Latest / Becker’s Healthcare Digital Health + Health IT / Dr. Don Casey, Associate Professor of Internal Medicine at Rush Medical College

Transcript

Dr. Don Casey, Associate Professor of Internal Medicine at Rush Medical College · Becker’s Healthcare Digital Health + Health IT

0:02
At athenahealth,
0:03
we know your ambulatory practice wants healthier, a
0:06
healthier business, healthier care teams, and healthier patients.
0:10
But the complexities of modern health care tech
0:12
make it hard for you and your care
0:14
teams to focus on what matters
0:16
most? That's where athenahealth can help. Our AI
0:19
native all in one solutions reduce administrative burdens,
0:22
streamline billing and payments, and deliver critical insights
0:25
when clinicians need it most. That means fewer
0:28
clicks, more time for patients, and stronger bottom
0:31
lines.
0:32
Practicing medicine is complex, but running a practice
0:35
can be that much simpler with athenahealth.
0:38
See how simpler is healthier@athenahealth.com.
0:44
Hi, everyone. This is Scott King with the
0:45
Becker's Healthcare podcast.
0:47
Brilled today to be joined by Donald Casey,
0:49
associate professor of internal medicine
0:52
at Rush Medical College. Donald, how are you
0:54
doing today? Thanks so much for joining us.
0:56
I'm good, Scott. Please call me Don.
0:58
You got it, Don. Will do.
1:00
Yeah. Appreciate you joining us. We're gonna have
1:01
a a great conversation today. A lot of
1:03
kind of big topics and trends,
1:05
in health care to discuss. But before we
1:08
get there, I was wondering if you could
1:09
just please tell us a little bit about
1:10
your background.
1:12
Sure thing, Scott. First of all, thanks for
1:14
the invitation. It's been a pleasure to be
1:16
working with Scott Becker's team,
1:19
over the years on on a lot of
1:21
great events and and a lot of great
1:22
insights, so I appreciate the opportunity to connect
1:25
with you.
1:26
I am by by training a general internist.
1:28
I trained at Rush,
1:30
in Chicago here
1:32
for first half of my career and then
1:34
later went into management, and then ultimately,
1:37
where I am now. I'm sort of a
1:39
hybrid of everything I've done in my career.
1:41
I am no longer practice internal medicine nor
1:44
am I in a health care delivery setting.
1:47
I'm really focused mainly
1:50
on the on the medical side as faculty
1:53
at Rush where I'm an associate professor of
1:55
internal medicine,
1:57
but also at the Thomas Jefferson College of
1:59
Population Health where I'm an adjunct
2:01
professor of
2:03
quality
2:04
patient safety
2:05
and also population health. And then,
2:09
I've also, for the past eleven years, been
2:11
on the faculty of the University of Minnesota
2:13
Institute for Healthcare Informatics
2:16
up in Minneapolis.
2:18
And that's been quite,
2:19
interesting,
2:20
in my career to be hanging out with
2:22
all these health care data scientists. So,
2:25
and then I'm a subject matter expert in
2:27
a number of a number of areas. I
2:28
was just in
2:30
Boston
2:32
this past, Friday with a group called ISER.
2:35
You may know about them.
2:37
Institute for Clinical and Economic Review. We're really
2:39
focused on,
2:40
as one example of, where I work, on
2:43
rational pricing for new,
2:46
drugs,
2:46
that are coming on the market,
2:48
especially those that are very expensive. So it's
2:51
it's a hybrid of a lot of things
2:53
at once, and, it's doing me well at
2:55
my point in time.
2:57
No. You you certainly are doing a lot
2:58
of things. I appreciate the the background info
3:00
and the specifics there, Don.
3:02
First topic I kinda wanted to get to
3:04
is, you know, nearly half
3:06
of medical practices reported using AI in some
3:09
capacity last year and remains a key topic
3:12
for health IT leaders.
3:13
From your perspective,
3:15
what are the use cases that are making
3:17
a difference right now
3:19
and how are you leveraging them in in
3:21
your organization?
