June 30, 2024 • 50 mins
Dr. Joseph Mailman is an intensive care physician at New York-Presbyterian Weill Cornell with a keen interest in technological innovation. In this lecture, he discusses the field of medical informatics and […]
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
It's my pleasure today to introduce, a friend
of mine, doctor Joe Mailman.
Doctor Mailman is an assistant professor of medicine
in the division of pulmonary and critical care
medicine at Cornell,
and he is gonna be talking today about
something that I know nothing about. And so
I'm really excited to hear what he has
to say about it. But he called this
hot topics in medicine, and this is some
about medical informatics.
(00:21):
And so I'm hoping you can educate us,
on this topic,
and we are really happy to have you
here. So thank you for joining us.
Yeah. Thanks so much.
I like to give this presentation because I
gave this for our department of medicine grand
rounds a few months ago and have subsequently
gone on the road with it because it's
actually become sort of my career path.
(00:42):
I think it's a very under
something is not known about to the general
physician
body,
and it's a great field. And I I
do this to make sure that people know
that this is a really great career option
and, you know, that there are more things
out there than just, you know, physician scientist
or clinical educator. So there's a lot of
(01:05):
things that people can do, and this is
one, and I I just like to talk
about it. A couple of the other things
that I do that tie into this is
one of the,
IMR divisions director of clinical informatics.
I,
am a New York Presbyterian clinical documentation physician
adviser,
and that ties into one of the projects
that I'll show you about.
(01:25):
And then
I completed a program,
that we have that includes classes at Columbia
Business School,
called the NYP Lead Academy, which kind of
escalated and launched
this huge project that I'll tell you about.
And through that, I also became,
what's called a certified Epic physician builder.
So I have super special powers in our
(01:46):
Epic system to design a lot of these
projects.
And so that's how I got involved, and,
and, you know, you'll see that throughout the
talk. I do have one,
disclosure, and that is I do do some
consulting for a company called Masimo,
which you may know, is in a battle
with Apple,
right now. But, I do consult with them.
(02:09):
So,
medical,
let me just I'm gonna give you a
background, and I'm gonna talk about some projects.
And, if if it's not too much over,
I I'll talk there's 4 there's, like, another
sort of health care, integration
that I have, but if it's you would
rather have time for questions, I can stop
before I do that. So I'll I'll ask
when I get to that point.
(02:30):
So medical informatics is this transdisciplinary
field, and it it studies and develops
using, you know, biomedical data,
IT
innovations,
knowledge, sci and and we're looking for scientific
inquiry,
problem solving, decision making, and the goal is
to ultimately improve human health,
(02:51):
you know, but through technology.
So there is a fundamental theorem of this,
and it it is elucidated by this man,
doctor Charles p Friedman.
And this person said that the goal of
informatics,
is to create an environment
of supported practice such as an intelligent person
(03:12):
is working in combination with information resources
and technology is better than the person without
such support. So in this case, here's your
person. Here's your resource. They they are in
the environment, which is the parentheses, and this
is supposed to be better than just this.
It it it's not saying that the computer
(03:32):
is better than the person. You need the
person plus the computer,
and that
synergy is better.
There are a couple of corollaries.
It's it's designed to be more about people
than technology.
It for this theorem to hold,
the resource has to offer something that a
person does not already know.
(03:54):
And whether this theorem holds depends on an
interaction between the person
and the resource, and you cannot predict the
results of that interaction in advance.
So just remember, it's not this. Okay?
But it is this. Informatics is
essentially people plus processes plus technology.
(04:18):
And, you know, every day,
people do benefit from informaticians.
I'm still getting used to calling myself that.
I I do got this way through experience
that one might say,
and, ability to accelerate, you know, health care's
transformation by collecting, analyzing, and applying data directly
to, care decisions.
(04:38):
And this is actually a graphic from,
the University of Texas at Arlington, and they
I they there's a lot of places that
have, you know, informatics departments and fellowships, and
and they have this multi interprofessional
center for health informatics. And and overall, it's
like a cycle. Right? You're looking at data
and information, and you're seeing how it impacts
(04:58):
your decisions.
But their belief is that with health informatics
from all this cycle, you can get better
information and make better decisions and improve, you
know, the health of people, ultimately.
And you really need to think about this
as an intersection between the, you know, the
work of the stakeholders across health and health
care delivery systems.
(05:20):
And the goal is you know, everyone has
many goals. Right? So one is some people
have the goal to improve outcomes. Some people
have a goal to improve cost. Others have
goals to improve safety
and, you know, use high quality services. But
it is frequently confused with data science and
big data
and health information management,
data analytics.
But, 1,
(05:41):
we want you to think of this as
the overarching study that pulls all of those
things together. It's it's a and and the
overarching field that is focused using, you know,
data and data science and all these different
things to improve health and health care.
And what I basically do is something called
clinical informatics, and it's
(06:03):
most relevant to,
clinicians and general, you know, medical practitioners. And
it's the interprofessional
practice that
blends medical practice with information technology and behavioral
management. So, you know, you're hearing, like, we're
looking at behavior and, you know, using computers
to do so.
It also,
has it's impractical because it improves patient outcomes.
(06:26):
It advances research
and increases the value of our delivery.
But,
a key goal is to understand that it
it
it successful evolution of health care is determined
not by the technical capability,
but how effectively we're using the technology that
we're designing and how well we're integrating into
our cultures and our processes and our institutional
(06:49):
workflows. And so we have a lot, but
are we using it well?
And it's been around a long time. It
but it really exploded with the Internet and,
you know, became achieving, like, a widespread consideration
application outside of the academic field. But prior
to this, it was mostly in academics.
And what happened was in the United States,
(07:11):
it was sort of driven by federal laws
that
really incentivize
health care, you know, information technology adoption,
because the belief at that time was that
this was gonna be the solution to the
soaring health care costs.
Whether it's not there is or it is
not, you know, probably, we don't know yet,
but it that that was why it accelerated
(07:32):
so fast, and it's cool.
So
there are a lot of,
opportunities
and organizations that people can get
involved in,
and, there's opportunities for board certification as well.
If anyone's interested and they have been doing
things that they may not realize are informatics,
there's an opportunity to get board certified for
(07:54):
the next 2 years,
without taking a fellowship. And that's through the
American Board of Preventive Medicine, actually.
The,
AMIA is sort of like the, you know,
chest of medical informatics. Right? It's the American
Medical Informatics Association. They have a conference, and
they do they provide the modules, the learning
(08:14):
modules, if you wanna take the boards for
it.
They HIMSS is a health care information society
system or something like that, and they they
also have a very, very large conference.
But I felt like I actually attended that
one, and I it seems very much focused
on more, like, business. Aimia seems more focused
on clinical and, medical practice.
(08:37):
So there's opportunities to do all those things.
So a little bit of background about how,
you know, I got involved in all this,
starting back in 2010,
to 12. I was in medical school, and
we had this. You know? Some of you
might recommend is the paper chart. Maybe Andy
also recognizes it. And this if anyone's familiar
(08:58):
I don't know if you have a do
you have a VA in Maryland?
Yeah.
So this is a this is CPRS,
and they it was a it's a computer
version of computerized orders and paper charting.
And that is, you know, that's what we
had. Right? We did it. We had it.
It was, like, clunky. You were writing as
fast as you could,
(09:20):
but we were advancing at that time. So
in 2012,
I went on to residency,
and we had
something called Cerner PowerChart, which is probably the
other largest competitor to Epic.
But it it
you know, interestingly, my residency program decided to,
(09:42):
that they were switching to paper notes when
we started from paper notes when my class
started as interns, but on July 9th.
So the 1st 9 days, we had to
write paper notes because they didn't wanna have
the new interns and then the, you know,
paper notes be computerized on the 1st day
they had new interns too. So they they
waited 9 days to do that. So that
was kind of interesting.
(10:04):
It it was not bad. It, you know,
it works. It it has its powerful tool.
But around 2015, I went to a fellowship
in, Rochester. I probably met Andy in 2015
or so,
and it was my first introduction to EPYC.
And this you know, I asked her before
be of you use EPYC because it's very
important for the rest of the presentation.
(10:25):
And what I noted was that it ran,
you know, quite smoothly.
It was fast. It didn't really crash. And
and that's where I sort of got it
interesting because you could start doing things. Like,
you could design phrases and customize things, and
so I started working on things like that.
And
then I took my first attending job here
at Cornell in 2018, and we had a
(10:46):
little back step. So we had to go
and I started here. We were on something
called Allscripts.
And I was promised that we were gonna
switch to epic, you know, in 6 months
and whatever that it happened eventually, but it
wasn't 6 months. But, anyway so I actually
brought all my smart phrases I had created
for as a fellow,
(11:06):
with me.
And, you know,
we had this sort of hybrid system where
our inpatient
operations were on Allscripts, and we have 10
campuses, but I could only see one of
the other campuses
even though we're all on the same system.
And our outpatient operation at Cornell was on
(11:27):
Epic, and outpatient operation at Columbia was on
something called CROWN.
So you couldn't see anything in this huge
health huge, you know, health care organization that
is supposed to be integrated.
And
so, it was very apparent that we had
inefficiencies,
and, you know, lack of communication between all
of our sites and campuses was
very, very problematic to patient care. So in
(11:50):
2020 is actually when we started our epic
conversion,
and we did go live we did 4
go lives. We did 2 to 3 campuses
at a time.
And I was our lead representative for this
at our campus in our division.
You know, there were a lot of meetings,
and we did planning sessions about how we're
gonna convert over and all of these things.
And we, you know, recruited people to be
(12:11):
super users. And,
we had a SWAT team where I went
around to the other hospitals and, like, made
sure that the super users knew what they
were doing and checked in them and reported
all the issues that the epic corporate people
that were here to help us with the
conversion.
And I was able to start converting the
old smart phrases, and and we continue to
this day to work on what we call
(12:32):
optimization
projects.
