Electronic Health Record Data for Substance Use Prevention Research and Intervention Delivery

Electronic Health Record Data for Substance Use Prevention Research and Intervention Delivery

Video length: 3:10:43

Transcript

1
00:00:00,320 --> 00:00:04,240
hey thanks i was just about to take it

2
00:00:02,320 --> 00:00:05,839
dave thank you so much steve for getting

3
00:00:04,240 --> 00:00:07,680
us set up with our housekeeping and

4
00:00:05,839 --> 00:00:09,599
thank you to everybody who's joining us

5
00:00:07,680 --> 00:00:11,120
today i have to say i guess both good

6
00:00:09,599 --> 00:00:13,280
morning and good afternoon depending on

7
00:00:11,120 --> 00:00:14,880
which post you're coming to us from my

8
00:00:13,280 --> 00:00:16,400
name is amy goldstein and i'm the

9
00:00:14,880 --> 00:00:18,160
prevention research branch chief at the

10
00:00:16,400 --> 00:00:20,080
national institute on drug abuse in our

11
00:00:18,160 --> 00:00:21,760
division of epidemiology services and

12
00:00:20,080 --> 00:00:24,080
prevention research and i'm pleased to

13
00:00:21,760 --> 00:00:25,519
welcome you all to this meeting um at

14
00:00:24,080 --> 00:00:27,359
nida and in the prevention research

15
00:00:25,519 --> 00:00:28,800
branch our goal is really to think about

16
00:00:27,359 --> 00:00:30,640
intervention development research that

17
00:00:28,800 --> 00:00:32,559
can be delivered in real world systems

18
00:00:30,640 --> 00:00:34,000
and settings so that there's a chance

19
00:00:32,559 --> 00:00:35,440
for long-term sustainability of

20
00:00:34,000 --> 00:00:37,840
interventions and we've been thinking a

21
00:00:35,440 --> 00:00:39,760
lot about the viability of identifying

22
00:00:37,840 --> 00:00:41,280
individuals at risk for substance use

23
00:00:39,760 --> 00:00:43,200
and delivering preventive interventions

24
00:00:41,280 --> 00:00:45,440
for substance use within primary care

25
00:00:43,200 --> 00:00:47,440
generally in pediatrics specifically and

26
00:00:45,440 --> 00:00:49,039
it's nearly impossible to think about

27
00:00:47,440 --> 00:00:50,559
that without talking about electronic

28
00:00:49,039 --> 00:00:52,559
health records and the role that they

29
00:00:50,559 --> 00:00:54,399
could play in either identifying

30
00:00:52,559 --> 00:00:55,680
individuals at risk or somehow

31
00:00:54,399 --> 00:00:57,440
delivering preventive intervention

32
00:00:55,680 --> 00:00:58,640
services so really the goal of this

33
00:00:57,440 --> 00:01:00,800
meeting today is to bring people

34
00:00:58,640 --> 00:01:02,399
together who are working either in

35
00:01:00,800 --> 00:01:04,320
substance use or in other fields of

36
00:01:02,399 --> 00:01:06,479
medicine and thinking about the ehr and

37
00:01:04,320 --> 00:01:08,640
its role in clinical research and to

38
00:01:06,479 --> 00:01:10,320
think about the potential for youth

39
00:01:08,640 --> 00:01:12,720
within substance use prevention research

40
00:01:10,320 --> 00:01:14,640
so i'm excited for the agenda that we've

41
00:01:12,720 --> 00:01:16,159
put together um for the talks that we're

42
00:01:14,640 --> 00:01:18,240
going to have today and with that i'll

43
00:01:16,159 --> 00:01:20,880
turn it over to united's deputy director

44
00:01:18,240 --> 00:01:22,240
dr wilson compton

45
00:01:20,880 --> 00:01:23,840
well thank you amy it's certainly a

46
00:01:22,240 --> 00:01:24,960
pleasure to be with you and with the

47
00:01:23,840 --> 00:01:26,640
group today

48
00:01:24,960 --> 00:01:28,640
you all will be wrestling with an

49
00:01:26,640 --> 00:01:30,720
important area for

50
00:01:28,640 --> 00:01:32,640
research and practice development and

51
00:01:30,720 --> 00:01:35,360
that's what we're talking about today is

52
00:01:32,640 --> 00:01:38,400
how can we use research at the boundary

53
00:01:35,360 --> 00:01:40,560
of practice and in this case how can we

54
00:01:38,400 --> 00:01:43,600
use electronic health records are these

55
00:01:40,560 --> 00:01:46,079
potentially very rich administrative

56
00:01:43,600 --> 00:01:48,320
sources of information to

57
00:01:46,079 --> 00:01:51,360
improve the quality of care and to

58
00:01:48,320 --> 00:01:54,000
improve the implementation of prevention

59
00:01:51,360 --> 00:01:57,520
interventions that can enhance the lives

60
00:01:54,000 --> 00:02:00,399
and well-being of broad swaths of our

61
00:01:57,520 --> 00:02:02,960
population across the united states

62
00:02:00,399 --> 00:02:05,119
now certainly a priority at nida is to

63
00:02:02,960 --> 00:02:07,360