3:23
Yeah. I I appreciate that question. It's it's
3:25
one that, I see often. And,
3:28
again, I with the disclaimer, I'm no longer
3:31
in medical practice, but I am a patient,
3:34
and I'm involved with number of health systems.
3:38
It's interesting that,
3:40
we talk about nearly half are, quote, unquote,
3:43
using AI.
3:45
And I would I would venture to say,
3:47
since AI is not a new topic, we've
3:49
been using AI, so to speak,
3:51
for years
3:53
in many, many capacities. Now in particular, I
3:56
think,
3:57
as borne out by the recent event that
3:59
I attended at, at Becker's.
4:02
Certainly,
4:03
medical practices from a business standpoint
4:06
are beginning to use a lot more
4:08
strategies
4:09
and,
4:10
and in information systems, for example, that deploy
4:14
AI techniques. And and those are, you know,
4:16
very vast in terms of,
4:18
the issues. I've said, for example, the big
4:20
one's reducing workflow burden. That's big one, you
4:23
know, using natural language processing
4:26
so that, physicians and other clinicians don't have
4:29
to spend time
4:30
standing in front of the,
4:32
electronic health record as much we hope.
4:35
So that that reducing that burden is certainly
4:38
front and center in terms of the applications
4:40
that are coming up, not just with the
4:42
large electronic health
4:44
care records, vendors, but also with, you know,
4:47
other groups that are trying to solve for
4:49
specific problems and specific,
4:51
specialties.
4:53
Obviously, revenue cycle is
4:55
a big deal too. You know, you gotta
4:57
keep the lights on, and revenue cycle
4:59
is a is a,
5:01
it's it's
5:02
it's a big game in terms of and
5:05
I'll use my that's my term. But, obviously,
5:07
with keeping up with all the rules, regulations,
5:09
and requirements of
5:11
submitting,
5:12
claims and then getting paid for them is
5:14
becoming,
5:15
I think, a big deal in terms of
5:18
just making
5:19
it easier, more accurate, better, obviously avoiding any
5:22
sort of
5:23
problems with fraud and abuse
5:25
and cybersecurity.
5:27
And then I think the the biggest part
5:30
that's exciting for me from patient care is
5:32
the access to
5:34
clinical information
5:35
in many fronts. Some of it is good
5:37
and some of it isn't so good. For
5:39
example, the alerts that come up in the
5:41
electronic health record,
5:42
we're trying to minimize
5:44
those that matter so that they don't burden
5:46
people. And yet on the other hand, the
5:48
amount of information that's much more readily available
5:52
through,
5:52
the use of, let's say, Copilot, which is
5:54
what I use. But there are others
5:57
compared to just using search strategies. I I
5:59
sort of think search is, for the most
6:01
part, unless you're trying to find a restaurant,
6:03
pretty much, out of out of favor and
6:06
that the ability to put more rational and
6:09
useful information together
6:11
on the informatics side is becoming,
6:14
very exciting for me. This obviously translates into
6:17
other issues such as clinical decision support. But
6:20
also, and and we're gonna get into this
6:22
a bit more,
6:24
Scott, is patients are using it
6:27
immensely and showing up in the in the
6:30
health care environments with lots of interesting and
6:33
useful information.
6:35
And, that's requiring the the clinicians at the
6:38
front end to
6:40
handle that with respect, but also,
6:42
to be sure that we're not getting us
6:44
going down the wrong pathway with with, let's
6:47
say, a hallucination about some strange drug that,
6:51
is used,
6:52
you know, in the woods, so to speak.
6:54
So,
6:55
we'll we'll leave it at that. I I
6:57
think the, the bottom line is it's it's
6:59
an exciting and particularly
7:01
growing,
7:03
challenge and opportunity.
7:05
Yeah. Don, that's really interesting because we we
7:08
talk obviously with a lot of systems
7:11
about using AI. But you mentioned patients coming
7:14
in and and, you know, kinda making,
7:16
the appointment or whatever, therefore, easier and more
7:18
efficient because they they have used AI and
7:20
are prepared for that meeting. What what uses
7:23
are you seeing from patients specifically? What are
7:25
what are they doing?
