Latest / Becker’s Healthcare Podcast / Clinical AI Use Cases Shaping the Future of Healthcare

Transcript

Clinical AI Use Cases Shaping the Future of Healthcare · Becker’s Healthcare Podcast

0:00
This is Scott Becker with the Becker's health
0:02
care podcast.
0:04
I am thrilled today to be joined by
0:05
your brilliant leader. We're gonna talk today about
0:08
clinical AI use cases that are shaping health
0:11
care's future.
0:13
We're excited to be joined by Marcus Perez.
0:15
Marcus, the president of Alterra Digital Health. He's
0:19
also by background. We shall be thankful to
0:21
him. He's a West Point graduate.
0:23
Marcus, so thrilled to have you on with
0:25
us today.
0:27
Can you take a moment
0:28
to introduce yourself and tell us a bit
0:30
about Alterra Digital Health?
0:32
Sure. Well, thanks for having me, Scott. I,
0:35
I've been with Alterra three and a half
0:37
years since we acquired these assets from Allscripts
0:40
in May 2022.
0:42
Been with Harris, which is the parent company
0:45
of Altera
0:46
since 2012 in various roles. Harris is the
0:49
wholly owned subsidiary of Constellation Software,
0:52
which is the publicly traded entity. We're a
0:55
vertical market
0:56
software organization, the sixteenth largest in the world,
1:00
and, we acquire and hold forever assets like
1:03
these. So we're doing our best over the
1:05
last three and a half years and going
1:07
forward to,
1:08
build
1:09
as strong,
1:10
portfolio of businesses in the health care space
1:12
as we can.
1:13
Marcus, let me ask you another question. Before
1:16
we dive into
1:17
data and technology
1:19
and in all of what you're doing and
1:20
what you're seeing out there and where you
1:22
see value being added, Can you take a
1:24
moment and just give us a snapshot of
1:26
what Alterra Digital Health does and specializes in
1:29
just so our audience has familiarity?
1:32
Sure. So, Scott, we we have a vision
1:34
and and that vision is,
1:36
we measure our success by our clients' outcomes.
1:39
What does that mean? Well, we really focus
1:41
on three areas. Right? We wanna enable clinicians,
1:45
to be able to do more. We wanna
1:47
empower patients to be able to drive their
1:50
own health care on some level, and we
1:52
wanna help organizations
1:54
transform.
1:55
What does that mean? Really, it's how do
1:57
you take advantage of the power of technology
1:59
to deliver better care as an organization.
2:02
Ultimately, we look at ourselves as a healthcare
2:04
company,
2:06
not a technology company. Yes, we have electronic
2:09
health records, we have data businesses, we have
2:11
hosting businesses, we do outsourcing. There's all sorts
2:15
of stuff we do,
2:16
but at the end of the day we
2:18
consider ourselves a health care company,
2:21
and that's where we think we're different in
2:22
this space.
2:24
Thank you very, very much. And and and
2:26
take a moment. There there's not a shortage
2:28
of data in health care. I mean, there's
2:30
so much data roaming around,
2:32
but organizations still struggle with data quality and
2:35
and then taking that data and turning it
2:37
into useful action.
2:40
And how was this mismatch between data and
2:42
use and how much data is out there
2:43
and and use and active usefulness of it?
2:47
How is that shaping the broader implications? What
2:50
are some of the broader implications
2:52
for clinical and financial operations?
2:54
No. Well, what I think is interesting, Scott,
2:56
is 97%
2:58
of the health care data out there today
3:00
isn't used at all.
3:02
And it's because of what you've described. There
3:04
are all kinds of challenges, particularly around
3:08
how the data is collated, the semantics around
3:10
it. There's a lack of standards.
3:12
And so if you think about, you know,
3:14
how do we maximize
3:15
AI?
3:16
AI is really driven by data. So data
3:19
quality, particularly trusted data,
3:22
becomes paramount to anything you wanna do in
3:24
healthcare with AI. And so if you think
3:26
about that,
3:27
right, the duplicative nature of data that's out
3:30
there, the the semantic issues,
3:33
the challenges around errors and misdiagnoses
3:35
that result from poor data or lack of
3:39
trusted data. I think that's one of the
3:41
things we're focused on more than anything
3:44
is trying to make sure that we create
3:47
a way to deliver trusted data to any
3:50
and every source possible.
