All Episodes

October 15, 2025 14 mins

This episode, recorded live at the 10th Annual Health IT + Digital Health + RCM Annual Meeting, features Dr. Bryon Frost, CMIO of McLeod Health. He shares how his team developed a bias-resistant, three-phase approach to selecting and scaling AI scribe technology, improving physician experience, reducing cognitive burden, and achieving measurable financial and clinical ROI.

This episode is sponsored by Suki.

Mark as Played
Transcript

Episode Transcript

Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Everyone. This is Lucas Voss with Becker's Healthcare.
Thanks so much for tuning in to the
Becker's Healthcare podcast series. Fantastic to have you.
Today, we're gonna talk about the three phases
to success, McLeod's Health's roadmap for scalable AI,
Scribe adoption. And joining me for that discussion
is doctor Brian Frost, CMIO at McLeod Health.
Doctor Frost, thanks so much for being here
today. It's great to have you. Thanks for

(00:21):
having me. Absolutely. I wanna start off with
introductions for our audience that might not know
you. If you could just introduce yourself and
give us a little bit of an overview
of your organization. So I am Brian Frost.
I'm a emergency room physician and chief medical
information officer for McLeod Health.
We are a seven hospital system in rural
South Carolina.
We're a non for profit

(00:41):
hospital and non academic, and I think that's
really important to kind of frame the discussion
here because
we do not unlike some of the larger
health systems out there, we don't have a
huge foundation,
and so innovation is
very dependent on finances. And so when we
pick a solution or a vendor,

(01:02):
we're very careful in selecting that vendor. The
ROI has to be there, and we really
cannot afford to purchase a solution
that is a lemon, so to speak. We've
gotta make sure that it is going to
live up to whatever the solution is that
we're the problem that we're trying to solve,
it really has to solve it. Which has
become very important, especially in a crowded AI
space. And we know especially Ambient AI has

(01:24):
really become
a strategic priority for a lot of organizations.
It's very important. It's risen to the top.
What factors from your perspective have made AI
scribes
so important for physicians, but also for organizations
from a strategic perspective?
We knew this was an area we wanted
to get into. And the problem that we're

(01:44):
solving, everybody solving,
is physician burnout. I really don't like that
term so much, but it's the cognitive burden
of practicing medicine. It's the thing that sucks
the life out of a physician who wants
to get back to return to the joy
of practicing medicine. And so we knew we
wanted to be in this space,
but we also one of the things I
was very aware of, so in medical school,

(02:05):
they teach you. Do not take,
you know, drug reps when they give you
free stuff, don't take it. And there's a
reason behind that because they want us to
be bias resistant. They want us to,
prescribe medications
not because the vendor bought you breakfast that
morning, but because they actually treat patients and
patients will get healthier.

(02:26):
I wanted to approach vendor selection in a
very similar fashion in this space.
I had seen a lot of the advertising.
I'd seen a lot of the demos,
and so we put together
what I'm gonna describe as a,
bias resistant,
multiphase approach to vendor selection in this space.

(02:46):
And I wanna talk about that approach a
little bit more in detail. Can you walk
us through it's three phases. It is. Can
you walk us through those three phases? And
you mentioned the bias piece. How is it
minimizing
that bias, and how is it improving those
workflows? Well,
the the experiment was fascinating to me personally
because
I didn't recognize the problem with the unconscious
bias is you're not conscious of it. Correct.

(03:07):
So I was pretty sure that I was
not biased, but I was also pretty sure
my organization was. And so I put this
experiment together to kind of prove to the,
organization
that this vendor that I thought was best
for this area was the one we should
select,
and boy was I wrong. It was very
humbling when I got to the other end
of this experiment. Mhmm. So the first part

(03:29):
of the experiment was actually pre phase, phase
zero, I'll call it. I spent about six
months vetting all the vendors in this space.
Mhmm. I went through Avia Marketplace.
I went to class.
I went to various demos, and I excluded,
a majority of the vendors based on scalability
concerns or financial viability.

(03:49):
I didn't feel like they'd still be there
in a year or two.
And I narrowed it down to the top
four vendors in this space,
and these are common vendors that everyone knows
and everyone is using,
and then entered those into
phase one of the experiment.
So and I can't believe I the the
vendors agreed to do this. So let me
just tell you what I did.

