1
00:00:00,640 --> 00:00:03,199
Philips is a health tech leader focused on

2
00:00:03,199 --> 00:00:06,000
innovation that improves the health and well-being of

3
00:00:06,000 --> 00:00:09,460
people. Our health care technology and informatics solutions

4
00:00:09,679 --> 00:00:12,559
help care teams diagnose, treat, and manage more

5
00:00:12,559 --> 00:00:16,244
patients with greater precision, speed, and confidence across

6
00:00:16,244 --> 00:00:19,225
the care journey. With Philips, clinicians are empowered

7
00:00:19,285 --> 00:00:21,765
with streamlined insights in the moments that matter

8
00:00:21,765 --> 00:00:24,904
for every patient. Better care for more people.

9
00:00:25,204 --> 00:00:25,704
Philips.

10
00:00:28,670 --> 00:00:30,910
This is Gracelyn Keller with the Becker's Healthcare

11
00:00:30,910 --> 00:00:33,149
podcast, and I am excited now to be

12
00:00:33,149 --> 00:00:35,229
joined by doctor Joseph Izzo, who is the

13
00:00:35,229 --> 00:00:38,030
chief medical information officer at San Joaquin General

14
00:00:38,030 --> 00:00:40,350
Hospital. So, doctor Izzo, thank you for joining

15
00:00:40,350 --> 00:00:41,090
me today.

16
00:00:41,465 --> 00:00:43,545
Let's get started by having you please share

17
00:00:43,545 --> 00:00:45,625
a bit about yourself and your background and

18
00:00:45,625 --> 00:00:47,004
your role in your organization.

19
00:00:47,945 --> 00:00:50,104
Yeah. Thank you for having me. So as

20
00:00:50,104 --> 00:00:52,344
you said, I'm the chief medical information officer

21
00:00:52,344 --> 00:00:53,324
for San Joaquin.

22
00:00:53,880 --> 00:00:55,640
We're a level 2 trauma center and a

23
00:00:55,640 --> 00:00:57,659
public hospital in the Central Valley of California.

24
00:00:58,120 --> 00:01:00,280
I'm also a practicing physician. I've worked in

25
00:01:00,280 --> 00:01:00,780
the

26
00:01:01,240 --> 00:01:03,640
emergency department for over 6 years. And I

27
00:01:03,640 --> 00:01:05,655
actually got my start, in computer science before

28
00:01:05,655 --> 00:01:07,974
I transitioned into medicine. And since that, I've

29
00:01:07,974 --> 00:01:10,295
been really working at the intersection between health

30
00:01:10,295 --> 00:01:11,034
care and technology.

31
00:01:11,495 --> 00:01:13,974
But particularly to my role, I oversee the

32
00:01:13,974 --> 00:01:16,855
health informatics and the clinical reporting teams. And

33
00:01:16,855 --> 00:01:18,454
you might say that our efforts are mostly

34
00:01:18,454 --> 00:01:22,140
directed at optimizing how clinicians engage with technology,

35
00:01:22,359 --> 00:01:25,479
fostering innovation, and and really ensuring clinical data

36
00:01:25,479 --> 00:01:27,239
stewardship. So I I do hope with that

37
00:01:27,239 --> 00:01:29,079
that my insights are valuable to the listeners.

38
00:01:29,079 --> 00:01:30,859
And again, I I appreciate you having me.

39
00:01:31,575 --> 00:01:33,575
Absolutely. Well, thank you for taking the time.

40
00:01:33,814 --> 00:01:36,614
And let's start our conversation today with AI

41
00:01:36,614 --> 00:01:38,694
adoption. So as we know, this is exploding

42
00:01:38,694 --> 00:01:40,694
in health care right now. So in your

43
00:01:40,694 --> 00:01:42,854
view, what is the most significant or promising

44
00:01:42,854 --> 00:01:44,314
application of this technology,

45
00:01:44,700 --> 00:01:47,259
and how is this informing your organization's innovation

46
00:01:47,259 --> 00:01:47,759
strategy?