3:52
Again, if we're all gonna take advantage of
3:54
the power of AI and our job
3:57
at Alterra is to serve,
3:59
our constituents, and that means doing all we
4:02
can do to help our constituents deliver the
4:04
best care possible.
4:05
Let's start and focus on that data. And
4:08
and we're working through that today with a
4:10
product called Care Intelligence
4:12
that really does ingest the data from any
4:15
source,
4:16
and then you've gotta aggregate that data. What
4:18
what does that mean, aggregate it? Really means
4:21
we gotta put it into a standard language.
4:23
So you get these, all these semantic issues,
4:26
then you now need to harmonize the data,
4:28
which is resolve all those semantic issues, have
4:31
this standard data set, but then you have
4:33
to apply governance to it, Scott. And the
4:35
governance is what really matters because you track
4:37
the lineage of that data and make sure,
4:40
that it's trusted. Well, how do you do
4:42
that? We've applied what we call a trust
4:44
score to the data to make sure that
4:46
any outliers or things that don't match or
4:48
things that look a little wonky get eliminated,
4:51
and so that the data that's left that
4:53
has been aggregated and harmonized
4:55
is now able to be used in any
4:57
way, shape, or form. And we're doing that
5:00
process through AI,
5:02
which means then that you can essentially
5:05
plug and play any AI engine or tool
5:08
or agent into
5:10
the system and be able to take advantage
5:12
of all that data.
5:13
In the absence of trusted data,
5:16
we're never really going to be able to
5:18
accelerate
5:19
change in health care through AI.
5:23
Thank you very, very much. And and, Marcus,
5:25
take us a step further.
5:27
We see more and more value based care
5:30
in different systems, different areas become a good
5:33
priority.
5:34
What health IT capabilities
5:36
should leaders be focusing on? And and where
5:39
does it hold the most promise for supporting
5:41
proactive and preventive care delivery? What are the
5:44
most important health IT capabilities?
5:46
Yeah. You know, I think if you think
5:48
about value based care and value based right?
5:52
What is that all about? It's about better
5:54
outcomes. I think, you know, at the end
5:56
of the
5:57
day, having this aggregated, normalized, harmonized data means
6:02
that you can then deliver better outcomes. And
6:04
again,
6:05
if you're
6:06
a if you're a, you know, an organization
6:08
who's trying to think about proactive
6:11
preventative care,
6:13
that again is about better outcomes. And so
6:15
these two things are, I I I think,
6:17
very congruent on a parallel track that the
6:20
better the data, the better the opportunity to
6:22
deliver better outcomes, which means,
6:25
that's what will drive value based care.
6:28
Thank you. And to follow-up on that, how
6:30
can leaders decide what are the most important
6:33
initiatives?
6:34
What's gonna bring measurable value? And and what
6:36
shouldn't they be doing in terms of health
6:38
IT priorities?
6:39
Yeah. You know, everybody likes to talk about
6:41
AI and and and driving AI initiatives, and
6:44
I think that's important.
6:46
But when it comes down to it, you
6:48
know, every community's got unique needs,
6:51
and this idea around reliable traceable data becomes
6:55
super critical.
6:57
If I was leading a healthcare organization today,
7:00
well, I would be absolutely focused on trying
7:03
to drive AI. I would be hyper focused
7:05
on where are we getting our data from,
7:08
how much can we trust that data, so
7:10
that we can make sure that whatever we
7:12
use that data for relative to AI or
7:15
processes, we're going to, you know, get great
7:17
results. We wanna make sure that our clinicians
7:20
have that they're they're most informed at the
7:22
point of care, so we can drive these
7:24
better outcomes.
7:26
That all, again, hinges back on on data.
7:30
Thank you. We're seeing the last couple years
7:33
and really accelerating.
7:35
Ambient Scribes and AgenTic AI continue to gain
7:39
traction across health care.
