1
00:00:07,350 --> 00:00:08,039
Hi everyone!
2
00:00:08,130 --> 00:00:11,520
Welcome to Hard Problems, Smart
Solutions, the Newfire Podcast.
3
00:00:11,880 --> 00:00:14,760
I'm Will Crawford, Head of
Advisory Services and CTO at
4
00:00:14,760 --> 00:00:17,759
Newfire Global Partners, and I'll
be your host for this episode.
5
00:00:18,659 --> 00:00:20,800
Today I'm talking with Dr. Paulo Pinho.
6
00:00:21,480 --> 00:00:24,840
Over his career, Paulo has straddled
the line between care delivery,
7
00:00:24,900 --> 00:00:26,610
payer technology and entrepreneurship
8
00:00:27,225 --> 00:00:30,675
including stints as Chief Medical
Officer of Discern Health, and his
9
00:00:30,675 --> 00:00:34,105
VP and Medical Director of Innovation
for Diameter Health and Availity.
10
00:00:34,695 --> 00:00:36,045
He's also a practicing physician.
11
00:00:36,885 --> 00:00:39,945
Today we're talking about the
evolution of healthcare standards
12
00:00:39,945 --> 00:00:42,795
and the importance of clinical
informatics for healthcare innovation.
13
00:00:43,455 --> 00:00:46,455
We'll explore how advances in
standards have enabled new kinds of
14
00:00:46,455 --> 00:00:50,685
applications, including ones that
leverage AI and some of the challenges
15
00:00:50,685 --> 00:00:51,855
that come with those changes.
16
00:00:52,605 --> 00:00:54,765
We're also going to talk about
the transition from clinical
17
00:00:54,765 --> 00:00:56,144
medicine to the corporate world.
18
00:00:56,504 --> 00:00:59,445
And how as a non-clinician
you can work more effectively
19
00:00:59,445 --> 00:01:00,584
with your clinical partners.
20
00:01:01,214 --> 00:01:02,745
Dr. Pinho, welcome to the podcast.
21
00:01:03,315 --> 00:01:03,855
Thanks, Will.
22
00:01:03,855 --> 00:01:07,875
I appreciate being here and, uh, looking
forward to, to digging in for sure.
23
00:01:09,005 --> 00:01:13,155
So before we dive into, uh, specific
topics, could you just tell us a little
24
00:01:13,155 --> 00:01:15,015
bit about your journey in healthcare?
25
00:01:15,375 --> 00:01:15,645
Uh,
26
00:01:16,155 --> 00:01:18,825
Yeah, so a, a little bit
of an unorthodox journey.
27
00:01:18,825 --> 00:01:22,215
I, um, you know, started out
in, um, in clinical practice
28
00:01:22,215 --> 00:01:23,445
right outta medical school.
29
00:01:23,745 --> 00:01:26,655
I'm a primary care, uh,
physician, board certified in
30
00:01:26,715 --> 00:01:28,485
internal medicine and pediatrics.
31
00:01:28,935 --> 00:01:32,175
Um, and worked in both disciplines,
had my own practice that I was able
32
00:01:32,175 --> 00:01:35,414
to build up and, you know, started to
experience many of the challenges that
33
00:01:35,414 --> 00:01:40,845
exist in healthcare delivery, um, from
a, a data standpoint, from a quality
34
00:01:40,845 --> 00:01:44,625
of care standpoint, the defining what
quality care was for individuals.
35
00:01:45,245 --> 00:01:49,775
Ended up selling my practice to a hospital
system and, uh, it was early on in the
36
00:01:49,775 --> 00:01:54,185
consolidation of hospital practices and
eventually I, um, uh, actually made the
37
00:01:54,185 --> 00:01:58,505
move to more executive health roles,
uh, initially with, uh, Prudential,
38
00:01:58,535 --> 00:02:02,905
where a lot of my work focused on things
like thoughtful benefits, benefits
39
00:02:02,905 --> 00:02:07,325
design, um, really creating sort of
an engaged and empowered workforce,
40
00:02:07,730 --> 00:02:11,420
you know, using, uh, benefits design
as the foundation for that and, you
41
00:02:11,420 --> 00:02:15,680
know, really using it to drive employee
engagement, et cetera, and then moved out
42
00:02:15,680 --> 00:02:19,950
into the healthcare data standards world,
predictive analytics world and AI world.
43
00:02:20,160 --> 00:02:23,549
So it took me a while to kind of realize
the scale and the opportunities that
44
00:02:23,549 --> 00:02:27,830
exist, uh, when you use technology to
enable the way healthcare is delivered.
45
00:02:28,310 --> 00:02:31,890
And I think that now more than ever,
it's, you know, becoming a need.
46
00:02:32,144 --> 00:02:35,250
I, I think you have to mitigate some
of that supply demand mismatch with
47
00:02:35,250 --> 00:02:37,440
thoughtful, uh, technology solutions.
48
00:02:37,470 --> 00:02:42,510
And so I'm excited to be part of this, you
know, new wave of tech-enabled healthcare.
49
00:02:44,040 --> 00:02:48,299
So for people listening who you know,
don't come from that clinical informatics
50
00:02:48,299 --> 00:02:53,459
background, uh, could you just talk a
little bit about why data standardization
51
00:02:53,459 --> 00:02:56,070
for these applications is so complicated?
52
00:02:56,910 --> 00:03:01,079
Yeah, so I think dating back to the
meaningful use days, it's uh, it's
53
00:03:01,079 --> 00:03:02,850
really been a significant challenge.
54
00:03:02,910 --> 00:03:06,900
Data in healthcare by definition
is multi-source and multi-format.
55
00:03:07,170 --> 00:03:10,140
It's been increase, there's an
increasing amount of medical data
56
00:03:10,140 --> 00:03:13,920
that's created for each individual with
nearly a thousand percent increase in,
57
00:03:13,920 --> 00:03:17,280
in data capture on individuals over
the course of the last 10, 10 years.
58
00:03:17,640 --> 00:03:20,790
Uh, and we also know that medicine
is becoming increasingly complex
59
00:03:20,970 --> 00:03:24,920
to the point that the, this full
body of medical knowledge is,
60
00:03:25,190 --> 00:03:28,070
uh, increasing at rapid paces.
61
00:03:28,070 --> 00:03:31,130
I mean, we know that, uh, medical
knowledge doubling, you know, as of
62
00:03:31,160 --> 00:03:34,760
2020 was at 73 days, and so it's only
shorter at this stage of the game.
63
00:03:35,240 --> 00:03:38,990
We know that there has been sort of
a challenge in healthcare in terms of
64
00:03:38,990 --> 00:03:41,740
evolving off of, uh, legacy systems.
65
00:03:42,140 --> 00:03:46,940
Uh, we know that despite the fact that
meaningful use was, uh, intended to drive
66
00:03:46,940 --> 00:03:52,710
data interoperability, it really created
a bunch of different formats of data could
67
00:03:52,710 --> 00:03:54,660
be captured within provider's offices.
68
00:03:54,660 --> 00:03:57,240
There were, uh, different
electronic health record platforms.
69
00:03:57,540 --> 00:04:00,210
Each had their different way
of sort of collecting data.
70
00:04:00,540 --> 00:04:04,140
Uh, each had their different key pieces
of information and sort of, and, and
71
00:04:04,140 --> 00:04:07,200
its storage in different areas of
the electronic health record chart.
72
00:04:07,590 --> 00:04:10,710
And then, you know, you could have
two separate implementations of the
73
00:04:10,710 --> 00:04:13,830
same electronic health record in
two different hospitals that were
74
00:04:13,830 --> 00:04:15,210
across the street from one another.
75
00:04:15,210 --> 00:04:17,430
And the way they captured
data was very, very different.
76
00:04:17,435 --> 00:04:21,334
So, um, despite the fact that
meaningful use intended to sort of
77
00:04:21,334 --> 00:04:25,500
wrangle some of this in, it really
created a bunch of different languages.
78
00:04:25,500 --> 00:04:27,840
And then on top of that, a bunch
of different dialects as well.
79
00:04:28,350 --> 00:04:31,380
Um, and we know that the language
of clinical data capture was very
80
00:04:31,380 --> 00:04:34,620
different when you're capturing data
for clinical intent versus when you're
81
00:04:34,620 --> 00:04:36,510
capturing data for billing purposes.
82
00:04:36,990 --> 00:04:40,320
Uh, we know that providers, some of
the data that we, we capture when we
83
00:04:40,320 --> 00:04:44,010
see patients tends to be, you know,
very structured codified data that
84
00:04:44,010 --> 00:04:45,780
we pulled out from pull down menus.
85
00:04:46,140 --> 00:04:49,770
Um, but in other times you really
can't tell a full narrative story,
86
00:04:50,039 --> 00:04:51,960
um, unless you actually type in pros.
87
00:04:51,960 --> 00:04:55,590
And so some of that's made data
standardization really, uh, complex.
88
00:04:55,950 --> 00:04:58,590
I think on top of that, we're
starting to see data capture in a
89
00:04:58,590 --> 00:05:01,740
bunch of different formats that we
never thought would be, uh, data,
90
00:05:01,919 --> 00:05:03,359
you know, methods of data capture.
91
00:05:03,750 --> 00:05:07,169
Uh, we know that data is captured,
you know, outside of electronic health
92
00:05:07,169 --> 00:05:10,080
records that's really meaningful to
healthcare delivery, for example.
93
00:05:10,355 --> 00:05:13,055
There's an increasing the amount
of wearables that that, that are
94
00:05:13,055 --> 00:05:16,085
out there, and we're seeing people
diagnosed with atrial fibrillation
95
00:05:16,085 --> 00:05:19,565
for the first time, you know, from
their Apple watch, or with sleep wake
96
00:05:19,565 --> 00:05:21,785
disturbances from a wearable ring.
97
00:05:22,025 --> 00:05:25,895
We know that there's social data aspects,
uh, that that, that, you know, can't
98
00:05:25,895 --> 00:05:27,485
necessarily be capturing the same way
99
00:05:27,924 --> 00:05:29,364
from electronic health records.
100
00:05:29,724 --> 00:05:34,164
And so ultimately, you know, the types of
data that really have become meaningful
101
00:05:34,164 --> 00:05:38,335
to today's healthcare delivery in a,
in a personalized fashion, um, have
102
00:05:38,335 --> 00:05:41,034
become too difficult to consolidate.
103
00:05:41,155 --> 00:05:45,800
There is a challenge in sort of
the volume, velocity, variety of
104
00:05:45,860 --> 00:05:49,099
clinical data and nonclinical data
that drive healthcare outcomes.
105
00:05:49,460 --> 00:05:52,640
Um, and, and that's I think probably
where the biggest challenges lie.
106
00:05:53,030 --> 00:05:56,840
And there is, there are increasing
challenges when it comes to, uh, data
107
00:05:56,840 --> 00:05:59,870
privacy, data security, um, et cetera.
108
00:06:00,260 --> 00:06:03,710
And so, you know, while there's a lot
of opportunity with the, the sheer
109
00:06:03,710 --> 00:06:07,515
amounts of data that we're collecting
in healthcare, I think the challenges
110
00:06:07,575 --> 00:06:10,844
and the complexities of the different
ways that data is, is, is collected,
111
00:06:10,844 --> 00:06:15,315
the different languages and dialects have
really made it so that interoperability,
112
00:06:15,554 --> 00:06:19,664
despite its most pure intent, has have,
has really been elusive for the healthcare
113
00:06:19,695 --> 00:06:21,435
industry at the stage of the game.
114
00:06:22,664 --> 00:06:28,005
So the mechanics are there, but the
sort of semantic agreement on what
115
00:06:28,005 --> 00:06:31,455
gets exchanged over those pipes,
there's still a lot of work to do.
116
00:06:32,159 --> 00:06:37,289
When you talk about medical knowledge
doubling every 73 days, just tell us
117
00:06:37,289 --> 00:06:39,030
a little bit more about what knowledge
118
00:06:39,030 --> 00:06:39,599
that is.
119
00:06:40,859 --> 00:06:41,940
We see this right now.
120
00:06:41,940 --> 00:06:46,710
I mean, I, I remember practicing
medicine while I was in residency.
121
00:06:47,400 --> 00:06:52,140
I would say that, when I first
started in residency, HIV, for
122
00:06:52,140 --> 00:06:53,940
example, was an inpatient disease.
123
00:06:54,090 --> 00:06:58,260
We managed, I would say that probably 70,
80% of the patients that I was taking care
124
00:06:58,260 --> 00:07:03,210
of on the wards, um, even in the pediatric
wards were HIV positive individuals
125
00:07:03,210 --> 00:07:04,830
that had an opportunistic infection.
126
00:07:05,430 --> 00:07:11,099
By the time I graduated residency, HIV had
shifted to largely in outpatient disease.
127
00:07:11,130 --> 00:07:15,460
Um, you know, focused on giving
people the right highly-active
128
00:07:15,480 --> 00:07:19,830
anti-retroviral therapy, um, to prevent
some of these opportunistic illnesses.
129
00:07:19,830 --> 00:07:22,140
And so, um, we didn't
have inpatient disease.
130
00:07:22,140 --> 00:07:23,370
It was a complete shift.
131
00:07:23,370 --> 00:07:26,940
I mean, I would say that if I
had 20% or less of my patients
132
00:07:26,940 --> 00:07:31,500
having, you know, HIV, um, that was
probably a, a large representation
133
00:07:31,500 --> 00:07:33,120
by the time presidency was over.
134
00:07:33,420 --> 00:07:34,770
We're seeing this with cancer care.
135
00:07:35,250 --> 00:07:39,240
You know, when you look at how cancer has
has been treated up until now, it's really
136
00:07:39,240 --> 00:07:41,280
been this sort of race against time.
137
00:07:41,640 --> 00:07:47,565
You know, we're providing the body with
a poison in the hopes that killing the
138
00:07:47,565 --> 00:07:52,065
cells that are responsible for this
tumor happens at a rate of speed that's
139
00:07:52,065 --> 00:07:55,605
quicker than killing the cells that you
know are keeping a human being alive.
140
00:07:55,995 --> 00:08:00,195
Um, and you know, now with more targeted
therapies that are, you know, molecular
141
00:08:00,195 --> 00:08:04,005
based and focused on cell surface
receptors and really personalized to
142
00:08:04,005 --> 00:08:06,565
the type of cancer that someone has,
143
00:08:06,965 --> 00:08:10,390
you know, we're really able to sort of
augment not only the quantity of life for
144
00:08:10,390 --> 00:08:14,109
individuals, um, who have these cancers,
but also the quality of life because
145
00:08:14,109 --> 00:08:16,090
that side effect burden just isn't there.
146
00:08:16,570 --> 00:08:22,180
And so, you know, I think every era
is defined by a clinical illness
147
00:08:22,180 --> 00:08:23,919
or a set of clinical illnesses.
148
00:08:24,135 --> 00:08:28,575
I would say that my residency training
was defined by sort of that rapid
149
00:08:28,575 --> 00:08:30,855
evolution in, in, in HIV treatment.
150
00:08:30,855 --> 00:08:34,544
And I think future generations is, are
gonna be the cancer personalization
151
00:08:34,544 --> 00:08:37,215
and the cardiovascular disease
personalization that exists.
152
00:08:37,515 --> 00:08:39,255
And these things are
happening very rapidly.
153
00:08:39,255 --> 00:08:42,765
I mean, when we look at cancer care,
you know, I think that we're gonna have
154
00:08:42,765 --> 00:08:47,325
a generation of clinicians, 10, 15, 20
years from now, they're gonna look back
155
00:08:47,325 --> 00:08:48,675
and say, wait, let me get this straight.
156
00:08:48,705 --> 00:08:50,205
This is how you used to treat cancer?
157
00:08:50,205 --> 00:08:52,829
You used to give people poison
and hope that it wouldn't kill
158
00:08:52,829 --> 00:08:54,719
them and kill the, kill the tumor.
159
00:08:54,990 --> 00:08:59,729
This is the kind of, you know, information
that's doubling and doing so very rapidly.
160
00:09:00,030 --> 00:09:01,670
It's not limited to cancer or HIV.
161
00:09:01,670 --> 00:09:05,579
We see it in cardiovascular disease
and, you know, we see the ability
162
00:09:05,579 --> 00:09:08,969
to personalize medicine, to do
it in minimally invasive ways.
163
00:09:09,240 --> 00:09:13,979
It's changing the access opportunities
and the quality of care that we deliver,
164
00:09:14,339 --> 00:09:18,780
um, in ways that we never thought
would be possible 10, 15, 20 years ago.
165
00:09:20,415 --> 00:09:21,765
So I, I think that's a great example.
166
00:09:21,765 --> 00:09:25,334
I mean, I, I have some people in my
life who have benefited from some of
167
00:09:25,334 --> 00:09:29,564
those targeted cancer therapies, and
it's, it's really been remarkable and
168
00:09:29,564 --> 00:09:33,375
it's been very grateful for that because
it would've been a very different
169
00:09:33,375 --> 00:09:35,115
story, even a, even a decade ago.
170
00:09:36,015 --> 00:09:40,214
But bringing those two themes together,
I, there's a lot more information around
171
00:09:40,214 --> 00:09:45,525
the patient and then there's a lot more
knowledge to apply that information to.
172
00:09:46,094 --> 00:09:48,495
So how are providers keeping
their head above water?
173
00:09:49,725 --> 00:09:55,545
Yeah, I think, um, medical education has
definitely needs to, to pivot and I think
174
00:09:55,545 --> 00:10:00,765
it already has started pivoting, obviously
providing sort of an evidence-based
175
00:10:00,765 --> 00:10:03,625
strategy for literature review.
176
00:10:03,895 --> 00:10:08,345
For how, and we saw this during the
COVID Pandemic, there were tons of
177
00:10:08,444 --> 00:10:12,425
articles that were released and some of
them were sort of in holding patterns,
178
00:10:12,425 --> 00:10:16,444
and we needed that data as clinicians
to be able to make informed decisions
179
00:10:16,444 --> 00:10:20,285
about our patients and how do we
do that in a way that's thoughtful?
180
00:10:20,465 --> 00:10:23,955
How do we do it to make sure that
what we are releasing to clinicians
181
00:10:23,955 --> 00:10:27,765
who are using these to kind of make
decisions about their patients is done
182
00:10:27,765 --> 00:10:31,515
so in an informed and an appropriate
way and that it's not quackery.
183
00:10:31,845 --> 00:10:35,955
And so I think that part of our medical
education system now has evolved to
184
00:10:35,955 --> 00:10:39,375
really be very critical of medical
literature and sort of some of the
185
00:10:39,375 --> 00:10:42,365
technology adoption that's happened.
186
00:10:42,795 --> 00:10:46,905
Um, and I think that was really
ushered in by the COVID Pandemic.
187
00:10:47,175 --> 00:10:50,985
Um, it's been ushered in by the
need to personalize, you know,
188
00:10:50,985 --> 00:10:52,485
healthcare for individuals.
189
00:10:53,235 --> 00:10:57,195
Physicians, nurses, nurse practitioners,
physician assistants, they all recognize
190
00:10:57,195 --> 00:11:01,125
at the stage of the game that, you know,
having that information accessible in
191
00:11:01,125 --> 00:11:02,805
the quickest possible way is important.
192
00:11:02,805 --> 00:11:05,295
So you see people using
smartphones on a regular basis,
193
00:11:05,295 --> 00:11:06,645
or tablets on a regular basis.
194
00:11:06,975 --> 00:11:10,125
You see electronic health records
embedded with solutions that allow
195
00:11:10,365 --> 00:11:15,210
providers to look for critical studies
that may help sort of guide care.
196
00:11:15,360 --> 00:11:18,780
You see clinical decision support systems
that are being embedded in electronic
197
00:11:18,780 --> 00:11:22,320
health record workflows that are guiding
clinicians to make the right decision
198
00:11:22,680 --> 00:11:26,400
and really sort of absolving them of
some of the administrative hurdles
199
00:11:26,400 --> 00:11:28,380
of the daily practice of medicine.
200
00:11:28,830 --> 00:11:31,680
And I think you see the advent of
solutions that are coming up that
201
00:11:31,740 --> 00:11:32,940
you know, obviously need to be done.
202
00:11:32,940 --> 00:11:38,810
So with a lot of responsible use
of comfort level with artificial
203
00:11:38,810 --> 00:11:44,060
intelligence, comfort level with things
like large language models, comfort level
204
00:11:44,060 --> 00:11:48,590
with things like ambient listening and how
that factors into the way data is captured
205
00:11:48,590 --> 00:11:50,210
from a patient at the point of care.