develop and test novel substance use and

64
00:02:05,119 --> 00:02:10,399
prevention strategies

65
00:02:07,360 --> 00:02:12,959
and uh being able to do this in the

66
00:02:10,399 --> 00:02:15,120
settings while they will be delivered uh

67
00:02:12,959 --> 00:02:17,280
really can help us decrease the research

68
00:02:15,120 --> 00:02:18,560
to practice gap and so we're looking for

69
00:02:17,280 --> 00:02:19,520
ways to

70
00:02:18,560 --> 00:02:23,120
to

71
00:02:19,520 --> 00:02:24,800
develop research within uh primary care

72
00:02:23,120 --> 00:02:26,720
settings that'll be reflected in the

73
00:02:24,800 --> 00:02:28,959
electronic health record and that's what

74
00:02:26,720 --> 00:02:31,040
you all can give us ideas on how best to

75
00:02:28,959 --> 00:02:33,200
do this

76
00:02:31,040 --> 00:02:34,879
now one of the barriers to

77
00:02:33,200 --> 00:02:36,400
primary care based screening and

78
00:02:34,879 --> 00:02:38,959
prevention

79
00:02:36,400 --> 00:02:40,800
is that providers may not know how to

80
00:02:38,959 --> 00:02:43,200
screen for substance use risk they don't

81
00:02:40,800 --> 00:02:44,400
have time to do this work and it's not

82
00:02:43,200 --> 00:02:46,879
really clear how they're going to get

83
00:02:44,400 --> 00:02:50,319
reimbursed for work and after all that's

84
00:02:46,879 --> 00:02:51,440
going to be an essential component of

85
00:02:50,319 --> 00:02:53,440
seeing this

86
00:02:51,440 --> 00:02:56,239
these practices widely used is that

87
00:02:53,440 --> 00:02:58,159
there's adequate reimburse reimbursement

88
00:02:56,239 --> 00:02:59,120
but there's also another key issue which

89
00:02:58,159 --> 00:03:02,239
is

90
00:02:59,120 --> 00:03:05,120
if you identify a youth

91
00:03:02,239 --> 00:03:08,400
uh a child who's at risk or an

92
00:03:05,120 --> 00:03:10,959
adolescent who's using substances or has

93
00:03:08,400 --> 00:03:12,879
significant major issues related to

94
00:03:10,959 --> 00:03:14,560
their substance use what do you do about

95
00:03:12,879 --> 00:03:16,319
that how do you refer them and where do

96
00:03:14,560 --> 00:03:18,319
you send them

97
00:03:16,319 --> 00:03:20,800
treatment services and intervention

98
00:03:18,319 --> 00:03:24,239
services are not always available and so

99
00:03:20,800 --> 00:03:26,400
that's a key barrier as well

100
00:03:24,239 --> 00:03:27,760
now one consideration is that

101
00:03:26,400 --> 00:03:30,480
reimbursement

102
00:03:27,760 --> 00:03:33,120
is at least partly dependent on the u.s

103
00:03:30,480 --> 00:03:36,319
preventive services task force practices

104
00:03:33,120 --> 00:03:38,040
given an a or b grade are reimbursable

105
00:03:36,319 --> 00:03:40,879
and right now

106
00:03:38,040 --> 00:03:43,519
uspstf has recommended screening in

107
00:03:40,879 --> 00:03:45,280
primary care for adults with a b grade

108
00:03:43,519 --> 00:03:47,519
but the evidence has been insufficient

109
00:03:45,280 --> 00:03:49,680
to remain to recommend screening for

110
00:03:47,519 --> 00:03:51,519
youth i would also point out that even

111
00:03:49,680 --> 00:03:53,280
in the recommendation for adults it's

112
00:03:51,519 --> 00:03:55,200
for screening it does not include a

113
00:03:53,280 --> 00:03:57,360
recommendation around brief

114
00:03:55,200 --> 00:03:59,439
interventions and so it's mostly

115
00:03:57,360 --> 00:04:01,280
screening with the idea of linkage to

116
00:03:59,439 --> 00:04:02,720
services and linkage to treatment for

117
00:04:01,280 --> 00:04:05,840
those

118
00:04:02,720 --> 00:04:07,439
with a significant involvement with

119
00:04:05,840 --> 00:04:11,200
substances and problems related to their

120
00:04:07,439 --> 00:04:13,439
substances is is the recommendation

121
00:04:11,200 --> 00:04:15,760
so there's insufficient evidence to

122
00:04:13,439 --> 00:04:17,840
recommend brief counseling

123
00:04:15,760 --> 00:04:19,600
in primary care for children adolescents

124
00:04:17,840 --> 00:04:22,400
and young adults these are

125
00:04:19,600 --> 00:04:23,759
an irating the insufficient evidence now

126
00:04:22,400 --> 00:04:26,240
when there's an eye rating that means

127
00:04:23,759 --> 00:04:26,960
that research might help to inform that

128
00:04:26,240 --> 00:04:29,280
so

129
00:04:26,960 --> 00:04:32,000
this is one of the areas where

130
00:04:29,280 --> 00:04:35,280
we want your thoughts and your help

131
00:04:32,000 --> 00:04:36,250
research to address the usps tf evidence

132
00:04:35,280 --> 00:04:37,520
gaps um

133
00:04:36,250 --> 00:04:39,280
[Music]