And after that,
I
a couple of things happened. So I became,
this
clinical documentation physician adviser,
and I started working with hospital
automation, and I took that program that I
mentioned called NYP Lita Academy.
And, you know, we had a project
(12:54):
that was designed to improve our mortality numbers,
which I'm gonna talk about. It's one of
the big projects.
And and this really is where I learned
how to probably get my really foot in
the informatics field and and design some of
these things that I'm gonna show you today.
So that's my path.
And I'm gonna tell you about 4 projects
(13:14):
and how you can think about,
you know,
informatics in ways you don't even realize you're
doing it and things that you can think
that maybe would be interesting for your organization.
So I'm gonna say a few terms. So
one is QPS, which probably all familiar is
quality and patient safety. So we have,
usually enterprise wide,
(13:36):
which means all of our campuses,
UPS goals.
And, we have some, you know, divisional goals,
and then we have,
you know, sort of, like, our own
center based goals. But,
there are 4 projects that I wanna talk
about.
The first of which was a QPS goal
a couple of years ago for the whole
(13:58):
organization, which was to reduce,
VTEs,
you know, inpatient VTEs.
And the second one is a a divisional
goal, which is to have a bronchoscopy
complications
reporting tool.
The third one is the largest project, which
remains ongoing,
and that was to reduce
(14:19):
adult mortality
through clinical documentation improvement.
That's quite a, you know,
task.
And then, hospital automation, which if we have
time, we'll talk about that, but
we'll go
from these 3 at least.
So
in order to understand, you know, what I
(14:40):
mentioned in the next slides, I need to
show you what I mean by the epic
dictionary. So
it's a word salad, basically. This is not
even comprehensive.
There are
a lot of the things that you might
have heard, there's a smart form and a
smart data element and a smart text and
a smart tool and a smart, you know,
(15:00):
all smart block and all these smart things.
And a lot of people couldn't say, you
know, that it's just your smart your dot
phrase.
There are different meanings to all these things.
It's not really important that, you know, like,
the minutiae there, but they the smart and
link and the smart list and the smart
text are all different things.
(15:21):
So but it for your purposes, that phrase
is fine.
And then there are several environments that once
you become a builder, you work in. So
one of them is is,
you know, the production environment, which is what
everyone works in to do their regular documentation
daily practices. Everyone uses this, you know, to
(15:42):
see patients, write your notes, do everything. That's
the real live environment.
We have another one called POC, which is
a proof of concept, and that's where we
do building. That's where we create all the
stuff that we move to the production environment
or we say push to production.
SUP or SUP is a a like a
SUP and TST
test are are test environment that are 24
(16:03):
hours delayed, and they have all the actual
real patients.
But it's delayed, and you can do everything
in there, but it doesn't affect the actual
patient.
POC does not. It has, like, fake patients,
and you have to add data and all
this stuff.
And it's, you know, it's it's complicated.
And then text is is, for those of
you who are of my generation or older,
(16:26):
it is like MS DOS.
It works like MS DOS, and it's from
the original epic, creation from the 19 seventies.
And there are still certain things you have
to go into text to modify to to
push or change into the into the system.
And then I don't know if you have
monitor, but it basically, like, shows your,
your
telemetry tracing and everything on any computer you
(16:48):
want as long as it's configured.
So that's the epic dictionary.
But the the first project I was gonna
tell you about was to reduce hospital acquired
DVT and PE by 10%. So, you know,
obviously, no DVT to PE. So how does
the organization approach this? We approach it, in
a pretty structured way, which what they like
to call a 3 thinking.
(17:10):
And it's it's designed it's called a 3
thinking because it started out on a a
3 piece of paper, and they want you
to get your whole project down on an
a 3 size piece of paper. Obviously, we
do this on the computer now, but this
is, you know, what it looks like. The
goal, the background,
the process, what are the root cut, all
of the things. And you're we work as
(17:31):
a group. We have a sponsor, a process
owner, and we would all go through this
together and, like, define all these issues.
This project was an enterprise wide project, and
it had representation from, you know, all campuses
and nursing, and we had data and analytics
support.
We have something called the NYP dashboard, dashboard,
which I'll show you a little bit.
And
(17:52):
what
is
interesting about it I mean, this is this
here is the actual
portion of the a 3 that we designed,
you know, that we've submitted for the project,
and is that we wanted to use
a way to find epic to show a
lot of this information to people in more,
like, real time and easier to find. So
(18:13):
this the dashboard here that I'm showing you
is from
a website. It's like a Tableau dashboard we
call, and you have to go to website
and you have to log in, and you
can click around and look. And the top
graph here shows the total visits of a
you know, with the DVT PE.
The bottom one shows the incidence of DVT
PE for 1,000 discharges.
(18:35):
Ultimately,
you know, the everything's been going down.
My point of showing this is that it's
a powerful tool, but it
it it's very powerful. It can be sorted
at patient level. We can organize it by
unit.
We can see who administered the medication,
but it's difficult to get to. It's hard.
You don't have to log in to data.nyp.org.
You have to go find this dashboard. You
(18:57):
have to let it load. It's slow. You
know? So it's not right readily available for
people
to look at.
So what did we do? So,
you we started with start smart phrases, and
we're empowering end users by making the this
data sort of readily available. Right? So,
our group designed a custom smart phrase, which
(19:18):
is dotcurrentvtemeds.
And the way that this works is if
you type that in, it will pull in
your
med, the the med the VTE med, the
last dose and time and date that it
was given.
And this can just every time, you could
just have this in your note, and it
will just populate every time.
It uses something called,
(19:39):
CER rules, which are just the
forget CR. It stands for common edit rule,
but it it's a rule that we designed
in the background in epic
to find this information and display it for
you.
And it has something called a smart data
element,
which is a way to track it. Okay?
That's a smart data element is a tracking
tool, and a rule is a way to,
(20:00):
like, find the information and display it. So
then we said, okay. Let's great. So we
can put this in our notes. How can
we display it in other ways? So we
designed what we call a custom column, and
this custom column can be added to a
patient list. This
and it will show you exactly what that
smart phrase showed you,
you know, that for each patient.
(20:22):
And you there's a function that you should
know about called hover to discover.
You may have heard about it. So if
you hover over the here, it will show
a pop up and show it clearer. And
sometimes there'll be more information,
for other things. In this case, it's the
same, but there might be more sometimes.
So
if anyone in the organization can add this
to their list and have this readily available
(20:44):
for them.
You know?
So that was another way. So then then
we said, okay. Now we wanna get this
find patients who are not receiving chemical DVT
prophylaxis.
So we said, let's build a report that
has some logic
that looks for patients that, you know, were
not administered chemical prophylaxis
(21:06):
in the last 24 hours.
And and then it says, okay. Did they
have it ordered?
And if it says yes, then did they
refuse?
If it says no, what is the contraindication?
And is mechanical prophylaxis
ordered?
And if a yes, what is the mechanical?
What is it?
(21:26):
So
we this, we need a data analytics support
to do, but it it is a customized
report by the unit. It shows,
you know,
that there's it looks a little cleaner now
that there's an order,
but,
that the patient and it shows the last
time, you know, the patient got it. Okay?
And then in this case, these patients were
(21:49):
not receiving it because
they had, like, something else going or they
had recent major bleeding or something, and they
were on SCDs. And then this person was
not on SCDs.
You know? And so that's what it shows,
and it goes out to every unit director
every day.
And these are some of the people I
worked with for this project. So that was
the first project.
Okay.
(22:09):
The second project is,
these are not in chronological order. They're in
order of how extensive they are.
Was from our pulmonary and QI committee,
and it was to develop a sustainable and
on demand way to track bronchoscopy complication rates.
So this should be interesting for our group
here because I'm sure you do bronchoscopies.
(22:30):
So we drafted in our committee a word
document, and we said, okay. What are some
of the complications we wanna track? What are
we gonna call them?
You know? And and go from there. So
this was just a word document, and we
said, okay. Now let's actualize this and and
put it into something that we can
track.
And what is an easiest to use
(22:52):
tool to do this? So the tool that
we
or, actually, I designed in this case is
a is a
cascading
lists
and smart text,
and it can be added to any note.
So it it's it looks like this, and
it will show. So it it begins,
by building,
all of these elements individually. So we start
(23:12):
with a list,
and then we build a link, and then
we build the text where it will print
in, and we put tracking elements on it,
and then we build a report so that
we can see if these things were being
captured. So I sort of built this list,
from the
from what they gave me.
And then,
you know, this is what the the list
starts, and then you start adding things to
(23:34):
list to make it trackable and recordable. So
what do you add? So you add these
things called smart data elements. Okay? So you
have to create these things
in the background, which you'll see in a
second, and they need sort of a way
to be tracked. So we track them by
if they file on the note level
and then what the value is. So in
this case, it's the value is the name,
(23:54):
so the hematoptysis.
Right?
Let's see. Okay.
And and and the context is, like, where
it searches. So the computer is gonna look
EPIC is gonna look for a note that
has this
NYC bronch complications,
you know, hemoptysis
smart data element filed.
(24:14):
And each answer choice has a unique
smart data element. So that's how we know
that we're recording a different thing for each
thing.
And then in order, you know, to create
this in the background, you need to do
a couple this is what the smart data
elements look like, and they live in this
thing called the smart data manager, and this
is what it looks like. You know? So
I had to create all these individual ones
(24:35):
to account for every possible complication
that we,
you know,
made for our our list. Now it is
not very difficult. It's just very, you know,
repetitive.
So and then then you have to build
2 more things, which I'm not gonna go
into too much g two, but extensions and
columns. And, basically, the extensions link the smart
(24:58):
data element to the column so you can
see
the complication that you wanna see in a
report. So you do extension, column, report, run
your report. You get exactly what the person
files on the note in your report.