206
00:11:50,569 --> 00:11:56,985
And so I think there's whereas medicine
training tended to be very apprenticeship
207
00:11:56,985 --> 00:12:01,395
based in my, in my days of training,
I think there's a lot more flexibility
208
00:12:01,395 --> 00:12:05,085
built into the way residents are trained
in today's day and age to not only sort
209
00:12:05,085 --> 00:12:09,255
of retain some of that apprenticeship
training, but also give people the
210
00:12:09,255 --> 00:12:13,035
flexibility to go out and sort of
experiment with new technologies of data
211
00:12:13,035 --> 00:12:17,655
capture, new technology, of, of grabbing
insights about diseases and treatments.
212
00:12:17,925 --> 00:12:20,830
And so I'm, I'm excited at the
evolution that we've seen in
213
00:12:20,830 --> 00:12:22,150
medical education for that reason.
214
00:12:23,110 --> 00:12:23,920
So I wanna go back to
215
00:12:23,920 --> 00:12:27,070
that concept of comfort
level, especially with AI.
216
00:12:28,030 --> 00:12:32,620
Medicine is a very historically
rigorous evidence-based culture.
217
00:12:33,460 --> 00:12:36,430
AI models, you can't always
tell what's going on inside.
218
00:12:36,850 --> 00:12:41,410
And yet there's so much value in
things like documents summation.
219
00:12:41,470 --> 00:12:45,480
I remember having people talking to
me 10, 15 years ago about the need to
220
00:12:45,480 --> 00:12:49,410
pull in a medical student to do a chart
biopsy on a new patient who came in
221
00:12:49,410 --> 00:12:53,460
and go reading through every previous
node, and I, I, I sort of became very
222
00:12:53,460 --> 00:12:55,380
enamored with that for his chart biopsy.
223
00:12:55,380 --> 00:12:59,520
And I was saying it too often for a
while, AI can do the chart biopsy.
224
00:13:00,210 --> 00:13:00,300
Mm-hmm.
225
00:13:00,600 --> 00:13:03,780
It could probably do the chart
biopsy as well as many medical
226
00:13:03,780 --> 00:13:08,190
students as the provider who's
making a decision about care.
227
00:13:08,820 --> 00:13:10,020
How do you get comfortable?
228
00:13:11,175 --> 00:13:13,574
Yeah, so I wanna approach it
from a couple different angles.
229
00:13:13,574 --> 00:13:17,025
There's a story that I, I routinely
tell that, and there's actually a really
230
00:13:17,025 --> 00:13:21,435
good article by Reinhardt, uh, that was
written in 2020, and it was really about
231
00:13:21,495 --> 00:13:25,844
the stethoscope and how the stethoscope
was ushered in and became something that
232
00:13:26,165 --> 00:13:31,145
physicians had comfort level with, and so
it was invented in the early 18 hundreds.
233
00:13:31,445 --> 00:13:35,675
And at this point, uh, as we think about
its adoption within the United States,
234
00:13:35,675 --> 00:13:39,995
American physicians really had to attend
lectures and, and clinics in Paris to
235
00:13:39,995 --> 00:13:44,225
really learn from some of the fathers
of the stethoscope and relaying it.
236
00:13:44,225 --> 00:13:47,285
René Laënnec was actually
the person who invented it.
237
00:13:47,575 --> 00:13:51,265
And the challenge was, is that it
was limited to very elite physicians
238
00:13:51,265 --> 00:13:54,685
who happened to work in academically
affiliated medical centers.
239
00:13:54,685 --> 00:13:57,985
And so it wasn't every doc who was seeing
every patient that was able to see it.
240
00:13:58,405 --> 00:14:00,895
Adoption was slow, and I think
there was a lot of reasons
241
00:14:00,895 --> 00:14:02,425
why adoption was challenged.
242
00:14:02,425 --> 00:14:04,465
There was a lack of formal education.
243
00:14:04,765 --> 00:14:06,865
You know, there wasn't really a
lot of bedside training that was
244
00:14:06,865 --> 00:14:08,185
available with the stethoscope.
245
00:14:08,545 --> 00:14:11,845
There was a complexity that existed
in the way interpretation happened.
246
00:14:12,215 --> 00:14:16,265
And how to sort of, what do we do with
this atory information that's coming?
247
00:14:16,475 --> 00:14:20,195
There was a hesitancy to create
this barrier between the patient
248
00:14:20,195 --> 00:14:23,195
and a and a clinician by putting
an instrument in between them.
249
00:14:23,465 --> 00:14:26,315
And then there wasn't really a lot of
opportunities for continuing medical
250
00:14:26,315 --> 00:14:31,235
education that physicians could join
into after they've left medical school.
251
00:14:31,235 --> 00:14:34,955
And so a lot of the learning tended to
stop, you know, when they left medical
252
00:14:34,955 --> 00:14:40,230
school and that was, so, it was probably
a little less than a full century
253
00:14:40,230 --> 00:14:45,390
before adoption was so commonplace and
it would be hard, you'd be hard pressed
254
00:14:45,390 --> 00:14:48,480
to find a physician that graduates
from medical school that doesn't
255
00:14:48,480 --> 00:14:50,700
own a stethoscope, hasn't used one.
256
00:14:50,760 --> 00:14:53,880
And granted, there's technology
that's evolved to replace a lot
257
00:14:53,880 --> 00:14:56,850
of the things that we did with
stethoscopes and do so better.
258
00:14:57,060 --> 00:15:00,660
But it's such a central concept
to how we trained as physicians.
259
00:15:01,200 --> 00:15:04,110
And I think we're going through
some of the same challenges with
260
00:15:04,200 --> 00:15:05,790
artificial intelligence right now.
261
00:15:06,120 --> 00:15:11,190
There's a lot of information that's
coming at us, a lot of noise that
262
00:15:11,190 --> 00:15:15,360
exists, a lot of bias, a lot of
reluctance to adopt beause of the fact
263
00:15:15,360 --> 00:15:19,020
that we're fearful that it creates
a, a wedge between us and patients.
264
00:15:19,320 --> 00:15:24,120
I think there's a lot of value to provide
that, that AI provides in doing some of
265
00:15:24,120 --> 00:15:26,020
the things that tend to be very root.
266
00:15:26,370 --> 00:15:29,010
Quite honestly, physicians don't
have the time to deal with.
267
00:15:29,370 --> 00:15:33,360
If you go to an emergency department and
you ask the average doc, how many pages
268
00:15:33,360 --> 00:15:37,440
do they sift through of a person's past
medical history before they abandon and
269
00:15:37,440 --> 00:15:39,280
ship, and then go in and ask the patients,
270
00:15:39,500 --> 00:15:42,465
I think it's, the number is
somewhere around like seven pages.
271
00:15:42,465 --> 00:15:42,645
Right?
272
00:15:42,645 --> 00:15:46,155
It just, there's so much information
that's gathered in that medical chart.
273
00:15:46,515 --> 00:15:50,235
Some, a lot of which is repetitious,
a lot of which may not even be correct
274
00:15:50,235 --> 00:15:51,615
in terms of the way it's captured.
275
00:15:51,855 --> 00:15:54,825
AI is a way of kind of sifting
through that and making it
276
00:15:54,825 --> 00:15:55,935
so that it's streamlined.
277
00:15:56,235 --> 00:15:58,605
There's so much nuance
that isn't captured.
278
00:15:58,605 --> 00:16:02,325
Like I, I used to think about the way
I used to write my clinical notes.
279
00:16:02,675 --> 00:16:05,405
I would go in and have an interview
with a patient and have a, it was a
280
00:16:05,405 --> 00:16:10,295
conversation and then translating that
conversation into something that would
281
00:16:10,505 --> 00:16:15,844
be necessary to, to document code and
bill, I'd lost so much of the nuance that
282
00:16:15,844 --> 00:16:20,464
was so critical for differentiating some
of the gray zones in people's diseases,
283
00:16:20,734 --> 00:16:22,625
and that information was lost forever.
284
00:16:22,625 --> 00:16:27,125
And so things like ambient listening,
while they may be something of
285
00:16:27,214 --> 00:16:30,005
uncomfort in the room at the
point that it's being captured.
286
00:16:30,344 --> 00:16:34,485
Oftentimes provide much more granularity
and much more nuance that's really
287
00:16:34,485 --> 00:16:37,365
important to ultimately making
the right decision for patients,
288
00:16:37,635 --> 00:16:40,395
making the right diagnosis, and
prescribing the right treatment.
289
00:16:40,490 --> 00:16:45,314
I, I do think that there is going to
be an adoption issue and a comfort
290
00:16:45,314 --> 00:16:49,755
level that people are gonna have to get
through, but there's no way that it's
291
00:16:49,755 --> 00:16:51,170
sustainable at this stage of the game.
292
00:16:51,720 --> 00:16:55,830
Where a clinician sees a patient, you
know, and for every hour that they
293
00:16:55,830 --> 00:17:00,600
spend seeing a patient, they document
for another half hour after that.
294
00:17:00,660 --> 00:17:00,930
Right.
295
00:17:00,930 --> 00:17:03,900
That's just not sustainable when
we're, it's sort of seeing the
296
00:17:03,900 --> 00:17:07,349
dwindling numbers of providers and
increased numbers of patients that
297
00:17:07,349 --> 00:17:08,910
are increasingly becoming complex.
298
00:17:08,910 --> 00:17:13,829
In terms of how we adopt it's
really the thoughtful learnings from
299
00:17:13,829 --> 00:17:17,880
technologies like the stethoscope that
we think about how we as a medical
300
00:17:17,880 --> 00:17:20,790
discipline start to adopt technology
in a way that's that's comfortable.
301
00:17:22,710 --> 00:17:26,879
Thinking a a little bit more about some
of the, you know, the natural language
302
00:17:26,879 --> 00:17:30,810
processing element you were just talking
about clinical documentation, do a little
303
00:17:30,810 --> 00:17:34,830
experiment, like if you get a letter back
from a, on a patient you sent out for
304
00:17:34,830 --> 00:17:39,450
a consult, and the first line is, thank
you for this very interesting consult.
305
00:17:39,794 --> 00:17:41,264
What do you read between the lines there?
306
00:17:42,764 --> 00:17:43,094
I've
307
00:17:43,094 --> 00:17:46,094
had that statement represent a
whole bunch of different things.
308
00:17:46,155 --> 00:17:49,905
One thing is, I don't know why you
sent this to us, because you probably
309
00:17:49,905 --> 00:17:51,195
could have figured it out on your own.
310
00:17:51,225 --> 00:17:55,304
You're a primary care doc and you
practiced internal medicine and
311
00:17:55,304 --> 00:17:56,534
you went to residency with me.
312
00:17:56,940 --> 00:17:59,520
I'm calling it interesting, but
it's a little tongue in cheek and
313
00:17:59,520 --> 00:18:01,470
this is probably something you
could have figured out on your own.
314
00:18:01,950 --> 00:18:05,340
I've been in academic medical centers
where there's some sincerity to that.
315
00:18:05,370 --> 00:18:08,610
There's there there's some, it's
a convoluted case that involved
316
00:18:08,880 --> 00:18:10,740
a lot of information gathering.
317
00:18:10,740 --> 00:18:14,700
Perhaps I, as the primary care doc
have gathered some of that information
318
00:18:14,700 --> 00:18:18,090
and synthesized it in a way that I
was able to send it to a specialist
319
00:18:18,090 --> 00:18:21,370
and the specialist pieced it all
together and because of their area of
320
00:18:21,370 --> 00:18:23,980
expertise came to an easier diagnosis.
321
00:18:24,100 --> 00:18:25,360
I've seen both of that, right?
322
00:18:25,389 --> 00:18:28,990
Both of those circumstances where on
one hand it's, hey, I just realized
323
00:18:28,990 --> 00:18:31,780
that the reason why you sent me
this patient is because I'm an ENT
324
00:18:31,780 --> 00:18:33,639
doc and this person had a nose.
325
00:18:33,850 --> 00:18:36,550
And then other times it's, hey,
this is a really complex case,
326
00:18:36,550 --> 00:18:39,340
you synthesized this very nicely
and I got a lot of information.
327
00:18:39,730 --> 00:18:43,750
But I think like I think when
you look at AI from a medical
328
00:18:43,750 --> 00:18:48,405
decisioning standpoint, some of that
information gets glossed over, right?
329
00:18:48,405 --> 00:18:53,025
What's really of essence is, hey, this
is a 53-year-old male with past medical
330
00:18:53,025 --> 00:18:55,425
history, significant for A, B, C.
331
00:18:55,785 --> 00:18:59,985
They're on these medications, these are
their allergies, these are social vices.
332
00:19:00,255 --> 00:19:03,315
They presented to you with
a complaint of X and this is
333
00:19:03,315 --> 00:19:04,755
what you saw on physical exam.
334
00:19:05,145 --> 00:19:07,305
I substantiated by finding this.
335
00:19:07,335 --> 00:19:10,035
I did a couple of extra tests
and in the end I've come to the
336
00:19:10,035 --> 00:19:13,155
conclusion that this is what the
diagnostic, the diagnosis is, right?
337
00:19:13,395 --> 00:19:17,524
It really, I think, a good consult
letter not only starts with those sort
338
00:19:17,524 --> 00:19:21,514
of entry statements and sort of exit
statements, but they really have a lot
339
00:19:21,514 --> 00:19:25,084
of information about the Socratic method
that medicine has practiced in, right?
340
00:19:25,355 --> 00:19:29,615
How we gather information from a
subjective objective, make an assessment,
341
00:19:29,645 --> 00:19:31,504
make a plan for what this patient has.
342
00:19:31,745 --> 00:19:34,534
And that's where I think AI has
become really good at pulling out of
343
00:19:34,715 --> 00:19:37,655
these consult notes to really help
make informed clinical decisions.
344
00:19:37,895 --> 00:19:40,745
And that was the theme that I was hoping
that we could dig in on a little bit
345
00:19:40,834 --> 00:19:46,140
because there is so much context and
things that are written and things
346
00:19:46,140 --> 00:19:48,000
that are not written in these notes.
347
00:19:48,300 --> 00:19:52,260
And you know, I remember working with
people years ago where simply pulling
348
00:19:52,260 --> 00:19:57,480
out someone's smoking status from
a, and this was before that was a
349
00:19:57,480 --> 00:20:01,230
required field as part of meaningful
use, was incredibly different.
350
00:20:01,290 --> 00:20:07,950
Difficult 'cause you only had the free
text node and non-smoker quit last
351
00:20:07,950 --> 00:20:10,650
year is very different than non-smoker.
352
00:20:11,070 --> 00:20:15,930
Which is very different than
12-pack-a-day smoker quit 10 years ago.
353
00:20:16,170 --> 00:20:22,110
And pulling that into a format where you
could make a, actually make a decision
354
00:20:22,140 --> 00:20:24,240
is this patient eligible for a trial?
355
00:20:24,240 --> 00:20:26,520
Are they in a real world evidence cohort?
356
00:20:26,520 --> 00:20:30,030
That's interesting, is very
challenging and I, I think while
357
00:20:30,060 --> 00:20:33,510
we have a lot more coded data now,
we have a lot more complexity that
358
00:20:33,510 --> 00:20:34,560
we need to work through as well.
359
00:20:35,850 --> 00:20:38,820
You know, I think you look at SNOMED
for example, it's become really
360
00:20:38,820 --> 00:20:43,949
expansive and you could really capture
a lot of the pros documentation
361
00:20:43,949 --> 00:20:45,840
in, in, in structured format.
362
00:20:46,050 --> 00:20:48,179
It's still not perfect and
it's not a hundred percent.
363
00:20:48,540 --> 00:20:50,879
And there's, I think the other
thing is that as we think
364
00:20:50,879 --> 00:20:53,189
about why physicians document,
365
00:20:54,855 --> 00:20:57,795
yes, some of it is to create
a historical medical record,
366
00:20:58,095 --> 00:21:00,105
but that really drives care.
367
00:21:00,375 --> 00:21:02,565
But there's a whole lot of other
reasons why they document to
368
00:21:02,565 --> 00:21:04,065
protect themselves medical legally.
369
00:21:04,065 --> 00:21:07,395
They document to establish
patient satisfaction.
370
00:21:07,395 --> 00:21:09,825
They document for the purposes
of building and coding.
371
00:21:10,095 --> 00:21:12,915
And so sometimes it's really
difficult to sift through that.
372
00:21:13,095 --> 00:21:16,035
And I think you can look at the
situation where in a consult letter,
373
00:21:16,035 --> 00:21:18,615
there's an intro sentence and
you can decide, all right, what's
374
00:21:18,615 --> 00:21:19,845
ultimately the meaning of that.
375
00:21:19,875 --> 00:21:25,620
But I think if someone, for example,
if you wanna allude to the fact that
376
00:21:25,620 --> 00:21:29,730
someone's, you believe that someone's
an alcoholic without allowing it
377
00:21:29,730 --> 00:21:33,720
to be a risk factor for you from
a, from a medical legal standpoint,
378
00:21:33,720 --> 00:21:35,010
because you didn't address it.
379
00:21:35,070 --> 00:21:38,400
While you also don't want the
patient to see their record where
380
00:21:38,400 --> 00:21:41,670
you're insinuating that they're an
alcoholic, so that you'll lose trust
381
00:21:41,670 --> 00:21:44,580
with a patient who you're trying to
essentially get to the bottom of things.
382
00:21:44,940 --> 00:21:48,630
It's a really dicey way about how to
appropriately document in a clinical
383
00:21:48,630 --> 00:21:51,420
loop, right, and still capture the nuance.
384
00:21:51,420 --> 00:21:55,240
Still capture, still protect yourself
from a medical legal standpoint and
385
00:21:55,240 --> 00:21:59,170
still establish and continue to establish
that trust with the patient because
386
00:21:59,170 --> 00:22:02,710
they ultimately get to see what's
written in there and dissecting that,
387
00:22:02,710 --> 00:22:07,120
from an, from a an NLP standpoint, that
becomes a bit of a challenge, right?
388
00:22:07,120 --> 00:22:09,400
Because you don't know what
the, what were the motives for
389
00:22:09,400 --> 00:22:11,280
documentation in this particular note?
390
00:22:12,660 --> 00:22:17,350
For companies that are developing, whether
it's ambient documentation systems or
391
00:22:17,620 --> 00:22:23,745
other AI assistive technologies, where
should they be focused if in terms
392
00:22:23,745 --> 00:22:25,755
of creating trust with physicians?
393
00:22:26,745 --> 00:22:27,015
Yeah.
394
00:22:27,015 --> 00:22:31,845
I mean, I, I think as I've educated in,
in informatics and what I think, I think a
395
00:22:31,845 --> 00:22:34,280
lot of the, a lot of the times physicians
396
00:22:35,520 --> 00:22:39,929
are afraid that the, these solutions
are coming to, to are being adopted, not
397
00:22:39,929 --> 00:22:44,370
necessarily to drive the improvement of
the patient experience or to diminish
398
00:22:44,370 --> 00:22:47,879
provider abrasion, rather to make
it so that workflows that are more
399
00:22:47,879 --> 00:22:51,600
administrative in healthcare that
lead to better billing and coding
400
00:22:51,600 --> 00:22:55,080
and reimbursement, things like that
are really the motivating factors.
401
00:22:55,110 --> 00:22:59,699
And while I think we always have to focus
on return on investment when it comes to
402
00:22:59,820 --> 00:23:04,145
a lot of these solutions, some of that
investment is in some of that return
403
00:23:04,145 --> 00:23:09,754
comes in the form of a life saved or
a medication error averted, physicians
404
00:23:09,754 --> 00:23:12,784
went to medical school and clinicians
went to their professional school.
405
00:23:13,095 --> 00:23:15,735
Because they really wanna make a
difference in, in people's lives.
406
00:23:15,765 --> 00:23:18,975
And I think we have to resonate with
them as we have these conversations with
407
00:23:19,065 --> 00:23:23,160
solutions that are meaningful and we have
to see whether or not it truly helps it.
408
00:23:23,165 --> 00:23:26,895
It helps diminish provider abrasion by
implementing some of these solutions.
409
00:23:27,105 --> 00:23:30,720
So we wanna make sure that we're
thoughtful about how, what are we
410
00:23:31,665 --> 00:23:35,445
have clinicians understand what the
motives for implementation are, that
411
00:23:35,445 --> 00:23:40,185
they're truly about helping improve the
physician-patient interaction and they're
412
00:23:40,185 --> 00:23:43,665
helping improve the way care is delivery,
and they're helping to reduce some of
413
00:23:43,665 --> 00:23:48,945
the challenges that we have as physicians
from a rote documentation or some of the
414
00:23:48,945 --> 00:23:50,925
administrative billing issues, et cetera.
415
00:23:50,985 --> 00:23:52,755
That's really what I
think physicians want.
416
00:23:54,795 --> 00:23:58,755
So a few minutes ago you mentioned
SNOMED, and that may not be a term
417
00:23:58,755 --> 00:24:01,785
that all the people listening to
the podcast are familiar with.