134
00:04:37,520 --> 00:04:41,360
and to inform the implementation of

135
00:04:39,280 --> 00:04:43,199
evidence-based interventions really can

136
00:04:41,360 --> 00:04:45,040
involve the electronic health records

137
00:04:43,199 --> 00:04:47,199
both to identify

138
00:04:45,040 --> 00:04:50,160
places where this might be conducted and

139
00:04:47,199 --> 00:04:52,560
also to provide the source of outcomes

140
00:04:50,160 --> 00:04:54,960
or the source of information to

141
00:04:52,560 --> 00:04:57,600
determine the impact and effectiveness

142
00:04:54,960 --> 00:04:59,199
of any interventions

143
00:04:57,600 --> 00:05:01,759
i will point out that something that i

144
00:04:59,199 --> 00:05:03,199
know that uh uh dr goldstein and

145
00:05:01,759 --> 00:05:04,960
colleagues in our prevention research

146
00:05:03,199 --> 00:05:08,160
brands have emphasized over the last few

147
00:05:04,960 --> 00:05:10,960
years is the potential value in

148
00:05:08,160 --> 00:05:13,120
uh implementing what had been either

149
00:05:10,960 --> 00:05:14,320
school or universal based prevention

150
00:05:13,120 --> 00:05:16,639
interventions

151
00:05:14,320 --> 00:05:18,080
within primary care settings and so

152
00:05:16,639 --> 00:05:19,759
seeing how some of those could be

153
00:05:18,080 --> 00:05:22,479
incorporated into

154
00:05:19,759 --> 00:05:25,600
broad-based screening approaches and

155
00:05:22,479 --> 00:05:26,960
and approaches that include the uh

156
00:05:25,600 --> 00:05:28,800
electronic health record and these

157
00:05:26,960 --> 00:05:31,039
large-scale administrator data sets will

158
00:05:28,800 --> 00:05:32,960
be an exciting and important

159
00:05:31,039 --> 00:05:35,360
uh next key step

160
00:05:32,960 --> 00:05:37,759
now you all have a wonderful set of

161
00:05:35,360 --> 00:05:39,919
opportunities to provide

162
00:05:37,759 --> 00:05:42,800
guidance to us through round tables over

163
00:05:39,919 --> 00:05:43,520
the next few hours and i'm very pleased

164
00:05:42,800 --> 00:05:46,560
with

165
00:05:43,520 --> 00:05:48,320
and excited by the way uh dr goldstein

166
00:05:46,560 --> 00:05:50,160
and dr blanco and colleagues in our

167
00:05:48,320 --> 00:05:53,440
prevention research branch

168
00:05:50,160 --> 00:05:55,199
here at nida have organized this session

169
00:05:53,440 --> 00:05:57,840
now i do want to say that i'm speaking

170
00:05:55,199 --> 00:05:59,919
you today somewhat unexpectedly we

171
00:05:57,840 --> 00:06:02,639
really did think our our director dr

172
00:05:59,919 --> 00:06:05,680
nora volkov would be able to welcome you

173
00:06:02,639 --> 00:06:08,240
uh unfortunately nora had a a a

174
00:06:05,680 --> 00:06:09,919
emergency and so she asked me to

175
00:06:08,240 --> 00:06:12,000
substitute for her today and it

176
00:06:09,919 --> 00:06:14,240
certainly is my pleasure to help welcome

177
00:06:12,000 --> 00:06:16,080
you on behalf of the national institute

178
00:06:14,240 --> 00:06:17,039
on drug abuse on behalf of our entire

179
00:06:16,080 --> 00:06:19,600
team

180
00:06:17,039 --> 00:06:21,360
at nida we really value your time this

181
00:06:19,600 --> 00:06:22,720
afternoon or this morning if you're on

182
00:06:21,360 --> 00:06:24,800
the west coast

183
00:06:22,720 --> 00:06:26,960
uh and we look forward to your best

184
00:06:24,800 --> 00:06:29,680
ideas for how we can develop these

185
00:06:26,960 --> 00:06:31,840
really important areas for prevention

186
00:06:29,680 --> 00:06:33,919
research and intervention delivery and

187
00:06:31,840 --> 00:06:37,280
with that i believe i turn it back to

188
00:06:33,919 --> 00:06:38,800
amy and will you turn it back to be uh

189
00:06:37,280 --> 00:06:40,000
coordinating from here thanks for

190
00:06:38,800 --> 00:06:41,520
watching thank you

191
00:06:40,000 --> 00:06:43,919
great thanks wilson

192
00:06:41,520 --> 00:06:46,720
now i'd like to introduce for another uh

193
00:06:43,919 --> 00:06:48,000
opening remark dr carlos blanco who's

194
00:06:46,720 --> 00:06:49,680
the director of our division of

195
00:06:48,000 --> 00:06:51,680
epidemiology services and prevention

196
00:06:49,680 --> 00:06:53,840
research

197
00:06:51,680 --> 00:06:55,680
thank you i mean so um

198
00:06:53,840 --> 00:06:57,039
one of the advantages of working at

199
00:06:55,680 --> 00:06:59,199
naida is that we are pretty much

200
00:06:57,039 --> 00:07:01,919
aligning our thinking so i'm going to

201
00:06:59,199 --> 00:07:03,039
spare you repeating what amy and wilson

202
00:07:01,919 --> 00:07:04,720
have said

203
00:07:03,039 --> 00:07:07,199
but i cannot help

204
00:07:04,720 --> 00:07:09,520
but think that today is march 21st so

205
00:07:07,199 --> 00:07:11,440
it's the first day of spring

206
00:07:09,520 --> 00:07:13,039
and and i think it is really fitting

207
00:07:11,440 --> 00:07:15,280
that we are having this workshop today

208
00:07:13,039 --> 00:07:17,759
because spring at least for me

209
00:07:15,280 --> 00:07:20,080
is a time of hope and particularly today

210
00:07:17,759 --> 00:07:22,720
which is very sunny at least where i am

211
00:07:20,080 --> 00:07:24,880
which is new york city and so for me the

212
00:07:22,720 --> 00:07:26,479
the sun and the spring are

213
00:07:24,880 --> 00:07:28,560
really a source of hope

214
00:07:26,479 --> 00:07:31,759
and and by the same token prevention i

215
00:07:28,560 --> 00:07:34,400
mean i uh by training i'm mostly a

216
00:07:31,759 --> 00:07:36,319
treating practitioner but as you all

217
00:07:34,400 --> 00:07:37,919
know most people with substance use

218
00:07:36,319 --> 00:07:39,440
disorders never seek treatment so

219
00:07:37,919 --> 00:07:41,440
prevention for me

220
00:07:39,440 --> 00:07:44,000
is crucial unless

221
00:07:41,440 --> 00:07:46,160
unless we really step up our work on

222
00:07:44,000 --> 00:07:48,160