It's it's it's too in the weeds. I
don't wanna bore you. So here's you know,
I'm just gonna move through this. This is
the columns that we that I built,
(25:20):
and
this is how I add the extension to
the column to link the tracking element to
the column.
And then we go in
and build a report, which has the smart
data elements and the columns, you know, the
columns that we custom designed and the smart
data elements that correlate with those columns.
(25:42):
And
it will,
you know, we we in order to do
you you have to keep records. My point
here is you have to keep a record.
So we keep a record in the background.
This is literally just an Excel spreadsheet that
I keep.
I then submit this to quality assurance,
and they move it from POC
(26:03):
to production when they approve that it works
properly, and it's okay.
So what does it look like for the
end user? It looks like this.
So you type the the smart phrase.
We're and why are we NYC?
Because in EPIC, we are the New York
consortium.
So we we're dot NYC.
We try to standardize the nomenclature,
(26:26):
And it it loads this smart tool,
and then you can go through and click
whatever you want and fill out some of
the, you know, the date.
And
it
some of the lists have additional list in
them,
and then it files like this.
And I have a video that's about oh,
(26:47):
and then you run the report, and the
report looks like this. You can see I
had a lot of bad complications whenever I
was doing this this week.
Oof. Bad. Not really. Fake patients. Everyone remember
fake patients.
So and then when you click on in
the report, it shows you the actual notes.
So you can see the full note
with the report, and each column shows, like,
the bleeding the bleeding, the gray each this
(27:09):
will show grade 1 through 4. This will
say they didn't have a MOPTIS. They didn't
have hypercarbia, whatever.
So it and it you can see all
the details of the note that filed that
that data element.
So
the this has sound, which is fun, but
I probably doesn't come through on Zoom for
some reason, but I'll narrate it so you
you can see what it looks like in
(27:30):
live action.
So you type the phrase.
You click through.
(27:55):
Whatever your complication was.
(28:23):
And there you go. We have a trackable
way to keep our prompt complication data.
And these are also I I work with
a lot of the same people. So all
the projects,
you might see similar names, especially
this person. She's one of my analysts,
(28:43):
k. So this is the biggest project, and
there's a couple of, like, lead ins that
I'll tell you about when we go into
it. And this is an ongoing project for
maybe now. It's in its 2nd year. NYPD
Lead Academy is a program I mentioned.
And, our our task, I had a partner
who is does the same thing that I
do at Columbia. She's a pulmonary and critical
care doctor as well,
(29:04):
and we worked together to,
reduce adult mortality through clinical documentation integrity. So,
you know, this also started with that a
3 map,
and we had enterprise
wide representation,
and we have a lot of data analytics
support now. I've never had so much data
and analytics support in my life.
And, you know,
(29:25):
our decision on how to take on this
project was that we decided to focus on
the capability of EPYC
to bring, information more easily to providers. Right?
So but
before I start with the project, I have
to explain what clinical documentation integrity is.
So it's our belief at our campus. I'm
sure your your hospital is the same that
(29:46):
you provide very high quality and complex care
for patients. Right?
But
our experience is that, you know, the documentation
doesn't generally reflect how complex they
are. And we feel that this documentation is
about professionalism
and that we should take pride in demonstrating
it,
And it's never going away. You know, this
(30:06):
is government mandates, and it's just never going
away. But the good news is we just
need to develop some good habits early and
minor changes, and we can make a really
big difference in in the things we need
to make a big difference in. I'll show
you what they are shortly.
So we started with a a little cute
acronym that's spelled incorrectly,
purposely called CHEETAs.
(30:29):
And CHITAS stands for this. It stands for
CKD.
CKD without a stage is not sufficient. You
really need to write the stage.
Heart failure, you need to write the chronicity,
you know, and the type. Encephalopathy,
it's really actually toxic or metabolic.
Acute kidney injury,
we're not so bad on that. Type 2
MI.
(30:50):
Anemia anemia in a big problem with acute,
you know, the type and the chronicity.
Believe it or not, this is very important.
Low k doesn't cut it. You know, low
k or
high. So high NA is not good enough.
You have to write the actual hyperkalemia
word.
(31:10):
And sepsis and septic shock
or or shock in general.
But we learned from this that just, you
know, education just doesn't cut it. Right? So
we had to think really on a largest
larger scale on how to get really capture
these diagnosis.
So
if anyone knows anything about New York Presbyterian,
they love the word amazing.
You know, we used to be amazing things
(31:32):
are happening here, and now we're stay amazing.
So we came up with this, acronym called
amazing care deserves amazing recognition.
And, basically,
accurate and complete documentation allows us to document,
you know, our and facilitate clinical care and
communication
and really reflect that acuity of the patients
that we see. We have what we have
(31:52):
very high complex transplant patients here all the
time. So, you know, we we see it
every day. I'm sure you do in Maryland
too.
You know? And it
improves hospital quality metrics,
such as expected mortality and expected length of
stay. Now this is our biggest competitor,
and you can see they do this a
lot better than we did.
It also improves rankings in safety and reputation,
(32:14):
and it does affect reimbursements.
So
we created
some so if you're not familiar, there is
something called the Busyant Risk Model, which is
an external company, and it uses
data from a 1,000 academic medical centers across
this country
(32:35):
to calculate,
page pay what
diagnoses
cap
increase the patient's expected length of stay and
expected mortality.
And, you know, these were some of the
ones that came up. Right? These 6. And
these were ones that all of our campuses
had issues capturing
at relatively low rates for whatever reason.
(32:56):
The important thing to understand about this is
that the high yield diagnoses have to be
captured as present on admission.
For us, present on admission means the time
the bed request order is put in. So
the patient is in the ER and not
in, you know, bed requested for
48 hours, all of that time in the
ER counts as something that can pop up
(33:16):
as present on admission,
just so people know. So if you're if
you have this similar issue, a little trick
to to know about.
And, really, they're they they're considered comorbidities for
the patient. And the number that you capture,
the more comorbidities you capture as present on
admission,
the higher expected
length of stay and expected mortality for the
(33:38):
patient. And that's what you wanna do. You
want those numbers, the denominator, to be higher
so that your observed to expected ratio is
lower. And that number that I showed you
about NYU versus, you know, NYP
is the ratio of observed to expected. Like,
patients actually dying
versus should they have died.
(33:59):
And many of these,
were picked because we could design rules
that were, you know, quickly quickly because they
were just based on numbers. Like, electrolytes, I
can show you a number.
And then we pick things based on using
problem lists and things like that. So how
do we do this?
So initial concept started with, the something like
this where, remember, I showed you the cheetahs.
(34:21):
Right? So I designed this phrase just for
myself called dotjm reminders,
and it's a disappearing smart phrase. If you
ever wanna make something disappear, all you need
to do is put the squiggly
and then put
the space colon one squiggly.
Okay? I love doing disappearing. Anyone can do
this. You can do this in any of
your smart phrases. It's my my favorite thing
to do.
(34:42):
And then when
the note is signed, that all goes away.
It just disappears. It's like it was never
there.
So
we said, that's great, but there was a
you know, like, I had to share that
with everyone. I can't share that with 10,000
doctors. You know, that's a real problem. Right?
So we needed to evolve this idea. So
how do we do that? We so I
(35:02):
then said, okay. Let me, like, create some
smart phrases with, like in a smart list
and, you know, designed to focus on fluid
and electrolyte disorders. So fluid and electrolyte disorders
were terrible. We're so bad.
And,
that was like I couldn't hear the word
fluid and I couldn't go 5 minutes without
hearing the word fluid and electrolytes with all
my, you know, documentation
(35:23):
people that I work with. So we started
with that, and this is what it you
know, I came up with these phrases called,
like, high electrolytes,
and it just said the diagnosis, and then
it gave you options to,
what you need to do
for the patient. Because you writing the diagnosis
alone is not enough. You need to,
say what you're doing for the patient.
(35:45):
And it was good,
except that I had to share each of
these lists with everyone who wanted to try
it. So, like, this is one list. This
is another list. This is another list. This
is another list. So you have to share
4 lists
with the person. And it didn't collect any
data
and really didn't provide anyone with any information
about,
like, if this diagnosis was true. It was
(36:05):
just if you saw it, you had to
you could add it, but it didn't show
you any supporting data right in front of
your face.
So
if you didn't use it, it disappeared as
well. You know, these were some of the
options. This is what it looked like,
and it prints like this. I I have
this special black and blue note that I'm,
like, famous for here, so my plan is
(36:26):
always in blue.
That's why it's in blue.
So
we said, okay. Let's, like, build
this out more
and start using rules
and show people data and and numbers and
and make it actionable
and and, you know, something that's easy to
do and fast.
So So what do we do? So we
(36:48):
started with something that this was my colleague
is a neuro ICU PA, and she's an
incredible physician builder.
And
she designed this first. This is the first
thing we tried. It was,
a smart phrase that that's why it says,
NEU ICU,
displays,
you know, morbid obesity or obesity if the
patient's BMI qualifies that and did that automatically
(37:10):
just in your note, and you just had
to, like,
say yes, basically.
So that was our first attempt.
And then we went crazy, and we designed,
like, thousands of these things, and we call
it the quality admission smart tool, which is
accumulation of all these, you know, rules and
lists and text and everything.
(37:31):
And the way that it works is it
shows you it it pops up in every
h and p
and every initial progress note within a certain
time period so that we don't miss because
all that time in the ER counts.
So we have a certain time period. We
have, like, 2 clocks, basically, and we call
it a time wrapper.
And and it suggests diagnosis based on what
(37:53):
it finds in the chart. So it so
if it shows that the potassium is low,
it suggests hypokalemia, and you can click through
a plan.
And then, you know, you get this when
you're done. And we it it's forced into
everyone's note.
And it
and here's how it looks,
like, you know, similar in a video.
(38:15):
So you you pick h and p.
You can you can add it to your
smart you can still add your template. We
did not delete templates. This was some of
the trigger criteria that we focused on initially.