47
00:01:49,099 --> 00:01:51,739
Right. You know, AI to me is excelling

48
00:01:51,739 --> 00:01:52,399
in spaces

49
00:01:52,700 --> 00:01:55,039
that are traditionally considered, I think, the administrative

50
00:01:55,099 --> 00:01:56,700
burden of medicine, especially,

51
00:01:57,665 --> 00:01:59,344
these areas that are very important to clinical

52
00:01:59,344 --> 00:02:00,784
medicine but are far enough,

53
00:02:01,665 --> 00:02:03,185
away from patient care where I think the

54
00:02:03,185 --> 00:02:04,784
margin for error is a little bit more

55
00:02:04,784 --> 00:02:06,385
forgiving. You know, I I don't think the

56
00:02:06,385 --> 00:02:07,905
tools are quite where they need to be

57
00:02:07,905 --> 00:02:09,584
right now for direct patient care. And, of

58
00:02:09,584 --> 00:02:11,425
course, not everyone may feel that way, and

59
00:02:11,425 --> 00:02:14,239
that's certainly an institutional decision. But these areas

60
00:02:14,299 --> 00:02:16,780
are traditionally burdensome to providers too, so it's

61
00:02:16,780 --> 00:02:18,219
it's a bit of a win win. I

62
00:02:18,219 --> 00:02:20,620
think the most pressing example that everyone's familiar

63
00:02:20,620 --> 00:02:22,699
with is AI scribes, but I will say

64
00:02:22,699 --> 00:02:24,299
that that really is the case. You know?

65
00:02:24,299 --> 00:02:26,085
In 2 of the depart emergency departments

66
00:02:27,205 --> 00:02:29,145
I work in, and

67
00:02:29,525 --> 00:02:29,661
even in our pilots at the hospital, the

68
00:02:29,661 --> 00:02:31,045
feedback is is just tremendous. You know, I've

69
00:02:31,045 --> 00:02:33,205
received paragraphs of gratitude from some of the

70
00:02:33,205 --> 00:02:35,365
physicians who have piloted the products, and it's

71
00:02:35,365 --> 00:02:36,724
more or less a day and night diff

72
00:02:36,805 --> 00:02:39,685
difference, for how their their charting changes and

73
00:02:39,685 --> 00:02:41,865
and how much administrative burden is on them.

74
00:02:42,389 --> 00:02:44,150
But I've also been really impressed with some

75
00:02:44,150 --> 00:02:46,790
of the recent work in in, medical record

76
00:02:46,790 --> 00:02:48,010
space, revenue cycle,

77
00:02:48,469 --> 00:02:52,469
clinical documentation improving, even automating nursing processes, and

78
00:02:52,469 --> 00:02:54,735
we've been exploring that at our hospital. And

79
00:02:54,895 --> 00:02:56,574
separately too from my role at San Joaquin,

80
00:02:56,574 --> 00:02:58,895
I've done some fractional work with large language

81
00:02:58,895 --> 00:03:01,534
models and agentic AI, and I've seen tremendous

82
00:03:01,534 --> 00:03:03,694
success too in areas of quality and safety

83
00:03:03,694 --> 00:03:05,775
auditing. And I've what I've noticed in all

84
00:03:05,775 --> 00:03:07,375
of these spaces, what it's doing is freeing

85
00:03:07,375 --> 00:03:09,694
up, staff and clinicians to do more creative

86
00:03:09,694 --> 00:03:11,420
and patient focused tasks.