418
00:24:02,235 --> 00:24:09,210
So for people who are joining us to learn
a little more about the shape of this
419
00:24:09,210 --> 00:24:13,004
healthcare standards and data world,
can you just give us the two or three
420
00:24:13,004 --> 00:24:15,195
minute tour through modern coding system?
421
00:24:15,885 --> 00:24:19,725
Yeah, so I think a lot of times, you
know, as if we're thinking about the
422
00:24:19,725 --> 00:24:22,965
lay people, there's probably two codes
that they, two types of codes that
423
00:24:22,965 --> 00:24:25,965
they see in their medical billing.
424
00:24:25,965 --> 00:24:29,715
For example, and that's usually the
ICD-10, which is the International
425
00:24:29,715 --> 00:24:33,645
Classification of Diseases in the
CPT, which is a, a list of medical
426
00:24:33,645 --> 00:24:37,425
procedures that are done, or, uh, the
types and intensities of visits that are
427
00:24:37,425 --> 00:24:39,284
captured at at, at the point of care.
428
00:24:39,584 --> 00:24:44,615
Those are largely transacted to be
able to drive thoughtful billing.
429
00:24:45,034 --> 00:24:48,935
But electronic health records have
become so much more complex in
430
00:24:48,935 --> 00:24:52,024
terms of the types of data that
sort of, you know, go into them.
431
00:24:52,415 --> 00:24:55,415
And, you know, there's a, a set
of terminology called LOINC that's
432
00:24:55,415 --> 00:25:00,495
responsible for how we capture things like
blood pressure and how we capture parts of
433
00:25:00,495 --> 00:25:04,274
a complete blood count or a comprehensive
metabolic panel or a blood culture.
434
00:25:04,635 --> 00:25:08,985
There's parts of medications that actually
have a set of terminologies that are
435
00:25:08,985 --> 00:25:10,635
associated with how they're captured.
436
00:25:10,875 --> 00:25:14,525
There's a coding standard that
exists for vaccinations, and so a
437
00:25:14,575 --> 00:25:17,780
lot of our vaccine registries are
set up with that coding standard.
438
00:25:18,050 --> 00:25:23,000
And SNOMED is really a, a sort of a
catchall set of terminology that really
439
00:25:23,000 --> 00:25:27,620
captures not only disease states and
their procedures, but also qualifying
440
00:25:27,620 --> 00:25:31,940
statements that tend to have adjective
and adverb qualities associated with
441
00:25:31,940 --> 00:25:34,220
them so that we can get more nuanced.
442
00:25:34,520 --> 00:25:40,310
Understand that ICD-10 only classifies
things in more broad brushes.
443
00:25:40,580 --> 00:25:44,295
I think you can get more specific
information about diseases if
444
00:25:44,295 --> 00:25:48,315
you start bringing in these other
coding standards and really creating
445
00:25:48,315 --> 00:25:51,585
sort of an amalgam of these coding
standards to help tell a story.
446
00:25:51,915 --> 00:25:54,645
And that's, I think the best way to
represent it is that each of these
447
00:25:54,645 --> 00:25:58,965
coding standards really is, represents a
different part of speech, let's just say
448
00:25:58,965 --> 00:26:04,395
in a sentence so that the full compendium
of the medical information that's captured
449
00:26:04,395 --> 00:26:09,165
in that chart has been codified in a
way that can be transacted by machines.
450
00:26:10,274 --> 00:26:13,695
Well, and we want our audience to be
able to sleep tonight so we are not
451
00:26:13,695 --> 00:26:17,085
gonna talk about NDC codes and how
that works with electronic prescribing.
452
00:26:17,264 --> 00:26:18,524
No, no, we're not.
453
00:26:19,485 --> 00:26:22,754
So pulling all this stuff together,
are there categories of applications
454
00:26:22,754 --> 00:26:26,895
that you think are just more feasible
now, thanks to all this progress,
455
00:26:26,895 --> 00:26:29,264
both in standards adoption and in AI?
456
00:26:29,925 --> 00:26:30,855
Like what are you excited about?
457
00:26:31,485 --> 00:26:32,235
Absolutely.
458
00:26:32,235 --> 00:26:36,719
I think, you know, there's been really an
evolution of a lot of different things,
459
00:26:36,719 --> 00:26:40,500
and I think just even through my, my
career, you know, I can come up with
460
00:26:40,500 --> 00:26:45,050
sort of four quick success stories as
different phases in, in, in my informatics
461
00:26:45,050 --> 00:26:47,460
career that have been really impactful.
462
00:26:47,550 --> 00:26:51,540
You know, I think COVID had a lot
to do with how standard adoption and
463
00:26:51,540 --> 00:26:54,990
electronic data capture, we became
more thoughtful about how to do it.
464
00:26:54,990 --> 00:26:58,860
And so I had the opportunity in one
of my former roles to work for a
465
00:26:58,860 --> 00:27:02,910
company that, um, happened to work in
a state at the public health level.
466
00:27:03,270 --> 00:27:06,270
One of the challenges that existed
from an interoperability standpoint
467
00:27:06,270 --> 00:27:09,630
is that there were over a hundred
different ways of capturing
468
00:27:10,004 --> 00:27:16,335
a simple concept like COVID-19 PCR test
and 40 different ways of representing that
469
00:27:16,335 --> 00:27:18,254
patient had a result that was negative.
470
00:27:18,524 --> 00:27:22,635
And so if you look at a hundred
different ways of capturing the test
471
00:27:22,695 --> 00:27:26,700
and 40 different ways of capturing
its negative value, there's 6,000
472
00:27:26,700 --> 00:27:29,640
different permutations that could
exist of how that data's captured.
473
00:27:30,120 --> 00:27:35,310
And so you could see where if you create
data standardization, it really has a huge
474
00:27:35,310 --> 00:27:37,650
ROI impact on public health initiatives.
475
00:27:37,650 --> 00:27:39,390
And again, this was only in one state.
476
00:27:39,395 --> 00:27:43,405
Imagine what this is across, you
know, this is 40 different labs.
477
00:27:43,405 --> 00:27:46,585
Imagine what this is across 400
different labs in 10 states.
478
00:27:46,585 --> 00:27:46,855
Right?
479
00:27:47,095 --> 00:27:51,505
So this, this is a, a real problem
that I think technology was aimed
480
00:27:51,505 --> 00:27:53,005
at solving and did a great job.
481
00:27:53,455 --> 00:27:57,865
I worked with a group of students
addressing how to take medication
482
00:27:57,865 --> 00:28:02,095
instructions and parse and parse
them using not only medication
483
00:28:02,125 --> 00:28:05,665
normalization or, or data
normalization, but then deploying
484
00:28:05,665 --> 00:28:08,290
novel solutions like NLP and Gen AI.
485
00:28:08,639 --> 00:28:14,520
To really get better F1 scores for some of
these solutions, and, you know, the amount
486
00:28:14,550 --> 00:28:18,870
of medication instructions that could
be parsed once normalization is added
487
00:28:18,870 --> 00:28:23,879
to to, to these NLP Gen AI solutions,
uh, makes it so that the capture of
488
00:28:23,879 --> 00:28:25,980
the information is almost near perfect.
489
00:28:25,980 --> 00:28:30,900
So that's, it's speaks to the fact that
we can really reduce medical errors,
490
00:28:30,900 --> 00:28:34,770
which are substantial impact in, into
why people stay in hospitals longer,
491
00:28:34,770 --> 00:28:38,430
and a significant cause of morbidity
and mortality in amongst patients.
492
00:28:39,690 --> 00:28:44,130
And then in sort of my most recent role,
it's really identifying individuals
493
00:28:44,370 --> 00:28:47,520
and pulling together some of this
multi-source and multi-format data,
494
00:28:47,880 --> 00:28:52,410
really pulling in a combination of
social patient reported outcomes
495
00:28:52,410 --> 00:28:55,950
data, pulling in a lot of the medical
data from electronic health records.
496
00:28:56,190 --> 00:28:59,775
A lot of the billing data and sometimes
the incongruence between what's
497
00:28:59,775 --> 00:29:03,765
captured in a chart and what's what's
billed is actually very insightful.
498
00:29:03,765 --> 00:29:07,925
That sort of absence of congruence
is actually very helpful in how
499
00:29:07,925 --> 00:29:09,585
you build machine learning models.
500
00:29:09,825 --> 00:29:12,675
And so we looked at
individuals primarily older age
501
00:29:12,975 --> 00:29:17,235
with clear modifiable risk factors
like things like polypharmacy or
502
00:29:17,235 --> 00:29:20,564
the way their homes were set up
that were at risk for falls and
503
00:29:20,564 --> 00:29:25,095
actually started targeting thoughtful
interventions to prevent those falls.
504
00:29:25,274 --> 00:29:29,085
And this is really showing that having
a combination of data standardization
505
00:29:29,415 --> 00:29:34,125
as the foundation for more advanced
analytics, predictive modeling and AI
506
00:29:34,385 --> 00:29:39,605
is really important to how we make the
job of treating patients and doing so
507
00:29:39,605 --> 00:29:43,845
in a way that's contextually appropriate
and thoughtful and diminishing in the
508
00:29:43,845 --> 00:29:46,745
provider's burden is really very valuable.
509
00:29:46,745 --> 00:29:52,685
So I think those are really four
good examples of how automation
510
00:29:52,745 --> 00:29:56,014
clinical decision support, and sort
of advanced analytics can be used to
511
00:29:56,014 --> 00:29:57,455
help drive healthcare improvement.
512
00:29:59,465 --> 00:29:59,735
So
513
00:29:59,855 --> 00:30:01,405
actually going back to SNOMED.
514
00:30:02,645 --> 00:30:05,910
Just listening to this conversation
as we've been having it, there's
515
00:30:05,910 --> 00:30:08,070
always been an adoption barrier there.
516
00:30:08,070 --> 00:30:09,420
It's obviously a lot of work.
517
00:30:09,870 --> 00:30:14,130
It's a lot more work to enter structured
information and you've gotta create
518
00:30:14,130 --> 00:30:17,700
an incredible user experience to
be able to do that versus being
519
00:30:17,700 --> 00:30:19,665
able to write a narrative note.
520
00:30:21,155 --> 00:30:25,350
With these a newer AI tools and summation
521
00:30:26,070 --> 00:30:30,480
have we been solving the wrong problem
in terms of designing code systems?
522
00:30:32,170 --> 00:30:36,465
Early on, I think the purpose of
coding systems was really automate
523
00:30:36,465 --> 00:30:40,635
the way hospitals worked, automate
the communication between different
524
00:30:40,635 --> 00:30:43,305
areas of the hospital, drive billing.
525
00:30:43,695 --> 00:30:48,135
I even think, by the way, I think that's
a very insightful, insightful question.
526
00:30:48,585 --> 00:30:50,505
I don't think it's even
about coding systems.
527
00:30:50,505 --> 00:30:53,865
I even think that it's about the
reason why the electronic health
528
00:30:53,865 --> 00:30:56,085
record evolved the way it did.
529
00:30:56,205 --> 00:30:56,865
It wasn't.
530
00:30:57,300 --> 00:31:00,570
I, I don't think it was created with
a thought in mind that this is the
531
00:31:00,570 --> 00:31:06,330
best way to capture data, to create a
chart, a living chart for a patient.
532
00:31:06,450 --> 00:31:10,110
It was really about how do we
create something that lends itself
533
00:31:10,110 --> 00:31:11,610
to billing and coding, right?
534
00:31:11,640 --> 00:31:13,950
And not necessarily how we
create something that tells this
535
00:31:13,950 --> 00:31:15,390
person's full medical story.
536
00:31:15,720 --> 00:31:19,650
And so I think the way coding standards
did evolve, or at least the way they were
537
00:31:19,650 --> 00:31:24,645
utilized, was not necessarily for the
primary intent of patient care delivery.
538
00:31:24,675 --> 00:31:29,715
I think we've done a good job of
starting to pivot and really being more
539
00:31:29,715 --> 00:31:35,025
thoughtful about how we use these tools to
identify opportunities and gaps in care.
540
00:31:35,235 --> 00:31:38,835
But I think we've been reactive in the,
in the way we've gone about it, because
541
00:31:38,835 --> 00:31:43,395
again, they weren't set up in a way to
drive true chronicling of medical care.
542
00:31:43,635 --> 00:31:46,305
They were invented for other
purposes, billing, coding,
543
00:31:46,305 --> 00:31:47,985
medical, legal, et cetera.
544
00:31:47,985 --> 00:31:51,990
And I do think that the technological
evolution that we're seeing now
545
00:31:52,290 --> 00:31:53,820
probably is more thoughtful.
546
00:31:54,030 --> 00:31:57,870
I think the role of the physician
and the clinician informaticist
547
00:31:58,169 --> 00:32:01,710
has made it so that we're being
more protective of our turf.
548
00:32:01,800 --> 00:32:05,100
We're not a technology, we're
not a technology person first.
549
00:32:05,100 --> 00:32:06,450
We're a clinician first, right?
550
00:32:06,810 --> 00:32:10,500
And we're building thoughtful technology
solutions that help augment the way we
551
00:32:10,500 --> 00:32:11,879
can deliver better care for patients.
552
00:32:13,455 --> 00:32:16,545
So in a way we're, we're really
almost going back to the beginning.
553
00:32:16,545 --> 00:32:20,835
People like Octo Barnett and Larry
Weed, who were right at the beginning
554
00:32:20,835 --> 00:32:25,815
of this electronic record keeping
revolution back in the sixties and
555
00:32:25,815 --> 00:32:31,065
seventies and had a very clinical,
very patient oriented historian
556
00:32:31,065 --> 00:32:32,835
approach to designing these tools.
557
00:32:33,270 --> 00:32:37,275
And then the process automation
juggernaut took over from that.
558
00:32:39,255 --> 00:32:43,784
I think we, we deviated a little bit
from that purity of intent early on,
559
00:32:43,875 --> 00:32:49,185
and I think that by having clinicians
involved in the process now, where I
560
00:32:49,185 --> 00:32:50,745
think we're getting back to it for sure.
561
00:32:52,034 --> 00:32:57,915
Continuing on on this theme of clinical
adoption, have you seen any organizations
562
00:32:57,915 --> 00:33:02,445
that you've felt have done a, a really
good job and of thinking about the ROI
563
00:33:02,445 --> 00:33:06,720
on some of these clinical informatics
programs and interviewing the work of
564
00:33:06,720 --> 00:33:09,600
their CMIOs into their strategic plans.
565
00:33:10,650 --> 00:33:14,070
Um, I think that there's a
lot of very forward thinking.
566
00:33:14,130 --> 00:33:18,460
I think there's forward thinking hospitals
that have done really good work around
567
00:33:18,730 --> 00:33:22,155
how are we incorporating some of these
solutions in a way that's not just
568
00:33:22,155 --> 00:33:28,034
about checking a box and really bringing
in the CMIO decision making, um, into
569
00:33:28,034 --> 00:33:29,895
the whole, into the whole process?
570
00:33:30,254 --> 00:33:33,524
I think that there's a couple of
very thoughtful states that really
571
00:33:33,524 --> 00:33:38,340
accelerated the sort of modernization
strategies during the COVID pandemic
572
00:33:38,340 --> 00:33:40,560
and did so in a way that was thoughtful.
573
00:33:41,070 --> 00:33:45,210
I think you're starting to see a lot
of tech enabled organizations that kind
574
00:33:45,210 --> 00:33:50,520
of have come out that are more maybe
concierge based or even concierge light.
575
00:33:50,520 --> 00:33:53,550
They don't demand sort of a lot
of, a lot of membership dues for
576
00:33:53,550 --> 00:33:59,450
patients, but ultimately really
deliver hugely tech-enabled solutions.
577
00:33:59,810 --> 00:34:02,690
I think that there's, I, you know,
look, I, one of the, one of the
578
00:34:02,690 --> 00:34:06,720
indictments that I have of meaningful
use is that I think that there wasn't
579
00:34:06,720 --> 00:34:09,929
necessarily the right carrot and there
wasn't necessarily the right stick.
580
00:34:09,960 --> 00:34:10,320
Right?
581
00:34:10,409 --> 00:34:13,949
It became a really big challenge
to get providers to engage because
582
00:34:13,949 --> 00:34:16,800
of the fact that it was such a
laborious process to get people on
583
00:34:16,800 --> 00:34:20,909
board, and then the penalties for
non-engagement were actually piddly,
584
00:34:20,909 --> 00:34:23,970
especially if you didn't necessarily
have a big Medicare or Medicaid
585
00:34:23,970 --> 00:34:26,490
population in your patient portfolio.
586
00:34:26,845 --> 00:34:31,155
And so I think that people adopted
and there was a lot of checking
587
00:34:31,155 --> 00:34:32,805
boxes to say that they adopted.
588
00:34:32,805 --> 00:34:37,395
A lot of the initial meaningful use
requirements were self-report, and I think
589
00:34:37,635 --> 00:34:42,675
when you have that kind of self-report
and you don't have the right incentives
590
00:34:42,735 --> 00:34:47,085
to get people to do it or disincentives to
prevent people from not doing it, I don't
591
00:34:47,085 --> 00:34:48,555
think that there's thoughtful adoption.
592
00:34:48,650 --> 00:34:52,155
One of the challenges that's existed
is, is that while the capacity exists
593
00:34:52,155 --> 00:34:56,284
to allow people to do that the education
doesn't necessarily exist, and I
594
00:34:56,284 --> 00:35:00,480
think there are thoughtful hospital
systems, thoughtful payers, thoughtful
595
00:35:00,480 --> 00:35:05,520
providers that are really making
patient education and patient navigation
596
00:35:05,520 --> 00:35:07,230
about how to use those solutions
597
00:35:07,440 --> 00:35:12,120
a central part, you need to get, I mean,
I think the two biggest adopters that need
598
00:35:12,120 --> 00:35:16,260
to drive this are your medical staff and
your clinical staff and hospitals, and
599
00:35:16,260 --> 00:35:19,860
then you gotta get patients on board with
understanding about how to do to do this.
600
00:35:19,860 --> 00:35:24,060
It's such a complex care delivery model
that unless they're fully bought in,
601
00:35:24,060 --> 00:35:26,130
you could have patient right of access.
602
00:35:26,430 --> 00:35:27,510
No one's accessing it.
603
00:35:27,510 --> 00:35:31,575
And so I think, I think that's really
what's required to have success.
604
00:35:31,575 --> 00:35:35,475
It's really the buy-in of both
the clinician staff as well as the
605
00:35:35,475 --> 00:35:39,225
patients, regardless of whether
it's a payer provider, hospital
606
00:35:39,225 --> 00:35:42,645
system, a concierge practice, a
tech enabled practice, et cetera.
607
00:35:42,945 --> 00:35:44,505
Those are the things that
are absolutely required.
608
00:35:45,735 --> 00:35:46,155
I'm
609
00:35:46,425 --> 00:35:51,555
reminded of a conversation from 15,
20 years ago with the uh, CIO of a
610
00:35:52,215 --> 00:35:57,325
hospital system that had one of the
first patient portals and he told me two
611
00:35:57,345 --> 00:36:01,605
things that people use it for refills and
appointments, and it's paid for out of the
612
00:36:01,605 --> 00:36:05,175
marketing budget, not out of the clinical
budget, not out of the core IT budget.
613
00:36:05,715 --> 00:36:09,225
And really they saw it as a tool for
retaining their patient population.
614
00:36:10,965 --> 00:36:15,135
What are some of the levers that, like
that convenience of being able to make
615
00:36:15,135 --> 00:36:19,725
an appointment or send a message to your
clinician that will help drive patient
616
00:36:19,725 --> 00:36:21,465
adoption of these new care models.
617
00:36:22,455 --> 00:36:24,915
So I, I think a lot of
it is really access.
618
00:36:24,975 --> 00:36:28,515
You know, you look at things like
mental health, there's some states,
619
00:36:28,515 --> 00:36:32,234
and we're not talking about states
that are necessarily significantly poor
620
00:36:32,234 --> 00:36:36,194
states, and we're talking about affluent
states that still have challenge with
621
00:36:36,404 --> 00:36:40,875
people being able to access crisis
care from a mental health standpoint.
622
00:36:40,875 --> 00:36:44,145
There, there's such an issue
that in one of my previous
623
00:36:44,145 --> 00:36:46,484
roles, I actually did a study on
624
00:36:46,785 --> 00:36:50,805
the incidents of somatization, so
people presenting with physical
625
00:36:50,805 --> 00:36:56,355
complaints for psychiatric illness has
gone up because of the fact that they
626
00:36:56,565 --> 00:37:01,065
oftentimes can access mental health
or can't access the appropriate type
627
00:37:01,065 --> 00:37:04,935
of mental health, or there's social
stigma as associated with mental health.
