. Flat, I'm gonna throw it over to you, get
the crystal ball out, take a peer into
it and tell me what's happening with emerging
technologies in the next five years. What's going to be
the big disruptor? Well, five years
is still pretty close, and
naturally, I'm biased because
I'm someone who focuses
more on C t and M R I.
And I will tell you, with both of those technologies,
I am really excited
what the near future is bringing.
When I look at what m
r. I is capable for the cardiovascular
realm what it's capable of doing, uh,
in looking at tissue characterization,
Uh, and more to the
point in the speed
at which it's being applied so
far for me. You know, a couple decades we've
had cardiac MRI examinations
that take a long time. It can be an
hour plus, uh, in many
places. But I think we are seeing
the advent of
examinations that
are Acquiring an incredible
breath of information with function
and tissue characterization and flow quantification,
uh, strain information.
All of this is going
to be acquired in 10 to 15 minutes,
and it will be so much faster, a better experience
for the patient and an incredible
wealth of information in this short period
of time.
And at the same time, I think about where
we are going in C t scanning
with these new, uh, crystals
that are, uh, photon counting
that are that are giving us greater
precision, greater detail,
both spatial resolution and tissue characterization
looking at arthroscopic
plaque in the coronary arteries. And
I think this sort of detailed information
combined with the AI and machine
learning that we've been talking about, is
going to get us to the point of finally
being
more predictive not just
for population science, but on
an individual patient basis,
really getting into the
people talk about precision medicine,
but into personalized medicine. And
I think we're going to get there, uh, soon
because we've got some great tools that are opening
some some great vistas for
us in patient care.
Interesting, I tie. Let me
allow you to peer into the crystal ball.
What do you see? So
first I want to join Dr Flan
with the whole shifting imaging mix.
I also believe that we're going to see more and more
cardiac cities and cardiac MRIs
being performed
on the expense of maybe
Auto Sounds and NM studies.
The second thing that I'm starting to
see is a better connectivity,
not just to the EMR but also
to data coming from outside of the hospital,
whether it's
consumer devices that
help is screening and closing the diagnostic
gap. So more patients actually
knows they have a condition earlier
in the process.
And the second I think also chronic care management,
home care management for people with
cardiac disease. And I think that
connection to the acute episode
is going to call stronger over time. So
those are the things I would,
you know,
with my money on
very good and you mentioned data.
So how important is data
consolidation?
Uh,
in the current environment, I
wonder if you could address that
I tie. I'll start with you, Go ahead.
So I think if we've learned
something today, is that data consolidation
is key for any future.
You know, success in cardiovascular
if you want to get insights If you want
to be able to improve patient care,
you have to start with being able to
understand the data you're getting and
turn that into insights
and be able to drive outcomes to understand
what is happening. So I think
one of the you know Dr Hall
talked about that. And Dr Frank talked about it several
times in this session that
one of the biggest challenges is
how do we get all this
data to preside in multiple places and actually
paint one full picture
that gives us, you know, understanding of the
patient health
and try to predict what it would look like
in the future. And by the way, we already
see some companies that are combining
SCG data within our data
and having some pretty good results
on predicting future onset
of cardiac disease such as heart
and problems, and even stole, so it's not
there yet, but I think
it's definitely a very important
key in enabling a lot of these things in the future.
We have another question from one of our viewers.
This is a question for Dr Hall.
Most of the software, the
question goes, most of the software provides very
loose
information in terms of patient risk assessment.
Uh, lose, because the data is from
mostly single databases rather than aggregations
from data storage locations, data
silos or from other vendors. So
the question is, do you believe artificial
intelligence can
definitely helped to pull the key data elements
about the patient
so that it automatically creates a risk
assessment from all these data
that are scattered in different locations
and still adhere to current guidelines
and and have some sort of graphical format?
I would
I would back up a little bit and say
that AI is not a magic
bullet.
And neither is any statistical
method. In fact, some of the older
statistical methods have outshone
ai and many methods,
Um, and in many different
papers of of late,
the key to all of it is the data
upon which you build
the artificial intelligence
models. And so if you have
and uh,
person with the question is absolutely
correct, if they're built on small
research studies or single research
studies.
Um, those risk prediction
models will be relatively,
um,
minor. You know, they're not gonna They're not
gonna have a great effect. It's not until we
begin to share data and use
data that includes
multiple
individuals, people from underrepresented
groups and minorities as
well as women.