(04:10):
We,
together with, some revenue cycle people at our
organization,
I put together 15 patient scripts,
and these are very detailed scripts. They were
describing
pretty,
bread and butter,
cases that you might see in medicine.
And I had three physicians. I had a
cardiologist,

(04:31):
a surgeon, and a family doctor.
Each of them had five encounters to see,
and we got we had,
basically, actors come in and play the part
of patients. And these were operational leaders throughout
the organization, like my CIO,
the chief of staff was,
on this our our our corporate CMO,

(04:51):
chief quality officer was there. We had a
lot of IT,
senior executives play the patients.
And these scripts were fun because they were
very complicated.
Their goal was to try to throw off
the AI. So they had interruptions, they they
had, you know, family member actors who would
interrupt the patient and contradict the patient. We

(05:11):
had like nurses would bust in the middle
of the encounter and give extraneous information on
a different patient. One of the encounters, I
I actually created the script and then right
at the last minute flipped it a little
bit. This script, I had the surgeon it
was all about the surgeon. I made the
surgeon be kind of a jerk to the
patient and dismissive of the referring physician, which

(05:31):
is funny because the surgeon is not like
that at all in real life. So it's
fun for all of us to watch this,
and
the goal of this experiment was to see
how well the note output
was generated.
And so the rules of the game were
that the vendors had to be there in
real time. They had to record these interactions
between the patient and the physician. They had

(05:52):
to send me the note output immediately following
the interaction.
Note.
We took those notes and we ran them
through three groups of people. We had revenue
cycle experts, we had physicians, and then we
had nonclinical patients who would review the notes
and rate them.
And this was the part in the experiment

(06:12):
where I became humbled
because the vendor that I chose to be
the one that I thought the organization should
go with finished last place.
Which is an incredible experience because you're again,
it it is it's a fast evolving space.
I I do wanna ask a follow-up to
this. Obviously, this is a very involved process.
How important was it to actually run this

(06:32):
exercise?
And if you're talking with other leaders, is
this something that you would recommend others do
as well to really make an informed decision?
It's hard to do, but if you can
do it, yes. Because and and the goal
is to get rid of unconscious bias.
Commonly, when you look at the various vendors
or any vendor, I don't care which space
we're talking about, It's so hard to remain

(06:54):
objective
and to look at they all say the
same things, unfortunately. A lot of the vendors
will say the exact same thing. It's not
until you get a little deeper under the
tech stack and you look at what is
under the hood, do you start to realize
there's some pretty significant differences between the vendors
and that's definitely true in this space. Yeah.
What while we did this,

(07:15):
evaluation
so let let's move into phase two of
the experiment. Phase two,
we had the two vendors that survived phase
one come back and present their
workflow to a group of a broad group
of physicians across McLeod Health.
And the focus was on physician
workflow, how well could you incorporate this into

(07:36):
your practice.
And so that's what they were evaluating. In
parallel,
we had the clinical informatics department and the
IT
security folks really look under the hood and
look at the tech stack, both from a
security perspective,
but also the just pure
what can this company do,
looking beyond Ambient. Because I'm pretty firm in

(07:56):
my belief
that if we're gonna fundamentally transform health care
delivery, we've gotta go way beyond just Ambient
documentation.
And then phase three is then the the
It was the pilot. Pick. Correct? And my
CEO was great about this because my initial
plan was to do a bake off between
the two surviving vendors.
She said and it was really the physicians.

(08:18):
They were so adamant that one vendor one
vendor performed 90%
compared to the other. She said, let's let's
not do a bake off. Let's just go
with the one that all the physicians liked.
If it doesn't work out, we'll switch. And
so we ended up going live with the
vendor, which happened to be Suki Mhmm. Which
I didn't I almost didn't even evaluate because

(08:39):
of the name. I just thought, I don't
know what that that's just kind of a
funny name. I Yeah. Yeah. Yeah. So
And it was it was significant.
The KPIs
that we were able to get out of
this were
very interesting. So,
we went into it with very firm understanding
that we were not going into this
space to solve a financial problem. We were

(09:00):
very clear with the doctors.
We don't wanna make money if we break
even. Great. We'll be fine if we just
break even. We want to mitigate burden. We
also told the doctors,
if you get that time back in your
day, don't use that time to see more
patients,
because the last thing I want to do
let's say you've got a family doctor who
is seeing 35 patients per day, and you

(09:22):
give him some time back in his day,
the last thing I want to do is
have that doctor pick up up five more
patients. So we we emphasized,
and and I'll tell you, I mean, there's
some doctors who are just starting their practice
and they do need to see more patients,
but, you know, if you're an established physician,
we want you to spend more time with
your patients, we want to have you spend
more time with your families after hours,

(09:42):
and so that's what we emphasized.
So my doctors were extremely happy, just like
all the solutions. Everybody says Ambient makes a
big difference, life changing difference for these doctors.
So from a physician perspective, I was happy.
My my medical staff was pleased, but I
didn't expect my CFO to be as happy
as he was.