628
00:37:05,355 --> 00:37:09,135
Or you even look at mental health
providers just being able to find a
629
00:37:09,135 --> 00:37:13,335
psychiatrist or a psychologist or a
social worker that's in your network
630
00:37:13,665 --> 00:37:17,595
that can see you any, anything
sooner than two months from now
631
00:37:17,834 --> 00:37:19,245
has been a significant issue.
632
00:37:19,245 --> 00:37:23,325
And you know, you can see where solutions
like telemedicine, you can see where the
633
00:37:23,325 --> 00:37:29,294
ability to ask patients' questions in a
way that's re recursive allows people,
634
00:37:29,294 --> 00:37:33,705
allows it to serve as a triage system
that can identify those people that
635
00:37:33,705 --> 00:37:36,734
truly need to be seen much more rapidly
636
00:37:36,765 --> 00:37:38,354
and those that perhaps can wait.
637
00:37:38,444 --> 00:37:41,805
You know, one of the stories I tell is
about a family member of mine that had
638
00:37:42,524 --> 00:37:47,145
respiratory complaints and ended up
getting admitted and nearly intubated
639
00:37:47,354 --> 00:37:50,504
because they went to an emergency
department at the same hospital three
640
00:37:50,504 --> 00:37:55,725
separate times complaining of shortness
of breath and on three separate times,
641
00:37:55,725 --> 00:38:00,915
got antibiotics as the prime modality of
treatment for a diagnosis of pneumonia.
642
00:38:01,275 --> 00:38:04,875
And it turns out that this person never
had pneumonia and if they just looked
643
00:38:04,875 --> 00:38:09,225
into their medical records and accessed
them in a way that was easier than
644
00:38:09,225 --> 00:38:13,185
having to look at pieces of paper or
duplicate charts, you know, where there
645
00:38:13,185 --> 00:38:18,424
was cut and paste, where you had an
active comprehensive problem list that
646
00:38:18,424 --> 00:38:20,705
was informatically coded into the chart
647
00:38:21,065 --> 00:38:25,384
you could really pull together information
and synthesize that information in a
648
00:38:25,384 --> 00:38:28,895
way to come up with a very thoughtful
differential diagnosis about what
649
00:38:28,895 --> 00:38:30,095
was going on with this patient.
650
00:38:30,095 --> 00:38:31,955
It turns out that this
patient didn't have pneumonia.
651
00:38:31,955 --> 00:38:36,004
They had pneumonitis, and if they just
went back to their gastroenterology
652
00:38:36,004 --> 00:38:38,825
records and they went back to their
neurology records, they would see
653
00:38:38,825 --> 00:38:42,694
that in addition to respiratory
complaints, they had GI complaints
654
00:38:42,694 --> 00:38:44,165
and they had neurologic complaints.
655
00:38:44,435 --> 00:38:47,160
And it turns out that this was the
first set of symptoms that this
656
00:38:47,160 --> 00:38:50,759
person was declaring themselves as
having a rheumatologic illness with.
657
00:38:51,029 --> 00:38:56,790
So I, I, I think the ability to put
information at the hands of clinicians in
658
00:38:56,790 --> 00:39:01,800
a thoughtful way, not putting the burden
on the patients to tell their story every
659
00:39:01,800 --> 00:39:06,060
time they go into the emergency department
or every time they see a new specialist.
660
00:39:06,390 --> 00:39:10,410
I mean, that's a huge absolution
for patients of having to do that
661
00:39:10,680 --> 00:39:12,240
every time they, they seek care.
662
00:39:12,420 --> 00:39:16,705
And think about it, that the, the more
complex you are as a patient, the more
663
00:39:16,705 --> 00:39:20,215
you're gonna likely have to become an
expert in your own diagnosis in order
664
00:39:20,215 --> 00:39:23,815
to basically transact your care because
your medical chart has grown to the
665
00:39:23,815 --> 00:39:27,565
point that there's thousands, tens of
thousands of pages, and you need to be
666
00:39:27,565 --> 00:39:30,805
able to synthesize and summarize that
so you know that the medications that
667
00:39:30,805 --> 00:39:34,675
you've been given have worked, and these
are the side effect profiles that you've
668
00:39:34,675 --> 00:39:38,035
had, and this is the medication that
didn't work, and this is what ended up
669
00:39:38,035 --> 00:39:39,384
happening when you took this medication.
670
00:39:39,565 --> 00:39:43,225
Like all of that information should
be in a way that's accessible so that
671
00:39:43,815 --> 00:39:47,775
it lends itself towards appropriate
clinical decisioning, be it automation
672
00:39:47,775 --> 00:39:53,775
by machines or improved rate of,
of speed by a clinician who's able
673
00:39:53,775 --> 00:39:57,765
to look at that information on one
screen and say, aha, this is what's
674
00:39:57,765 --> 00:39:58,845
going on with this individual.
675
00:39:58,845 --> 00:40:01,845
It looks like pneumonitis
because they have GI issues and
676
00:40:01,845 --> 00:40:03,045
they have neurologic issues.
677
00:40:04,605 --> 00:40:09,615
So that's a good segue into talking a
little bit more about equity and access.
678
00:40:10,095 --> 00:40:15,495
So, how both with improvements that
we have made around standardization
679
00:40:15,495 --> 00:40:19,485
of healthcare data and of course
these newer AI enabled technologies,
680
00:40:20,115 --> 00:40:23,895
how do you see that helping bridge
some of the gaps in healthcare access?
681
00:40:23,925 --> 00:40:26,505
Like we talked about it a little bit
in the mental health context, but
682
00:40:26,835 --> 00:40:28,015
I'm sure there's quite a few others.
683
00:40:29,004 --> 00:40:32,370
Yeah, I mean, I, I think it comes
with some pitfalls that I think
684
00:40:32,370 --> 00:40:33,690
we need to be concerned about.
685
00:40:33,750 --> 00:40:38,440
I think one of the things that I did
in a study at one of my, recent job
686
00:40:38,460 --> 00:40:45,180
was I, I looked at data quality as it
varied by demographic features, how as
687
00:40:45,180 --> 00:40:51,595
it varied by socioeconomic status, race,
ethnicity, languages spoken, et cetera.
688
00:40:52,410 --> 00:40:55,859
We have a significant challenge when
it comes to data equity, um, in the
689
00:40:55,859 --> 00:40:59,580
United States, and not everybody
has their data captured at the point
690
00:40:59,580 --> 00:41:02,129
of care in an equivalent fashion.
691
00:41:02,220 --> 00:41:06,149
And a lot of this has to do with ways
that people utilize the healthcare system.
692
00:41:06,480 --> 00:41:09,390
I think in areas where
there is provider mistrust
693
00:41:09,629 --> 00:41:12,150
I think the emergency department
tends to be a big source.
694
00:41:12,180 --> 00:41:16,740
I think you fail to create a longitudinal
record that tells a full story.
695
00:41:17,040 --> 00:41:21,060
Even for chronic illness it tends
to be an episodic care delivery
696
00:41:21,060 --> 00:41:22,680
model for these chronic illnesses.
697
00:41:22,950 --> 00:41:26,460
And so a lot of times you're not really
getting a robust data capture, and
698
00:41:26,670 --> 00:41:30,270
that's, it's one of the many reasons
why you see such social disparities
699
00:41:30,270 --> 00:41:33,450
in the treatment of chronic illness
is that some people receive their
700
00:41:33,450 --> 00:41:36,180
chronic illness care as a series of
701
00:41:36,315 --> 00:41:39,645
independent acute illness representations.
702
00:41:39,645 --> 00:41:42,255
They go to the emergency department
when they feel dizzy because their
703
00:41:42,255 --> 00:41:43,875
blood sugars are running 400.
704
00:41:44,295 --> 00:41:47,445
Instead of having proactive care
that drives, that helps them so
705
00:41:47,445 --> 00:41:49,125
that their sugars are never 400.
706
00:41:49,125 --> 00:41:53,145
So we need to identify that there's
a substantial data equity piece.
707
00:41:53,835 --> 00:41:57,465
When we're looking at build, you know,
machine learning models, one of the
708
00:41:57,465 --> 00:42:00,944
challenges that exist there is that,
you know, your features may be the
709
00:42:00,944 --> 00:42:04,665
same for two diverse populations, but
the way diagnoses are captured may
710
00:42:04,665 --> 00:42:08,265
be vastly different to the point that
we may underestimate the prevalence
711
00:42:08,265 --> 00:42:12,810
in a population of a certain disease
entity, or because of social stigma, it
712
00:42:12,855 --> 00:42:17,865
may actually present with a different
ICD-10 code than depression, right?
713
00:42:17,865 --> 00:42:23,805
They may have a series of somatic
presentations of, uh, mental illness
714
00:42:23,805 --> 00:42:28,935
that we should have built-in ways in
machine learning models that identify,
715
00:42:29,115 --> 00:42:33,825
hey, this person's now been seen for
palpitations, headache, abdominal pain,
716
00:42:33,915 --> 00:42:35,685
alternating diarrhea and constipation.
717
00:42:36,299 --> 00:42:38,850
Perhaps there's a mental health
issue that we need to address.
718
00:42:38,850 --> 00:42:43,170
Perhaps that should be built into the way
that we think about how this population
719
00:42:43,410 --> 00:42:47,700
in this geography or in this community
tends to represent the way it, you
720
00:42:47,700 --> 00:42:49,799
know, mental health gets represented.
721
00:42:50,220 --> 00:42:53,430
And so I think the data
equity piece is a big one.
722
00:42:53,640 --> 00:42:55,020
We need to understand that.
723
00:42:55,080 --> 00:42:59,919
We need to make sure that we're
considering the cultural context, we're
724
00:42:59,950 --> 00:43:02,140
considering broadened data sources.
725
00:43:02,470 --> 00:43:06,520
You know, not only some of the medical
data, but also some of the social data is
726
00:43:06,520 --> 00:43:08,379
really important for certain populations.
727
00:43:08,799 --> 00:43:12,250
Um, we wanna do even stratified
sampling within some of those
728
00:43:12,250 --> 00:43:16,120
populations to make sure that we're
doing things in a way that's fair
729
00:43:16,120 --> 00:43:17,950
from a, an algorithmic standpoint.
730
00:43:18,370 --> 00:43:21,160
And at times where there's
data that's missing, we need to
731
00:43:21,435 --> 00:43:25,755
understand disease prevalence in
certain populations and even do data
732
00:43:25,755 --> 00:43:29,715
augmentation until we get to a point
where data capture is seen as more
733
00:43:29,715 --> 00:43:31,206
equitable for certain communities.
734
00:43:32,126 --> 00:43:35,636
So you started your career on
the practice side and then made
735
00:43:36,101 --> 00:43:38,951
transition into clinical informatics,
but never really stepped away
736
00:43:38,951 --> 00:43:40,541
from the practice element either.
737
00:43:41,141 --> 00:43:48,371
So for other physicians who are interested
in getting more deeply involved with
738
00:43:48,671 --> 00:43:52,931
building digital healthcare technology
and wanna make that transition either
739
00:43:52,931 --> 00:43:57,821
partially or completely into the
healthcare technology world, what
740
00:43:57,941 --> 00:43:59,736
would you suggest they keep in mind?
741
00:44:01,001 --> 00:44:04,661
Medical education's evolved and has
needed to evolve because of how much
742
00:44:04,661 --> 00:44:06,521
technology that exists in healthcare.
743
00:44:06,971 --> 00:44:09,761
You know, I think that different
schools have evolved in different cases.
744
00:44:10,211 --> 00:44:13,481
Um, I would say that those students of
medicine that are kind of going through
745
00:44:13,506 --> 00:44:17,711
the, the, the programming now take the
opportunity to learn with people that
746
00:44:17,711 --> 00:44:19,121
are doing different things in healthcare.
747
00:44:19,481 --> 00:44:22,571
And they'll feel that because of the
fact that you went to medical school,
748
00:44:22,571 --> 00:44:25,961
that there's only one traditional way of
practicing medicine and delivering, you
749
00:44:25,961 --> 00:44:28,451
know, value from a healthcare standpoint.
750
00:44:28,481 --> 00:44:31,061
I think there's a lot that
needs to be fixed in healthcare.
751
00:44:31,391 --> 00:44:34,571
Um, there's a lot of policy
issues that that PD addressed.
752
00:44:34,691 --> 00:44:37,271
Um, there's a lot of the different
ways we transac healthcare.
753
00:44:37,571 --> 00:44:40,841
There's a lot of gaps when it comes
to care and ation, and it's one of the
754
00:44:40,841 --> 00:44:44,141
things that I was kind of mentioning
is previously, uh, patients don't know
755
00:44:44,141 --> 00:44:47,471
how you use healthcare in the United
States, like I had the opportunity to
756
00:44:47,471 --> 00:44:52,391
work in an urgent care center like it,
it was a mixture of people that truly
757
00:44:52,391 --> 00:44:55,631
needed to be there because they had
something that acquired urgent care.
758
00:44:56,036 --> 00:45:00,176
And then I would say the other half of
the people were people that probably
759
00:45:00,176 --> 00:45:04,256
either to be an emergency department and
passed six of them along the way, uh, in
760
00:45:04,256 --> 00:45:05,876
order to get to this urgent care center.
761
00:45:06,236 --> 00:45:08,486
And there were, there were people
that would come in and they had
762
00:45:08,486 --> 00:45:11,186
six months worth of the complaints
that probably could be better dealt
763
00:45:11,186 --> 00:45:12,356
with with a primary care doctor.
764
00:45:12,986 --> 00:45:14,906
People don't know how to
use the healthcare system.
765
00:45:14,906 --> 00:45:19,226
I think that we as clinicians can
be empowered as of getting people
766
00:45:19,226 --> 00:45:22,616
to, you know, use it in, in a way
that's contextually appropriate.
767
00:45:23,126 --> 00:45:26,786
And then obviously, you know, in many
of the conversations that we had, you've
768
00:45:26,786 --> 00:45:31,166
had, there's a gap in, in, in terms
of the resources that are necessary
769
00:45:31,166 --> 00:45:36,116
to provide care to whatever complex
and ever aging, you know, population.
770
00:45:36,116 --> 00:45:39,566
And so technology's gonna have
be the way to, to mitigate that.
771
00:45:39,566 --> 00:45:46,346
And so really identifying that you can
provide value with your clinical know-how,
772
00:45:46,616 --> 00:45:51,536
in a way that's non-traditional is really
very, very valuable for, for, you know,
773
00:45:51,536 --> 00:45:53,126
clinicians of medicine to understand.
774
00:45:54,656 --> 00:45:59,006
And then the opposite of that, for
all the non-clinicians in the audience
775
00:45:59,066 --> 00:46:04,946
who want to get involved in healthcare
technology and improving the way that
776
00:46:04,946 --> 00:46:09,206
we deliver care, whether it's in,
in this country or globally, where
777
00:46:09,206 --> 00:46:12,986
should they focus, you know, as they
build their foundational knowledge of
778
00:46:13,361 --> 00:46:16,421
clinical informatics and
interoperability in healthcare, AI?
779
00:46:17,291 --> 00:46:21,801
Some of the more successful companies,
um, when, when it comes to tech, new
780
00:46:21,821 --> 00:46:25,361
technology companies, some of the more
successful companies in healthcare
781
00:46:25,931 --> 00:46:30,161
are first and foremost the health
solutions company, and they have the,
782
00:46:30,191 --> 00:46:33,201
the, the key clinical pieces in place.
783
00:46:33,791 --> 00:46:35,381
Healthcare is nuanced, right?
784
00:46:35,381 --> 00:46:40,421
It's, it's not like there's a binary
outcome for every compilation of tests.
785
00:46:40,811 --> 00:46:44,501
Not every test combination
adds up to the same diagnosis.
786
00:46:44,501 --> 00:46:49,241
There's a lot of many to one, one to many
relationships that exist in healthcare.
787
00:46:49,511 --> 00:46:52,721
There's a lot of interrelationships
that exist and there's no one
788
00:46:52,721 --> 00:46:54,281
way that a diagnosis presents.
789
00:46:54,281 --> 00:46:56,111
There's a lot of nuance
to how they presents.
790
00:46:56,141 --> 00:47:01,426
There's not a one size fits all treatment
when it comes to, you know, hypertension.
791
00:47:01,806 --> 00:47:07,181
We often look at people with hypertension
who have other comorbid illnesses, may
792
00:47:07,181 --> 00:47:11,036
respond to different anti-hypertensive
medications, may actually have those
793
00:47:11,036 --> 00:47:14,456
anti-hypertensive medications treat
some of their other comorbid illnesses
794
00:47:14,666 --> 00:47:16,406
if the right medication is selected.
795
00:47:16,766 --> 00:47:22,436
So as people that are studying technology,
studying data science, studying machine
796
00:47:22,436 --> 00:47:27,386
learning, AI, et cetera, understand
that medicine is binary, it's nuanced.
797
00:47:27,446 --> 00:47:30,956
Um, there's a lot of drivers of medicine
that are not just simply, purely based
798
00:47:30,986 --> 00:47:34,586
on, you know, how someone presents what
diagnosis they have of what treatment
799
00:47:34,586 --> 00:47:38,606
they receive, but there's a whole lot of
interplay from social factors, et cetera.
800
00:47:39,056 --> 00:47:45,056
Having clinicians on board to learn,
knowing, knowing how to talk the talk and
801
00:47:45,056 --> 00:47:48,806
walk the walk of, of, of the clinician
is important to adoption, right?
802
00:47:48,806 --> 00:47:53,311
Because there are stakeholders that
oftentimes may get overlooked and these,
803
00:47:53,576 --> 00:47:57,416
these technology solutions get implemented
in these, you know, hospital systems and
804
00:47:57,416 --> 00:48:01,136
provider offices, et cetera, and they're
the ones that could be part of your
805
00:48:01,136 --> 00:48:03,066
greatest success on a go forward basis.
806
00:48:03,066 --> 00:48:09,791
So, you know, I would say understand
the limitations of AI, the biases
807
00:48:09,791 --> 00:48:14,171
that exist in AI for something as
convoluted as as the medical sciences
808
00:48:14,561 --> 00:48:16,751
and learn from those medical scientists.
809
00:48:16,781 --> 00:48:20,981
The healthiest organizations I've been
a part of, I've taught as much medicine
810
00:48:20,981 --> 00:48:22,661
as I've, as I've learned data science.
811
00:48:22,991 --> 00:48:25,631
I've taught as much medicine as
I've learned standards, right?
812
00:48:25,991 --> 00:48:30,251
Um, and I think that that bi-directional
exchange of information really makes
813
00:48:30,251 --> 00:48:34,431
growing organizations and really
deliver, delivers the best ROIs
814
00:48:35,561 --> 00:48:40,541
So, uh, if our listeners could take, uh,
one thing away from this conversation
815
00:48:40,541 --> 00:48:43,451
that we've had over the last hour or
so, what would you like that to be?
816
00:48:44,831 --> 00:48:46,271
So, I think the time is now.
817
00:48:46,541 --> 00:48:50,441
Setting aside adoption, setting aside,
you know, some of the challenges that
818
00:48:50,441 --> 00:48:55,601
exist from the bias standpoint, setting
aside some of the, uh, challenges that
819
00:48:55,841 --> 00:49:02,471
exists from patients and providers
and technology solutions, et cetera.
820
00:49:03,176 --> 00:49:07,406
We need to reform the way healthcare
is delivered in a way that's equitable,
821
00:49:07,406 --> 00:49:09,986
in a way that's thoughtful and
a way makes the right diagnosis,
822
00:49:09,986 --> 00:49:11,126
prescribes the right treatment.
823
00:49:11,546 --> 00:49:14,186
At the end of the day, technology
enablement are really related.
824
00:49:14,816 --> 00:49:21,326
Um, and having the right organizational
mixture of clinicians and technologists
825
00:49:21,536 --> 00:49:26,881
at a company is vital key to success,
not only of that company, but to the
826
00:49:26,881 --> 00:49:28,081
entire healthcare delivery model.
827
00:49:28,331 --> 00:49:28,991
Thank you very much.
828
00:49:30,431 --> 00:49:31,011
Dr. Pinho.
829
00:49:31,031 --> 00:49:33,161
Thank you so much for being
with us this afternoon.
830
00:49:33,551 --> 00:49:36,521
I think that this conversation is
gonna provide a lot of great context
831
00:49:36,521 --> 00:49:38,981
for listeners as they continue
their own journeys in health IT..
832
00:49:40,241 --> 00:49:43,451
This has been Hard Problems, Smart
Solutions, the Newfire podcast.
833
00:49:43,841 --> 00:49:46,571
Thanks for listening and come back
next time when we'll be diving
834
00:49:46,571 --> 00:49:50,051
into interoperability strategy and
digital health product management
835
00:49:50,291 --> 00:49:51,671
with podcaster Omar Muza.