That will have better insights that
we can begin to really break down
what's happening. And then when we have
that large data set, then we can do
things with artificial intelligence and keep
in mind bias and things. Thanks
for the question. Yeah,
I've done a number of interviews regarding artificial intelligence
and machine learning, and the recurring
theme is always
you need the big what we used to say Big data.
You need lots of data for it.
In other words, you know, if you don't have
that base of data,
it just won't be as effective.
Here's a question from Laura.
Um, it's a question concerning how much
effort it takes from clinicians to
learn to operate new software. Is
it better designing the user interface
to look and follow the same logic as
existing ones?
Four.
Try to truly understand the user flows
in the inter place and optimize the experience.
Uh, and taking the risk
of the interface looking so different from the existing
ones. Well, something that looks different,
the intimidating who approach,
uh, in your opinion, even though
you know, if all the user testing in and heat maps
show evidence that they work better.
And since you're in product management,
I ty I'm putting this one on
your desk.
Yeah, thank you. So, by the way, I would definitely
love to hear Dr Flam after
the You know the clinician's perspective
on that because I can, you know,
bubble on the, uh
what what we've seen and what our
in theory, you can walk in with US
companies and UX people.
At the end of the day again, it's a balance
between how do we drive adoption? What is the adoption
period? Because
ultimately, of course, it's better to
have the work for terror
to the you extra to
to serve the work for best, even if it's very
different from today.
But there is a learning curve, and you
have to get above a certain threshold
to be able to get adoption.
And if you won't get above that threshold. It's
hard, tough, and traditions are very busy,
and I do think it's intimidating
to switch.
So I think if you want to do the switch, you have
to be very mindful of what it means
to clinicians. And you have to provide the right tools
in the right level of support
to get them over that barrier. I think in
the long run, it's probably the right
thing to do is to focus on what's best
for the workflow, but what we think
is the best user experience.
But we have again to be very mindful
of what is the journey from
today to that new user
experience? It's not gonna happen in
a single day and it's not gonna happen.
It's not gonna be pain free,
so we have to provide the right tools
to get them from Point A to
point B. Dr.
Flam Uh, if you'd like, I can recap
the question for you unless you know
I'm good, ready
to go on that one, and I think
it is really
an individualized answer. I
certainly have colleagues and no people
who continue to use one tool
because they just
don't want to change their dug in. They
know how to use that tool, and it provides
them enough information
and they just don't want to change their
kind of dug in.
And, uh, there are
other people like me who are sort of in
the middle, where
I'm perfectly comfortable learning
new tools and new interfaces
as long as it's not onerous.
Uh, if it's an onerous
proposition, then you know
who wants to do that. It becomes too difficult,
no matter how good the tool is. And
so you really need to design
tools that have a at least a reasonable
interface. I mean, certainly you'd like
it to be simple, intuitive,
obvious, uh, and something
that you can embrace very, very easily
and obviously, But, uh, most
Most of these
complex
tools available now to do
post processing of a variety
of imaging modalities have
a complexity to them because
the imaging that we're acquiring is complex,
with a variety of dimensions
to it. Uh, so it
is challenging. Uh, you know, I
think that
a reasonable person can move
to a another
tool walks tool set with
without it being too much of a process.
Um, but, uh, people
who are developing these tools do have
to be mindful and make sure
that the transition
process is one that is not
too difficult, uh, and
something that people really embrace
and would like to move forward with,
because the tool available is better.
Well, folks, I have to say that the
sun and the moon and the stars must have aligned
just perfectly because
we have run out of questions at the same
moment that we've run out of time. So
I'd like to thank our Panelists.
Dr Jennifer Hall
from the American Heart Association, Dr Scott
Flam from the Cleveland Clinic Imaging Center
and I tie Galilee of Change
Healthcare. Thank you very much. Uh,
for those of you watching, we will follow up
with the recording of this session
so that you can share it with others
for later viewing. If you're interested
in learning more about change healthcare,
cardiovascular solutions,
please head to info dot
change healthcare dot com
forward slash cardiology Again.
That's info dot change healthcare
dot com forward slash cardiology.
Thank you. We hope you have a great day
Related resources
White paper
By applying analytics solutions, states can create a comprehensive roadmap to better health equity.
Case study
Using digital integration to ease administrative burdens can help you spend more time with patients and deliver positive patient outcomes.
White paper
Read the white paper for strategies on controlling fixed costs, efficiency and organizational flexibility.