There is a ton more now.
(38:38):
And you build a plan.
And we had you know, you could say
this isn't present. If it was it shouldn't
be an error, but there are opportunities to
say it's not present. Like, more complicated diagnosis
in our numbers.
And that's how it works.
And it's it's
here's some data to show how well it
(39:00):
has worked. So we had a go live
period,
from 426
to 5:15.
In that time period, we had it was
optional. So there were, like, 2024100
uses of this. There were that was across
about 1900 encounters, and we captured
5,500
more diagnoses,
using our report that we built. Right? And
(39:21):
you can see that fluid electrolytes
was obviously captured the most, and that's what
we wanted to capture the most at that
time. And it's spread across, you know, all
of our campuses. WCM is Cornell. CUIMC is
Columbia. BMH is Brooklyn Method. You know, whatever.
All the campuses. Right?
So it was showing that, yes, we have
this tool. It's capturing more diagnoses,
(39:41):
but we don't it basically proved that the
tool did what it was supposed to. It
was supposed to file a diagnosis, and it
did. But that that doesn't tell us whether
it's actually helping our
numbers or anything. Right? So how did how
do we do with that?
Before I show you that, I'll explain is
we created that this is something not done
(40:02):
by me. This is done by our analytics
team, but this is what they call a
funnel chart. You don't really need to worry
about the top three columns. The bottom 3
are the most important. And what this shows
is that back then,
the
note, every single HMP note, because we're only
doing HMP notes that time, was populating
60% of the HMP notes were populating with
(40:22):
our template that included all of those diagnoses.
And then only about 45 of the top
45%
of that
60% did they have a trigger, meaning, like,
they had, you know, hyperkalemia
or something.
And people are only answering it 20% of
the time.
So we're like, oh, darn. This isn't a
very effective tool.
(40:43):
The analytics team created another
map, which looks crazy, but I'm gonna explain
it. So it's called a tree map.
And, basically or you might think of it
as, like, a heat map. What it is
is this version
shows
how many
how much the template is being populated. So
in the blue, it's the bigger the more
(41:05):
blue and the bigger, the more it's populating.
So you could see in medicine,
it was populating a lot. But in surgical
services, it was basically not populating. Right? So
we said, okay. We have to see what
is the problem with surgical services
and why it's not populating as surgical services.
This was for across all of our campuses.
It was not separated out by campus at
(41:26):
that time.
So they created a correlating map that says,
hey.
Like, how many people are deleting that tool?
So you could see medicine. Remember, in population,
it was very blue, but a lot of
people were deleting it. And then the other
services are blue because it wasn't populating, so
they weren't deleting it.
So
(41:48):
we had to do a few things to
make sure that compliance
improved. So we added
something called a smart block,
which basically is an information resource
that pops up when the tool fires. It
shows you all the criteria,
but it makes it undeletable.
So people cannot delete the tool anymore.
(42:09):
And then people had aesthetic issues with it
because it files at the bottom of your
note. So we designed a customized
header that designs that collapses this. So it
just says quality review, and you can click
the down arrows to expand
it. Because people are like, I don't like
these things that I know. It's not relevant.
You know? All these different things.
(42:29):
And then we,
designed a progress note version for the initial
progress notes.
After we
implemented all of those things and we looked
at some of the surgical services and found
that there were,
sub templates that the profile sub profiles we
that we were missing and added the template
(42:52):
to the sub profiles,
we can see that the template populates nearly
a 100% of the time with 80% triggers
present and
almost the same, you know, answering. People are
answering in 80 basically, a 100% of the
time. Okay? And that's true today as well.
This is a little bit older, but it
hasn't really changed.
We're we could never really get up to,
(43:12):
like, a 100% triggers present, but
some of the patients don't have them.
So it's much improved with all those those
tools we've added.
And you can see that we changed the
tree map so that it's sorted by campus
and by service and by type of provider.
So I can look I mean, you can
look down on any of these things to
the person level, the actual
(43:34):
person doing it and see, like, how well
they're using the tool. So it's quite it's
very interactive. This is a very static version
of it, but you click and the box
gets bigger, and then you click here, and
then the box gets bigger. And it's it's
pretty cool. It's very impressive
work. It's so,
big that, apparently, it crashes the analyst's computer
that's designed it every time she runs it.
(43:54):
So it's so they only run it about
once a month now.
So what you know, next steps for this
project are,
you know, we're we're able to capture this
usage
and and and,
deletion rate and diagnosis,
but, we're working to capture
really,
like, the Visiyan metrics, right, and and our
(44:16):
coding data. And so we have DAO design
coding, and we're expanding. We're expanding it and,
you know, going forward with that. But what
I can tell you is that the coding
data has is able to be captured on
a subsequent report, and everything has gone up
on coding. And then the Visian reports have
gone up. So everything is trending up since
the implementation of this cool this tool.
(44:38):
But our o to e ratio is going
down, which is exactly what we wanted.
And, you know, there are still issues, and
people still have many
problems with it, but it's it's a fairly
useful tool. And now that people have had
it for about almost a year, they're they're
they don't it doesn't bother them anymore. It's
it's not that,
cumbersome.
(44:59):
And this was a, like, a epic learning,
elearning that we had designed that you goes
through it that's available at our campuses.
And, again, the same people.
You know, we need to do,
ICUs need to adapt, right, for the future.
We need to have
(45:20):
new more services. We need to adjust the
economic pressures. We need innovation.
And here at Cornell, especially, we're doing a
a new device integration
project in hospital automation.
So
you might recognize this. This might look like
one of your rooms. Right? We have all
these devices with the monitor. We have the
pumps. We have high flow,
(45:41):
you know, bypass,
vent, whatever. Right? ECMO sometimes,
dialysis.
We have these EPOC machines. I don't know
if you have these, but they're little point
of care, like, blood glucose blood gas machines.
They're in every room in the ICU, which
is wonderful.
And, they're pretty cool.
And then,
a lot of cords and wires.
(46:01):
So much so if anyone remembers something called
Gomer blog, there was an article that's one
of my favorites called ICU team excited to
untangle lines after patient dropped off from OR.
But,
you know, we need a state that's more
like a seamless communication, right, a crosstalk.
But right now, we're just doing flow sheets
and typing, and it looks like this. Right?
(46:24):
Nurses are sitting looking at these flow sheets
all day long. Right? Writing soap notes or
whatever. And she looks happy, but she's probably
not.
And I looked down into an admission. You
know, I looked into some patients here,
and I found that, 15 hours into an
admission, the patient there was a patient had
15 notes.
A 2 day admission had 38 notes.
(46:44):
The 13 day admission had 233
notes,
and a 2 month admission had 700 notes.
One admission note was 1580
words,
which is 10,407
characters.
And, ultimately, like, instead of working for,
our technology, we really should have the technology
work for us, I think.
(47:06):
So how can we do that? You know,
we need to talk about what this cost
too. So we everything is becoming more extensive.
Right? I I don't I could not find
more recent information, but just a day in
IC with mechanical ventilation back,
you know, 20 years ago almost cost,
like, $11,000.
And medical error is very expensive. Right? 20,000,000,000.
(47:28):
It's very costly in deaths.
And
this is the biggest problem. Right? This is
the biggest problem, and all that documentation
is causing this problem.
Okay?
5,000,000,000
annually in health care spending.
Physician turnover because of this is this is
more than this, for sure. But it it
(47:48):
was estimated for a number that I found,
500,000
per physician turnover and about 50,000 per nurse
turnover.
It just makes people very unhappy.
And so between, you know, 2,021,010,
critical care medicine
costs went up 92%
from 56,000,000,000
to a 108,000,000,000
just for critical care costs. And, you know,
(48:10):
that represent about 13%
of hospital costs or 4.1%
of national health care expenditures,
which is
almost 1% of the country's gross domestic product.
Critical
care medicine costs.
So, you know, COVID came. It showed that
we have cracks in the health care system,
(48:32):
but it actually showed that we could do
things rather quickly. And, we went from, like,
2%
of telemedicine to, like, 99% in about 2
days here,
you know, towards hospital automation.
And it showed that the benefits of, you
know, remote monitoring and wireless technology as well.
So we believe that this hospital automation may
cut cost and improve status satisfaction
(48:55):
and audit you know, is to automate patient
care by using intelligent medical,
integrating them. Right? So and letting doctors and
nurses and everyone go back to take care
of the patient. So we have a pilot
that's starting that's gonna transmit our ventilator. We
already do this with the ventilators, but we're
changing companies
to a different to directly to Epic,
(49:15):
along with our CRT and, eventually, other devices.
Other devices will do this too. And this
is a pilot that we're running in the
ICU starting actually next week. It's been in
development for about 2 years, but it's gonna
start next week.
And and IV pumps soon as well. IV
pumps will could potentially also transmit automatically.
So the nurses don't have to type all
(49:36):
that stuff in.
You know, and I create them with this
little slogan. We can take our our organization
from burnout to bliss,
hopefully.
And,
hopefully, one day, we'll get an ICU that
looks as beautiful as this, and, you know,
hospital automation will
need to actually focus more than just on
the patient. It'll need to be comprehensive and
integrate patients, families, and staff. Staff is often
(49:59):
left out of this.
And, you know, we think of patients and
families, but the staff is providing the care
for the patients and the families, and we
need to do that better
to think about them and how things are
designed. You know, maybe we can get to
the ICU of the future.
So some final thoughts are
my colleagues that worked on this project
(50:20):
are that medical informatics is rapidly evolving, and
it's just intertwined with our practice and and
continuously advancing science of medicine.
And we're just starting to scratch the surface.
So I think it's a really great enticing
career choice for maybe some of your fellows
who might be interested in this or some
of the younger attendings or anyone, actually.
(50:40):
And final word is that what is a
hot topic today will be a core topic
tomorrow,
and thank you very much.