prevention we're never going to to solve

223
00:07:46,160 --> 00:07:50,879
the substance use

224
00:07:48,160 --> 00:07:52,800
crisis so i'm very grateful to to sarah

225
00:07:50,879 --> 00:07:55,520
steverman and to amy

226
00:07:52,800 --> 00:07:58,479
goldstein for working on this workshop

227
00:07:55,520 --> 00:08:00,479
as already amy and wilson have said one

228
00:07:58,479 --> 00:08:02,160
of our interest is how to use electronic

229
00:08:00,479 --> 00:08:03,280
health records and also how do we create

230
00:08:02,160 --> 00:08:05,680
evidence

231
00:08:03,280 --> 00:08:08,000
to really move some of the interventions

232
00:08:05,680 --> 00:08:10,240
to be grades a or b

233
00:08:08,000 --> 00:08:11,440
to the uspstf because one of the things

234
00:08:10,240 --> 00:08:13,919
that we want to do it's not just

235
00:08:11,440 --> 00:08:15,280
research that appears on journals but

236
00:08:13,919 --> 00:08:17,840
but research that can really be

237
00:08:15,280 --> 00:08:20,720
sustained as interventions so i'm going

238
00:08:17,840 --> 00:08:22,400
to to stop here i hope that

239
00:08:20,720 --> 00:08:24,400
that we will have a mixture in our

240
00:08:22,400 --> 00:08:26,560
audience and presenters of optimists and

241
00:08:24,400 --> 00:08:28,639
pessimists think optimists are very good

242
00:08:26,560 --> 00:08:31,280
because we really need hope

243
00:08:28,639 --> 00:08:33,680
but we also need people like me who

244
00:08:31,280 --> 00:08:36,080
always finds problems in everything that

245
00:08:33,680 --> 00:08:38,000
other people present so we can address

246
00:08:36,080 --> 00:08:39,599
those problems before we bring this to

247
00:08:38,000 --> 00:08:42,000
the field so i hope that there's a sort

248
00:08:39,599 --> 00:08:44,159
of debate between people being hopeful

249
00:08:42,000 --> 00:08:46,000
but also be people being critical and

250
00:08:44,159 --> 00:08:47,440
finding problems so that at the end we

251
00:08:46,000 --> 00:08:49,680
get the best product

252
00:08:47,440 --> 00:08:51,360
so amy and sarah thanks for organizing

253
00:08:49,680 --> 00:08:53,360
this i think this is i mean i always say

254
00:08:51,360 --> 00:08:54,880
i i i don't know why they pay me to do

255
00:08:53,360 --> 00:08:57,200
this job like i should pay to do this

256
00:08:54,880 --> 00:08:58,880
work no so this is for me better than

257
00:08:57,200 --> 00:09:00,800
going to the movies and i'm looking

258
00:08:58,880 --> 00:09:03,519
forward to what you have prepared for

259
00:09:00,800 --> 00:09:05,519
for us today so thank you

260
00:09:03,519 --> 00:09:06,959
great thank you carlos yeah i wonder if

261
00:09:05,519 --> 00:09:08,160
our agenda our meeting should be

262
00:09:06,959 --> 00:09:09,760
entitled like challenges and

263
00:09:08,160 --> 00:09:11,360
opportunities because i agree with you

264
00:09:09,760 --> 00:09:12,800
there's a lot of potential but there

265
00:09:11,360 --> 00:09:14,560
might also be some limitations to what

266
00:09:12,800 --> 00:09:16,880
we're going to talk about so just to

267
00:09:14,560 --> 00:09:19,279
orient everybody to how the afternoon or

268
00:09:16,880 --> 00:09:20,800
morning is going to be structured we

269
00:09:19,279 --> 00:09:23,040
have the time divided into three

270
00:09:20,800 --> 00:09:25,120
sessions so session one will be more of

271
00:09:23,040 --> 00:09:26,959
an overview of the potential uses

272
00:09:25,120 --> 00:09:29,600
challenges and opportunities for ehr

273
00:09:26,959 --> 00:09:30,720
data and primary care and pediatrics and

274
00:09:29,600 --> 00:09:32,399
our speakers are going to provide the

275
00:09:30,720 --> 00:09:35,120
perspectives of some of our key

276
00:09:32,399 --> 00:09:37,279
stakeholders so the uspsdf pediatric

277
00:09:35,120 --> 00:09:38,640
practices and health care systems

278
00:09:37,279 --> 00:09:40,080
and then our second session will be

279
00:09:38,640 --> 00:09:41,920
joined by four researchers who are

280
00:09:40,080 --> 00:09:44,000
working across different systems so

281
00:09:41,920 --> 00:09:46,160
mental health pediatrics substance use

282
00:09:44,000 --> 00:09:48,080
and veterans administration to learn

283
00:09:46,160 --> 00:09:50,160
about how they've used ehr data in their

284
00:09:48,080 --> 00:09:52,320
research and some of the challenges that

285
00:09:50,160 --> 00:09:54,480
they've encountered and lastly we'll

286
00:09:52,320 --> 00:09:56,880
hear about two knight-of-funded projects

287
00:09:54,480 --> 00:09:58,800
both of which are using ehr data either

288
00:09:56,880 --> 00:10:01,040
for risk prediction or to track outcomes

289
00:09:58,800 --> 00:10:03,040
in their research so in terms of how

290
00:10:01,040 --> 00:10:04,399
we've structured the time

291
00:10:03,040 --> 00:10:05,680
we'll take five-minute breaks between

292
00:10:04,399 --> 00:10:07,200
each sessions just to give people a

293
00:10:05,680 --> 00:10:08,880
stretch break and a little bit of screen

294
00:10:07,200 --> 00:10:10,800
break um and we'll do all the

295
00:10:08,880 --> 00:10:13,440
presentations in block and then have

296
00:10:10,800 --> 00:10:15,440
time for q a at the end so as we said in

297
00:10:13,440 --> 00:10:16,959
the beginning feel free to use the chat

298
00:10:15,440 --> 00:10:19,279
to put in questions and comments and

299
00:10:16,959 --> 00:10:20,880
we'll be tracking those as we go

300
00:10:19,279 --> 00:10:22,480
so with that i'm going to turn it over

301
00:10:20,880 --> 00:10:23,839
to sarah steverman who's a program

302
00:10:22,480 --> 00:10:26,240
officer in the prevention research

303
00:10:23,839 --> 00:10:27,760
branch and oversees our research on

304
00:10:26,240 --> 00:10:29,279
healthcare and healthcare systems and

305
00:10:27,760 --> 00:10:31,200
has really been the brains behind this

306