00:00:07,350 --> 00:00:08,039
Hi everyone!
2
00:00:08,130 --> 00:00:11,520
Welcome to Hard Problems, Smart
Solutions, the Newfire Podcast.
3
00:00:11,880 --> 00:00:14,760
I'm Will Crawford, Head of
Advisory Services and CTO at
4
00:00:14,760 --> 00:00:17,759
Newfire Global Partners, and I'll
be your host for this episode.
5
00:00:18,659 --> 00:00:20,800
Today I'm talking with Dr. Paulo Pinho.
6
00:00:21,480 --> 00:00:24,840
Over his career, Paulo has straddled
the line between care delivery,
7
00:00:24,900 --> 00:00:26,610
payer technology and entrepreneurship
8
00:00:27,225 --> 00:00:30,675
including stints as Chief Medical
Officer of Discern Health, and his
9
00:00:30,675 --> 00:00:34,105
VP and Medical Director of Innovation
for Diameter Health and Availity.
10
00:00:34,695 --> 00:00:36,045
He's also a practicing physician.
11
00:00:36,885 --> 00:00:39,945
Today we're talking about the
evolution of healthcare standards
12
00:00:39,945 --> 00:00:42,795
and the importance of clinical
informatics for healthcare innovation.
13
00:00:43,455 --> 00:00:46,455
We'll explore how advances in
standards have enabled new kinds of
14
00:00:46,455 --> 00:00:50,685
applications, including ones that
leverage AI and some of the challenges
15
00:00:50,685 --> 00:00:51,855
that come with those changes.
16
00:00:52,605 --> 00:00:54,765
We're also going to talk about
the transition from clinical
17
00:00:54,765 --> 00:00:56,144
medicine to the corporate world.
18
00:00:56,504 --> 00:00:59,445
And how as a non-clinician
you can work more effectively
19
00:00:59,445 --> 00:01:00,584
with your clinical partners.
20
00:01:01,214 --> 00:01:02,745
Dr. Pinho, welcome to the podcast.
21
00:01:03,315 --> 00:01:03,855
Thanks, Will.
22
00:01:03,855 --> 00:01:07,875
I appreciate being here and, uh, looking
forward to, to digging in for sure.
23
00:01:09,005 --> 00:01:13,155
So before we dive into, uh, specific
topics, could you just tell us a little
24
00:01:13,155 --> 00:01:15,015
bit about your journey in healthcare?
25
00:01:15,375 --> 00:01:15,645
Uh,
26
00:01:16,155 --> 00:01:18,825
Yeah, so a, a little bit
of an unorthodox journey.
27
00:01:18,825 --> 00:01:22,215
I, um, you know, started out
in, um, in clinical practice
28
00:01:22,215 --> 00:01:23,445
right outta medical school.
29
00:01:23,745 --> 00:01:26,655
I'm a primary care, uh,
physician, board certified in
30
00:01:26,715 --> 00:01:28,485
internal medicine and pediatrics.
31
00:01:28,935 --> 00:01:32,175
Um, and worked in both disciplines,
had my own practice that I was able
32
00:01:32,175 --> 00:01:35,414
to build up and, you know, started to
experience many of the challenges that
33
00:01:35,414 --> 00:01:40,845
exist in healthcare delivery, um, from
a, a data standpoint, from a quality
34
00:01:40,845 --> 00:01:44,625
of care standpoint, the defining what
quality care was for individuals.
35
00:01:45,245 --> 00:01:49,775
Ended up selling my practice to a hospital
system and, uh, it was early on in the
36
00:01:49,775 --> 00:01:54,185
consolidation of hospital practices and
eventually I, um, uh, actually made the
37
00:01:54,185 --> 00:01:58,505
move to more executive health roles,
uh, initially with, uh, Prudential,
38
00:01:58,535 --> 00:02:02,905
where a lot of my work focused on things
like thoughtful benefits, benefits
39
00:02:02,905 --> 00:02:07,325
design, um, really creating sort of
an engaged and empowered workforce,
40
00:02:07,730 --> 00:02:11,420
you know, using, uh, benefits design
as the foundation for that and, you
41
00:02:11,420 --> 00:02:15,680
know, really using it to drive employee
engagement, et cetera, and then moved out
42
00:02:15,680 --> 00:02:19,950
into the healthcare data standards world,
predictive analytics world and AI world.
43
00:02:20,160 --> 00:02:23,549
So it took me a while to kind of realize
the scale and the opportunities that
44
00:02:23,549 --> 00:02:27,830
exist, uh, when you use technology to
enable the way healthcare is delivered.
45
00:02:28,310 --> 00:02:31,890
And I think that now more than ever,
it's, you know, becoming a need.
46
00:02:32,144 --> 00:02:35,250
I, I think you have to mitigate some
of that supply demand mismatch with
47
00:02:35,250 --> 00:02:37,440
thoughtful, uh, technology solutions.
48
00:02:37,470 --> 00:02:42,510
And so I'm excited to be part of this, you
know, new wave of tech-enabled healthcare.
49
00:02:44,040 --> 00:02:48,299
So for people listening who you know,
don't come from that clinical informatics
50
00:02:48,299 --> 00:02:53,459
background, uh, could you just talk a
little bit about why data standardization
51
00:02:53,459 --> 00:02:56,070
for these applications is so complicated?
52
00:02:56,910 --> 00:03:01,079
Yeah, so I think dating back to the
meaningful use days, it's uh, it's
53
00:03:01,079 --> 00:03:02,850
really been a significant challenge.
54
00:03:02,910 --> 00:03:06,900
Data in healthcare by definition
is multi-source and multi-format.
55
00:03:07,170 --> 00:03:10,140
It's been increase, there's an
increasing amount of medical data
56
00:03:10,140 --> 00:03:13,920
that's created for each individual with
nearly a thousand percent increase in,
57
00:03:13,920 --> 00:03:17,280
in data capture on individuals over
the course of the last 10, 10 years.
58
00:03:17,640 --> 00:03:20,790
Uh, and we also know that medicine
is becoming increasingly complex
59
00:03:20,970 --> 00:03:24,920
to the point that the, this full
body of medical knowledge is,
60
00:03:25,190 --> 00:03:28,070
uh, increasing at rapid paces.
61
00:03:28,070 --> 00:03:31,130
I mean, we know that, uh, medical
knowledge doubling, you know, as of
62
00:03:31,160 --> 00:03:34,760
2020 was at 73 days, and so it's only
shorter at this stage of the game.
63
00:03:35,240 --> 00:03:38,990
We know that there has been sort of
a challenge in healthcare in terms of
64
00:03:38,990 --> 00:03:41,740
evolving off of, uh, legacy systems.
65
00:03:42,140 --> 00:03:46,940
Uh, we know that despite the fact that
meaningful use was, uh, intended to drive
66
00:03:46,940 --> 00:03:52,710
data interoperability, it really created
a bunch of different formats of data could
67
00:03:52,710 --> 00:03:54,660
be captured within provider's offices.
68
00:03:54,660 --> 00:03:57,240
There were, uh, different
electronic health record platforms.
69
00:03:57,540 --> 00:04:00,210
Each had their different way
of sort of collecting data.
70
00:04:00,540 --> 00:04:04,140
Uh, each had their different key pieces
of information and sort of, and, and
71
00:04:04,140 --> 00:04:07,200
its storage in different areas of
the electronic health record chart.
72
00:04:07,590 --> 00:04:10,710
And then, you know, you could have
two separate implementations of the
73
00:04:10,710 --> 00:04:13,830
same electronic health record in
two different hospitals that were
74
00:04:13,830 --> 00:04:15,210
across the street from one another.
75
00:04:15,210 --> 00:04:17,430
And the way they captured
data was very, very different.
76
00:04:17,435 --> 00:04:21,334
So, um, despite the fact that
meaningful use intended to sort of
77
00:04:21,334 --> 00:04:25,500
wrangle some of this in, it really
created a bunch of different languages.
78
00:04:25,500 --> 00:04:27,840
And then on top of that, a bunch
of different dialects as well.
79
00:04:28,350 --> 00:04:31,380
Um, and we know that the language
of clinical data capture was very
80
00:04:31,380 --> 00:04:34,620
different when you're capturing data
for clinical intent versus when you're
81
00:04:34,620 --> 00:04:36,510
capturing data for billing purposes.
82
00:04:36,990 --> 00:04:40,320
Uh, we know that providers, some of
the data that we, we capture when we
83
00:04:40,320 --> 00:04:44,010
see patients tends to be, you know,
very structured codified data that
84
00:04:44,010 --> 00:04:45,780
we pulled out from pull down menus.
85
00:04:46,140 --> 00:04:49,770
Um, but in other times you really
can't tell a full narrative story,
86
00:04:50,039 --> 00:04:51,960
um, unless you actually type in pros.
87
00:04:51,960 --> 00:04:55,590
And so some of that's made data
standardization really, uh, complex.
88
00:04:55,950 --> 00:04:58,590
I think on top of that, we're
starting to see data capture in a
89
00:04:58,590 --> 00:05:01,740
bunch of different formats that we
never thought would be, uh, data,
90
00:05:01,919 --> 00:05:03,359
you know, methods of data capture.
91
00:05:03,750 --> 00:05:07,169
Uh, we know that data is captured,
you know, outside of electronic health
92
00:05:07,169 --> 00:05:10,080
records that's really meaningful to
healthcare delivery, for example.
93
00:05:10,355 --> 00:05:13,055
There's an increasing the amount
of wearables that that, that are
94
00:05:13,055 --> 00:05:16,085
out there, and we're seeing people
diagnosed with atrial fibrillation
95
00:05:16,085 --> 00:05:19,565
for the first time, you know, from
their Apple watch, or with sleep wake
96
00:05:19,565 --> 00:05:21,785
disturbances from a wearable ring.
97
00:05:22,025 --> 00:05:25,895
We know that there's social data aspects,
uh, that that, that, you know, can't
98
00:05:25,895 --> 00:05:27,485
necessarily be capturing the same way
99
00:05:27,924 --> 00:05:29,364
from electronic health records.
100
00:05:29,724 --> 00:05:34,164
And so ultimately, you know, the types of
data that really have become meaningful
101
00:05:34,164 --> 00:05:38,335
to today's healthcare delivery in a,
in a personalized fashion, um, have
102
00:05:38,335 --> 00:05:41,034
become too difficult to consolidate.
103
00:05:41,155 --> 00:05:45,800
There is a challenge in sort of
the volume, velocity, variety of
104
00:05:45,860 --> 00:05:49,099
clinical data and nonclinical data
that drive healthcare outcomes.
105
00:05:49,460 --> 00:05:52,640
Um, and, and that's I think probably
where the biggest challenges lie.
106
00:05:53,030 --> 00:05:56,840
And there is, there are increasing
challenges when it comes to, uh, data
107
00:05:56,840 --> 00:05:59,870
privacy, data security, um, et cetera.
108
00:06:00,260 --> 00:06:03,710
And so, you know, while there's a lot
of opportunity with the, the sheer
109
00:06:03,710 --> 00:06:07,515
amounts of data that we're collecting
in healthcare, I think the challenges
110
00:06:07,575 --> 00:06:10,844
and the complexities of the different
ways that data is, is, is collected,
111
00:06:10,844 --> 00:06:15,315
the different languages and dialects have
really made it so that interoperability,
112
00:06:15,554 --> 00:06:19,664
despite its most pure intent, has have,
has really been elusive for the healthcare
113
00:06:19,695 --> 00:06:21,435
industry at the stage of the game.
114
00:06:22,664 --> 00:06:28,005
So the mechanics are there, but the
sort of semantic agreement on what
115
00:06:28,005 --> 00:06:31,455
gets exchanged over those pipes,
there's still a lot of work to do.
116
00:06:32,159 --> 00:06:37,289
When you talk about medical knowledge
doubling every 73 days, just tell us
117
00:06:37,289 --> 00:06:39,030
a little bit more about what knowledge
118
00:06:39,030 --> 00:06:39,599
that is.
119
00:06:40,859 --> 00:06:41,940
We see this right now.
120
00:06:41,940 --> 00:06:46,710
I mean, I, I remember practicing
medicine while I was in residency.
121
00:06:47,400 --> 00:06:52,140
I would say that, when I first
started in residency, HIV, for
122
00:06:52,140 --> 00:06:53,940
example, was an inpatient disease.
123
00:06:54,090 --> 00:06:58,260
We managed, I would say that probably 70,
80% of the patients that I was taking care
124
00:06:58,260 --> 00:07:03,210
of on the wards, um, even in the pediatric
wards were HIV positive individuals
125
00:07:03,210 --> 00:07:04,830
that had an opportunistic infection.
126
00:07:05,430 --> 00:07:11,099
By the time I graduated residency, HIV had
shifted to largely in outpatient disease.
127
00:07:11,130 --> 00:07:15,460
Um, you know, focused on giving
people the right highly-active
128
00:07:15,480 --> 00:07:19,830
anti-retroviral therapy, um, to prevent
some of these opportunistic illnesses.
129
00:07:19,830 --> 00:07:22,140
And so, um, we didn't
have inpatient disease.
130
00:07:22,140 --> 00:07:23,370
It was a complete shift.
131
00:07:23,370 --> 00:07:26,940
I mean, I would say that if I
had 20% or less of my patients
132
00:07:26,940 --> 00:07:31,500
having, you know, HIV, um, that was
probably a, a large representation
133
00:07:31,500 --> 00:07:33,120
by the time presidency was over.
134
00:07:33,420 --> 00:07:34,770
We're seeing this with cancer care.
135
00:07:35,250 --> 00:07:39,240
You know, when you look at how cancer has
has been treated up until now, it's really
136
00:07:39,240 --> 00:07:41,280
been this sort of race against time.
137
00:07:41,640 --> 00:07:47,565
You know, we're providing the body with
a poison in the hopes that killing the
138
00:07:47,565 --> 00:07:52,065
cells that are responsible for this
tumor happens at a rate of speed that's
139
00:07:52,065 --> 00:07:55,605
quicker than killing the cells that you
know are keeping a human being alive.
140
00:07:55,995 --> 00:08:00,195
Um, and you know, now with more targeted
therapies that are, you know, molecular
141
00:08:00,195 --> 00:08:04,005
based and focused on cell surface
receptors and really personalized to
142
00:08:04,005 --> 00:08:06,565
the type of cancer that someone has,
143
00:08:06,965 --> 00:08:10,390
you know, we're really able to sort of
augment not only the quantity of life for
144
00:08:10,390 --> 00:08:14,109
individuals, um, who have these cancers,
but also the quality of life because
145
00:08:14,109 --> 00:08:16,090
that side effect burden just isn't there.
146
00:08:16,570 --> 00:08:22,180
And so, you know, I think every era
is defined by a clinical illness
147
00:08:22,180 --> 00:08:23,919
or a set of clinical illnesses.
148
00:08:24,135 --> 00:08:28,575
I would say that my residency training
was defined by sort of that rapid
149
00:08:28,575 --> 00:08:30,855
evolution in, in, in HIV treatment.
150
00:08:30,855 --> 00:08:34,544
And I think future generations is, are
gonna be the cancer personalization
151
00:08:34,544 --> 00:08:37,215
and the cardiovascular disease
personalization that exists.
152
00:08:37,515 --> 00:08:39,255
And these things are
happening very rapidly.
153
00:08:39,255 --> 00:08:42,765
I mean, when we look at cancer care,
you know, I think that we're gonna have
154
00:08:42,765 --> 00:08:47,325
a generation of clinicians, 10, 15, 20
years from now, they're gonna look back
155
00:08:47,325 --> 00:08:48,675
and say, wait, let me get this straight.
156
00:08:48,705 --> 00:08:50,205
This is how you used to treat cancer?
157
00:08:50,205 --> 00:08:52,829
You used to give people poison
and hope that it wouldn't kill
158
00:08:52,829 --> 00:08:54,719
them and kill the, kill the tumor.
159
00:08:54,990 --> 00:08:59,729
This is the kind of, you know, information
that's doubling and doing so very rapidly.
160
00:09:00,030 --> 00:09:01,670
It's not limited to cancer or HIV.
161
00:09:01,670 --> 00:09:05,579
We see it in cardiovascular disease
and, you know, we see the ability
162
00:09:05,579 --> 00:09:08,969
to personalize medicine, to do
it in minimally invasive ways.
163
00:09:09,240 --> 00:09:13,979
It's changing the access opportunities
and the quality of care that we deliver,
164
00:09:14,339 --> 00:09:18,780
um, in ways that we never thought
would be possible 10, 15, 20 years ago.
165
00:09:20,415 --> 00:09:21,765
So I, I think that's a great example.
166
00:09:21,765 --> 00:09:25,334
I mean, I, I have some people in my
life who have benefited from some of
167
00:09:25,334 --> 00:09:29,564
those targeted cancer therapies, and
it's, it's really been remarkable and
168
00:09:29,564 --> 00:09:33,375
it's been very grateful for that because
it would've been a very different
169
00:09:33,375 --> 00:09:35,115
story, even a, even a decade ago.
170
00:09:36,015 --> 00:09:40,214
But bringing those two themes together,
I, there's a lot more information around
171
00:09:40,214 --> 00:09:45,525
the patient and then there's a lot more
knowledge to apply that information to.
172
00:09:46,094 --> 00:09:48,495
So how are providers keeping
their head above water?
173
00:09:49,725 --> 00:09:55,545
Yeah, I think, um, medical education has
definitely needs to, to pivot and I think
174
00:09:55,545 --> 00:10:00,765
it already has started pivoting, obviously
providing sort of an evidence-based
175
00:10:00,765 --> 00:10:03,625
strategy for literature review.
176
00:10:03,895 --> 00:10:08,345
For how, and we saw this during the
COVID Pandemic, there were tons of
177
00:10:08,444 --> 00:10:12,425
articles that were released and some of
them were sort of in holding patterns,
178
00:10:12,425 --> 00:10:16,444
and we needed that data as clinicians
to be able to make informed decisions
179
00:10:16,444 --> 00:10:20,285
about our patients and how do we
do that in a way that's thoughtful?
180
00:10:20,465 --> 00:10:23,955
How do we do it to make sure that
what we are releasing to clinicians
181
00:10:23,955 --> 00:10:27,765
who are using these to kind of make
decisions about their patients is done
182
00:10:27,765 --> 00:10:31,515
so in an informed and an appropriate
way and that it's not quackery.
183
00:10:31,845 --> 00:10:35,955
And so I think that part of our medical
education system now has evolved to
184
00:10:35,955 --> 00:10:39,375
really be very critical of medical
literature and sort of some of the
185
00:10:39,375 --> 00:10:42,365
technology adoption that's happened.
186
00:10:42,795 --> 00:10:46,905
Um, and I think that was really
ushered in by the COVID Pandemic.
187
00:10:47,175 --> 00:10:50,985
Um, it's been ushered in by the
need to personalize, you know,
188
00:10:50,985 --> 00:10:52,485
healthcare for individuals.
189
00:10:53,235 --> 00:10:57,195
Physicians, nurses, nurse practitioners,
physician assistants, they all recognize
190
00:10:57,195 --> 00:11:01,125
at the stage of the game that, you know,
having that information accessible in
191
00:11:01,125 --> 00:11:02,805
the quickest possible way is important.
192
00:11:02,805 --> 00:11:05,295
So you see people using
smartphones on a regular basis,
193
00:11:05,295 --> 00:11:06,645
or tablets on a regular basis.
194
00:11:06,975 --> 00:11:10,125
You see electronic health records
embedded with solutions that allow
195
00:11:10,365 --> 00:11:15,210
providers to look for critical studies
that may help sort of guide care.
196
00:11:15,360 --> 00:11:18,780
You see clinical decision support systems
that are being embedded in electronic
197
00:11:18,780 --> 00:11:22,320
health record workflows that are guiding
clinicians to make the right decision
198
00:11:22,680 --> 00:11:26,400
and really sort of absolving them of
some of the administrative hurdles
199
00:11:26,400 --> 00:11:28,380
of the daily practice of medicine.
200
00:11:28,830 --> 00:11:31,680
And I think you see the advent of
solutions that are coming up that
201
00:11:31,740 --> 00:11:32,940
you know, obviously need to be done.
202
00:11:32,940 --> 00:11:38,810
So with a lot of responsible use
of comfort level with artificial
203
00:11:38,810 --> 00:11:44,060
intelligence, comfort level with things
like large language models, comfort level
204
00:11:44,060 --> 00:11:48,590
with things like ambient listening and how
that factors into the way data is captured
205
00:11:48,590 --> 00:11:50,210
from a patient at the point of care.
