1
00:00:00,275 --> 00:00:02,485
- This is Chris SOA with the
Becker's Healthcare Podcast.

2
00:00:03,185 --> 00:00:04,365
I'm thrilled to be joined by Dr.

3
00:00:04,645 --> 00:00:07,725
Hamel gda, president of the
UPMC Heart Vascular Institute.

4
00:00:08,355 --> 00:00:09,975
Dr. gta, thank you for joining us today.

5
00:00:11,275 --> 00:00:12,815
- Oh, thank you so much,
Chris, for having me.

6
00:00:14,085 --> 00:00:15,825
- Of course. Could you
please introduce yourself

7
00:00:15,825 --> 00:00:16,985
and tell us about the Heart Vascular

8
00:00:16,985 --> 00:00:18,065
Institute in particular?

9
00:00:19,485 --> 00:00:22,415
- Yeah, I am, uh,
interventional cardiologist

10
00:00:22,595 --> 00:00:26,175
and so I'm, I'm used to doing
less invasive, uh, heart

11
00:00:26,435 --> 00:00:30,055
and vascular procedures on
patients that require, uh,

12
00:00:30,255 --> 00:00:33,095
a whole host of different
levels of management.

13
00:00:33,195 --> 00:00:38,095
Uh, but I oversee a very
specific sector of the UPMC,

14
00:00:38,095 --> 00:00:40,535
heart and Vascular Institute
here in central Pennsylvania.

15
00:00:41,235 --> 00:00:43,855
And, uh, we're located
with our hub in Harrisburg,

16
00:00:43,955 --> 00:00:48,295
but we have seven hospitals
that kind of serve the hub,

17
00:00:48,755 --> 00:00:53,095
uh, in about a 50 mile radius
around Harrisburg proper.

18
00:00:53,755 --> 00:00:57,935
In addition to that, we have
20 plus cardiology offices in

19
00:00:58,175 --> 00:01:01,535
that area, and we're part
of the larger UPMC system,

20
00:01:01,535 --> 00:01:03,975
which is of course based in
Pittsburgh, Pennsylvania.

21
00:01:04,515 --> 00:01:07,975
Uh, but we have kind of tentacles
that reach north, south,

22
00:01:08,085 --> 00:01:10,735
east, and west all throughout
the state of Pennsylvania.

23
00:01:11,115 --> 00:01:13,495
So we're very large healthcare system, one

24
00:01:13,495 --> 00:01:14,935
of the largest in the United States.

25
00:01:17,055 --> 00:01:19,915
- Doctor, thanks for
detailing that. I should say.

26
00:01:20,375 --> 00:01:23,395
Uh, today we have you on talk
about cardiovascular disease

27
00:01:23,395 --> 00:01:25,515
and how that sorts with ai.

28
00:01:25,705 --> 00:01:28,755
Alright. My first question
would be why is early detection

29
00:01:28,755 --> 00:01:30,795
of cardiovascular disease so important

30
00:01:31,055 --> 00:01:32,675
and can you briefly explain how AI helps

31
00:01:32,675 --> 00:01:34,355
cardiologists with early detection?

32
00:01:35,865 --> 00:01:38,155
- Yeah, I think we're, we're
kind of framing the discussion

33
00:01:38,255 --> 00:01:39,395
to talk more about

34
00:01:39,625 --> 00:01:41,635
what would be considered
preventative healthcare,

35
00:01:41,735 --> 00:01:45,115
and of course a problem
is best extinguished

36
00:01:45,215 --> 00:01:46,355
if you never have it.

37
00:01:46,895 --> 00:01:50,115
So we here at UPMC are
committed to working continually

38
00:01:50,575 --> 00:01:52,315
to eliminate or reduce the risk

39
00:01:52,315 --> 00:01:53,875
of preventable harm to our patients.

40
00:01:54,535 --> 00:01:57,035
Um, we definitely take a
proactive approach across our

41
00:01:57,035 --> 00:01:59,715
healthcare system that focuses
on creating a care plan

42
00:02:00,105 --> 00:02:03,235
with the best possible
outcomes for all patients.

43
00:02:03,875 --> 00:02:07,295
Cardiovascular diseases
clearly a very large

44
00:02:07,295 --> 00:02:08,335
threat to human health.

45
00:02:08,685 --> 00:02:10,775
It's the leading cause
of death nationwide.