00:10:29,279 --> 00:10:33,760
operation in putting this meeting

307
00:10:31,200 --> 00:10:36,720
together and cultivating our speakers so

308
00:10:33,760 --> 00:10:38,640
sarah with that i'll turn it over to you

309
00:10:36,720 --> 00:10:41,519
great thanks amy

310
00:10:38,640 --> 00:10:43,760
um and thanks everybody for um

311
00:10:41,519 --> 00:10:45,680
for joining us i am gonna

312
00:10:43,760 --> 00:10:48,800
quickly move to our first session and

313
00:10:45,680 --> 00:10:52,079
introduce um introduce our speakers

314
00:10:48,800 --> 00:10:54,480
um as amy said this panel um we're gonna

315
00:10:52,079 --> 00:10:57,600
start to orient ourselves to

316
00:10:54,480 --> 00:11:00,240
um to to our um our challenges and our

317
00:10:57,600 --> 00:11:01,920
opportunities for using ehr

318
00:11:00,240 --> 00:11:04,800
data and research as well as an

319
00:11:01,920 --> 00:11:07,519
implementation so um the first speaker

320
00:11:04,800 --> 00:11:08,320
is um bob mcnellis

321
00:11:07,519 --> 00:11:11,360
he

322
00:11:08,320 --> 00:11:14,079
has newly joined us in nih's office of

323
00:11:11,360 --> 00:11:16,000
disease prevention as a senior advisor

324
00:11:14,079 --> 00:11:17,920
and there he leads the effort to

325
00:11:16,000 --> 00:11:20,000
identify prevention research areas for

326
00:11:17,920 --> 00:11:22,320
investment across nih

327
00:11:20,000 --> 00:11:24,160
and also serves as the nih

328
00:11:22,320 --> 00:11:26,000
liaison to the u.s preventive services

329
00:11:24,160 --> 00:11:28,079
task force and the community preventive

330
00:11:26,000 --> 00:11:30,720
services task force he's also the

331
00:11:28,079 --> 00:11:32,000
scientific advisor for the office of

332
00:11:30,720 --> 00:11:33,760
disease prevention pathways to

333
00:11:32,000 --> 00:11:36,160
prevention program

334
00:11:33,760 --> 00:11:37,519
and he recently came to nih from the

335
00:11:36,160 --> 00:11:41,440
agency for healthcare research and

336
00:11:37,519 --> 00:11:43,279
quality where he worked on a portfolio

337
00:11:41,440 --> 00:11:45,040
related to primary care research and

338
00:11:43,279 --> 00:11:47,040
also served as a medical officer for the

339
00:11:45,040 --> 00:11:49,120
uspstf

340
00:11:47,040 --> 00:11:51,760
and then after

341
00:11:49,120 --> 00:11:54,639
after bob we will turn to dr mario

342
00:11:51,760 --> 00:11:56,399
tehran a physician and clinical

343
00:11:54,639 --> 00:11:58,399
informaticist for the division of

344
00:11:56,399 --> 00:12:00,399
digital healthcare research in the

345
00:11:58,399 --> 00:12:02,399
center for evidence and practice

346
00:12:00,399 --> 00:12:04,639
improvement at hrq

347
00:12:02,399 --> 00:12:06,399
um dr tehran is

348
00:12:04,639 --> 00:12:08,240
has overseen the implementation of ehrs

349
00:12:06,399 --> 00:12:10,079
across multiple health systems and led

350
00:12:08,240 --> 00:12:12,639
efforts to improve physician ehr

351
00:12:10,079 --> 00:12:14,639
optimization and satisfaction

352
00:12:12,639 --> 00:12:17,200
and then finally we're going to

353
00:12:14,639 --> 00:12:19,279
move to dr brian jensen

354
00:12:17,200 --> 00:12:20,720
a physician researcher at

355
00:12:19,279 --> 00:12:22,000
chop at children's hospital of

356
00:12:20,720 --> 00:12:23,920
philadelphia

357
00:12:22,000 --> 00:12:26,320
and dr jensen is a faculty member at

358
00:12:23,920 --> 00:12:27,760
chop's policy lab an assistant professor

359
00:12:26,320 --> 00:12:29,279
in the department of pediatrics at

360
00:12:27,760 --> 00:12:31,279
university of pennsylvania and the

361
00:12:29,279 --> 00:12:32,480
medical director for value-based care

362
00:12:31,279 --> 00:12:33,760
for chops

363
00:12:32,480 --> 00:12:34,560
care network

364
00:12:33,760 --> 00:12:37,760
um

365
00:12:34,560 --> 00:12:39,200
he is a board-certified pediatrician and

366
00:12:37,760 --> 00:12:41,360
informaticist

367
00:12:39,200 --> 00:12:44,000
um and he's going to discuss

368
00:12:41,360 --> 00:12:46,000
his experience with optimizing ehr data

369
00:12:44,000 --> 00:12:47,519
and then their use by pediatricians and

370
00:12:46,000 --> 00:12:50,959
their health system

371
00:12:47,519 --> 00:12:53,680
so i am very thankful for for these um

372
00:12:50,959 --> 00:12:55,040
three experts for for you all joining us

373
00:12:53,680 --> 00:12:56,480
and at this point i'm going to turn over

374
00:12:55,040 --> 00:12:57,440
to bob we'll go

375
00:12:56,480 --> 00:12:58,480
um

376
00:12:57,440 --> 00:12:59,839
we'll go

377
00:12:58,480 --> 00:13:02,399
through our presentations and then we'll

378
00:12:59,839 --> 00:13:04,399
have time at the end for questions so

379
00:13:02,399 --> 00:13:06,240
please in the meantime as questions come

380
00:13:04,399 --> 00:13:08,079
up go ahead and put them in the chat and

381
00:13:06,240 --> 00:13:09,279
we'll start to compile them

382
00:13:08,079 --> 00:13:10,160
and then we'll get to the discussion at

383
00:13:09,279 --> 00:13:11,920
the end

384
00:13:10,160 --> 00:13:13,920
thank you

385
00:13:11,920 --> 00:13:15,519
great thank you sarah and good morning

386
00:13:13,920 --> 00:13:17,279
and good afternoon to all of you it's a

387
00:13:15,519 --> 00:13:19,440
pleasure to be here special thanks to dr

388
00:13:17,279 --> 00:13:21,760
goldstein and blanco and compton for

389
00:13:19,440 --> 00:13:23,680
inviting odp to participate uh in this

390
00:13:21,760 --> 00:13:25,519
session as sarah mentioned i'm bob

391
00:13:23,680 --> 00:13:27,600
mcnallis and i'm starting my sixth week

392
00:13:25,519 --> 00:13:30,320
at odp which officially means i know

393
00:13:27,600 --> 00:13:32,160
enough to be dangerous um but honestly i

394
00:13:30,320 --> 00:13:35,040
am really thrilled to be here today and

395
00:13:32,160 --> 00:13:37,519
while i'm new to odp um today's topic is

396
00:13:35,040 --> 00:13:39,600