87
00:03:11,879 --> 00:03:13,639
But, to your other question, for us to

88
00:03:13,639 --> 00:03:15,560
stay strategic, we really first had to start

89
00:03:15,560 --> 00:03:18,620
with, a robust executive governance structure and policy,

90
00:03:18,840 --> 00:03:20,360
and we really do vet any of these

91
00:03:20,360 --> 00:03:22,920
AI tools, for the reasons I mentioned. And

92
00:03:22,920 --> 00:03:24,935
I I personally authored the policy for a

93
00:03:24,935 --> 00:03:26,935
hospital, and I've been very proactive in getting

94
00:03:26,935 --> 00:03:28,954
clinician leaders involved in the dialogue,

95
00:03:29,254 --> 00:03:31,974
particularly with a physician innovation group who vets

96
00:03:31,974 --> 00:03:33,974
the products, try to attend many of the

97
00:03:33,974 --> 00:03:35,674
vendor conferences, including Becker's.

98
00:03:36,189 --> 00:03:38,509
Take attend local meetups and societies, and really

99
00:03:38,509 --> 00:03:39,949
stay informed so that we can make the

100
00:03:39,949 --> 00:03:42,030
best decisions on this technology. But either way,

101
00:03:42,030 --> 00:03:44,289
our strategy is that we are moving forward,

102
00:03:44,349 --> 00:03:46,109
and we wanna take advantage of these tools,

103
00:03:46,430 --> 00:03:47,250
that are promising.

104
00:03:48,364 --> 00:03:48,864
Absolutely.

105
00:03:49,324 --> 00:03:51,245
And kind of going off of this AI

106
00:03:51,245 --> 00:03:52,144
AI adoption,

107
00:03:52,444 --> 00:03:55,564
technological advancements within health care, we see that

108
00:03:55,564 --> 00:03:58,364
leaders are managing greater volumes of data and

109
00:03:58,364 --> 00:04:00,844
more devices on a growing number of care

110
00:04:00,844 --> 00:04:03,759
settings and populations. And this is a complex

111
00:04:03,900 --> 00:04:06,780
environment. So what clinical data integration tools or

112
00:04:06,780 --> 00:04:08,959
practices are you seeing drive improvements

113
00:04:09,580 --> 00:04:12,139
in patient outcomes? And can you share some

114
00:04:12,139 --> 00:04:13,360
examples of this?

115
00:04:14,219 --> 00:04:16,274
Yeah. Certainly. The truth is, I I think

116
00:04:16,274 --> 00:04:18,595
our organization is is actually limited in this

117
00:04:18,595 --> 00:04:20,355
space with very small teams of analysts. You

118
00:04:20,355 --> 00:04:21,954
know, we're a public hospital, like I mentioned,

119
00:04:21,954 --> 00:04:24,115
and that's not too uncommon from talking to

120
00:04:24,115 --> 00:04:25,735
other individuals from those spaces.

121
00:04:26,194 --> 00:04:28,355
And I end up, helping out with a

122
00:04:28,355 --> 00:04:30,514
fair number of the coding, the report writing

123
00:04:30,514 --> 00:04:31,014
myself.

124
00:04:31,370 --> 00:04:32,649
But on the plus side, I think that

125
00:04:32,649 --> 00:04:34,169
means that our teams are really close knit

126
00:04:34,169 --> 00:04:35,930
and really understand the data and are able

127
00:04:35,930 --> 00:04:37,050
to make changes so,

128
00:04:37,529 --> 00:04:39,629
that we do drive these improvements you mentioned.

129
00:04:40,009 --> 00:04:41,930
And in that context, what's what's really worked

130
00:04:41,930 --> 00:04:42,990
for us is being

131
00:04:43,495 --> 00:04:44,235
quite intentional,

132
00:04:44,535 --> 00:04:47,014
about reconciling all of these sources of data

133
00:04:47,014 --> 00:04:47,834
that you mentioned

134
00:04:48,214 --> 00:04:49,754
and kind of ingesting them

135
00:04:50,375 --> 00:04:52,294
into really only 1 or 2 sources of

136
00:04:52,294 --> 00:04:54,634
of truth. You know, if there's a census

137
00:04:54,694 --> 00:04:56,555
that says a 170 patients

138
00:04:56,935 --> 00:04:59,050
were at in the hospital at any given

139
00:04:59,050 --> 00:05:00,970
moment, it should be true across any data

140
00:05:00,970 --> 00:05:01,930
we query. And,

141
00:05:02,649 --> 00:05:04,810
for us, we do use our our EMR's

142
00:05:04,810 --> 00:05:07,449
cloud enterprise data warehouse for that purpose. But