46
00:02:10,775 --> 00:02:12,535
And of course in the
central Pennsylvania region

47
00:02:12,535 --> 00:02:14,215
that we service, um,

48
00:02:14,345 --> 00:02:16,815
heart disease deaths nationwide increased

49
00:02:17,275 --> 00:02:19,935
and have been increasing
on an annual basis since

50
00:02:19,955 --> 00:02:21,095
the late 20 teens.

51
00:02:21,875 --> 00:02:24,495
Um, so we need better
strategies to diagnose

52
00:02:24,495 --> 00:02:26,455
and treat cardiovascular
disease in the future.

53
00:02:27,095 --> 00:02:30,535
I think artificial intelligence
detection algorithms are

54
00:02:30,535 --> 00:02:33,815
going to have a lot of
potential to solve this problem.

55
00:02:34,595 --> 00:02:38,655
And so really the whole
concept here is to use the data

56
00:02:38,655 --> 00:02:40,935
that we have in the
electronic medical record,

57
00:02:41,895 --> 00:02:43,535
identify patients at risk

58
00:02:43,755 --> 00:02:46,815
of certain cardiovascular
conditions, inaugurate

59
00:02:46,835 --> 00:02:48,735
and advance their care within the system

60
00:02:49,035 --> 00:02:51,735
to prevent them from
having burdensome outcomes.

61
00:02:54,815 --> 00:02:56,245
- Thank you for that Dr. Gada.

62
00:02:56,315 --> 00:02:57,805
Next, uh, I would ask you,

63
00:02:57,905 --> 00:03:00,965
can you tell us about the AI
tools that you're using at UPMC

64
00:03:01,225 --> 00:03:02,525
and explain how they're used?

65
00:03:04,105 --> 00:03:07,635
- Yeah. We've used several
different, um, AI tools,

66
00:03:07,695 --> 00:03:10,795
if you will, and a lot of
them are using our electronic

67
00:03:10,795 --> 00:03:13,755
medical record and all of
the power that it gives us

68
00:03:13,775 --> 00:03:17,795
to really look into our
individual health profiles as far

69
00:03:17,795 --> 00:03:19,115
as our patients are concerned

70
00:03:19,495 --> 00:03:22,915
and discern any risks that
they have that would lead them

71
00:03:22,975 --> 00:03:24,955
to develop more severe conditions.

72
00:03:25,335 --> 00:03:28,395
And some of that is just
disease detection, if you will.

73
00:03:28,455 --> 00:03:30,435
And so, uh, I work a lot

74
00:03:30,465 --> 00:03:32,315
with valvular heart disease patients.

75
00:03:32,635 --> 00:03:34,355
Patients that have problems
with their heart valves

76
00:03:34,855 --> 00:03:38,675
and detecting these problems
before they cause symptoms

77
00:03:38,975 --> 00:03:42,515
or in patients who have
milder symptoms is something

78
00:03:42,515 --> 00:03:45,555
that's not overtly done
with clinical visits.

79
00:03:45,735 --> 00:03:47,755
You know, primary care providers,

80
00:03:47,785 --> 00:03:49,355
cardiologists are really strapped

81
00:03:49,355 --> 00:03:52,555
for time in these visits slots
that are literally like 15

82
00:03:52,555 --> 00:03:53,675
to 20 minutes long.

83
00:03:54,335 --> 00:03:57,675
And so having a very powerful
electronic medical record like

84
00:03:57,675 --> 00:03:59,155
we have that is able

85
00:03:59,155 --> 00:04:02,755
to garner information
from cardiac ultrasounds

86
00:04:02,985 --> 00:04:06,195
that the patient may have had
other types of ultrasounds

87
00:04:06,375 --> 00:04:10,555
or non uh, invasive assessments,
we take that information

88
00:04:11,175 --> 00:04:13,315
and we're able to basically qualify

89
00:04:13,655 --> 00:04:17,075
or quantify the severity
of certain valve diseases.

90
00:04:17,415 --> 00:04:20,155
And so then we're able to get
those patients seen by one

91
00:04:20,155 --> 00:04:23,595
of our providers earlier than
they otherwise would have

92
00:04:24,055 --> 00:04:26,755
and advance therapeutic options in a way

93
00:04:27,025 --> 00:04:29,035
that they wouldn't have
had, you know, the advantage

94
00:04:29,135 --> 00:04:33,835
to partake in if it wasn't
for that AI algorithm kind

95
00:04:33,835 --> 00:04:36,715
of kicking in for this
type of disease detection.