not new to me um as sarah mentioned i

397
00:13:37,519 --> 00:13:40,800
came to odp from the agency of

398
00:13:39,600 --> 00:13:43,040
healthcare research and quality where i

399
00:13:40,800 --> 00:13:45,440
was senior advisor for primary care for

400
00:13:43,040 --> 00:13:47,680
the past 10 years and we led several dni

401
00:13:45,440 --> 00:13:49,680
initiatives that really relied on ehr

402
00:13:47,680 --> 00:13:50,959
data so this one is near and dear to me

403
00:13:49,680 --> 00:13:52,399
and then of course i did serve as a

404
00:13:50,959 --> 00:13:54,320
medical officer about five years for the

405
00:13:52,399 --> 00:13:55,680
us preventive services task force and

406
00:13:54,320 --> 00:13:58,720
today what i'm going to talk about just

407
00:13:55,680 --> 00:14:01,920
briefly is odp's role at nih as a

408
00:13:58,720 --> 00:14:04,240
liaison to the task force and share at a

409
00:14:01,920 --> 00:14:05,519
very high level um the task force's

410
00:14:04,240 --> 00:14:07,120
considerations for making

411
00:14:05,519 --> 00:14:10,560
recommendations on clinical preventive

412
00:14:07,120 --> 00:14:12,560
services um like screening for drug use

413
00:14:10,560 --> 00:14:14,800
i i can't speak to how the task force

414
00:14:12,560 --> 00:14:16,959
would view studies that use ehr data but

415
00:14:14,800 --> 00:14:18,720
i think understanding their process and

416
00:14:16,959 --> 00:14:21,440
the rigor of their methods can really

417
00:14:18,720 --> 00:14:23,360
inform thinking uh on this approach

418
00:14:21,440 --> 00:14:25,600
so first let me just start with odp i

419
00:14:23,360 --> 00:14:27,120
sort of nuts and bolts odps mission is

420
00:14:25,600 --> 00:14:29,680
to improve public health by increasing

421
00:14:27,120 --> 00:14:31,199
the scope and quality the dissemination

422
00:14:29,680 --> 00:14:33,600
impact of prevention research that's

423
00:14:31,199 --> 00:14:36,160
supported by nih we actually also work

424
00:14:33,600 --> 00:14:39,199
across all nih to help to coordinate and

425
00:14:36,160 --> 00:14:42,240
facilitate and make aware um uh research

426
00:14:39,199 --> 00:14:45,360
gaps for uh other institutes and centers

427
00:14:42,240 --> 00:14:46,959
to to include in their portfolios

428
00:14:45,360 --> 00:14:49,279
we sit in the office of the director at

429
00:14:46,959 --> 00:14:50,880
nih uh within the division of program

430
00:14:49,279 --> 00:14:52,320
coordination planning and strategic

431
00:14:50,880 --> 00:14:54,079
initiatives and actually it's nice

432
00:14:52,320 --> 00:14:56,240
because it gives us we don't not we're

433
00:14:54,079 --> 00:14:58,160
agnostic around disorders and diseases

434
00:14:56,240 --> 00:15:00,000
and conditions uh and gives us a little

435
00:14:58,160 --> 00:15:02,399
bit of credibility for working broadly

436
00:15:00,000 --> 00:15:04,560
across nih and with other agencies

437
00:15:02,399 --> 00:15:07,040
across hhs

438
00:15:04,560 --> 00:15:08,399
at odp we have uh some strategic

439
00:15:07,040 --> 00:15:10,320
priorities that we've been working on

440
00:15:08,399 --> 00:15:12,320
over the last several years and we've

441
00:15:10,320 --> 00:15:13,839
got six of them listed here we conduct

442
00:15:12,320 --> 00:15:16,639
portfolio analysis and impact

443
00:15:13,839 --> 00:15:19,120
assessments we look for research gaps uh

444
00:15:16,639 --> 00:15:20,880
work on prevention research methods uh

445
00:15:19,120 --> 00:15:23,360
promote collaborative prevention

446
00:15:20,880 --> 00:15:25,839
research and then we actually work to

447
00:15:23,360 --> 00:15:27,680
advance tobacco regulatory science and

448
00:15:25,839 --> 00:15:30,320
communicate all of these things and just

449
00:15:27,680 --> 00:15:32,560
recently uh within the past year we've

450
00:15:30,320 --> 00:15:34,240
added now a seventh strategic priority

451
00:15:32,560 --> 00:15:36,320
uh working on health disparities used to

452
00:15:34,240 --> 00:15:38,399
be a cross-cutting theme and now it's

453
00:15:36,320 --> 00:15:40,399
actually what a strategic priority in

454
00:15:38,399 --> 00:15:42,000
and of itself to promote and coordinate

455
00:15:40,399 --> 00:15:44,079
prevention research that addresses

456
00:15:42,000 --> 00:15:46,240
health disparities

457
00:15:44,079 --> 00:15:48,560
i work on team two i'm new the new team

458
00:15:46,240 --> 00:15:49,839
two lead uh and still working my way

459
00:15:48,560 --> 00:15:51,600
through this but you can see here's the

460
00:15:49,839 --> 00:15:54,079
strategic priority for team two and that

461
00:15:51,600 --> 00:15:56,160
is to identify prevention research areas

462
00:15:54,079 --> 00:15:57,839
for investment or expanded effort by the

463
00:15:56,160 --> 00:15:59,600
national institutes of health then we

464
00:15:57,839 --> 00:16:01,440
have these three objectives we work with

465
00:15:59,600 --> 00:16:03,120
a variety of stakeholders to identify

466
00:16:01,440 --> 00:16:04,639
needs and these stakeholders include

467
00:16:03,120 --> 00:16:06,480
people like the us preventive services

468
00:16:04,639 --> 00:16:08,240
task force the community preventive

469
00:16:06,480 --> 00:16:10,160
services task force which is located at

470
00:16:08,240 --> 00:16:12,480
the cdc and then the larger healthy

471
00:16:10,160 --> 00:16:14,880
people initiative that's across nih and

472
00:16:12,480 --> 00:16:17,600
actually across many government agencies

473
00:16:14,880 --> 00:16:18,959
and then we compare um

474
00:16:17,600 --> 00:16:21,360
prevention research with the current

475
00:16:18,959 --> 00:16:23,279
portfolio to identify some gaps and then

476
00:16:21,360 --> 00:16:25,120
we work across institutes and centers

477
00:16:23,279 --> 00:16:27,360
and other stakeholders to try to find

478
00:16:25,120 --> 00:16:29,040