143
00:05:07,449 --> 00:05:09,610
I've seen other successful strategies where,

144
00:05:10,035 --> 00:05:11,475
some of our clinics have done their own

145
00:05:11,475 --> 00:05:14,115
homegrown versions and similar efforts like that. But

146
00:05:14,115 --> 00:05:15,634
like you mentioned, there there's so many,

147
00:05:16,194 --> 00:05:18,275
different areas to pull data from and and,

148
00:05:18,595 --> 00:05:20,915
sources of data. And another thing in that

149
00:05:20,915 --> 00:05:22,675
in that vein is is actually being really

150
00:05:22,675 --> 00:05:23,814
realistic with workflows

151
00:05:24,115 --> 00:05:26,529
because I I think, everyone's familiar that most

152
00:05:26,529 --> 00:05:28,449
EMRs have a number of ways to maybe

153
00:05:28,449 --> 00:05:30,449
document or perform the same task. And so

154
00:05:30,449 --> 00:05:32,529
if there's, say, 5 different ways to document

155
00:05:32,529 --> 00:05:35,029
a fall risk, then that's effectively 5 potentially

156
00:05:35,410 --> 00:05:37,250
different locations on the back end to find

157
00:05:37,250 --> 00:05:40,495
that data. And for us, standardizing these workflows

158
00:05:40,495 --> 00:05:41,634
has been really important.

159
00:05:42,495 --> 00:05:44,495
Now, of course, there's a delicate balance between

160
00:05:44,495 --> 00:05:46,915
forcing folks to do something that is laborious,

161
00:05:46,975 --> 00:05:49,694
but produces the perfect data, and then optimizing

162
00:05:49,694 --> 00:05:51,615
their workflows. And I think we've learned really

163
00:05:51,615 --> 00:05:52,675
to walk that line.

164
00:05:53,375 --> 00:05:55,620
The example that really comes to mind in

165
00:05:55,620 --> 00:05:57,939
in this space is actually about using interpreter

166
00:05:57,939 --> 00:05:58,439
service,

167
00:05:58,819 --> 00:06:00,740
for our patient. And, you know, we know

168
00:06:00,740 --> 00:06:02,339
we're doing it, but a problem is really

169
00:06:02,339 --> 00:06:04,500
being able to demonstrate that we are, I

170
00:06:04,500 --> 00:06:06,100
say, for the case of auditing or for

171
00:06:06,100 --> 00:06:08,384
quality improvement. Because like you mentioned, this is

172
00:06:08,384 --> 00:06:10,865
so much disparate data. There's data from the

173
00:06:10,865 --> 00:06:13,105
health plan. There's data from interpreter and service

174
00:06:13,105 --> 00:06:16,644
networks, from registration, from nursing notes, from physician,

175
00:06:17,425 --> 00:06:19,665
progress notes, and so on. And however, by

176
00:06:19,665 --> 00:06:21,839
ingesting this all into one place, and then

177
00:06:21,839 --> 00:06:23,060
coming up with standard,

178
00:06:23,519 --> 00:06:26,479
reasonable, but expected workflows, we've actually been able

179
00:06:26,479 --> 00:06:27,779
to address this problem

180
00:06:28,479 --> 00:06:30,339
and make very good visualizations

181
00:06:30,800 --> 00:06:32,719
for the units to see how their staff

182
00:06:32,719 --> 00:06:34,855
are at documenting this data. And we've seen

183
00:06:34,855 --> 00:06:38,394
tremendous progress. Some departments actually went from ostensibly