96
00:04:38,215 --> 00:04:41,235
- Gotcha. Uh, looking a
little further into this, Dr.

97
00:04:41,585 --> 00:04:44,005
Gada, could you tell us
about some of the patient

98
00:04:44,605 --> 00:04:45,565
outcomes you've seen while using

99
00:04:45,565 --> 00:04:46,765
these machine learning models?

100
00:04:48,475 --> 00:04:50,795
- I think what we've noticed,
Chris, is that we've been able

101
00:04:50,795 --> 00:04:53,275
to avoid emergent

102
00:04:53,375 --> 00:04:55,755
and urgent hospitalizations in,

103
00:04:55,815 --> 00:04:57,435
in a really, you know, great way.

104
00:04:57,935 --> 00:05:00,755
And so when we started using
the electronic medical record

105
00:05:00,905 --> 00:05:03,035
like this, and this is now over five years

106
00:05:03,265 --> 00:05:05,275
that we've been doing, you
probably have more experience

107
00:05:05,505 --> 00:05:07,035
with electronic medical records

108
00:05:07,215 --> 00:05:09,795
and artificial intelligence
therein than any other health

109
00:05:09,795 --> 00:05:13,075
system in the state of
Pennsylvania that we've been able

110
00:05:13,095 --> 00:05:16,595
to avoid some harmful consequences such

111
00:05:16,595 --> 00:05:18,715
as sudden emergency department visits,

112
00:05:19,515 --> 00:05:21,315
patients developing critical conditions

113
00:05:21,825 --> 00:05:24,315
that would've been prevented
had they been detected

114
00:05:24,315 --> 00:05:25,675
earlier in milder forms.

115
00:05:26,175 --> 00:05:29,195
And so being able to
stave off those emergent

116
00:05:29,375 --> 00:05:31,755
life-threatening presentations is kind

117
00:05:31,755 --> 00:05:33,395
of the most important, I think, benefit

118
00:05:33,505 --> 00:05:36,075
that we've garnered from our artificial

119
00:05:36,075 --> 00:05:37,515
intelligence algorithms here.

120
00:05:37,935 --> 00:05:40,275
Uh, but in addition to
that, you know, kind of

121
00:05:41,395 --> 00:05:45,015
making the patient aware
that we are always on top

122
00:05:45,015 --> 00:05:48,175
of their health and are using
all the tools that we can

123
00:05:48,995 --> 00:05:51,615
to make sure that they get the
best longitudinal follow up.

124
00:05:51,955 --> 00:05:55,215
You know, they don't necessarily
need to see us in person

125
00:05:55,875 --> 00:05:58,685
in order for us to be able
to track what's going on

126
00:05:59,115 --> 00:06:00,645
with some aspects of their health.

127
00:06:01,105 --> 00:06:04,485
And so that lag between
getting diagnostic information

128
00:06:05,115 --> 00:06:08,005
from a test and having
that conveyed to a patient

129
00:06:08,385 --> 00:06:09,645
and then a follow-up visit

130
00:06:09,735 --> 00:06:12,045
where the patient hears
more about it, we can make

131
00:06:12,045 --> 00:06:16,085
that much more efficient if
we've got flags in their chart

132
00:06:16,185 --> 00:06:20,125
and in the electronic medical
record that alert us to

133
00:06:20,885 --> 00:06:23,445
critical, severe, moderate
conditions that they develop

134
00:06:23,865 --> 00:06:26,085
and then we can kind of
troubleshoot those right away

135
00:06:26,465 --> 00:06:28,765
and save the patient and
the system a lot of time

136
00:06:29,185 --> 00:06:30,845
by getting those patients seen earlier

137
00:06:31,185 --> 00:06:33,445
and getting them triaged more effectively.

138
00:06:35,035 --> 00:06:37,365
- Yeah, as you said, doctor,
hopefully patients well,

139
00:06:37,365 --> 00:06:39,765
I imagine patients are very
grateful when they're able to,

140
00:06:40,425 --> 00:06:43,675
uh, learn about what is needed

141
00:06:43,775 --> 00:06:45,515
for their treatment
sooner rather than later.