It's my pleasure today to introduce, a friend
of mine, doctor Joe Mailman.
Doctor Mailman is an assistant professor of medicine
in the division of pulmonary and critical care
medicine at Cornell,
and he is gonna be talking today about
something that I know nothing about. And so
I'm really excited to hear what he has
to say about it. But he called this
hot topics in medicine, and this is some
about medical informatics.
(00:21):
And so I'm hoping you can educate us,
on this topic,
and we are really happy to have you
here. So thank you for joining us.
Yeah. Thanks so much.
I like to give this presentation because I
gave this for our department of medicine grand
rounds a few months ago and have subsequently
gone on the road with it because it's
actually become sort of my career path.
(00:42):
I think it's a very under
something is not known about to the general
physician
body,
and it's a great field. And I I
do this to make sure that people know
that this is a really great career option
and, you know, that there are more things
out there than just, you know, physician scientist
or clinical educator. So there's a lot of
(01:05):
things that people can do, and this is
one, and I I just like to talk
about it. A couple of the other things
that I do that tie into this is
one of the,
IMR divisions director of clinical informatics.
I,
am a New York Presbyterian clinical documentation physician
adviser,
and that ties into one of the projects
that I'll show you about.
(01:25):
And then
I completed a program,
that we have that includes classes at Columbia
Business School,
called the NYP Lead Academy, which kind of
escalated and launched
this huge project that I'll tell you about.
And through that, I also became,
what's called a certified Epic physician builder.
So I have super special powers in our
(01:46):
Epic system to design a lot of these
projects.
And so that's how I got involved, and,
and, you know, you'll see that throughout the
talk. I do have one,
disclosure, and that is I do do some
consulting for a company called Masimo,
which you may know, is in a battle
with Apple,
right now. But, I do consult with them.
(02:09):
So,
medical,
let me just I'm gonna give you a
background, and I'm gonna talk about some projects.
And, if if it's not too much over,
I I'll talk there's 4 there's, like, another
sort of health care, integration
that I have, but if it's you would
rather have time for questions, I can stop
before I do that. So I'll I'll ask
when I get to that point.
(02:30):
So medical informatics is this transdisciplinary
field, and it it studies and develops
using, you know, biomedical data,
IT
innovations,
knowledge, sci and and we're looking for scientific
inquiry,
problem solving, decision making, and the goal is
to ultimately improve human health,
(02:51):
you know, but through technology.
So there is a fundamental theorem of this,
and it it is elucidated by this man,
doctor Charles p Friedman.
And this person said that the goal of
informatics,
is to create an environment
of supported practice such as an intelligent person
(03:12):
is working in combination with information resources
and technology is better than the person without
such support. So in this case, here's your
person. Here's your resource. They they are in
the environment, which is the parentheses, and this
is supposed to be better than just this.
It it it's not saying that the computer
(03:32):
is better than the person. You need the
person plus the computer,
and that
synergy is better.
There are a couple of corollaries.
It's it's designed to be more about people
than technology.
It for this theorem to hold,
the resource has to offer something that a
person does not already know.
(03:54):
And whether this theorem holds depends on an
interaction between the person
and the resource, and you cannot predict the
results of that interaction in advance.
So just remember, it's not this. Okay?
But it is this. Informatics is
essentially people plus processes plus technology.
(04:18):
And, you know, every day,
people do benefit from informaticians.
I'm still getting used to calling myself that.
I I do got this way through experience
that one might say,
and, ability to accelerate, you know, health care's
transformation by collecting, analyzing, and applying data directly
to, care decisions.
(04:38):
And this is actually a graphic from,
the University of Texas at Arlington, and they
I they there's a lot of places that
have, you know, informatics departments and fellowships, and
and they have this multi interprofessional
center for health informatics. And and overall, it's
like a cycle. Right? You're looking at data
and information, and you're seeing how it impacts
(04:58):
your decisions.
But their belief is that with health informatics
from all this cycle, you can get better
information and make better decisions and improve, you
know, the health of people, ultimately.
And you really need to think about this
as an intersection between the, you know, the
work of the stakeholders across health and health
care delivery systems.
(05:20):
And the goal is you know, everyone has
many goals. Right? So one is some people
have the goal to improve outcomes. Some people
have a goal to improve cost. Others have
goals to improve safety
and, you know, use high quality services. But
it is frequently confused with data science and
big data
and health information management,
data analytics.
But, 1,
(05:41):
we want you to think of this as
the overarching study that pulls all of those
things together. It's it's a and and the
overarching field that is focused using, you know,
data and data science and all these different
things to improve health and health care.
And what I basically do is something called
clinical informatics, and it's
(06:03):
most relevant to,
clinicians and general, you know, medical practitioners. And
it's the interprofessional
practice that
blends medical practice with information technology and behavioral
management. So, you know, you're hearing, like, we're
looking at behavior and, you know, using computers
to do so.
It also,
has it's impractical because it improves patient outcomes.
(06:26):
It advances research
and increases the value of our delivery.
But,
a key goal is to understand that it
it
it successful evolution of health care is determined
not by the technical capability,
but how effectively we're using the technology that
we're designing and how well we're integrating into
our cultures and our processes and our institutional
(06:49):
workflows. And so we have a lot, but
are we using it well?
And it's been around a long time. It
but it really exploded with the Internet and,
you know, became achieving, like, a widespread consideration
application outside of the academic field. But prior
to this, it was mostly in academics.
And what happened was in the United States,
(07:11):
it was sort of driven by federal laws
that
really incentivize
health care, you know, information technology adoption,
because the belief at that time was that
this was gonna be the solution to the
soaring health care costs.
Whether it's not there is or it is
not, you know, probably, we don't know yet,
but it that that was why it accelerated
(07:32):
so fast, and it's cool.
So
there are a lot of,
opportunities
and organizations that people can get
involved in,
and, there's opportunities for board certification as well.
If anyone's interested and they have been doing
things that they may not realize are informatics,
there's an opportunity to get board certified for
(07:54):
the next 2 years,
without taking a fellowship. And that's through the
American Board of Preventive Medicine, actually.
The,
AMIA is sort of like the, you know,
chest of medical informatics. Right? It's the American
Medical Informatics Association. They have a conference, and
they do they provide the modules, the learning
(08:14):
modules, if you wanna take the boards for
it.
They HIMSS is a health care information society
system or something like that, and they they
also have a very, very large conference.
But I felt like I actually attended that
one, and I it seems very much focused
on more, like, business. Aimia seems more focused
on clinical and, medical practice.
(08:37):
So there's opportunities to do all those things.
So a little bit of background about how,
you know, I got involved in all this,
starting back in 2010,
to 12. I was in medical school, and
we had this. You know? Some of you
might recommend is the paper chart. Maybe Andy
also recognizes it. And this if anyone's familiar
(08:58):
I don't know if you have a do
you have a VA in Maryland?
Yeah.
So this is a this is CPRS,
and they it was a it's a computer
version of computerized orders and paper charting.
And that is, you know, that's what we
had. Right? We did it. We had it.
It was, like, clunky. You were writing as
fast as you could,
(09:20):
but we were advancing at that time. So
in 2012,
I went on to residency,
and we had
something called Cerner PowerChart, which is probably the
other largest competitor to Epic.
But it it
you know, interestingly, my residency program decided to,
(09:42):
that they were switching to paper notes when
we started from paper notes when my class
started as interns, but on July 9th.
So the 1st 9 days, we had to
write paper notes because they didn't wanna have
the new interns and then the, you know,
paper notes be computerized on the 1st day
they had new interns too. So they they
waited 9 days to do that. So that
was kind of interesting.
(10:04):
It it was not bad. It, you know,
it works. It it has its powerful tool.
But around 2015, I went to a fellowship
in, Rochester. I probably met Andy in 2015
or so,
and it was my first introduction to EPYC.
And this you know, I asked her before
be of you use EPYC because it's very
important for the rest of the presentation.
(10:25):
And what I noted was that it ran,
you know, quite smoothly.
It was fast. It didn't really crash. And
and that's where I sort of got it
interesting because you could start doing things. Like,
you could design phrases and customize things, and
so I started working on things like that.
And
then I took my first attending job here
at Cornell in 2018, and we had a
(10:46):
little back step. So we had to go
and I started here. We were on something
called Allscripts.
And I was promised that we were gonna
switch to epic, you know, in 6 months
and whatever that it happened eventually, but it
wasn't 6 months. But, anyway so I actually
brought all my smart phrases I had created
for as a fellow,
(11:06):
with me.
And, you know,
we had this sort of hybrid system where
our inpatient
operations were on Allscripts, and we have 10
campuses, but I could only see one of
the other campuses
even though we're all on the same system.
And our outpatient operation at Cornell was on
(11:27):
Epic, and outpatient operation at Columbia was on
something called CROWN.
So you couldn't see anything in this huge
health huge, you know, health care organization that
is supposed to be integrated.
And
so, it was very apparent that we had
inefficiencies,
and, you know, lack of communication between all
of our sites and campuses was
very, very problematic to patient care. So in
(11:50):
2020 is actually when we started our epic
conversion,
and we did go live we did 4
go lives. We did 2 to 3 campuses
at a time.
And I was our lead representative for this
at our campus in our division.
You know, there were a lot of meetings,
and we did planning sessions about how we're
gonna convert over and all of these things.
And we, you know, recruited people to be
(12:11):
super users. And,
we had a SWAT team where I went
around to the other hospitals and, like, made
sure that the super users knew what they
were doing and checked in them and reported
all the issues that the epic corporate people
that were here to help us with the
conversion.
And I was able to start converting the
old smart phrases, and and we continue to
this day to work on what we call
(12:32):
optimization
projects.
And after that,
I
a couple of things happened. So I became,
this
clinical documentation physician adviser,
and I started working with hospital
automation, and I took that program that I
mentioned called NYP Lita Academy.