7:41
From your perspective,
7:42
what are the most compelling AI use cases
7:44
right now?
7:46
And you could can you share an example
7:47
that illustrates why these things are driving impact
7:50
and where they're driving impact?
7:52
Sure.
7:53
I I think, you know, for us anyway,
7:57
AI is an asset
7:58
that connects with human intelligence. Right? We wanna
8:01
connect AI to the human side of care.
8:04
So you mentioned, you know, scribes, and and
8:07
I think that's important, right, to have this,
8:09
AI ambient listening machine learning that that can
8:12
drive,
8:13
you know, alleviate clinician burden, which is certainly
8:16
a problem.
8:17
But at the end of the day, what
8:18
you don't want is just glorified dictation.
8:21
Right? You think about,
8:23
again back to the data and, you know,
8:25
the ability to read unstructured notes is super
8:28
critical.
8:29
And so here we are, you know, with
8:31
this dictation, essentially,
8:33
this ambient listening that's supposed to alleviate burden,
8:36
but really all it does is is data
8:38
dump into the unstructured note. And you gotta
8:41
be able to read that unstructured note, and
8:43
you've gotta be able to process the data
8:45
inside it. What we're doing is taking that
8:48
ambient listening and machine learning and AI to
8:50
a different level, our note plus product in
8:53
in our Touchworks
8:54
business.
8:55
What it does is not only
8:58
do the ambient listening,
8:59
it then populates all of the clinical documentation
9:03
to include notes, datasets, referrals,
9:06
even CPT codes. What that does now is
9:09
it really does allow the clinician to review
9:12
that note as soon as the encounter is
9:15
over and make sure that everything is there,
9:17
then there's no documentation at all to be
9:20
done later.
9:21
Think about the difference between
9:24
dictation, which just puts notes in a box,
9:29
and a tool
9:30
that not only listens,
9:32
but then populates all the fields in a
9:35
clinical documentation,
9:37
including the note and the CPT codes and
9:39
the referrals and all of the order sets.
9:41
And now that's a full note that doesn't
9:43
need to be worked on later.
9:45
The alleviation of burden that that could do
9:48
for clinicians, I think, is monumental.
9:51
Thank you. No. We and we love that,
9:53
and that is so important in the era
9:54
of burnout and change as well.
9:57
Talk for a second about change management. We
10:00
constantly hear critical change management is,
10:03
for moving IT initiatives from pilot to wide
10:06
scale adoption.
10:07
How are you seeing clinicians
10:09
and operational teams adopt to the pace of
10:12
change in health care,
10:13
especially as technology plays a much bigger role
10:16
in everyday workflows? What are you seeing there?
10:20
You know, we love to I love to
10:21
hear folks talk about health care revolutions.
10:24
There there there's no revolution, Scott. Right? The
10:26
the the folks aren't coming with the pitchforks
10:29
and the torches.
10:30
It's an evolution, and I think you've gotta
10:32
find places
10:34
that you can get incremental
10:36
and sustainable progress, right? And really this is
10:39
around to me meaningful
10:41
workflow improvement.
10:43
And this is where AI can really augment
10:45
that improvement, right? It's never gonna replace the
10:48
expertise
10:49
of clinicians, but it can
10:51
augment them in terms of meaningful workflow improvement.
10:55
And let's look at a use case like
10:57
say the ED,
10:58
right? It's a great example
11:00
of where if you have
11:03
this aggregated harmonized,
11:05
data that you've gotten from all these different
11:07
sources,
11:08
now a clinician is able to, at the
11:11
point of care,
11:12
have all that data, including all the claim
11:15
history
11:16
to help alleviate denials, to make sure that
11:18
they process,
11:20
what they're going to do and get it,
11:22
you know, get it paid for. They get
11:24
all the clinical data that they need on
11:26
that patient at the point of care, and
11:28
they can make sure that they can deliver
11:30
great care that they know they'll get paid
11:32
to do, and the patient will have a
11:33
better outcome. That's that's the win that everybody's
11:36
looking for, but candidly, today just doesn't exist.