184
00:06:38,535 --> 00:06:40,935
very low compliance in in in documenting that

185
00:06:40,935 --> 00:06:43,014
they were using interpreter service to near a

186
00:06:43,014 --> 00:06:45,194
100% just by able to have visibility

187
00:06:45,574 --> 00:06:47,750
and by having continuous quality improvement. And the

188
00:06:47,750 --> 00:06:49,189
truth is a lot of our efforts are

189
00:06:49,189 --> 00:06:50,490
structured just like this.

190
00:06:51,509 --> 00:06:52,009
Absolutely.

191
00:06:52,710 --> 00:06:54,870
And kind of going off of both of

192
00:06:54,870 --> 00:06:55,850
those points,

193
00:06:56,230 --> 00:06:57,509
I would love to know how you think

194
00:06:57,509 --> 00:07:00,089
that health care organizations can better support

195
00:07:00,574 --> 00:07:02,975
IT and clinical teams as we carry out

196
00:07:02,975 --> 00:07:03,475
innovation

197
00:07:03,855 --> 00:07:05,375
efforts. And then what are some of the

198
00:07:05,375 --> 00:07:07,235
common pitfalls that you're seeing here?

199
00:07:08,175 --> 00:07:10,095
Sure. I think, like I mentioned, we found

200
00:07:10,095 --> 00:07:12,699
the the most important first step is really

201
00:07:12,699 --> 00:07:14,779
to have that governance structure in place to

202
00:07:14,779 --> 00:07:16,779
protect these teams and gives them that space

203
00:07:16,779 --> 00:07:19,019
to to really explore. You want to be

204
00:07:19,019 --> 00:07:20,779
on the same page as an institution as

205
00:07:20,779 --> 00:07:21,759
you've got new technology.

206
00:07:22,220 --> 00:07:23,979
And I think it's also important to really

207
00:07:23,979 --> 00:07:24,879
foster communication

208
00:07:25,419 --> 00:07:26,479
between these teams.

209
00:07:27,464 --> 00:07:29,305
It really isn't helpful if these teams are

210
00:07:29,305 --> 00:07:31,705
working in silos and pursuing technology that may

211
00:07:31,705 --> 00:07:33,324
not be best for the whole organization.

212
00:07:34,024 --> 00:07:36,745
As a example, the clinician innovation group I

213
00:07:36,745 --> 00:07:39,305
mentioned regularly communicates with IT and the executive

214
00:07:39,305 --> 00:07:41,165
teams to make sure we're in agreement.

215
00:07:42,240 --> 00:07:44,079
On the flip side, actually, though, there is

216
00:07:44,079 --> 00:07:46,000
a tendency I've seen, maybe being a little

217
00:07:46,000 --> 00:07:47,699
bit too bureaucratic or overprotective,

218
00:07:48,079 --> 00:07:49,360
and I think it's something to be aware

219
00:07:49,360 --> 00:07:51,759
of. It's probably the most common pitfall I've

220
00:07:51,759 --> 00:07:53,139
seen, in my experiences.

221
00:07:53,680 --> 00:07:55,245
And to be fair, you know, a lot

222
00:07:55,245 --> 00:07:57,084
of these tools, particularly the AI ones, are

223
00:07:57,084 --> 00:07:59,084
a little intimidating at first glance, and that

224
00:07:59,084 --> 00:08:01,084
probably is what motivates the desire to be

225
00:08:01,084 --> 00:08:01,584
protective.