206
00:11:50,569 --> 00:11:56,985
And so I think there's whereas medicine
training tended to be very apprenticeship
207
00:11:56,985 --> 00:12:01,395
based in my, in my days of training,
I think there's a lot more flexibility
208
00:12:01,395 --> 00:12:05,085
built into the way residents are trained
in today's day and age to not only sort
209
00:12:05,085 --> 00:12:09,255
of retain some of that apprenticeship
training, but also give people the
210
00:12:09,255 --> 00:12:13,035
flexibility to go out and sort of
experiment with new technologies of data
211
00:12:13,035 --> 00:12:17,655
capture, new technology, of, of grabbing
insights about diseases and treatments.
212
00:12:17,925 --> 00:12:20,830
And so I'm, I'm excited at the
evolution that we've seen in
213
00:12:20,830 --> 00:12:22,150
medical education for that reason.
214
00:12:23,110 --> 00:12:23,920
So I wanna go back to
215
00:12:23,920 --> 00:12:27,070
that concept of comfort
level, especially with AI.
216
00:12:28,030 --> 00:12:32,620
Medicine is a very historically
rigorous evidence-based culture.
217
00:12:33,460 --> 00:12:36,430
AI models, you can't always
tell what's going on inside.
218
00:12:36,850 --> 00:12:41,410
And yet there's so much value in
things like documents summation.
219
00:12:41,470 --> 00:12:45,480
I remember having people talking to
me 10, 15 years ago about the need to
220
00:12:45,480 --> 00:12:49,410
pull in a medical student to do a chart
biopsy on a new patient who came in
221
00:12:49,410 --> 00:12:53,460
and go reading through every previous
node, and I, I, I sort of became very
222
00:12:53,460 --> 00:12:55,380
enamored with that for his chart biopsy.
223
00:12:55,380 --> 00:12:59,520
And I was saying it too often for a
while, AI can do the chart biopsy.
224
00:13:00,210 --> 00:13:00,300
Mm-hmm.
225
00:13:00,600 --> 00:13:03,780
It could probably do the chart
biopsy as well as many medical
226
00:13:03,780 --> 00:13:08,190
students as the provider who's
making a decision about care.
227
00:13:08,820 --> 00:13:10,020
How do you get comfortable?
228
00:13:11,175 --> 00:13:13,574
Yeah, so I wanna approach it
from a couple different angles.
229
00:13:13,574 --> 00:13:17,025
There's a story that I, I routinely
tell that, and there's actually a really
230
00:13:17,025 --> 00:13:21,435
good article by Reinhardt, uh, that was
written in 2020, and it was really about
231
00:13:21,495 --> 00:13:25,844
the stethoscope and how the stethoscope
was ushered in and became something that
232
00:13:26,165 --> 00:13:31,145
physicians had comfort level with, and so
it was invented in the early 18 hundreds.
233
00:13:31,445 --> 00:13:35,675
And at this point, uh, as we think about
its adoption within the United States,
234
00:13:35,675 --> 00:13:39,995
American physicians really had to attend
lectures and, and clinics in Paris to
235
00:13:39,995 --> 00:13:44,225
really learn from some of the fathers
of the stethoscope and relaying it.
236
00:13:44,225 --> 00:13:47,285
René Laënnec was actually
the person who invented it.
237
00:13:47,575 --> 00:13:51,265
And the challenge was, is that it
was limited to very elite physicians
238
00:13:51,265 --> 00:13:54,685
who happened to work in academically
affiliated medical centers.
239
00:13:54,685 --> 00:13:57,985
And so it wasn't every doc who was seeing
every patient that was able to see it.
240
00:13:58,405 --> 00:14:00,895
Adoption was slow, and I think
there was a lot of reasons
241
00:14:00,895 --> 00:14:02,425
why adoption was challenged.
242
00:14:02,425 --> 00:14:04,465
There was a lack of formal education.
243
00:14:04,765 --> 00:14:06,865
You know, there wasn't really a
lot of bedside training that was
244
00:14:06,865 --> 00:14:08,185
available with the stethoscope.
245
00:14:08,545 --> 00:14:11,845
There was a complexity that existed
in the way interpretation happened.
246
00:14:12,215 --> 00:14:16,265
And how to sort of, what do we do with
this atory information that's coming?
247
00:14:16,475 --> 00:14:20,195
There was a hesitancy to create
this barrier between the patient
248
00:14:20,195 --> 00:14:23,195
and a and a clinician by putting
an instrument in between them.
249
00:14:23,465 --> 00:14:26,315
And then there wasn't really a lot of
opportunities for continuing medical
250
00:14:26,315 --> 00:14:31,235
education that physicians could join
into after they've left medical school.
251
00:14:31,235 --> 00:14:34,955
And so a lot of the learning tended to
stop, you know, when they left medical
252
00:14:34,955 --> 00:14:40,230
school and that was, so, it was probably
a little less than a full century
253
00:14:40,230 --> 00:14:45,390
before adoption was so commonplace and
it would be hard, you'd be hard pressed
254
00:14:45,390 --> 00:14:48,480
to find a physician that graduates
from medical school that doesn't
255
00:14:48,480 --> 00:14:50,700
own a stethoscope, hasn't used one.
256
00:14:50,760 --> 00:14:53,880
And granted, there's technology
that's evolved to replace a lot
257
00:14:53,880 --> 00:14:56,850
of the things that we did with
stethoscopes and do so better.
258
00:14:57,060 --> 00:15:00,660
But it's such a central concept
to how we trained as physicians.
259
00:15:01,200 --> 00:15:04,110
And I think we're going through
some of the same challenges with
260
00:15:04,200 --> 00:15:05,790
artificial intelligence right now.
261
00:15:06,120 --> 00:15:11,190
There's a lot of information that's
coming at us, a lot of noise that
262
00:15:11,190 --> 00:15:15,360
exists, a lot of bias, a lot of
reluctance to adopt beause of the fact
263
00:15:15,360 --> 00:15:19,020
that we're fearful that it creates
a, a wedge between us and patients.
264
00:15:19,320 --> 00:15:24,120
I think there's a lot of value to provide
that, that AI provides in doing some of
265
00:15:24,120 --> 00:15:26,020
the things that tend to be very root.
266
00:15:26,370 --> 00:15:29,010
Quite honestly, physicians don't
have the time to deal with.
267
00:15:29,370 --> 00:15:33,360
If you go to an emergency department and
you ask the average doc, how many pages
268
00:15:33,360 --> 00:15:37,440
do they sift through of a person's past
medical history before they abandon and
269
00:15:37,440 --> 00:15:39,280
ship, and then go in and ask the patients,
270
00:15:39,500 --> 00:15:42,465
I think it's, the number is
somewhere around like seven pages.
271
00:15:42,465 --> 00:15:42,645
Right?
272
00:15:42,645 --> 00:15:46,155
It just, there's so much information
that's gathered in that medical chart.
273
00:15:46,515 --> 00:15:50,235
Some, a lot of which is repetitious,
a lot of which may not even be correct
274
00:15:50,235 --> 00:15:51,615
in terms of the way it's captured.
275
00:15:51,855 --> 00:15:54,825
AI is a way of kind of sifting
through that and making it
276
00:15:54,825 --> 00:15:55,935
so that it's streamlined.
277
00:15:56,235 --> 00:15:58,605
There's so much nuance
that isn't captured.
278
00:15:58,605 --> 00:16:02,325
Like I, I used to think about the way
I used to write my clinical notes.
279
00:16:02,675 --> 00:16:05,405
I would go in and have an interview
with a patient and have a, it was a
280
00:16:05,405 --> 00:16:10,295
conversation and then translating that
conversation into something that would
281
00:16:10,505 --> 00:16:15,844
be necessary to, to document code and
bill, I'd lost so much of the nuance that
282
00:16:15,844 --> 00:16:20,464
was so critical for differentiating some
of the gray zones in people's diseases,
283
00:16:20,734 --> 00:16:22,625
and that information was lost forever.
284
00:16:22,625 --> 00:16:27,125
And so things like ambient listening,
while they may be something of
285
00:16:27,214 --> 00:16:30,005
uncomfort in the room at the
point that it's being captured.
286
00:16:30,344 --> 00:16:34,485
Oftentimes provide much more granularity
and much more nuance that's really
287
00:16:34,485 --> 00:16:37,365
important to ultimately making
the right decision for patients,
288
00:16:37,635 --> 00:16:40,395
making the right diagnosis, and
prescribing the right treatment.
289
00:16:40,490 --> 00:16:45,314
I, I do think that there is going to
be an adoption issue and a comfort
290
00:16:45,314 --> 00:16:49,755
level that people are gonna have to get
through, but there's no way that it's
291
00:16:49,755 --> 00:16:51,170
sustainable at this stage of the game.
292
00:16:51,720 --> 00:16:55,830
Where a clinician sees a patient, you
know, and for every hour that they
293
00:16:55,830 --> 00:17:00,600
spend seeing a patient, they document
for another half hour after that.
294
00:17:00,660 --> 00:17:00,930
Right.
295
00:17:00,930 --> 00:17:03,900
That's just not sustainable when
we're, it's sort of seeing the
296
00:17:03,900 --> 00:17:07,349
dwindling numbers of providers and
increased numbers of patients that
297
00:17:07,349 --> 00:17:08,910
are increasingly becoming complex.
298
00:17:08,910 --> 00:17:13,829
In terms of how we adopt it's
really the thoughtful learnings from
299
00:17:13,829 --> 00:17:17,880
technologies like the stethoscope that
we think about how we as a medical
300
00:17:17,880 --> 00:17:20,790
discipline start to adopt technology
in a way that's that's comfortable.
301
00:17:22,710 --> 00:17:26,879
Thinking a a little bit more about some
of the, you know, the natural language
302
00:17:26,879 --> 00:17:30,810
processing element you were just talking
about clinical documentation, do a little
303
00:17:30,810 --> 00:17:34,830
experiment, like if you get a letter back
from a, on a patient you sent out for
304
00:17:34,830 --> 00:17:39,450
a consult, and the first line is, thank
you for this very interesting consult.
305
00:17:39,794 --> 00:17:41,264
What do you read between the lines there?
306
00:17:42,764 --> 00:17:43,094
I've
307
00:17:43,094 --> 00:17:46,094
had that statement represent a
whole bunch of different things.
308
00:17:46,155 --> 00:17:49,905
One thing is, I don't know why you
sent this to us, because you probably
309
00:17:49,905 --> 00:17:51,195
could have figured it out on your own.
310
00:17:51,225 --> 00:17:55,304
You're a primary care doc and you
practiced internal medicine and
311
00:17:55,304 --> 00:17:56,534
you went to residency with me.
312
00:17:56,940 --> 00:17:59,520
I'm calling it interesting, but
it's a little tongue in cheek and
313
00:17:59,520 --> 00:18:01,470
this is probably something you
could have figured out on your own.
314
00:18:01,950 --> 00:18:05,340
I've been in academic medical centers
where there's some sincerity to that.
315
00:18:05,370 --> 00:18:08,610
There's there there's some, it's
a convoluted case that involved
316
00:18:08,880 --> 00:18:10,740
a lot of information gathering.
317
00:18:10,740 --> 00:18:14,700
Perhaps I, as the primary care doc
have gathered some of that information
318
00:18:14,700 --> 00:18:18,090
and synthesized it in a way that I
was able to send it to a specialist
319
00:18:18,090 --> 00:18:21,370
and the specialist pieced it all
together and because of their area of
320
00:18:21,370 --> 00:18:23,980
expertise came to an easier diagnosis.
321
00:18:24,100 --> 00:18:25,360
I've seen both of that, right?
322
00:18:25,389 --> 00:18:28,990
Both of those circumstances where on
one hand it's, hey, I just realized
323
00:18:28,990 --> 00:18:31,780
that the reason why you sent me
this patient is because I'm an ENT
324
00:18:31,780 --> 00:18:33,639
doc and this person had a nose.
325
00:18:33,850 --> 00:18:36,550
And then other times it's, hey,
this is a really complex case,
326
00:18:36,550 --> 00:18:39,340
you synthesized this very nicely
and I got a lot of information.
327
00:18:39,730 --> 00:18:43,750
But I think like I think when
you look at AI from a medical
328
00:18:43,750 --> 00:18:48,405
decisioning standpoint, some of that
information gets glossed over, right?
329
00:18:48,405 --> 00:18:53,025
What's really of essence is, hey, this
is a 53-year-old male with past medical
330
00:18:53,025 --> 00:18:55,425
history, significant for A, B, C.
331
00:18:55,785 --> 00:18:59,985
They're on these medications, these are
their allergies, these are social vices.
332
00:19:00,255 --> 00:19:03,315
They presented to you with
a complaint of X and this is
333
00:19:03,315 --> 00:19:04,755
what you saw on physical exam.
334
00:19:05,145 --> 00:19:07,305
I substantiated by finding this.
335
00:19:07,335 --> 00:19:10,035
I did a couple of extra tests
and in the end I've come to the
336
00:19:10,035 --> 00:19:13,155
conclusion that this is what the
diagnostic, the diagnosis is, right?
337
00:19:13,395 --> 00:19:17,524
It really, I think, a good consult
letter not only starts with those sort
338
00:19:17,524 --> 00:19:21,514
of entry statements and sort of exit
statements, but they really have a lot
339
00:19:21,514 --> 00:19:25,084
of information about the Socratic method
that medicine has practiced in, right?
340
00:19:25,355 --> 00:19:29,615
How we gather information from a
subjective objective, make an assessment,
341
00:19:29,645 --> 00:19:31,504
make a plan for what this patient has.
342
00:19:31,745 --> 00:19:34,534
And that's where I think AI has
become really good at pulling out of
343
00:19:34,715 --> 00:19:37,655
these consult notes to really help
make informed clinical decisions.
344
00:19:37,895 --> 00:19:40,745
And that was the theme that I was hoping
that we could dig in on a little bit
345
00:19:40,834 --> 00:19:46,140
because there is so much context and
things that are written and things
346
00:19:46,140 --> 00:19:48,000
that are not written in these notes.
347
00:19:48,300 --> 00:19:52,260
And you know, I remember working with
people years ago where simply pulling
348
00:19:52,260 --> 00:19:57,480
out someone's smoking status from
a, and this was before that was a
349
00:19:57,480 --> 00:20:01,230
required field as part of meaningful
use, was incredibly different.
350
00:20:01,290 --> 00:20:07,950
Difficult 'cause you only had the free
text node and non-smoker quit last
351
00:20:07,950 --> 00:20:10,650
year is very different than non-smoker.
352
00:20:11,070 --> 00:20:15,930
Which is very different than
12-pack-a-day smoker quit 10 years ago.
353
00:20:16,170 --> 00:20:22,110
And pulling that into a format where you
could make a, actually make a decision
354
00:20:22,140 --> 00:20:24,240
is this patient eligible for a trial?
355
00:20:24,240 --> 00:20:26,520
Are they in a real world evidence cohort?
356
00:20:26,520 --> 00:20:30,030
That's interesting, is very
challenging and I, I think while
357
00:20:30,060 --> 00:20:33,510
we have a lot more coded data now,
we have a lot more complexity that
358
00:20:33,510 --> 00:20:34,560
we need to work through as well.
359
00:20:35,850 --> 00:20:38,820
You know, I think you look at SNOMED
for example, it's become really
360
00:20:38,820 --> 00:20:43,949
expansive and you could really capture
a lot of the pros documentation
361
00:20:43,949 --> 00:20:45,840
in, in, in structured format.
362
00:20:46,050 --> 00:20:48,179
It's still not perfect and
it's not a hundred percent.
363
00:20:48,540 --> 00:20:50,879
And there's, I think the other
thing is that as we think
364
00:20:50,879 --> 00:20:53,189
about why physicians document,
365
00:20:54,855 --> 00:20:57,795
yes, some of it is to create
a historical medical record,
366
00:20:58,095 --> 00:21:00,105
but that really drives care.
367
00:21:00,375 --> 00:21:02,565
But there's a whole lot of other
reasons why they document to
368
00:21:02,565 --> 00:21:04,065
protect themselves medical legally.
369
00:21:04,065 --> 00:21:07,395
They document to establish
patient satisfaction.
370
00:21:07,395 --> 00:21:09,825
They document for the purposes
of building and coding.
371
00:21:10,095 --> 00:21:12,915
And so sometimes it's really
difficult to sift through that.
372
00:21:13,095 --> 00:21:16,035
And I think you can look at the
situation where in a consult letter,
373
00:21:16,035 --> 00:21:18,615
there's an intro sentence and
you can decide, all right, what's
374
00:21:18,615 --> 00:21:19,845
ultimately the meaning of that.
375
00:21:19,875 --> 00:21:25,620
But I think if someone, for example,
if you wanna allude to the fact that
376
00:21:25,620 --> 00:21:29,730
someone's, you believe that someone's
an alcoholic without allowing it
377
00:21:29,730 --> 00:21:33,720
to be a risk factor for you from
a, from a medical legal standpoint,
378
00:21:33,720 --> 00:21:35,010
because you didn't address it.
379
00:21:35,070 --> 00:21:38,400
While you also don't want the
patient to see their record where
380
00:21:38,400 --> 00:21:41,670
you're insinuating that they're an
alcoholic, so that you'll lose trust
381
00:21:41,670 --> 00:21:44,580
with a patient who you're trying to
essentially get to the bottom of things.
382
00:21:44,940 --> 00:21:48,630
It's a really dicey way about how to
appropriately document in a clinical
383
00:21:48,630 --> 00:21:51,420
loop, right, and still capture the nuance.
384
00:21:51,420 --> 00:21:55,240
Still capture, still protect yourself
from a medical legal standpoint and
385
00:21:55,240 --> 00:21:59,170
still establish and continue to establish
that trust with the patient because
386
00:21:59,170 --> 00:22:02,710
they ultimately get to see what's
written in there and dissecting that,
387
00:22:02,710 --> 00:22:07,120
from an, from a an NLP standpoint, that
becomes a bit of a challenge, right?
388
00:22:07,120 --> 00:22:09,400
Because you don't know what
the, what were the motives for
389
00:22:09,400 --> 00:22:11,280
documentation in this particular note?
390
00:22:12,660 --> 00:22:17,350
For companies that are developing, whether
it's ambient documentation systems or
391
00:22:17,620 --> 00:22:23,745
other AI assistive technologies, where
should they be focused if in terms
392
00:22:23,745 --> 00:22:25,755
of creating trust with physicians?
393
00:22:26,745 --> 00:22:27,015
Yeah.
394
00:22:27,015 --> 00:22:31,845
I mean, I, I think as I've educated in,
in informatics and what I think, I think a
395
00:22:31,845 --> 00:22:34,280
lot of the, a lot of the times physicians
396
00:22:35,520 --> 00:22:39,929
are afraid that the, these solutions
are coming to, to are being adopted, not
397
00:22:39,929 --> 00:22:44,370
necessarily to drive the improvement of
the patient experience or to diminish
398
00:22:44,370 --> 00:22:47,879
provider abrasion, rather to make
it so that workflows that are more
399
00:22:47,879 --> 00:22:51,600
administrative in healthcare that
lead to better billing and coding
400
00:22:51,600 --> 00:22:55,080
and reimbursement, things like that
are really the motivating factors.
401
00:22:55,110 --> 00:22:59,699
And while I think we always have to focus
on return on investment when it comes to
402
00:22:59,820 --> 00:23:04,145
a lot of these solutions, some of that
investment is in some of that return
403
00:23:04,145 --> 00:23:09,754
comes in the form of a life saved or
a medication error averted, physicians
404
00:23:09,754 --> 00:23:12,784
went to medical school and clinicians
went to their professional school.
405
00:23:13,095 --> 00:23:15,735
Because they really wanna make a
difference in, in people's lives.
406
00:23:15,765 --> 00:23:18,975
And I think we have to resonate with
them as we have these conversations with
407
00:23:19,065 --> 00:23:23,160
solutions that are meaningful and we have
to see whether or not it truly helps it.
408
00:23:23,165 --> 00:23:26,895
It helps diminish provider abrasion by
implementing some of these solutions.
409
00:23:27,105 --> 00:23:30,720
So we wanna make sure that we're
thoughtful about how, what are we
410
00:23:31,665 --> 00:23:35,445
have clinicians understand what the
motives for implementation are, that
411
00:23:35,445 --> 00:23:40,185
they're truly about helping improve the
physician-patient interaction and they're
412
00:23:40,185 --> 00:23:43,665
helping improve the way care is delivery,
and they're helping to reduce some of
413
00:23:43,665 --> 00:23:48,945
the challenges that we have as physicians
from a rote documentation or some of the
414
00:23:48,945 --> 00:23:50,925
administrative billing issues, et cetera.
415
00:23:50,985 --> 00:23:52,755
That's really what I
think physicians want.
416
00:23:54,795 --> 00:23:58,755
So a few minutes ago you mentioned
SNOMED, and that may not be a term
417
00:23:58,755 --> 00:24:01,785
that all the people listening to
the podcast are familiar with.