And, you know, we had a project
(12:54):
that was designed to improve our mortality numbers,
which I'm gonna talk about. It's one of
the big projects.
And and this really is where I learned
how to probably get my really foot in
the informatics field and and design some of
these things that I'm gonna show you today.
So that's my path.
And I'm gonna tell you about 4 projects
(13:14):
and how you can think about,
you know,
informatics in ways you don't even realize you're
doing it and things that you can think
that maybe would be interesting for your organization.
So I'm gonna say a few terms. So
one is QPS, which probably all familiar is
quality and patient safety. So we have,
usually enterprise wide,
(13:36):
which means all of our campuses,
UPS goals.
And, we have some, you know, divisional goals,
and then we have,
you know, sort of, like, our own
center based goals. But,
there are 4 projects that I wanna talk
about.
The first of which was a QPS goal
a couple of years ago for the whole
(13:58):
organization, which was to reduce,
VTEs,
you know, inpatient VTEs.
And the second one is a a divisional
goal, which is to have a bronchoscopy
complications
reporting tool.
The third one is the largest project, which
remains ongoing,
and that was to reduce
(14:19):
adult mortality
through clinical documentation improvement.
That's quite a, you know,
task.
And then, hospital automation, which if we have
time, we'll talk about that, but
we'll go
from these 3 at least.
So
in order to understand, you know, what I
(14:40):
mentioned in the next slides, I need to
show you what I mean by the epic
dictionary. So
it's a word salad, basically. This is not
even comprehensive.
There are
a lot of the things that you might
have heard, there's a smart form and a
smart data element and a smart text and
a smart tool and a smart, you know,
(15:00):
all smart block and all these smart things.
And a lot of people couldn't say, you
know, that it's just your smart your dot
phrase.
There are different meanings to all these things.
It's not really important that, you know, like,
the minutiae there, but they the smart and
link and the smart list and the smart
text are all different things.
(15:21):
So but it for your purposes, that phrase
is fine.
And then there are several environments that once
you become a builder, you work in. So
one of them is is,
you know, the production environment, which is what
everyone works in to do their regular documentation
daily practices. Everyone uses this, you know, to
(15:42):
see patients, write your notes, do everything. That's
the real live environment.
We have another one called POC, which is
a proof of concept, and that's where we
do building. That's where we create all the
stuff that we move to the production environment
or we say push to production.
SUP or SUP is a a like a
SUP and TST
test are are test environment that are 24
(16:03):
hours delayed, and they have all the actual
real patients.
But it's delayed, and you can do everything
in there, but it doesn't affect the actual
patient.
POC does not. It has, like, fake patients,
and you have to add data and all
this stuff.
And it's, you know, it's it's complicated.
And then text is is, for those of
you who are of my generation or older,
(16:26):
it is like MS DOS.
It works like MS DOS, and it's from
the original epic, creation from the 19 seventies.
And there are still certain things you have
to go into text to modify to to
push or change into the into the system.
And then I don't know if you have
monitor, but it basically, like, shows your,
your
telemetry tracing and everything on any computer you
(16:48):
want as long as it's configured.
So that's the epic dictionary.
But the the first project I was gonna
tell you about was to reduce hospital acquired
DVT and PE by 10%. So, you know,
obviously, no DVT to PE. So how does
the organization approach this? We approach it, in
a pretty structured way, which what they like
to call a 3 thinking.
(17:10):
And it's it's designed it's called a 3
thinking because it started out on a a
3 piece of paper, and they want you
to get your whole project down on an
a 3 size piece of paper. Obviously, we
do this on the computer now, but this
is, you know, what it looks like. The
goal, the background,
the process, what are the root cut, all
of the things. And you're we work as
(17:31):
a group. We have a sponsor, a process
owner, and we would all go through this
together and, like, define all these issues.
This project was an enterprise wide project, and
it had representation from, you know, all campuses
and nursing, and we had data and analytics
support.
We have something called the NYP dashboard, dashboard,
which I'll show you a little bit.
And
(17:52):
what
is
interesting about it I mean, this is this
here is the actual
portion of the a 3 that we designed,
you know, that we've submitted for the project,
and is that we wanted to use
a way to find epic to show a
lot of this information to people in more,
like, real time and easier to find. So
(18:13):
this the dashboard here that I'm showing you
is from
a website. It's like a Tableau dashboard we
call, and you have to go to website
and you have to log in, and you
can click around and look. And the top
graph here shows the total visits of a
you know, with the DVT PE.
The bottom one shows the incidence of DVT
PE for 1,000 discharges.
(18:35):
Ultimately,
you know, the everything's been going down.
My point of showing this is that it's
a powerful tool, but it
it it's very powerful. It can be sorted
at patient level. We can organize it by
unit.
We can see who administered the medication,
but it's difficult to get to. It's hard.
You don't have to log in to data.nyp.org.
You have to go find this dashboard. You
(18:57):
have to let it load. It's slow. You
know? So it's not right readily available for
people
to look at.
So what did we do? So,
you we started with start smart phrases, and
we're empowering end users by making the this
data sort of readily available. Right? So,
our group designed a custom smart phrase, which
(19:18):
is dotcurrentvtemeds.
And the way that this works is if
you type that in, it will pull in
your
med, the the med the VTE med, the
last dose and time and date that it
was given.
And this can just every time, you could
just have this in your note, and it
will just populate every time.
It uses something called,
(19:39):
CER rules, which are just the
forget CR. It stands for common edit rule,
but it it's a rule that we designed
in the background in epic
to find this information and display it for
you.
And it has something called a smart data
element,
which is a way to track it. Okay?
That's a smart data element is a tracking
tool, and a rule is a way to,
(20:00):
like, find the information and display it. So
then we said, okay. Let's great. So we
can put this in our notes. How can
we display it in other ways? So we
designed what we call a custom column, and
this custom column can be added to a
patient list. This
and it will show you exactly what that
smart phrase showed you,
you know, that for each patient.
(20:22):
And you there's a function that you should
know about called hover to discover.
You may have heard about it. So if
you hover over the here, it will show
a pop up and show it clearer. And
sometimes there'll be more information,
for other things. In this case, it's the
same, but there might be more sometimes.
So
if anyone in the organization can add this
to their list and have this readily available
(20:44):
for them.
You know?
So that was another way. So then then
we said, okay. Now we wanna get this
find patients who are not receiving chemical DVT
prophylaxis.
So we said, let's build a report that
has some logic
that looks for patients that, you know, were
not administered chemical prophylaxis
(21:06):
in the last 24 hours.
And and then it says, okay. Did they
have it ordered?
And if it says yes, then did they
refuse?
If it says no, what is the contraindication?
And is mechanical prophylaxis
ordered?
And if a yes, what is the mechanical?
What is it?
(21:26):
So
we this, we need a data analytics support
to do, but it it is a customized
report by the unit. It shows,
you know,
that there's it looks a little cleaner now
that there's an order,
but,
that the patient and it shows the last
time, you know, the patient got it. Okay?
And then in this case, these patients were
(21:49):
not receiving it because
they had, like, something else going or they
had recent major bleeding or something, and they
were on SCDs. And then this person was
not on SCDs.
You know? And so that's what it shows,
and it goes out to every unit director
every day.
And these are some of the people I
worked with for this project. So that was
the first project.
Okay.
(22:09):
The second project is,
these are not in chronological order. They're in
order of how extensive they are.
Was from our pulmonary and QI committee,
and it was to develop a sustainable and
on demand way to track bronchoscopy complication rates.
So this should be interesting for our group
here because I'm sure you do bronchoscopies.
(22:30):
So we drafted in our committee a word
document, and we said, okay. What are some
of the complications we wanna track? What are
we gonna call them?
You know? And and go from there. So
this was just a word document, and we
said, okay. Now let's actualize this and and
put it into something that we can
track.
And what is an easiest to use
(22:52):
tool to do this? So the tool that
we
or, actually, I designed in this case is
a is a
cascading
lists
and smart text,
and it can be added to any note.
So it it's it looks like this, and
it will show. So it it begins,
by building,
all of these elements individually. So we start
(23:12):
with a list,
and then we build a link, and then
we build the text where it will print
in, and we put tracking elements on it,
and then we build a report so that
we can see if these things were being
captured. So I sort of built this list,
from the
from what they gave me.
And then,
you know, this is what the the list
starts, and then you start adding things to
(23:34):
list to make it trackable and recordable. So
what do you add? So you add these
things called smart data elements. Okay? So you
have to create these things
in the background, which you'll see in a
second, and they need sort of a way
to be tracked. So we track them by
if they file on the note level
and then what the value is. So in
this case, it's the value is the name,
(23:54):
so the hematoptysis.
Right?
Let's see. Okay.
And and and the context is, like, where
it searches. So the computer is gonna look
EPIC is gonna look for a note that
has this
NYC bronch complications,
you know, hemoptysis
smart data element filed.
(24:14):
And each answer choice has a unique
smart data element. So that's how we know
that we're recording a different thing for each
thing.
And then in order, you know, to create
this in the background, you need to do
a couple this is what the smart data
elements look like, and they live in this
thing called the smart data manager, and this
is what it looks like. You know? So
I had to create all these individual ones
(24:35):
to account for every possible complication
that we,
you know,
made for our our list. Now it is
not very difficult. It's just very, you know,
repetitive.
So and then then you have to build
2 more things, which I'm not gonna go
into too much g two, but extensions and
columns. And, basically, the extensions link the smart
(24:58):
data element to the column so you can
see
the complication that you wanna see in a
report. So you do extension, column, report, run
your report. You get exactly what the person
files on the note in your report.
It's it's it's too in the weeds. I
don't wanna bore you. So here's you know,
I'm just gonna move through this. This is
the columns that we that I built,
(25:20):
and
this is how I add the extension to
the column to link the tracking element to
the column.