ways to invest in filling some of those

479
00:16:27,360 --> 00:16:31,759
gaps

480
00:16:29,040 --> 00:16:33,279
specifically for today's talk what's i

481
00:16:31,759 --> 00:16:34,880
think most important is that we're the

482
00:16:33,279 --> 00:16:36,320
liaison office

483
00:16:34,880 --> 00:16:37,519
for the us preventive services task

484
00:16:36,320 --> 00:16:39,279
force and the community preventive

485
00:16:37,519 --> 00:16:40,959
services task force as we as we

486
00:16:39,279 --> 00:16:43,120
mentioned and we have some specific

487
00:16:40,959 --> 00:16:45,440
things that we do to address some of the

488
00:16:43,120 --> 00:16:48,639
research gaps um that come out of those

489
00:16:45,440 --> 00:16:51,040
task forces with nih's 27 institutes and

490
00:16:48,639 --> 00:16:53,680
centers to develop activities to address

491
00:16:51,040 --> 00:16:55,600
some of those gaps

492
00:16:53,680 --> 00:16:56,959
so i think i'll just pause and and now

493
00:16:55,600 --> 00:16:58,800
talk a little about the us preventive

494
00:16:56,959 --> 00:17:00,160
services task force and again i don't

495
00:16:58,800 --> 00:17:02,560
don't mean to speak for the task force

496
00:17:00,160 --> 00:17:04,240
but in our liaison role we are very much

497
00:17:02,560 --> 00:17:06,079
aware of what the task force does in

498
00:17:04,240 --> 00:17:07,760
their processes and i can reflect on my

499
00:17:06,079 --> 00:17:10,240
own previous experience as a medical

500
00:17:07,760 --> 00:17:11,919
officer to give you some authoritative

501
00:17:10,240 --> 00:17:14,319
information around this and and first

502
00:17:11,919 --> 00:17:16,160
it's to say that um the us preventive

503
00:17:14,319 --> 00:17:18,000
service test force is an independent

504
00:17:16,160 --> 00:17:20,079
volunteer panel of clinical and

505
00:17:18,000 --> 00:17:22,480
behavioral health experts and they have

506
00:17:20,079 --> 00:17:25,039
expertise in primary care in prevention

507
00:17:22,480 --> 00:17:27,600
research uh in in prevention research

508
00:17:25,039 --> 00:17:30,000
methodology uh and evidence-based

509
00:17:27,600 --> 00:17:32,000
medicine and so it's that about 16

510
00:17:30,000 --> 00:17:33,840
members who bring together a wide

511
00:17:32,000 --> 00:17:36,320
variety of skills

512
00:17:33,840 --> 00:17:38,000
to consider the evidence and make

513
00:17:36,320 --> 00:17:39,600
recommendations the agency for

514
00:17:38,000 --> 00:17:40,960
healthcare research and quality i'll

515
00:17:39,600 --> 00:17:43,039
hear from my colleague

516
00:17:40,960 --> 00:17:44,400
at arc in a moment provides scientific

517
00:17:43,039 --> 00:17:46,559
technical administration and

518
00:17:44,400 --> 00:17:49,280
dissemination support it is although the

519
00:17:46,559 --> 00:17:52,559
us pstf is not a federal agency it is

520
00:17:49,280 --> 00:17:54,880
supported by ahrq and the task force as

521
00:17:52,559 --> 00:17:56,960
you know and dr compton mentioned issues

522
00:17:54,880 --> 00:17:59,600
recommendations about preventive

523
00:17:56,960 --> 00:18:02,080
services in primary care that can be

524
00:17:59,600 --> 00:18:05,360
generalized to a primary care population

525
00:18:02,080 --> 00:18:07,520
who don't generally have symptoms

526
00:18:05,360 --> 00:18:09,679
of care so it's got to be a broad group

527
00:18:07,520 --> 00:18:12,240
who are not selected usually for these

528
00:18:09,679 --> 00:18:15,039
services based on symptomatology or

529
00:18:12,240 --> 00:18:16,880
previous diseases

530
00:18:15,039 --> 00:18:18,320
uh the task force's visibility went up

531
00:18:16,880 --> 00:18:20,240
quite a lot with the passage of the

532
00:18:18,320 --> 00:18:22,160
affordable care act um and that the

533
00:18:20,240 --> 00:18:23,919
affordable care act mandates coverage of

534
00:18:22,160 --> 00:18:26,000
grade a and b services without cost

535
00:18:23,919 --> 00:18:27,360
sharing that definitely

536
00:18:26,000 --> 00:18:28,960
increased the pressure i think on the

537
00:18:27,360 --> 00:18:30,240
task force around a and b

538
00:18:28,960 --> 00:18:32,320
recommendations

539
00:18:30,240 --> 00:18:33,840
um because there are lots of people who

540
00:18:32,320 --> 00:18:35,600
are interested in having those services

541
00:18:33,840 --> 00:18:37,440
paid for

542
00:18:35,600 --> 00:18:40,400
however as you can see the list of

543
00:18:37,440 --> 00:18:43,919
uspstf recommendation grades those are a

544
00:18:40,400 --> 00:18:47,440
b c d and i um specifically i think

545
00:18:43,919 --> 00:18:49,760
we're focused today on i statements and

546
00:18:47,440 --> 00:18:51,360
mostly the task force looks at um

547
00:18:49,760 --> 00:18:53,440
evidence and then comes to two

548
00:18:51,360 --> 00:18:55,840
considerations and one is about the

549
00:18:53,440 --> 00:18:57,679
magnitude of net benefit for for

550
00:18:55,840 --> 00:19:00,880
delivering a particular service and then

551
00:18:57,679 --> 00:19:02,960
the certainty around that benefit um

552
00:19:00,880 --> 00:19:05,039
mostly based on the quality of evidence

553
00:19:02,960 --> 00:19:06,640
itself and if there's high certainty

554
00:19:05,039 --> 00:19:07,919
that there's a substantial net benefit

555
00:19:06,640 --> 00:19:10,080
then they'll give a service in a

556
00:19:07,919 --> 00:19:12,640
recommendation if there's high certainty

557
00:19:10,080 --> 00:19:14,000
that the net benefit is moderate or or

558
00:19:12,640 --> 00:19:16,880
on the other hand it's only moderate

559
00:19:14,000 --> 00:19:18,640
certainty about a significant

560
00:19:16,880 --> 00:19:20,480
benefit significant net benefit then

561
00:19:18,640 --> 00:19:22,480
they'll give it a b recommendation when

562
00:19:20,480 --> 00:19:24,960
the balance is tricky about the net

563
00:19:22,480 --> 00:19:27,120
benefit um uh or either around certainty