418
00:24:02,235 --> 00:24:09,210
So for people who are joining us to learn
a little more about the shape of this
419
00:24:09,210 --> 00:24:13,004
healthcare standards and data world,
can you just give us the two or three
420
00:24:13,004 --> 00:24:15,195
minute tour through modern coding system?
421
00:24:15,885 --> 00:24:19,725
Yeah, so I think a lot of times, you
know, as if we're thinking about the
422
00:24:19,725 --> 00:24:22,965
lay people, there's probably two codes
that they, two types of codes that
423
00:24:22,965 --> 00:24:25,965
they see in their medical billing.
424
00:24:25,965 --> 00:24:29,715
For example, and that's usually the
ICD-10, which is the International
425
00:24:29,715 --> 00:24:33,645
Classification of Diseases in the
CPT, which is a, a list of medical
426
00:24:33,645 --> 00:24:37,425
procedures that are done, or, uh, the
types and intensities of visits that are
427
00:24:37,425 --> 00:24:39,284
captured at at, at the point of care.
428
00:24:39,584 --> 00:24:44,615
Those are largely transacted to be
able to drive thoughtful billing.
429
00:24:45,034 --> 00:24:48,935
But electronic health records have
become so much more complex in
430
00:24:48,935 --> 00:24:52,024
terms of the types of data that
sort of, you know, go into them.
431
00:24:52,415 --> 00:24:55,415
And, you know, there's a, a set
of terminology called LOINC that's
432
00:24:55,415 --> 00:25:00,495
responsible for how we capture things like
blood pressure and how we capture parts of
433
00:25:00,495 --> 00:25:04,274
a complete blood count or a comprehensive
metabolic panel or a blood culture.
434
00:25:04,635 --> 00:25:08,985
There's parts of medications that actually
have a set of terminologies that are
435
00:25:08,985 --> 00:25:10,635
associated with how they're captured.
436
00:25:10,875 --> 00:25:14,525
There's a coding standard that
exists for vaccinations, and so a
437
00:25:14,575 --> 00:25:17,780
lot of our vaccine registries are
set up with that coding standard.
438
00:25:18,050 --> 00:25:23,000
And SNOMED is really a, a sort of a
catchall set of terminology that really
439
00:25:23,000 --> 00:25:27,620
captures not only disease states and
their procedures, but also qualifying
440
00:25:27,620 --> 00:25:31,940
statements that tend to have adjective
and adverb qualities associated with
441
00:25:31,940 --> 00:25:34,220
them so that we can get more nuanced.
442
00:25:34,520 --> 00:25:40,310
Understand that ICD-10 only classifies
things in more broad brushes.
443
00:25:40,580 --> 00:25:44,295
I think you can get more specific
information about diseases if
444
00:25:44,295 --> 00:25:48,315
you start bringing in these other
coding standards and really creating
445
00:25:48,315 --> 00:25:51,585
sort of an amalgam of these coding
standards to help tell a story.
446
00:25:51,915 --> 00:25:54,645
And that's, I think the best way to
represent it is that each of these
447
00:25:54,645 --> 00:25:58,965
coding standards really is, represents a
different part of speech, let's just say
448
00:25:58,965 --> 00:26:04,395
in a sentence so that the full compendium
of the medical information that's captured
449
00:26:04,395 --> 00:26:09,165
in that chart has been codified in a
way that can be transacted by machines.
450
00:26:10,274 --> 00:26:13,695
Well, and we want our audience to be
able to sleep tonight so we are not
451
00:26:13,695 --> 00:26:17,085
gonna talk about NDC codes and how
that works with electronic prescribing.
452
00:26:17,264 --> 00:26:18,524
No, no, we're not.
453
00:26:19,485 --> 00:26:22,754
So pulling all this stuff together,
are there categories of applications
454
00:26:22,754 --> 00:26:26,895
that you think are just more feasible
now, thanks to all this progress,
455
00:26:26,895 --> 00:26:29,264
both in standards adoption and in AI?
456
00:26:29,925 --> 00:26:30,855
Like what are you excited about?
457
00:26:31,485 --> 00:26:32,235
Absolutely.
458
00:26:32,235 --> 00:26:36,719
I think, you know, there's been really an
evolution of a lot of different things,
459
00:26:36,719 --> 00:26:40,500
and I think just even through my, my
career, you know, I can come up with
460
00:26:40,500 --> 00:26:45,050
sort of four quick success stories as
different phases in, in, in my informatics
461
00:26:45,050 --> 00:26:47,460
career that have been really impactful.
462
00:26:47,550 --> 00:26:51,540
You know, I think COVID had a lot
to do with how standard adoption and
463
00:26:51,540 --> 00:26:54,990
electronic data capture, we became
more thoughtful about how to do it.
464
00:26:54,990 --> 00:26:58,860
And so I had the opportunity in one
of my former roles to work for a
465
00:26:58,860 --> 00:27:02,910
company that, um, happened to work in
a state at the public health level.
466
00:27:03,270 --> 00:27:06,270
One of the challenges that existed
from an interoperability standpoint
467
00:27:06,270 --> 00:27:09,630
is that there were over a hundred
different ways of capturing
468
00:27:10,004 --> 00:27:16,335
a simple concept like COVID-19 PCR test
and 40 different ways of representing that
469
00:27:16,335 --> 00:27:18,254
patient had a result that was negative.
470
00:27:18,524 --> 00:27:22,635
And so if you look at a hundred
different ways of capturing the test
471
00:27:22,695 --> 00:27:26,700
and 40 different ways of capturing
its negative value, there's 6,000
472
00:27:26,700 --> 00:27:29,640
different permutations that could
exist of how that data's captured.
473
00:27:30,120 --> 00:27:35,310
And so you could see where if you create
data standardization, it really has a huge
474
00:27:35,310 --> 00:27:37,650
ROI impact on public health initiatives.
475
00:27:37,650 --> 00:27:39,390
And again, this was only in one state.
476
00:27:39,395 --> 00:27:43,405
Imagine what this is across, you
know, this is 40 different labs.
477
00:27:43,405 --> 00:27:46,585
Imagine what this is across 400
different labs in 10 states.
478
00:27:46,585 --> 00:27:46,855
Right?
479
00:27:47,095 --> 00:27:51,505
So this, this is a, a real problem
that I think technology was aimed
480
00:27:51,505 --> 00:27:53,005
at solving and did a great job.
481
00:27:53,455 --> 00:27:57,865
I worked with a group of students
addressing how to take medication
482
00:27:57,865 --> 00:28:02,095
instructions and parse and parse
them using not only medication
483
00:28:02,125 --> 00:28:05,665
normalization or, or data
normalization, but then deploying
484
00:28:05,665 --> 00:28:08,290
novel solutions like NLP and Gen AI.
485
00:28:08,639 --> 00:28:14,520
To really get better F1 scores for some of
these solutions, and, you know, the amount
486
00:28:14,550 --> 00:28:18,870
of medication instructions that could
be parsed once normalization is added
487
00:28:18,870 --> 00:28:23,879
to to, to these NLP Gen AI solutions,
uh, makes it so that the capture of
488
00:28:23,879 --> 00:28:25,980
the information is almost near perfect.
489
00:28:25,980 --> 00:28:30,900
So that's, it's speaks to the fact that
we can really reduce medical errors,
490
00:28:30,900 --> 00:28:34,770
which are substantial impact in, into
why people stay in hospitals longer,
491
00:28:34,770 --> 00:28:38,430
and a significant cause of morbidity
and mortality in amongst patients.
492
00:28:39,690 --> 00:28:44,130
And then in sort of my most recent role,
it's really identifying individuals
493
00:28:44,370 --> 00:28:47,520
and pulling together some of this
multi-source and multi-format data,
494
00:28:47,880 --> 00:28:52,410
really pulling in a combination of
social patient reported outcomes
495
00:28:52,410 --> 00:28:55,950
data, pulling in a lot of the medical
data from electronic health records.
496
00:28:56,190 --> 00:28:59,775
A lot of the billing data and sometimes
the incongruence between what's
497
00:28:59,775 --> 00:29:03,765
captured in a chart and what's what's
billed is actually very insightful.
498
00:29:03,765 --> 00:29:07,925
That sort of absence of congruence
is actually very helpful in how
499
00:29:07,925 --> 00:29:09,585
you build machine learning models.
500
00:29:09,825 --> 00:29:12,675
And so we looked at
individuals primarily older age
501
00:29:12,975 --> 00:29:17,235
with clear modifiable risk factors
like things like polypharmacy or
502
00:29:17,235 --> 00:29:20,564
the way their homes were set up
that were at risk for falls and
503
00:29:20,564 --> 00:29:25,095
actually started targeting thoughtful
interventions to prevent those falls.
504
00:29:25,274 --> 00:29:29,085
And this is really showing that having
a combination of data standardization
505
00:29:29,415 --> 00:29:34,125
as the foundation for more advanced
analytics, predictive modeling and AI
506
00:29:34,385 --> 00:29:39,605
is really important to how we make the
job of treating patients and doing so
507
00:29:39,605 --> 00:29:43,845
in a way that's contextually appropriate
and thoughtful and diminishing in the
508
00:29:43,845 --> 00:29:46,745
provider's burden is really very valuable.
509
00:29:46,745 --> 00:29:52,685
So I think those are really four
good examples of how automation
510
00:29:52,745 --> 00:29:56,014
clinical decision support, and sort
of advanced analytics can be used to
511
00:29:56,014 --> 00:29:57,455
help drive healthcare improvement.
512
00:29:59,465 --> 00:29:59,735
So
513
00:29:59,855 --> 00:30:01,405
actually going back to SNOMED.
514
00:30:02,645 --> 00:30:05,910
Just listening to this conversation
as we've been having it, there's
515
00:30:05,910 --> 00:30:08,070
always been an adoption barrier there.
516
00:30:08,070 --> 00:30:09,420
It's obviously a lot of work.
517
00:30:09,870 --> 00:30:14,130
It's a lot more work to enter structured
information and you've gotta create
518
00:30:14,130 --> 00:30:17,700
an incredible user experience to
be able to do that versus being
519
00:30:17,700 --> 00:30:19,665
able to write a narrative note.
520
00:30:21,155 --> 00:30:25,350
With these a newer AI tools and summation
521
00:30:26,070 --> 00:30:30,480
have we been solving the wrong problem
in terms of designing code systems?
522
00:30:32,170 --> 00:30:36,465
Early on, I think the purpose of
coding systems was really automate
523
00:30:36,465 --> 00:30:40,635
the way hospitals worked, automate
the communication between different
524
00:30:40,635 --> 00:30:43,305
areas of the hospital, drive billing.
525
00:30:43,695 --> 00:30:48,135
I even think, by the way, I think that's
a very insightful, insightful question.
526
00:30:48,585 --> 00:30:50,505
I don't think it's even
about coding systems.
527
00:30:50,505 --> 00:30:53,865
I even think that it's about the
reason why the electronic health
528
00:30:53,865 --> 00:30:56,085
record evolved the way it did.
529
00:30:56,205 --> 00:30:56,865
It wasn't.
530
00:30:57,300 --> 00:31:00,570
I, I don't think it was created with
a thought in mind that this is the
531
00:31:00,570 --> 00:31:06,330
best way to capture data, to create a
chart, a living chart for a patient.
532
00:31:06,450 --> 00:31:10,110
It was really about how do we
create something that lends itself
533
00:31:10,110 --> 00:31:11,610
to billing and coding, right?
534
00:31:11,640 --> 00:31:13,950
And not necessarily how we
create something that tells this
535
00:31:13,950 --> 00:31:15,390
person's full medical story.
536
00:31:15,720 --> 00:31:19,650
And so I think the way coding standards
did evolve, or at least the way they were
537
00:31:19,650 --> 00:31:24,645
utilized, was not necessarily for the
primary intent of patient care delivery.
538
00:31:24,675 --> 00:31:29,715
I think we've done a good job of
starting to pivot and really being more
539
00:31:29,715 --> 00:31:35,025
thoughtful about how we use these tools to
identify opportunities and gaps in care.
540
00:31:35,235 --> 00:31:38,835
But I think we've been reactive in the,
in the way we've gone about it, because
541
00:31:38,835 --> 00:31:43,395
again, they weren't set up in a way to
drive true chronicling of medical care.
542
00:31:43,635 --> 00:31:46,305
They were invented for other
purposes, billing, coding,
543
00:31:46,305 --> 00:31:47,985
medical, legal, et cetera.
544
00:31:47,985 --> 00:31:51,990
And I do think that the technological
evolution that we're seeing now
545
00:31:52,290 --> 00:31:53,820
probably is more thoughtful.
546
00:31:54,030 --> 00:31:57,870
I think the role of the physician
and the clinician informaticist
547
00:31:58,169 --> 00:32:01,710
has made it so that we're being
more protective of our turf.
548
00:32:01,800 --> 00:32:05,100
We're not a technology, we're
not a technology person first.
549
00:32:05,100 --> 00:32:06,450
We're a clinician first, right?
550
00:32:06,810 --> 00:32:10,500
And we're building thoughtful technology
solutions that help augment the way we
551
00:32:10,500 --> 00:32:11,879
can deliver better care for patients.
552
00:32:13,455 --> 00:32:16,545
So in a way we're, we're really
almost going back to the beginning.
553
00:32:16,545 --> 00:32:20,835
People like Octo Barnett and Larry
Weed, who were right at the beginning
554
00:32:20,835 --> 00:32:25,815
of this electronic record keeping
revolution back in the sixties and
555
00:32:25,815 --> 00:32:31,065
seventies and had a very clinical,
very patient oriented historian
556
00:32:31,065 --> 00:32:32,835
approach to designing these tools.
557
00:32:33,270 --> 00:32:37,275
And then the process automation
juggernaut took over from that.
558
00:32:39,255 --> 00:32:43,784
I think we, we deviated a little bit
from that purity of intent early on,
559
00:32:43,875 --> 00:32:49,185
and I think that by having clinicians
involved in the process now, where I
560
00:32:49,185 --> 00:32:50,745
think we're getting back to it for sure.
561
00:32:52,034 --> 00:32:57,915
Continuing on on this theme of clinical
adoption, have you seen any organizations
562
00:32:57,915 --> 00:33:02,445
that you've felt have done a, a really
good job and of thinking about the ROI
563
00:33:02,445 --> 00:33:06,720
on some of these clinical informatics
programs and interviewing the work of
564
00:33:06,720 --> 00:33:09,600
their CMIOs into their strategic plans.
565
00:33:10,650 --> 00:33:14,070
Um, I think that there's a
lot of very forward thinking.
566
00:33:14,130 --> 00:33:18,460
I think there's forward thinking hospitals
that have done really good work around
567
00:33:18,730 --> 00:33:22,155
how are we incorporating some of these
solutions in a way that's not just
568
00:33:22,155 --> 00:33:28,034
about checking a box and really bringing
in the CMIO decision making, um, into
569
00:33:28,034 --> 00:33:29,895
the whole, into the whole process?
570
00:33:30,254 --> 00:33:33,524
I think that there's a couple of
very thoughtful states that really
571
00:33:33,524 --> 00:33:38,340
accelerated the sort of modernization
strategies during the COVID pandemic
572
00:33:38,340 --> 00:33:40,560
and did so in a way that was thoughtful.
573
00:33:41,070 --> 00:33:45,210
I think you're starting to see a lot
of tech enabled organizations that kind
574
00:33:45,210 --> 00:33:50,520
of have come out that are more maybe
concierge based or even concierge light.
575
00:33:50,520 --> 00:33:53,550
They don't demand sort of a lot
of, a lot of membership dues for
576
00:33:53,550 --> 00:33:59,450
patients, but ultimately really
deliver hugely tech-enabled solutions.
577
00:33:59,810 --> 00:34:02,690
I think that there's, I, you know,
look, I, one of the, one of the
578
00:34:02,690 --> 00:34:06,720
indictments that I have of meaningful
use is that I think that there wasn't
579
00:34:06,720 --> 00:34:09,929
necessarily the right carrot and there
wasn't necessarily the right stick.
580
00:34:09,960 --> 00:34:10,320
Right?
581
00:34:10,409 --> 00:34:13,949
It became a really big challenge
to get providers to engage because
582
00:34:13,949 --> 00:34:16,800
of the fact that it was such a
laborious process to get people on
583
00:34:16,800 --> 00:34:20,909
board, and then the penalties for
non-engagement were actually piddly,
584
00:34:20,909 --> 00:34:23,970
especially if you didn't necessarily
have a big Medicare or Medicaid
585
00:34:23,970 --> 00:34:26,490
population in your patient portfolio.
586
00:34:26,845 --> 00:34:31,155
And so I think that people adopted
and there was a lot of checking
587
00:34:31,155 --> 00:34:32,805
boxes to say that they adopted.
588
00:34:32,805 --> 00:34:37,395
A lot of the initial meaningful use
requirements were self-report, and I think
589
00:34:37,635 --> 00:34:42,675
when you have that kind of self-report
and you don't have the right incentives
590
00:34:42,735 --> 00:34:47,085
to get people to do it or disincentives to
prevent people from not doing it, I don't
591
00:34:47,085 --> 00:34:48,555
think that there's thoughtful adoption.
592
00:34:48,650 --> 00:34:52,155
One of the challenges that's existed
is, is that while the capacity exists
593
00:34:52,155 --> 00:34:56,284
to allow people to do that the education
doesn't necessarily exist, and I
594
00:34:56,284 --> 00:35:00,480
think there are thoughtful hospital
systems, thoughtful payers, thoughtful
595
00:35:00,480 --> 00:35:05,520
providers that are really making
patient education and patient navigation
596
00:35:05,520 --> 00:35:07,230
about how to use those solutions
597
00:35:07,440 --> 00:35:12,120
a central part, you need to get, I mean,
I think the two biggest adopters that need
598
00:35:12,120 --> 00:35:16,260
to drive this are your medical staff and
your clinical staff and hospitals, and
599
00:35:16,260 --> 00:35:19,860
then you gotta get patients on board with
understanding about how to do to do this.
600
00:35:19,860 --> 00:35:24,060
It's such a complex care delivery model
that unless they're fully bought in,
601
00:35:24,060 --> 00:35:26,130
you could have patient right of access.
602
00:35:26,430 --> 00:35:27,510
No one's accessing it.
603
00:35:27,510 --> 00:35:31,575
And so I think, I think that's really
what's required to have success.
604
00:35:31,575 --> 00:35:35,475
It's really the buy-in of both
the clinician staff as well as the
605
00:35:35,475 --> 00:35:39,225
patients, regardless of whether
it's a payer provider, hospital
606
00:35:39,225 --> 00:35:42,645
system, a concierge practice, a
tech enabled practice, et cetera.
607
00:35:42,945 --> 00:35:44,505
Those are the things that
are absolutely required.
608
00:35:45,735 --> 00:35:46,155
I'm
609
00:35:46,425 --> 00:35:51,555
reminded of a conversation from 15,
20 years ago with the uh, CIO of a
610
00:35:52,215 --> 00:35:57,325
hospital system that had one of the
first patient portals and he told me two
611
00:35:57,345 --> 00:36:01,605
things that people use it for refills and
appointments, and it's paid for out of the
612
00:36:01,605 --> 00:36:05,175
marketing budget, not out of the clinical
budget, not out of the core IT budget.
613
00:36:05,715 --> 00:36:09,225
And really they saw it as a tool for
retaining their patient population.
614
00:36:10,965 --> 00:36:15,135
What are some of the levers that, like
that convenience of being able to make
615
00:36:15,135 --> 00:36:19,725
an appointment or send a message to your
clinician that will help drive patient
616
00:36:19,725 --> 00:36:21,465
adoption of these new care models.
617
00:36:22,455 --> 00:36:24,915
So I, I think a lot of
it is really access.
618
00:36:24,975 --> 00:36:28,515
You know, you look at things like
mental health, there's some states,
619
00:36:28,515 --> 00:36:32,234
and we're not talking about states
that are necessarily significantly poor
620
00:36:32,234 --> 00:36:36,194
states, and we're talking about affluent
states that still have challenge with
621
00:36:36,404 --> 00:36:40,875
people being able to access crisis
care from a mental health standpoint.
622
00:36:40,875 --> 00:36:44,145
There, there's such an issue
that in one of my previous
623
00:36:44,145 --> 00:36:46,484
roles, I actually did a study on
624
00:36:46,785 --> 00:36:50,805
the incidents of somatization, so
people presenting with physical
625
00:36:50,805 --> 00:36:56,355
complaints for psychiatric illness has
gone up because of the fact that they
626
00:36:56,565 --> 00:37:01,065
oftentimes can access mental health
or can't access the appropriate type
627
00:37:01,065 --> 00:37:04,935
of mental health, or there's social
stigma as associated with mental health.