And then we go in
and build a report, which has the smart
data elements and the columns, you know, the
columns that we custom designed and the smart
data elements that correlate with those columns.
(25:42):
And
it will,
you know, we we in order to do
you you have to keep records. My point
here is you have to keep a record.
So we keep a record in the background.
This is literally just an Excel spreadsheet that
I keep.
I then submit this to quality assurance,
and they move it from POC
(26:03):
to production when they approve that it works
properly, and it's okay.
So what does it look like for the
end user? It looks like this.
So you type the the smart phrase.
We're and why are we NYC?
Because in EPIC, we are the New York
consortium.
So we we're dot NYC.
We try to standardize the nomenclature,
(26:26):
And it it loads this smart tool,
and then you can go through and click
whatever you want and fill out some of
the, you know, the date.
And
it
some of the lists have additional list in
them,
and then it files like this.
And I have a video that's about oh,
(26:47):
and then you run the report, and the
report looks like this. You can see I
had a lot of bad complications whenever I
was doing this this week.
Oof. Bad. Not really. Fake patients. Everyone remember
fake patients.
So and then when you click on in
the report, it shows you the actual notes.
So you can see the full note
with the report, and each column shows, like,
the bleeding the bleeding, the gray each this
(27:09):
will show grade 1 through 4. This will
say they didn't have a MOPTIS. They didn't
have hypercarbia, whatever.
So it and it you can see all
the details of the note that filed that
that data element.
So
the this has sound, which is fun, but
I probably doesn't come through on Zoom for
some reason, but I'll narrate it so you
you can see what it looks like in
(27:30):
live action.
So you type the phrase.
You click through.
(27:55):
Whatever your complication was.
(28:23):
And there you go. We have a trackable
way to keep our prompt complication data.
And these are also I I work with
a lot of the same people. So all
the projects,
you might see similar names, especially
this person. She's one of my analysts,
(28:43):
k. So this is the biggest project, and
there's a couple of, like, lead ins that
I'll tell you about when we go into
it. And this is an ongoing project for
maybe now. It's in its 2nd year. NYPD
Lead Academy is a program I mentioned.
And, our our task, I had a partner
who is does the same thing that I
do at Columbia. She's a pulmonary and critical
care doctor as well,
(29:04):
and we worked together to,
reduce adult mortality through clinical documentation integrity. So,
you know, this also started with that a
3 map,
and we had enterprise
wide representation,
and we have a lot of data analytics
support now. I've never had so much data
and analytics support in my life.
And, you know,
(29:25):
our decision on how to take on this
project was that we decided to focus on
the capability of EPYC
to bring, information more easily to providers. Right?
So but
before I start with the project, I have
to explain what clinical documentation integrity is.
So it's our belief at our campus. I'm
sure your your hospital is the same that
(29:46):
you provide very high quality and complex care
for patients. Right?
But
our experience is that, you know, the documentation
doesn't generally reflect how complex they
are. And we feel that this documentation is
about professionalism
and that we should take pride in demonstrating
it,
And it's never going away. You know, this
(30:06):
is government mandates, and it's just never going
away. But the good news is we just
need to develop some good habits early and
minor changes, and we can make a really
big difference in in the things we need
to make a big difference in. I'll show
you what they are shortly.
So we started with a a little cute
acronym that's spelled incorrectly,
purposely called CHEETAs.
(30:29):
And CHITAS stands for this. It stands for
CKD.
CKD without a stage is not sufficient. You
really need to write the stage.
Heart failure, you need to write the chronicity,
you know, and the type. Encephalopathy,
it's really actually toxic or metabolic.
Acute kidney injury,
we're not so bad on that. Type 2
MI.
(30:50):
Anemia anemia in a big problem with acute,
you know, the type and the chronicity.
Believe it or not, this is very important.
Low k doesn't cut it. You know, low
k or
high. So high NA is not good enough.
You have to write the actual hyperkalemia
word.
(31:10):
And sepsis and septic shock
or or shock in general.
But we learned from this that just, you
know, education just doesn't cut it. Right? So
we had to think really on a largest
larger scale on how to get really capture
these diagnosis.
So
if anyone knows anything about New York Presbyterian,
they love the word amazing.
You know, we used to be amazing things
(31:32):
are happening here, and now we're stay amazing.
So we came up with this, acronym called
amazing care deserves amazing recognition.
And, basically,
accurate and complete documentation allows us to document,
you know, our and facilitate clinical care and
communication
and really reflect that acuity of the patients
that we see. We have what we have
(31:52):
very high complex transplant patients here all the
time. So, you know, we we see it
every day. I'm sure you do in Maryland
too.
You know? And it
improves hospital quality metrics,
such as expected mortality and expected length of
stay. Now this is our biggest competitor,
and you can see they do this a
lot better than we did.
It also improves rankings in safety and reputation,
(32:14):
and it does affect reimbursements.
So
we created
some so if you're not familiar, there is
something called the Busyant Risk Model, which is
an external company, and it uses
data from a 1,000 academic medical centers across
this country
(32:35):
to calculate,
page pay what
diagnoses
cap
increase the patient's expected length of stay and
expected mortality.
And, you know, these were some of the
ones that came up. Right? These 6. And
these were ones that all of our campuses
had issues capturing
at relatively low rates for whatever reason.
(32:56):
The important thing to understand about this is
that the high yield diagnoses have to be
captured as present on admission.
For us, present on admission means the time
the bed request order is put in. So
the patient is in the ER and not
in, you know, bed requested for
48 hours, all of that time in the
ER counts as something that can pop up
(33:16):
as present on admission,
just so people know. So if you're if
you have this similar issue, a little trick
to to know about.
And, really, they're they they're considered comorbidities for
the patient. And the number that you capture,
the more comorbidities you capture as present on
admission,
the higher expected
length of stay and expected mortality for the
(33:38):
patient. And that's what you wanna do. You
want those numbers, the denominator, to be higher
so that your observed to expected ratio is
lower. And that number that I showed you
about NYU versus, you know, NYP
is the ratio of observed to expected. Like,
patients actually dying
versus should they have died.
(33:59):
And many of these,
were picked because we could design rules
that were, you know, quickly quickly because they
were just based on numbers. Like, electrolytes, I
can show you a number.
And then we pick things based on using
problem lists and things like that. So how
do we do this?
So initial concept started with, the something like
this where, remember, I showed you the cheetahs.
(34:21):
Right? So I designed this phrase just for
myself called dotjm reminders,
and it's a disappearing smart phrase. If you
ever wanna make something disappear, all you need
to do is put the squiggly
and then put
the space colon one squiggly.
Okay? I love doing disappearing. Anyone can do
this. You can do this in any of
your smart phrases. It's my my favorite thing
to do.
(34:42):
And then when
the note is signed, that all goes away.
It just disappears. It's like it was never
there.
So
we said, that's great, but there was a
you know, like, I had to share that
with everyone. I can't share that with 10,000
doctors. You know, that's a real problem. Right?
So we needed to evolve this idea. So
how do we do that? We so I
(35:02):
then said, okay. Let me, like, create some
smart phrases with, like in a smart list
and, you know, designed to focus on fluid
and electrolyte disorders. So fluid and electrolyte disorders
were terrible. We're so bad.
And,
that was like I couldn't hear the word
fluid and I couldn't go 5 minutes without
hearing the word fluid and electrolytes with all
my, you know, documentation
(35:23):
people that I work with. So we started
with that, and this is what it you
know, I came up with these phrases called,
like, high electrolytes,
and it just said the diagnosis, and then
it gave you options to,
what you need to do
for the patient. Because you writing the diagnosis
alone is not enough. You need to,
say what you're doing for the patient.
(35:45):
And it was good,
except that I had to share each of
these lists with everyone who wanted to try
it. So, like, this is one list. This
is another list. This is another list. This
is another list. So you have to share
4 lists
with the person. And it didn't collect any
data
and really didn't provide anyone with any information
about,
like, if this diagnosis was true. It was
(36:05):
just if you saw it, you had to
you could add it, but it didn't show
you any supporting data right in front of
your face.
So
if you didn't use it, it disappeared as
well. You know, these were some of the
options. This is what it looked like,
and it prints like this. I I have
this special black and blue note that I'm,
like, famous for here, so my plan is
(36:26):
always in blue.
That's why it's in blue.
So
we said, okay. Let's, like, build
this out more
and start using rules
and show people data and and numbers and
and make it actionable
and and, you know, something that's easy to
do and fast.
So So what do we do? So we
(36:48):
started with something that this was my colleague
is a neuro ICU PA, and she's an
incredible physician builder.
And
she designed this first. This is the first
thing we tried. It was,
a smart phrase that that's why it says,
NEU ICU,
displays,
you know, morbid obesity or obesity if the
patient's BMI qualifies that and did that automatically
(37:10):
just in your note, and you just had
to, like,
say yes, basically.
So that was our first attempt.
And then we went crazy, and we designed,
like, thousands of these things, and we call
it the quality admission smart tool, which is
accumulation of all these, you know, rules and
lists and text and everything.
(37:31):
And the way that it works is it
shows you it it pops up in every
h and p
and every initial progress note within a certain
time period so that we don't miss because
all that time in the ER counts.
So we have a certain time period. We
have, like, 2 clocks, basically, and we call
it a time wrapper.
And and it suggests diagnosis based on what
(37:53):
it finds in the chart. So it so
if it shows that the potassium is low,
it suggests hypokalemia, and you can click through
a plan.
And then, you know, you get this when
you're done. And we it it's forced into
everyone's note.
And it
and here's how it looks,
like, you know, similar in a video.
(38:15):
So you you pick h and p.
You can you can add it to your
smart you can still add your template. We
did not delete templates. This was some of
the trigger criteria that we focused on initially.
There is a ton more now.
(38:38):
And you build a plan.
And we had you know, you could say
this isn't present. If it was it shouldn't
be an error, but there are opportunities to
say it's not present. Like, more complicated diagnosis
in our numbers.