142
00:06:45,895 --> 00:06:47,755
Uh, the last thing I'll leave you with Dr.

143
00:06:47,905 --> 00:06:50,475
Gada is what in your field you
think deserves the brightest

144
00:06:50,635 --> 00:06:51,675
spotlight than it's getting at the moment?

145
00:06:53,525 --> 00:06:55,165
- I, I think it would be that, and,

146
00:06:55,165 --> 00:06:58,045
and I think that the pathways
of care that have kind

147
00:06:58,045 --> 00:07:00,605
of been historic and have led to a lot

148
00:07:00,605 --> 00:07:01,965
of inefficiency in healthcare

149
00:07:02,505 --> 00:07:06,285
and maybe a lot of frustration
both from providers as well

150
00:07:06,285 --> 00:07:09,045
as the patients that we can stem a lot of,

151
00:07:09,045 --> 00:07:11,565
that we can cure a lot of
that, alleviate all that burden

152
00:07:12,025 --> 00:07:16,365
and create more efficient way
of, of providing healthcare

153
00:07:17,185 --> 00:07:21,165
by invoking these
artificial intelligence, uh,

154
00:07:21,165 --> 00:07:22,965
type approaches, especially those

155
00:07:22,965 --> 00:07:24,845
that are embedded in our
electronic medical record

156
00:07:25,065 --> 00:07:28,085
for early detection of critical diseases

157
00:07:28,345 --> 00:07:29,605
or burdensome conditions

158
00:07:30,025 --> 00:07:32,645
and being able to get patients
through the system quicker.

159
00:07:33,065 --> 00:07:35,485
And so I do think that we need
to put more emphasis on that.

160
00:07:36,025 --> 00:07:37,045
And honestly, I think

161
00:07:37,045 --> 00:07:40,165
that our payment policy
structure should be engineered

162
00:07:40,625 --> 00:07:43,925
to provide, you know,
further, I would say impetus

163
00:07:44,025 --> 00:07:45,045
for physicians

164
00:07:45,105 --> 00:07:47,325
and for healthcare systems
to partake in that.

165
00:07:47,905 --> 00:07:49,005
Um, this is something

166
00:07:49,005 --> 00:07:51,285
that I don't think is being
done adequately at the moment.

167
00:07:51,665 --> 00:07:53,725
And so there needs to
be adequate incentive

168
00:07:54,185 --> 00:07:56,885
for healthcare systems to invest time

169
00:07:57,265 --> 00:08:00,085
and for physicians to
invest time in order to

170
00:08:00,695 --> 00:08:02,805
inaugurate these types of, of technologies

171
00:08:03,145 --> 00:08:05,205
and again, make, uh, better healthcare,

172
00:08:05,395 --> 00:08:07,565
more efficient healthcare
for our patient populations.

173
00:08:09,125 --> 00:08:11,105
- Dr. Got it. Thank you
for being so generous

174
00:08:11,105 --> 00:08:12,545
with your time and insights today.

175
00:08:12,645 --> 00:08:14,425
We cannot wait to share
them with our audience

176
00:08:14,425 --> 00:08:16,065
and we look forward to
connecting with you soon.

177
00:08:17,035 --> 00:08:18,895
- Thanks so much, Chris.
I really enjoyed this.

178
00:08:21,685 --> 00:08:24,095
- It's so important for leaders
at the top of organizations

179
00:08:24,095 --> 00:08:26,575
to keep learning, stay
sharp, grow their networks,

180
00:08:27,045 --> 00:08:29,655
help our audience better do
this in a more simplified,

181
00:08:29,655 --> 00:08:31,335
personalized, and meaningful way.

182
00:08:31,895 --> 00:08:34,375
Becker's Healthcare has launched my BHC,

183
00:08:34,845 --> 00:08:36,855
it's your trusted Becker's
healthcare experience

184
00:08:36,855 --> 00:08:39,015
and more with content, connections, events

185
00:08:39,075 --> 00:08:40,095
and learning opportunities.

186
00:08:40,805 --> 00:08:41,855
Join the community free

187
00:08:41,855 --> 00:08:46,695
of charge@www.my dot
becker's hospital review.com

188
00:08:46,795 --> 00:08:47,655
and we'll see you there.