628
00:37:05,355 --> 00:37:09,135
Or you even look at mental health
providers just being able to find a
629
00:37:09,135 --> 00:37:13,335
psychiatrist or a psychologist or a
social worker that's in your network
630
00:37:13,665 --> 00:37:17,595
that can see you any, anything
sooner than two months from now
631
00:37:17,834 --> 00:37:19,245
has been a significant issue.
632
00:37:19,245 --> 00:37:23,325
And you know, you can see where solutions
like telemedicine, you can see where the
633
00:37:23,325 --> 00:37:29,294
ability to ask patients' questions in a
way that's re recursive allows people,
634
00:37:29,294 --> 00:37:33,705
allows it to serve as a triage system
that can identify those people that
635
00:37:33,705 --> 00:37:36,734
truly need to be seen much more rapidly
636
00:37:36,765 --> 00:37:38,354
and those that perhaps can wait.
637
00:37:38,444 --> 00:37:41,805
You know, one of the stories I tell is
about a family member of mine that had
638
00:37:42,524 --> 00:37:47,145
respiratory complaints and ended up
getting admitted and nearly intubated
639
00:37:47,354 --> 00:37:50,504
because they went to an emergency
department at the same hospital three
640
00:37:50,504 --> 00:37:55,725
separate times complaining of shortness
of breath and on three separate times,
641
00:37:55,725 --> 00:38:00,915
got antibiotics as the prime modality of
treatment for a diagnosis of pneumonia.
642
00:38:01,275 --> 00:38:04,875
And it turns out that this person never
had pneumonia and if they just looked
643
00:38:04,875 --> 00:38:09,225
into their medical records and accessed
them in a way that was easier than
644
00:38:09,225 --> 00:38:13,185
having to look at pieces of paper or
duplicate charts, you know, where there
645
00:38:13,185 --> 00:38:18,424
was cut and paste, where you had an
active comprehensive problem list that
646
00:38:18,424 --> 00:38:20,705
was informatically coded into the chart
647
00:38:21,065 --> 00:38:25,384
you could really pull together information
and synthesize that information in a
648
00:38:25,384 --> 00:38:28,895
way to come up with a very thoughtful
differential diagnosis about what
649
00:38:28,895 --> 00:38:30,095
was going on with this patient.
650
00:38:30,095 --> 00:38:31,955
It turns out that this
patient didn't have pneumonia.
651
00:38:31,955 --> 00:38:36,004
They had pneumonitis, and if they just
went back to their gastroenterology
652
00:38:36,004 --> 00:38:38,825
records and they went back to their
neurology records, they would see
653
00:38:38,825 --> 00:38:42,694
that in addition to respiratory
complaints, they had GI complaints
654
00:38:42,694 --> 00:38:44,165
and they had neurologic complaints.
655
00:38:44,435 --> 00:38:47,160
And it turns out that this was the
first set of symptoms that this
656
00:38:47,160 --> 00:38:50,759
person was declaring themselves as
having a rheumatologic illness with.
657
00:38:51,029 --> 00:38:56,790
So I, I, I think the ability to put
information at the hands of clinicians in
658
00:38:56,790 --> 00:39:01,800
a thoughtful way, not putting the burden
on the patients to tell their story every
659
00:39:01,800 --> 00:39:06,060
time they go into the emergency department
or every time they see a new specialist.
660
00:39:06,390 --> 00:39:10,410
I mean, that's a huge absolution
for patients of having to do that
661
00:39:10,680 --> 00:39:12,240
every time they, they seek care.
662
00:39:12,420 --> 00:39:16,705
And think about it, that the, the more
complex you are as a patient, the more
663
00:39:16,705 --> 00:39:20,215
you're gonna likely have to become an
expert in your own diagnosis in order
664
00:39:20,215 --> 00:39:23,815
to basically transact your care because
your medical chart has grown to the
665
00:39:23,815 --> 00:39:27,565
point that there's thousands, tens of
thousands of pages, and you need to be
666
00:39:27,565 --> 00:39:30,805
able to synthesize and summarize that
so you know that the medications that
667
00:39:30,805 --> 00:39:34,675
you've been given have worked, and these
are the side effect profiles that you've
668
00:39:34,675 --> 00:39:38,035
had, and this is the medication that
didn't work, and this is what ended up
669
00:39:38,035 --> 00:39:39,384
happening when you took this medication.
670
00:39:39,565 --> 00:39:43,225
Like all of that information should
be in a way that's accessible so that
671
00:39:43,815 --> 00:39:47,775
it lends itself towards appropriate
clinical decisioning, be it automation
672
00:39:47,775 --> 00:39:53,775
by machines or improved rate of,
of speed by a clinician who's able
673
00:39:53,775 --> 00:39:57,765
to look at that information on one
screen and say, aha, this is what's
674
00:39:57,765 --> 00:39:58,845
going on with this individual.
675
00:39:58,845 --> 00:40:01,845
It looks like pneumonitis
because they have GI issues and
676
00:40:01,845 --> 00:40:03,045
they have neurologic issues.
677
00:40:04,605 --> 00:40:09,615
So that's a good segue into talking a
little bit more about equity and access.
678
00:40:10,095 --> 00:40:15,495
So, how both with improvements that
we have made around standardization
679
00:40:15,495 --> 00:40:19,485
of healthcare data and of course
these newer AI enabled technologies,
680
00:40:20,115 --> 00:40:23,895
how do you see that helping bridge
some of the gaps in healthcare access?
681
00:40:23,925 --> 00:40:26,505
Like we talked about it a little bit
in the mental health context, but
682
00:40:26,835 --> 00:40:28,015
I'm sure there's quite a few others.
683
00:40:29,004 --> 00:40:32,370
Yeah, I mean, I, I think it comes
with some pitfalls that I think
684
00:40:32,370 --> 00:40:33,690
we need to be concerned about.
685
00:40:33,750 --> 00:40:38,440
I think one of the things that I did
in a study at one of my, recent job
686
00:40:38,460 --> 00:40:45,180
was I, I looked at data quality as it
varied by demographic features, how as
687
00:40:45,180 --> 00:40:51,595
it varied by socioeconomic status, race,
ethnicity, languages spoken, et cetera.
688
00:40:52,410 --> 00:40:55,859
We have a significant challenge when
it comes to data equity, um, in the
689
00:40:55,859 --> 00:40:59,580
United States, and not everybody
has their data captured at the point
690
00:40:59,580 --> 00:41:02,129
of care in an equivalent fashion.
691
00:41:02,220 --> 00:41:06,149
And a lot of this has to do with ways
that people utilize the healthcare system.
692
00:41:06,480 --> 00:41:09,390
I think in areas where
there is provider mistrust
693
00:41:09,629 --> 00:41:12,150
I think the emergency department
tends to be a big source.
694
00:41:12,180 --> 00:41:16,740
I think you fail to create a longitudinal
record that tells a full story.
695
00:41:17,040 --> 00:41:21,060
Even for chronic illness it tends
to be an episodic care delivery
696
00:41:21,060 --> 00:41:22,680
model for these chronic illnesses.
697
00:41:22,950 --> 00:41:26,460
And so a lot of times you're not really
getting a robust data capture, and
698
00:41:26,670 --> 00:41:30,270
that's, it's one of the many reasons
why you see such social disparities
699
00:41:30,270 --> 00:41:33,450
in the treatment of chronic illness
is that some people receive their
700
00:41:33,450 --> 00:41:36,180
chronic illness care as a series of
701
00:41:36,315 --> 00:41:39,645
independent acute illness representations.
702
00:41:39,645 --> 00:41:42,255
They go to the emergency department
when they feel dizzy because their
703
00:41:42,255 --> 00:41:43,875
blood sugars are running 400.
704
00:41:44,295 --> 00:41:47,445
Instead of having proactive care
that drives, that helps them so
705
00:41:47,445 --> 00:41:49,125
that their sugars are never 400.
706
00:41:49,125 --> 00:41:53,145
So we need to identify that there's
a substantial data equity piece.
707
00:41:53,835 --> 00:41:57,465
When we're looking at build, you know,
machine learning models, one of the
708
00:41:57,465 --> 00:42:00,944
challenges that exist there is that,
you know, your features may be the
709
00:42:00,944 --> 00:42:04,665
same for two diverse populations, but
the way diagnoses are captured may
710
00:42:04,665 --> 00:42:08,265
be vastly different to the point that
we may underestimate the prevalence
711
00:42:08,265 --> 00:42:12,810
in a population of a certain disease
entity, or because of social stigma, it
712
00:42:12,855 --> 00:42:17,865
may actually present with a different
ICD-10 code than depression, right?
713
00:42:17,865 --> 00:42:23,805
They may have a series of somatic
presentations of, uh, mental illness
714
00:42:23,805 --> 00:42:28,935
that we should have built-in ways in
machine learning models that identify,
715
00:42:29,115 --> 00:42:33,825
hey, this person's now been seen for
palpitations, headache, abdominal pain,
716
00:42:33,915 --> 00:42:35,685
alternating diarrhea and constipation.
717
00:42:36,299 --> 00:42:38,850
Perhaps there's a mental health
issue that we need to address.
718
00:42:38,850 --> 00:42:43,170
Perhaps that should be built into the way
that we think about how this population
719
00:42:43,410 --> 00:42:47,700
in this geography or in this community
tends to represent the way it, you
720
00:42:47,700 --> 00:42:49,799
know, mental health gets represented.
721
00:42:50,220 --> 00:42:53,430
And so I think the data
equity piece is a big one.
722
00:42:53,640 --> 00:42:55,020
We need to understand that.
723
00:42:55,080 --> 00:42:59,919
We need to make sure that we're
considering the cultural context, we're
724
00:42:59,950 --> 00:43:02,140
considering broadened data sources.
725
00:43:02,470 --> 00:43:06,520
You know, not only some of the medical
data, but also some of the social data is
726
00:43:06,520 --> 00:43:08,379
really important for certain populations.
727
00:43:08,799 --> 00:43:12,250
Um, we wanna do even stratified
sampling within some of those
728
00:43:12,250 --> 00:43:16,120
populations to make sure that we're
doing things in a way that's fair
729
00:43:16,120 --> 00:43:17,950
from a, an algorithmic standpoint.
730
00:43:18,370 --> 00:43:21,160
And at times where there's
data that's missing, we need to
731
00:43:21,435 --> 00:43:25,755
understand disease prevalence in
certain populations and even do data
732
00:43:25,755 --> 00:43:29,715
augmentation until we get to a point
where data capture is seen as more
733
00:43:29,715 --> 00:43:31,206
equitable for certain communities.
734
00:43:32,126 --> 00:43:35,636
So you started your career on
the practice side and then made
735
00:43:36,101 --> 00:43:38,951
transition into clinical informatics,
but never really stepped away
736
00:43:38,951 --> 00:43:40,541
from the practice element either.
737
00:43:41,141 --> 00:43:48,371
So for other physicians who are interested
in getting more deeply involved with
738
00:43:48,671 --> 00:43:52,931
building digital healthcare technology
and wanna make that transition either
739
00:43:52,931 --> 00:43:57,821
partially or completely into the
healthcare technology world, what
740
00:43:57,941 --> 00:43:59,736
would you suggest they keep in mind?
741
00:44:01,001 --> 00:44:04,661
Medical education's evolved and has
needed to evolve because of how much
742
00:44:04,661 --> 00:44:06,521
technology that exists in healthcare.
743
00:44:06,971 --> 00:44:09,761
You know, I think that different
schools have evolved in different cases.
744
00:44:10,211 --> 00:44:13,481
Um, I would say that those students of
medicine that are kind of going through
745
00:44:13,506 --> 00:44:17,711
the, the, the programming now take the
opportunity to learn with people that
746
00:44:17,711 --> 00:44:19,121
are doing different things in healthcare.
747
00:44:19,481 --> 00:44:22,571
And they'll feel that because of the
fact that you went to medical school,
748
00:44:22,571 --> 00:44:25,961
that there's only one traditional way of
practicing medicine and delivering, you
749
00:44:25,961 --> 00:44:28,451
know, value from a healthcare standpoint.
750
00:44:28,481 --> 00:44:31,061
I think there's a lot that
needs to be fixed in healthcare.
751
00:44:31,391 --> 00:44:34,571
Um, there's a lot of policy
issues that that PD addressed.
752
00:44:34,691 --> 00:44:37,271
Um, there's a lot of the different
ways we transac healthcare.
753
00:44:37,571 --> 00:44:40,841
There's a lot of gaps when it comes
to care and ation, and it's one of the
754
00:44:40,841 --> 00:44:44,141
things that I was kind of mentioning
is previously, uh, patients don't know
755
00:44:44,141 --> 00:44:47,471
how you use healthcare in the United
States, like I had the opportunity to
756
00:44:47,471 --> 00:44:52,391
work in an urgent care center like it,
it was a mixture of people that truly
757
00:44:52,391 --> 00:44:55,631
needed to be there because they had
something that acquired urgent care.
758
00:44:56,036 --> 00:45:00,176
And then I would say the other half of
the people were people that probably
759
00:45:00,176 --> 00:45:04,256
either to be an emergency department and
passed six of them along the way, uh, in
760
00:45:04,256 --> 00:45:05,876
order to get to this urgent care center.
761
00:45:06,236 --> 00:45:08,486
And there were, there were people
that would come in and they had
762
00:45:08,486 --> 00:45:11,186
six months worth of the complaints
that probably could be better dealt
763
00:45:11,186 --> 00:45:12,356
with with a primary care doctor.
764
00:45:12,986 --> 00:45:14,906
People don't know how to
use the healthcare system.
765
00:45:14,906 --> 00:45:19,226
I think that we as clinicians can
be empowered as of getting people
766
00:45:19,226 --> 00:45:22,616
to, you know, use it in, in a way
that's contextually appropriate.
767
00:45:23,126 --> 00:45:26,786
And then obviously, you know, in many
of the conversations that we had, you've
768
00:45:26,786 --> 00:45:31,166
had, there's a gap in, in, in terms
of the resources that are necessary
769
00:45:31,166 --> 00:45:36,116
to provide care to whatever complex
and ever aging, you know, population.
770
00:45:36,116 --> 00:45:39,566
And so technology's gonna have
be the way to, to mitigate that.
771
00:45:39,566 --> 00:45:46,346
And so really identifying that you can
provide value with your clinical know-how,
772
00:45:46,616 --> 00:45:51,536
in a way that's non-traditional is really
very, very valuable for, for, you know,
773
00:45:51,536 --> 00:45:53,126
clinicians of medicine to understand.
774
00:45:54,656 --> 00:45:59,006
And then the opposite of that, for
all the non-clinicians in the audience
775
00:45:59,066 --> 00:46:04,946
who want to get involved in healthcare
technology and improving the way that
776
00:46:04,946 --> 00:46:09,206
we deliver care, whether it's in,
in this country or globally, where
777
00:46:09,206 --> 00:46:12,986
should they focus, you know, as they
build their foundational knowledge of
778
00:46:13,361 --> 00:46:16,421
clinical informatics and
interoperability in healthcare, AI?
779
00:46:17,291 --> 00:46:21,801
Some of the more successful companies,
um, when, when it comes to tech, new
780
00:46:21,821 --> 00:46:25,361
technology companies, some of the more
successful companies in healthcare
781
00:46:25,931 --> 00:46:30,161
are first and foremost the health
solutions company, and they have the,
782
00:46:30,191 --> 00:46:33,201
the, the key clinical pieces in place.
783
00:46:33,791 --> 00:46:35,381
Healthcare is nuanced, right?
784
00:46:35,381 --> 00:46:40,421
It's, it's not like there's a binary
outcome for every compilation of tests.
785
00:46:40,811 --> 00:46:44,501
Not every test combination
adds up to the same diagnosis.
786
00:46:44,501 --> 00:46:49,241
There's a lot of many to one, one to many
relationships that exist in healthcare.
787
00:46:49,511 --> 00:46:52,721
There's a lot of interrelationships
that exist and there's no one
788
00:46:52,721 --> 00:46:54,281
way that a diagnosis presents.
789
00:46:54,281 --> 00:46:56,111
There's a lot of nuance
to how they presents.
790
00:46:56,141 --> 00:47:01,426
There's not a one size fits all treatment
when it comes to, you know, hypertension.
791
00:47:01,806 --> 00:47:07,181
We often look at people with hypertension
who have other comorbid illnesses, may
792
00:47:07,181 --> 00:47:11,036
respond to different anti-hypertensive
medications, may actually have those
793
00:47:11,036 --> 00:47:14,456
anti-hypertensive medications treat
some of their other comorbid illnesses
794
00:47:14,666 --> 00:47:16,406
if the right medication is selected.
795
00:47:16,766 --> 00:47:22,436
So as people that are studying technology,
studying data science, studying machine
796
00:47:22,436 --> 00:47:27,386
learning, AI, et cetera, understand
that medicine is binary, it's nuanced.
797
00:47:27,446 --> 00:47:30,956
Um, there's a lot of drivers of medicine
that are not just simply, purely based
798
00:47:30,986 --> 00:47:34,586
on, you know, how someone presents what
diagnosis they have of what treatment
799
00:47:34,586 --> 00:47:38,606
they receive, but there's a whole lot of
interplay from social factors, et cetera.
800
00:47:39,056 --> 00:47:45,056
Having clinicians on board to learn,
knowing, knowing how to talk the talk and
801
00:47:45,056 --> 00:47:48,806
walk the walk of, of, of the clinician
is important to adoption, right?
802
00:47:48,806 --> 00:47:53,311
Because there are stakeholders that
oftentimes may get overlooked and these,
803
00:47:53,576 --> 00:47:57,416
these technology solutions get implemented
in these, you know, hospital systems and
804
00:47:57,416 --> 00:48:01,136
provider offices, et cetera, and they're
the ones that could be part of your
805
00:48:01,136 --> 00:48:03,066
greatest success on a go forward basis.
806
00:48:03,066 --> 00:48:09,791
So, you know, I would say understand
the limitations of AI, the biases
807
00:48:09,791 --> 00:48:14,171
that exist in AI for something as
convoluted as as the medical sciences
808
00:48:14,561 --> 00:48:16,751
and learn from those medical scientists.
809
00:48:16,781 --> 00:48:20,981
The healthiest organizations I've been
a part of, I've taught as much medicine
810
00:48:20,981 --> 00:48:22,661
as I've, as I've learned data science.
811
00:48:22,991 --> 00:48:25,631
I've taught as much medicine as
I've learned standards, right?
812
00:48:25,991 --> 00:48:30,251
Um, and I think that that bi-directional
exchange of information really makes
813
00:48:30,251 --> 00:48:34,431
growing organizations and really
deliver, delivers the best ROIs
814
00:48:35,561 --> 00:48:40,541
So, uh, if our listeners could take, uh,
one thing away from this conversation
815
00:48:40,541 --> 00:48:43,451
that we've had over the last hour or
so, what would you like that to be?
816
00:48:44,831 --> 00:48:46,271
So, I think the time is now.
817
00:48:46,541 --> 00:48:50,441
Setting aside adoption, setting aside,
you know, some of the challenges that
818
00:48:50,441 --> 00:48:55,601
exist from the bias standpoint, setting
aside some of the, uh, challenges that
819
00:48:55,841 --> 00:49:02,471
exists from patients and providers
and technology solutions, et cetera.
820
00:49:03,176 --> 00:49:07,406
We need to reform the way healthcare
is delivered in a way that's equitable,
821
00:49:07,406 --> 00:49:09,986
in a way that's thoughtful and
a way makes the right diagnosis,
822
00:49:09,986 --> 00:49:11,126
prescribes the right treatment.
823
00:49:11,546 --> 00:49:14,186
At the end of the day, technology
enablement are really related.
824
00:49:14,816 --> 00:49:21,326
Um, and having the right organizational
mixture of clinicians and technologists
825
00:49:21,536 --> 00:49:26,881
at a company is vital key to success,
not only of that company, but to the
826
00:49:26,881 --> 00:49:28,081
entire healthcare delivery model.
827
00:49:28,331 --> 00:49:28,991
Thank you very much.
828
00:49:30,431 --> 00:49:31,011
Dr. Pinho.
829
00:49:31,031 --> 00:49:33,161
Thank you so much for being
with us this afternoon.
830
00:49:33,551 --> 00:49:36,521
I think that this conversation is
gonna provide a lot of great context
831
00:49:36,521 --> 00:49:38,981
for listeners as they continue
their own journeys in health IT..
832
00:49:40,241 --> 00:49:43,451
This has been Hard Problems, Smart
Solutions, the Newfire podcast.
833
00:49:43,841 --> 00:49:46,571
Thanks for listening and come back
next time when we'll be diving
834
00:49:46,571 --> 00:49:50,051
into interoperability strategy and
digital health product management
835
00:49:50,291 --> 00:49:51,671
with podcaster Omar Muza.