226
00:08:02,125 --> 00:08:03,805
But what we realize is even if we

227
00:08:03,805 --> 00:08:05,904
think we're being protected by being too strict,

228
00:08:05,964 --> 00:08:08,444
that will potentially limit innovation. And so as

229
00:08:08,444 --> 00:08:10,045
we've grown in this journey, we've learned to

230
00:08:10,045 --> 00:08:12,419
better strike that balance. But ultimately, I think

231
00:08:12,419 --> 00:08:14,259
it's important really to trust these teams. And

232
00:08:14,259 --> 00:08:15,779
more importantly, I I would say listen to

233
00:08:15,779 --> 00:08:17,459
them. You know? Give them that leeway to

234
00:08:17,459 --> 00:08:19,620
explore, but also that infrastructure to ensure the

235
00:08:19,620 --> 00:08:20,120
success.

236
00:08:21,060 --> 00:08:23,894
Absolutely. And as we wrap our conversation up

237
00:08:24,055 --> 00:08:26,375
today, I'd love to end on your top

238
00:08:26,375 --> 00:08:28,615
piece of advice for healthcare leaders as they

239
00:08:28,615 --> 00:08:29,435
prepare to

240
00:08:29,974 --> 00:08:32,695
take on further advancements in technology and greater

241
00:08:32,695 --> 00:08:34,394
demands for care in the future.

242
00:08:35,575 --> 00:08:37,170
Sure. I think it's really important,

243
00:08:37,629 --> 00:08:39,870
to stay informed of the technology and these

244
00:08:39,870 --> 00:08:41,330
use cases and the applications,

245
00:08:42,029 --> 00:08:43,870
and really be aware of the limitations. You

246
00:08:43,870 --> 00:08:45,710
know, it's something I try to work into

247
00:08:45,710 --> 00:08:47,790
my role daily, because it's really easy to

248
00:08:47,790 --> 00:08:49,550
fall into a trap maybe of pursuing this

249
00:08:49,550 --> 00:08:50,050
technology,

250
00:08:50,985 --> 00:08:53,144
like AI for the sake of AI, which

251
00:08:53,144 --> 00:08:55,065
is probably another common pitfall to your previous

252
00:08:55,065 --> 00:08:57,485
question, as well as maybe products that overpromise

253
00:08:57,625 --> 00:09:00,024
and don't necessarily deliver. And I think I

254
00:09:00,024 --> 00:09:01,465
found the best way for me and my

255
00:09:01,465 --> 00:09:03,144
team to insulate against that is to really

256
00:09:03,144 --> 00:09:05,690
be as informed as possible. And I think

257
00:09:05,690 --> 00:09:06,669
the time is now

258
00:09:07,049 --> 00:09:08,649
to do that, especially if you're listening to

259
00:09:08,649 --> 00:09:09,389
this podcast.

260
00:09:09,690 --> 00:09:11,049
And I feel like the barrier to entry

261
00:09:11,049 --> 00:09:12,409
is a lot lower than it's ever been

262
00:09:12,409 --> 00:09:14,649
with some of the general available tools such

263
00:09:14,649 --> 00:09:16,970
as large language models. But, just like your

264
00:09:16,970 --> 00:09:18,730
last question, ultimately, I I keep coming back

265
00:09:18,730 --> 00:09:20,174
to the goal should really be to build

266
00:09:20,174 --> 00:09:22,575
up these clinical and technology leaders, listen to

267
00:09:22,575 --> 00:09:24,254
their guidance, and give them that space to

268
00:09:24,254 --> 00:09:26,174
to work with these tools within the proper

269
00:09:26,174 --> 00:09:28,034
guardrails and governance to feel protected.

270
00:09:29,054 --> 00:09:31,391
Well, doctor Izzo, thank you so much for

271
00:09:31,391 --> 00:09:33,071
taking the time today to join me on

272
00:09:33,071 --> 00:09:35,871
the Becker's HealthCare podcast and share your insights

273
00:09:35,871 --> 00:09:36,771
with our listeners.

274
00:09:37,711 --> 00:09:39,071
Yeah. Thank you again for having me. I

275
00:09:39,071 --> 00:09:40,051
really appreciate it.