7:26
Well, I I I would say,
7:29
let's put it together. I mean, we have
7:31
the the doctor
7:32
slash clinician relationship that includes,
7:35
other other,
7:37
practitioners,
7:38
who may or may not have doctor in
7:40
their name. Yeah. But, you know, you can
7:42
look around,
7:43
Scott, and let's just say this has been
7:45
going on for a while. I can name
7:47
I wrote down a few specialties for which
7:50
the amount of information
7:52
requires
7:54
sort of these data science machine learning applications.
7:57
I I don't like to use the term
7:58
AI because it gets too hyped up, and
8:01
I think it sounds
8:02
to a lot of people
8:04
who aren't familiar with this like it's something
8:06
new. I mean, it actually
8:08
first was coined almost about,
8:11
seventy years ago at my alma mater at
8:13
Dartmouth, had a had a meeting. The first
8:15
that was the when the first term
8:17
artificial intelligence appeared. But if you look at
8:20
let's let's just tick off some specialties. Is
8:22
that right? Radiology.
8:25
Huge opportunities here to
8:28
look at, for example,
8:30
reading by,
8:32
a nonhuman
8:33
on top of
8:35
what radiology experts are using to enhance, not
8:38
to
8:39
not to substitute,
8:42
better insights.
8:44
Pathology, you know, that's a huge area where
8:47
the the for example, the tissues that are
8:49
being
8:50
analyzed are subject to tremendous amounts of computing
8:53
power
8:54
in terms of learning. You know, just pick
8:56
the cancer cells, for example.
8:58
Dermatology.
8:59
Right? That's another area that
9:01
is is really being used on a daily
9:03
basis. So patients can walk into their dermatologist
9:06
and have, let's say,
9:08
a problem with risk of melanoma and be
9:10
subject to cameras that will
9:12
be used using artificial intelligence to help guide,
9:16
not only the presence of melanoma, but perhaps
9:19
the risk of it. Oncology is a huge
9:21
issue,
9:23
something that I've been focused on a lot.
9:25
Cardiac electrophysiology.
9:27
I mean, you if someone goes in for
9:29
atrial fibrillation, for example,
9:31
the amount of data in these electrophysiologic,
9:35
mappings when ablation is done is huge. You
9:37
know, there's biosensing
9:39
in in the world of, cardiac cardiovascular medicine
9:43
all over the place. Ophthalmology,
9:45
you know, getting your eyes checked
9:47
with a camera that then can automatically, without
9:50
even seeing a doctor, be, subject to an
9:52
analysis that tells you if you're at risk
9:54
for, let's say, early
9:56
macular degeneration or diabetic eye disease. I mean,
9:59
these things are in practice now, Scott. So
10:01
what I'm what I'm trying to say is
10:03
that,
10:05
there are,
10:06
you know, a growing growing large number of
10:09
use cases, so to speak, from the standpoint
10:11
of patient care. Certainly on the business side,
10:14
we we talk about it a lot. But
10:16
from the standpoint of what matters in health
10:17
care, that to me is where
10:20
the new evolution and discovery are very, very
10:22
exciting.
10:25
Absolutely. And, Don, as virtual care expands from
10:29
AI enabled tools
10:30
and remote monitoring to broader digital health platforms,
10:33
introducing new technology
10:35
brings new challenges. So what advice do you
10:38
have for leaders navigating everything from governance to
10:42
patient engagement?
10:44
And
10:45
can you share an example
10:46
from your experience that's kind of balanced innovation
10:49
with operational constraints?
10:52
Well, obviously,
10:54
there's the monetary aspect of this. I mean,
10:56
you can't you don't you don't have limited
10:59
unlimited budgets like some large health systems
11:02
do. I mean, Mayo Clinic's got a whole
11:04
office building with IBM across across the way
11:07
from it.
11:08
But that's obviously,
11:10
a business strategy that is also enhancing Mayo's,
11:13
you know, supportive
11:14
care.
11:16
That's an extreme example. But from day to
11:17
day, the average health system doesn't have that
11:20
type of act,
11:21
access, so to speak.