(10:02):
So let me just tell you a mistake
we made.
When we set out on this, pilot, we
initially made the decision to only include the
physicians who were at the, seventieth percentile of
efficiency.
And that decision was born out of fear.
So most of the vendors in this space,
they issue a
subscription model. So you sign up for 500

(10:25):
licenses, and you pay a fixed fee for
those licenses.
The problem with that model
is if I didn't have a 100% adoption,
I've got some licenses that are unseated that
I'm paying for, and I didn't want that.
And so when I get down to the
total cost of ownership,
I consider not only the licenses I'm not
using, but also if I have a doctor,
let's say, who enjoys using it but doesn't

(10:47):
fully adopt it. You know, they may just
use 10 times a month.
Could I handle, you know, the cost of
$400
per provider per month?
No. So what we did,
we went live. And by the way, that
was a dumb decision because
the doctor's practice who's super efficient doesn't really
have much pajama time. I mean, there's not

(11:08):
a lot of after hours time that they're
spending.
And so a little further into the, pilot,
probably month two, we negotiated with Suki to
come up with a new model. So we
came up with a encounter
based pricing model.
That unlocked something that I didn't expect. So
what that allowed us to do is to

(11:28):
simply pay a small fee per encounter
with a cap that was actually a little
lower than the other vendors we were looking
at. And
now I don't have to worry about who
gets a license or not. So one of
the things a lot of CMIOs are now
learning that they're now in the business of
license swapping,
where I have to pull a license from
one doc because they're not really using it

(11:50):
and try to give it to another,
or maybe they like it, but they just
don't use it much. And I gotta make
this decision. Am I gonna pull this from
that doctor and give it to another? I
don't have to worry about that.
I've actually provisioned all of our doctors with
licenses. The other thing I'll say is you
really can't
predict who will adopt and not. I had
several doctors that I was absolutely certain were

(12:11):
gonna love the product and didn't adopt, and
then I had some I thought weren't going
to adopt and they loved it. So
I didn't have to worry about that. The
second thing is it aligns the interest between
the vendor and the health system.
So I wasn't
you know, it was they went at risk,
essentially. If we didn't adopt and utilize it,

(12:33):
they didn't get paid. And so they were
very vested in making sure we
scaled the the product
currently. And this is what I'm really excited
about.
Net, and the I'm gonna tell you our
financial ROI on this. Just looking at CPT
changes, so looking at level four
charges went up, level three charges went down,

(12:54):
and we had probably about an eight to
10% increase in volume.
Just those changes alone, net, after subscription costs,
we're currently right now
making $2,636
per provider per month net every month.
And I don't think you can get there
without a utilization based model,

(13:14):
without some creative math.
You have to kinda look past the fact
that you're paying for licenses
that aren't being fully utilized. Yeah. Absolutely. It
certainly makes an impact on it has an
impact on physicians. It has an impact financially
on the organization as you've highlighted, which is
really important.
Doctor Frost, thanks so much for being here.
What a great conversation. Anything else that you
wanna share, that might be important for for

(13:35):
our listeners to understand about the project itself
or in general your feelings towards Ambient AI?
Any final thoughts on on the topic? I'll
just say so I'm very optimistic about how
these platforms will transform health care delivery. They
are the thing that is now in the
room with the patient and the doctor and
I feel like they're a platform to deliver
a lot more than just ambient documentation.

(14:05):
Specialty specific
summary of the patient is gonna make a
huge difference. Revenue cycle. There's a lot of
things that this platform can deliver to the
provider to make us better providers and, at
the end of the day, provide better care
for patients. So And that's the key. Better
outcomes, better care for patients. Doctor Frost, thanks
so much for taking the time. It's great
to have you. Thanks for the time. Yeah.

(14:25):
Absolutely.
And you can certainly turn into more podcasts
from Becker's Healthcare at beckershospitalreview.com.
And we also want to thank our podcast
sponsor,
Suki.
Advertise With Us

Popular Podcasts

CrimeLess: Hillbilly Heist

CrimeLess: Hillbilly Heist

It’s 1996 in rural North Carolina, and an oddball crew makes history when they pull off America’s third largest cash heist. But it’s all downhill from there. Join host Johnny Knoxville as he unspools a wild and woolly tale about a group of regular ‘ol folks who risked it all for a chance at a better life. CrimeLess: Hillbilly Heist answers the question: what would you do with 17.3 million dollars? The answer includes diamond rings, mansions, velvet Elvis paintings, plus a run for the border, murder-for-hire-plots, and FBI busts.

Crime Junkie

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.

Stuff You Should Know

Stuff You Should Know

If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.

Music, radio and podcasts, all free. Listen online or download the iHeart App.

Connect

© 2025 iHeartMedia, Inc.