And that's how it works.
And it's it's
here's some data to show how well it
(39:00):
has worked. So we had a go live
period,
from 426
to 5:15.
In that time period, we had it was
optional. So there were, like, 2024100
uses of this. There were that was across
about 1900 encounters, and we captured
5,500
more diagnoses,
using our report that we built. Right? And
(39:21):
you can see that fluid electrolytes
was obviously captured the most, and that's what
we wanted to capture the most at that
time. And it's spread across, you know, all
of our campuses. WCM is Cornell. CUIMC is
Columbia. BMH is Brooklyn Method. You know, whatever.
All the campuses. Right?
So it was showing that, yes, we have
this tool. It's capturing more diagnoses,
(39:41):
but we don't it basically proved that the
tool did what it was supposed to. It
was supposed to file a diagnosis, and it
did. But that that doesn't tell us whether
it's actually helping our
numbers or anything. Right? So how did how
do we do with that?
Before I show you that, I'll explain is
we created that this is something not done
(40:02):
by me. This is done by our analytics
team, but this is what they call a
funnel chart. You don't really need to worry
about the top three columns. The bottom 3
are the most important. And what this shows
is that back then,
the
note, every single HMP note, because we're only
doing HMP notes that time, was populating
60% of the HMP notes were populating with
(40:22):
our template that included all of those diagnoses.
And then only about 45 of the top
45%
of that
60% did they have a trigger, meaning, like,
they had, you know, hyperkalemia
or something.
And people are only answering it 20% of
the time.
So we're like, oh, darn. This isn't a
very effective tool.
(40:43):
The analytics team created another
map, which looks crazy, but I'm gonna explain
it. So it's called a tree map.
And, basically or you might think of it
as, like, a heat map. What it is
is this version
shows
how many
how much the template is being populated. So
in the blue, it's the bigger the more
(41:05):
blue and the bigger, the more it's populating.
So you could see in medicine,
it was populating a lot. But in surgical
services, it was basically not populating. Right? So
we said, okay. We have to see what
is the problem with surgical services
and why it's not populating as surgical services.
This was for across all of our campuses.
It was not separated out by campus at
(41:26):
that time.
So they created a correlating map that says,
hey.
Like, how many people are deleting that tool?
So you could see medicine. Remember, in population,
it was very blue, but a lot of
people were deleting it. And then the other
services are blue because it wasn't populating, so
they weren't deleting it.
So
(41:48):
we had to do a few things to
make sure that compliance
improved. So we added
something called a smart block,
which basically is an information resource
that pops up when the tool fires. It
shows you all the criteria,
but it makes it undeletable.
So people cannot delete the tool anymore.
(42:09):
And then people had aesthetic issues with it
because it files at the bottom of your
note. So we designed a customized
header that designs that collapses this. So it
just says quality review, and you can click
the down arrows to expand
it. Because people are like, I don't like
these things that I know. It's not relevant.
You know? All these different things.
(42:29):
And then we,
designed a progress note version for the initial
progress notes.
After we
implemented all of those things and we looked
at some of the surgical services and found
that there were,
sub templates that the profile sub profiles we
that we were missing and added the template
(42:52):
to the sub profiles,
we can see that the template populates nearly
a 100% of the time with 80% triggers
present and
almost the same, you know, answering. People are
answering in 80 basically, a 100% of the
time. Okay? And that's true today as well.
This is a little bit older, but it
hasn't really changed.
We're we could never really get up to,
(43:12):
like, a 100% triggers present, but
some of the patients don't have them.
So it's much improved with all those those
tools we've added.
And you can see that we changed the
tree map so that it's sorted by campus
and by service and by type of provider.
So I can look I mean, you can
look down on any of these things to
the person level, the actual
(43:34):
person doing it and see, like, how well
they're using the tool. So it's quite it's
very interactive. This is a very static version
of it, but you click and the box
gets bigger, and then you click here, and
then the box gets bigger. And it's it's
pretty cool. It's very impressive
work. It's so,
big that, apparently, it crashes the analyst's computer
that's designed it every time she runs it.
(43:54):
So it's so they only run it about
once a month now.
So what you know, next steps for this
project are,
you know, we're we're able to capture this
usage
and and and,
deletion rate and diagnosis,
but, we're working to capture
really,
like, the Visiyan metrics, right, and and our
(44:16):
coding data. And so we have DAO design
coding, and we're expanding. We're expanding it and,
you know, going forward with that. But what
I can tell you is that the coding
data has is able to be captured on
a subsequent report, and everything has gone up
on coding. And then the Visian reports have
gone up. So everything is trending up since
the implementation of this cool this tool.
(44:38):
But our o to e ratio is going
down, which is exactly what we wanted.
And, you know, there are still issues, and
people still have many
problems with it, but it's it's a fairly
useful tool. And now that people have had
it for about almost a year, they're they're
they don't it doesn't bother them anymore. It's
it's not that,
cumbersome.
(44:59):
And this was a, like, a epic learning,
elearning that we had designed that you goes
through it that's available at our campuses.
And, again, the same people.
You know, we need to do,
ICUs need to adapt, right, for the future.
We need to have
(45:20):
new more services. We need to adjust the
economic pressures. We need innovation.
And here at Cornell, especially, we're doing a
a new device integration
project in hospital automation.
So
you might recognize this. This might look like
one of your rooms. Right? We have all
these devices with the monitor. We have the
pumps. We have high flow,
(45:41):
you know, bypass,
vent, whatever. Right? ECMO sometimes,
dialysis.
We have these EPOC machines. I don't know
if you have these, but they're little point
of care, like, blood glucose blood gas machines.
They're in every room in the ICU, which
is wonderful.
And, they're pretty cool.
And then,
a lot of cords and wires.
(46:01):
So much so if anyone remembers something called
Gomer blog, there was an article that's one
of my favorites called ICU team excited to
untangle lines after patient dropped off from OR.
But,
you know, we need a state that's more
like a seamless communication, right, a crosstalk.
But right now, we're just doing flow sheets
and typing, and it looks like this. Right?
(46:24):
Nurses are sitting looking at these flow sheets
all day long. Right? Writing soap notes or
whatever. And she looks happy, but she's probably
not.
And I looked down into an admission. You
know, I looked into some patients here,
and I found that, 15 hours into an
admission, the patient there was a patient had
15 notes.
A 2 day admission had 38 notes.
(46:44):
The 13 day admission had 233
notes,
and a 2 month admission had 700 notes.
One admission note was 1580
words,
which is 10,407
characters.
And, ultimately, like, instead of working for,
our technology, we really should have the technology
work for us, I think.
(47:06):
So how can we do that? You know,
we need to talk about what this cost
too. So we everything is becoming more extensive.
Right? I I don't I could not find
more recent information, but just a day in
IC with mechanical ventilation back,
you know, 20 years ago almost cost,
like, $11,000.
And medical error is very expensive. Right? 20,000,000,000.
(47:28):
It's very costly in deaths.
And
this is the biggest problem. Right? This is
the biggest problem, and all that documentation
is causing this problem.
Okay?
5,000,000,000
annually in health care spending.
Physician turnover because of this is this is
more than this, for sure. But it it
(47:48):
was estimated for a number that I found,
500,000
per physician turnover and about 50,000 per nurse
turnover.
It just makes people very unhappy.
And so between, you know, 2,021,010,
critical care medicine
costs went up 92%
from 56,000,000,000
to a 108,000,000,000
just for critical care costs. And, you know,
(48:10):
that represent about 13%
of hospital costs or 4.1%
of national health care expenditures,
which is
almost 1% of the country's gross domestic product.
Critical
care medicine costs.
So, you know, COVID came. It showed that
we have cracks in the health care system,
(48:32):
but it actually showed that we could do
things rather quickly. And, we went from, like,
2%
of telemedicine to, like, 99% in about 2
days here,
you know, towards hospital automation.
And it showed that the benefits of, you
know, remote monitoring and wireless technology as well.
So we believe that this hospital automation may
cut cost and improve status satisfaction
(48:55):
and audit you know, is to automate patient
care by using intelligent medical,
integrating them. Right? So and letting doctors and
nurses and everyone go back to take care
of the patient. So we have a pilot
that's starting that's gonna transmit our ventilator. We
already do this with the ventilators, but we're
changing companies
to a different to directly to Epic,
(49:15):
along with our CRT and, eventually, other devices.
Other devices will do this too. And this
is a pilot that we're running in the
ICU starting actually next week. It's been in
development for about 2 years, but it's gonna
start next week.
And and IV pumps soon as well. IV
pumps will could potentially also transmit automatically.
So the nurses don't have to type all
(49:36):
that stuff in.
You know, and I create them with this
little slogan. We can take our our organization
from burnout to bliss,
hopefully.
And,
hopefully, one day, we'll get an ICU that
looks as beautiful as this, and, you know,
hospital automation will
need to actually focus more than just on
the patient. It'll need to be comprehensive and
integrate patients, families, and staff. Staff is often
(49:59):
left out of this.
And, you know, we think of patients and
families, but the staff is providing the care
for the patients and the families, and we
need to do that better
to think about them and how things are
designed. You know, maybe we can get to
the ICU of the future.
So some final thoughts are
my colleagues that worked on this project
(50:20):
are that medical informatics is rapidly evolving, and
it's just intertwined with our practice and and
continuously advancing science of medicine.
And we're just starting to scratch the surface.
So I think it's a really great enticing
career choice for maybe some of your fellows
who might be interested in this or some
of the younger attendings or anyone, actually.
(50:40):
And final word is that what is a
hot topic today will be a core topic
tomorrow,
and thank you very much.
Popular Podcasts
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.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Special Summer Offer: Exclusively on Apple Podcasts, try our Dateline Premium subscription completely free for one month! With Dateline Premium, you get every episode ad-free plus exclusive bonus content.
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!