11:23
I'd say, let's back out of that question
11:25
for a minute if I might. I mentioned
11:27
sensing.
11:29
I mean, sensing devices that patients are wearing,
11:32
and I don't like the term wearables because
11:34
not all sensing devices are wearable. But,
11:37
I'm working, for example, on
11:39
a device that can measure your blood pressure
11:41
every beat.
11:43
And, if your heart rate is is 60
11:46
per minute, that's 3,600
11:48
blood pressures taken in an hour.
11:51
So whether we like it or not and
11:53
these are this is one of
11:56
many dozens of different types of devices around
11:58
the world.
11:59
And part of it is we don't know,
12:01
you know, what's real and what isn't. The
12:03
the ability to validate these devices
12:06
you know, we're just getting started on it.
12:08
But, certainly, you're gonna need, let's say, that
12:10
we use the term artificial intelligence to sort
12:13
out
12:14
what these new devices are actually telling us.
12:17
It isn't just, you know, my blood pressure
12:19
was one forty anymore.
12:21
So I think that patients
12:23
are are
12:24
already using more and more AI in their
12:26
daily lives.
12:28
And,
12:29
you know, how how do we decide what's
12:31
what's good and what's bad? And that's that's
12:33
a an interface that, health systems have to
12:38
really start working much more hard on. This
12:40
is this is a new game. Certainly, in
12:42
in the inpatient side,
12:45
hospitals are using,
12:46
you know, machine learning
12:48
and predictive modeling and all these other
12:51
things for people with really sick,
12:55
who are really sick in, let's say, intensive
12:57
care units trying to find out who's at
12:59
risk for sepsis, etcetera. I mean, there's a
13:01
lots lots going on in that regard that
13:03
apply to the daily practice.
13:06
And with that comes responsibility again to be
13:09
sure that we're not causing harm.
13:12
And my colleague, Constantina Alferas, who's the head
13:14
of IHI, the Institute for Healthcare Informatics
13:18
at University of Minnesota, has a book out
13:20
about 450 pages saying the,
13:23
it's it's called the,
13:25
the strengths, so to speak, and the perils
13:28
of artificial intelligence. And it's it's a book
13:30
I recommend to everyone. It it's
13:33
available. I'm not pitching the book. I'm just
13:35
telling you that's 450
13:36
pages.
13:37
So we have a lot a lot a
13:39
lot of, thinking to do about
13:42
how we, go down this road.
13:45
I also think that in the health system,
13:49
to your point,
13:50
I I think that,
13:52
leaders can't just shoot from the hip and
13:54
delegate and say, wow. Here's an it's something
13:56
with AI that we're doing.
13:58
I think they have to get down in
13:59
the weeds every day
14:01
and continuously
14:03
get hands on with what's happening. Alright? Because
14:06
clinicians, as even some of the specialties I
14:09
mentioned that are relevant to health systems,
14:12
are already doing this. And so the other
14:14
the other lens is governance can no longer
14:17
be passive.
14:18
They have to show up and be, ready
14:20
to support
14:21
the health system as well. So it's a
14:23
it's a it's a hybrid of looking at
14:27
the organizational
14:28
structure
14:29
from the standpoint of who exactly is accountable.
14:33
And I think the leaders have to hold
14:35
themselves accountable as well. I mean, hospitals, as
14:37
we know, will
14:38
more and more become what I call, centers
14:41
for advanced sick care.
14:44
That's without question, not gonna change.
14:47
But as we know, you know, there are
14:49
new areas where, you know,
14:52
patients are using,
14:54
different aspects of this, including hospital at home.
14:57
That's one paradigm. It isn't the only paradigm
15:00
to get out of the hospital and not
15:02
be in there. So I think that,
15:05
from my standpoint,
15:07
the best thing they can do also is
15:08
e even within their health system is engage
15:10
the clinical experts in the specialties.
15:13
As one example that I spoke about,
15:16
with the use cases that show, and this
15:18
is the important word, Scott, impact
15:21
on population health, not just, wow, it did
15:24
x, y, and z, and someone likes us.