11:40
But I think your point on this being
11:42
an evolution
11:43
versus a revolution is is right on. And
11:46
it seems to be sort of cascading
11:48
waves, some of them quicker, some of them
11:51
slower. We all know in the long term,
11:53
technology is gonna have a bigger and bigger
11:55
impact in input, but it really is an
11:57
evolution. It's not like a a a switch
11:59
goes off and it all changes. It really
12:01
is an evolution
12:03
and teams and people gotta get used to
12:04
it as well. I think you're right on
12:06
there. I love that. That's right. Yeah. And
12:08
if you think about
12:10
what's unique about Alterra is that we actually
12:13
can work at the pace of change for
12:15
an individual entity,
12:17
and their workflows, because workflows are very unique,
12:20
and and organizations
12:23
all have a uniqueness to them that can't
12:25
be addressed by a one size fits all
12:29
approach. And so
12:30
if you think about trying to pull people
12:32
along this spectrum
12:34
of change, this evolution, if you will, you've
12:37
gotta have the ability to meet them where
12:39
they're at.
12:40
And where they're at is in the workflows,
12:42
and so I I think that's what makes
12:44
us, quite unique is that we can work
12:46
at the pace of change that folks are
12:48
comfortable with. Some organizations are trying to accelerate,
12:52
others are really sort of dipping their toe
12:54
in the water and looking for
12:56
wins, simple wins that drive
12:59
more energy and enthusiasm towards
13:01
doing more things to evolve. I think that's
13:03
the joy of of what we do is
13:05
we're willing to meet folks where they are.
13:07
Thank you, Noah. And I and I and
13:09
I love that. I'm gonna come back to
13:10
it with another question about Terra in a
13:11
moment. Be be before we go further, let
13:14
me ask you this question.
13:15
Looking ahead,
13:17
what technologies or innovative approaches
13:20
do you believe will have the greatest value
13:21
for health organizations over the next few years?
13:25
Yeah. I I to me, it still distills
13:27
down to the data, and I know I
13:29
sound a bit like a data wonk right
13:30
now, but it really does when you think
13:33
that 97%
13:34
of health care data goes unused.
13:36
And, you know, clean data with robust governance
13:40
is really the linchpin
13:41
for success.
13:43
I think you know, imagine a world, right,
13:45
it's one of the things I love to
13:46
do is sort of, you know, ask people
13:48
to close their eyes and imagine a world
13:50
where, right, imagine a world where you could
13:52
have a single source of reliable,
13:54
true data
13:55
fueled by strong governance, right, that enables clinicians
13:58
and organizations
14:00
to unlock the full potential
14:01
of both the data
14:03
and AI.
14:04
And that's where the technology really does start
14:07
to,
14:08
be a force multiplier
14:10
because it still is the linchpin of the
14:13
data. And so if I was, you know,
14:15
if I was an organization,
14:17
and I was thinking over the next few
14:19
years, where where do I invest my time,
14:21
energy, attention, and resources?
14:24
I would be investing in how do we
14:26
get tools that ensure that we can take
14:29
advantage of the power
14:31
that lies in this data. And that's not
14:33
just power to treat patients, but again it's
14:35
it's the power to
14:37
improve,
14:38
right denials
14:40
and denial rates because you've now got all
14:42
this payer data. There's
14:44
so many things. There's predictive analytics,
14:47
there's proactive care, there's all kinds of things
14:50
you can do
14:52
when you unlock the power of the data.
14:54
And if you think about
14:56
how much health care data is out there
14:58
today,
15:00
it's almost mind numbing to realize how much
15:02
is out there, and to know that 97%
15:05
of that goes unused
15:07
is,
15:08
I I think it's quite disappointing because
15:10
what could we do
15:12
if we could access all that data and
15:14
then
15:15
apply
15:15
the technological
15:16
tools like AI and machine learning to it?
15:19
The advances could be vast.
15:22
Marcus, I think just fantastic what you're doing
15:25
in Alterra Digital Health. Again, we're joined today
15:28
by Marcus Perez,
15:29
the president of Alterra Digital Health. Thank you
15:32
so much for joining us today on the
15:33
Becker's Healthcare podcast.
15:35
Thanks for having me, Scott. Really enjoyed it.