15:27
I mean, that's important. But on the other
15:29
hand,
15:30
there's a lot of this that's embedded in
15:32
the technology and will become,
15:34
heavier and heavier
15:36
as we go along. And so I I
15:37
do think it requires,
15:39
leaders to, in essence,
15:41
get out of the house and learn as
15:42
much as they can.
15:44
And as I said, read books like, doctor
15:46
Alifaras and find out what the what the
15:48
risks are, not just the benefits, because there
15:51
are risks.
15:52
Absolutely. And it's certainly a very
15:55
timely topic that everyone does need to kind
15:57
of stay current on education with,
16:00
of course.
16:01
How are you seeing recent legislation,
16:04
both state and federal, affect health care organizations
16:07
and health care IT
16:09
specifically.
16:10
Have you adjusted
16:11
strategies in response?
16:14
Well, I I think that the question
16:17
there there's a lot in there.
16:19
I I do know because I watch c
16:21
span a lot that certainly if you just
16:22
look at the, commerce department, there's a whole
16:25
lot centered in that one area of the
16:28
government
16:29
that's focused on this large scale
16:32
discussion
16:33
of not just AI, but the advancement of
16:37
again, I don't even like the term digital
16:39
technology
16:40
around the world. I mean, there are all
16:42
sorts of issues with cybersecurity that we can't
16:44
even imagine right now, right, as one issue.
16:48
But, you know, that's important
16:50
to pay attention to from a legislative standpoint
16:54
to to not just
16:56
wait for a bill to come out, but
16:57
to listen to these discussions that are occurring.
16:59
I I find, quite frankly, people get upset
17:01
about politics.
17:03
If you listen to the commerce committee in
17:05
the house,
17:06
there are a lot of smart people there
17:08
who actually have like minds about some of
17:10
these issues such as cybersecurity.
17:12
So,
17:14
I I think it's, if you're you're just
17:15
gonna learn about it watching the news or
17:17
reading the the New York Times, forget about
17:19
it.
17:20
I do do spend a lot of time,
17:23
specifically on the regulatory side. The FDA,
17:27
the Food and Drug Administration has something which
17:29
we nicknamed Cheddar,
17:31
c d r h, Center for Devices and
17:33
Radiologic Health,
17:35
is really
17:36
evolving,
17:37
its regular regulatory approach, not just to the
17:40
devices themselves, but something called,
17:42
and this is their term, software as a
17:44
medical device. So you can imagine
17:46
let's say you've got an app on your
17:48
phone that will tell you if
17:50
you have signs of depression and that you
17:52
should seek help.
17:54
Those things have to be regulated from a
17:56
from the standpoint of being sure that these
17:58
aren't,
17:59
you know, hooting and hollering about nothing and
18:02
just sort of stuff. I mean, we're talking
18:04
about peace people's health and people's lives now.
18:08
And they're just beginning,
18:09
to look at, they've been they've been at
18:11
this for several years. I I would bet
18:13
you most, health care leaders are not even
18:16
aware of what CDRH is or does. But
18:18
I think they're at least looking at things
18:21
like AI enabled device software functions,
18:24
which includes,
18:25
how
18:26
manufacturers
18:27
should handle updates,
18:29
monitoring, and transparency.
18:32
You know, then these things are
18:34
always,
18:36
subject to lots of change on a regular
18:39
basis given
18:40
the,
18:42
evaluate the evolution of the software, but also
18:45
the evolution of information that comes out of
18:47
the work.
18:48
So that, you know, if we're just talking
18:50
about, let's say, algorithms,
18:51
how are we learning and improving over time?
18:54
And then the amount of data that we're
18:56
generating
18:57
is enormous. It it's beyond belief.
19:01
And the use of real what we call
19:03
real world evidence is also
19:05
coming about so that these devices, for example,
19:08
that are occurring everywhere
19:10
contain information. We have to analyze those
19:13
and,
19:14
and use,
19:16
you know, them to
19:17
not just evaluate
19:19
the
19:21
device itself from a post market surveillance standpoint,
19:24
but for effectiveness evaluation. Right? That's really what
19:26
we're what we're looking for.
19:29
And, you know, here in Chicago, if I
19:31
can put a pitch in, Scott,
19:33
we have an exciting time because
19:36
and this is a little bit off message,
19:38
but I do think it's going to be
19:39
applicable to health care.
19:41
We have a giant effort in quantum computing
19:45
down on the lake that's begin just beginning
19:47
to start up. I'm sure you've heard about
19:49
this.
19:50
And it's amazing to me to see that
19:52
University of Illinois, University of Chicago Northwestern, Illinois
19:55
Tech, which used to be IIT and some
19:57
other,
19:58
local universities here are going together
20:01
out to Silicon Valley to raise money for
20:03
this. There's also IBM, Google.
20:06
You know, you can you can throw all
20:07
the big names into that.
20:09
And quantum computing is going
20:12
to obviously take a lot of energy, but
20:13
it's also gonna increase the speed
20:16
by which we, are now using the so
20:19
called AI enabled devices here,
20:22
which means that,
20:24
we have a long way to go in
20:25
terms of
20:26
learning as much as we can about, what
20:29
works. And so that to me is kind
20:30
of exciting for the future and an example
20:33
of why
20:34
Chicago is, still a pretty exciting place to
20:36
be.
20:38
It certainly is.
20:39
Don, you shared so so many great thoughts
20:42
on where technology is at in health care
20:44
and so many great practical use cases. I
20:47
what I wanna ask for your last question
20:49
is, what's your top piece of advice for
20:51
health care leaders as they prepare for further
20:53
advancements in technology
20:55
and rising demands for care?
20:57
Okay. Well, now you've now you've got me
20:59
right in my sweet spot. My number one,
21:01
number two, number three, number 10, big dot
21:04
is high blood pressure.
21:06
Yeah. It's not anything that anyone pays attention
21:09
to in the C Suite or in the
21:10
Governance Room. But to me, it's the largest
21:13
problem in the world.
21:15
And I'm wearing a device.
21:17
I wore it last night, but I'm wearing
21:18
a device that measures my blood pressure every
21:21
beat. I told you about that.
21:23
And I think that,
21:25
I I would advise
21:27
each of these,
21:28
leaders to get something that's validated
21:31
and learn more about their blood pressure, not
21:34
I went to the doctor's office, you know,
21:36
in April, and it was one thirty eight
21:38
over 70.
21:39
I mean, that's like saying the temperature in
21:41
Chicago in April
21:42
was
21:44
62.
21:46
And therefore, the temperature in Chicago the whole
21:48
year is 62. It's just just not correct.
21:51
Right? So I I would say,
21:53
learn learn some of these. Get get to
21:55
get to learn some of these things. Get
21:57
down in the weeds. Go go into the
21:58
radiology
21:59
department and see how the radiologists are using
22:02
this. Talk to the,
22:04
cardiac electrophysiologist.
22:05
I'm I'm dealing with specialists here, but, obviously,
22:09
I think the frontline clinicians in general internal
22:12
medicine and family medicine can easily
22:14
tell you and show you how they're beginning
22:16
to to use this type of information. Certainly,
22:19
we wanna worry about revenue cycle too. So
22:21
maybe they can have a conversation about documentation
22:24
and coding as well. But I I think
22:26
it's just get out of the house and
22:27
get in the weeds.
22:29
Don, thanks so much for joining us. It
22:31
was a great conversation. I look forward to
22:32
working with you again soon. Thank you, sir.
22:37
At athenahealth,
22:38
we know your ambulatory practice wants healthier,
22:41
a healthier business, healthier care teams, and healthier
22:44
patients.
22:45
But the complexities of modern health care tech
22:47
make it hard for you and your care
22:49
teams to focus on what matters most.
22:51
That's where athenahealth can help.
22:53
Our AI native all in one solutions reduce
22:56
administrative burdens,
22:57
streamline billing and payments, and deliver critical insights
23:00
when clinicians need it most. That means fewer
23:03
clicks, more time for patients, and stronger bottom
23:06
lines.
23:07
Practicing medicine is complex, but running a practice
23:10
can be that much simpler with athenahealth.
23:13
See how simpler is healthier at athenahealth.com.