564
00:19:24,960 --> 00:19:28,880
and that magnitude then then really a c

565
00:19:27,120 --> 00:19:30,799
recommendation and that really focuses

566
00:19:28,880 --> 00:19:32,240
on discussions between patients and

567
00:19:30,799 --> 00:19:34,160
their clinicians about whether that

568
00:19:32,240 --> 00:19:36,880
service is right for them and then there

569
00:19:34,160 --> 00:19:38,559
are um services that are are can be

570
00:19:36,880 --> 00:19:40,080
harmful or at least have no benefit at

571
00:19:38,559 --> 00:19:41,760
all and those are typically given a d

572
00:19:40,080 --> 00:19:43,600
recommendation and then as we mentioned

573
00:19:41,760 --> 00:19:45,520
the eye is issued when evidence is

574
00:19:43,600 --> 00:19:47,120
insufficient to balance those benefits

575
00:19:45,520 --> 00:19:48,240
and harms or there's a great deal of

576
00:19:47,120 --> 00:19:50,000
uncertainty

577
00:19:48,240 --> 00:19:51,760
and we've figured that nih is well

578
00:19:50,000 --> 00:19:53,440
positioned to support research needed to

579
00:19:51,760 --> 00:19:54,960
close some of those evidence gaps around

580
00:19:53,440 --> 00:19:56,880
i statements especially if they're in

581
00:19:54,960 --> 00:19:59,200
line with some of the missions of the

582
00:19:56,880 --> 00:20:00,960
institutes and centers

583
00:19:59,200 --> 00:20:03,039
there is a long process that the task

584
00:20:00,960 --> 00:20:05,200
force goes through starting from topic

585
00:20:03,039 --> 00:20:08,080
nominations all the way to their final

586
00:20:05,200 --> 00:20:09,440
recommendation um statements and then

587
00:20:08,080 --> 00:20:11,039
they have several opportunities for

588
00:20:09,440 --> 00:20:13,039
public comment along the way first they

589
00:20:11,039 --> 00:20:14,799
begin with topic nominations and in

590
00:20:13,039 --> 00:20:16,400
addition to adding in new topics they're

591
00:20:14,799 --> 00:20:18,159
always reconsidering topics that they've

592
00:20:16,400 --> 00:20:20,799
had from the past to try to keep them

593
00:20:18,159 --> 00:20:22,000
updated every five years once they've

594
00:20:20,799 --> 00:20:23,840
identified the topic that they're going

595
00:20:22,000 --> 00:20:25,280
to work on to begin to develop a draft

596
00:20:23,840 --> 00:20:27,200
research plan and that draft research

597
00:20:25,280 --> 00:20:28,960
plan is usually based on a peacock's

598
00:20:27,200 --> 00:20:32,000
kind of framework that looks at the

599
00:20:28,960 --> 00:20:34,960
patient population uh interventions uh

600
00:20:32,000 --> 00:20:38,320
looks at various types of treatments as

601
00:20:34,960 --> 00:20:40,880
well as um uh timing of those services

602
00:20:38,320 --> 00:20:42,480
um and uh and and then the settings that

603
00:20:40,880 --> 00:20:44,080
those are in and

604
00:20:42,480 --> 00:20:46,400
as well as the comparisons for many of

605
00:20:44,080 --> 00:20:48,640
those things and that the that work is

606
00:20:46,400 --> 00:20:50,960
done by an evidence-based practice

607
00:20:48,640 --> 00:20:53,039
center or epc so if we talk about epc

608
00:20:50,960 --> 00:20:54,880
reports they're the ones who synthesize

609
00:20:53,039 --> 00:20:56,799
the evidence across things and actually

610
00:20:54,880 --> 00:20:58,159
they help develop that research plan

611
00:20:56,799 --> 00:21:00,640
that research plan has an opportunity

612
00:20:58,159 --> 00:21:02,240
for public comment and and then it gets

613
00:21:00,640 --> 00:21:04,400
finalized and then based on that

614
00:21:02,240 --> 00:21:06,960
research plan the task force will begin

615
00:21:04,400 --> 00:21:08,640
an evidence review that when will lead

616
00:21:06,960 --> 00:21:10,480
to development of their final

617
00:21:08,640 --> 00:21:11,760
recommendation with an opportunity for

618
00:21:10,480 --> 00:21:13,440
public comment for that draft

619
00:21:11,760 --> 00:21:15,039
recommendation and the evidence report

620
00:21:13,440 --> 00:21:16,799
in between

621
00:21:15,039 --> 00:21:18,640
here's just a standard template the task

622
00:21:16,799 --> 00:21:21,200
force uses an analytic framework to do

623
00:21:18,640 --> 00:21:23,120
this work um uh and and this is a

624
00:21:21,200 --> 00:21:24,480
screening one so it may not apply to

625
00:21:23,120 --> 00:21:26,240
behavioral counseling interventions

626
00:21:24,480 --> 00:21:28,640
where there's a slightly different uh

627
00:21:26,240 --> 00:21:30,240
framework as well as for preventive

628
00:21:28,640 --> 00:21:31,840
medications also a slightly different

629
00:21:30,240 --> 00:21:33,520
framework but

630
00:21:31,840 --> 00:21:35,039
as you can see there's linkages of

631
00:21:33,520 --> 00:21:37,679
evidence across the way here and it's

632
00:21:35,039 --> 00:21:40,000
pretty rare that you get an overarching

633
00:21:37,679 --> 00:21:42,000
um pathway from say for screening for

634
00:21:40,000 --> 00:21:43,760
instance all the way to outcomes in one

635
00:21:42,000 --> 00:21:45,679
study typically what you do have to have

636
00:21:43,760 --> 00:21:47,280
to build a chain of evidence over time

637
00:21:45,679 --> 00:21:49,200
to connect screening with early

638
00:21:47,280 --> 00:21:50,640
detection then with treatment with

639
00:21:49,200 --> 00:21:52,559
potentially intermediate outcomes and

640
00:21:50,640 --> 00:21:53,679
then at least have a sense for um how

641
00:21:52,559 --> 00:21:55,440
those intermediate outcomes are

642
00:21:53,679 --> 00:21:57,840
associated with final outcomes around

643
00:21:55,440 --> 00:21:59,200
morbidity and mortality they also look

644
00:21:57,840 --> 00:22:01,840
carefully first

645
00:21:59,200 --> 00:22:03,919
side effects adverse effects either from

646
00:22:01,840 --> 00:22:05,039
screening or treatment or counseling as

647
00:22:03,919 --> 00:22:07,919
it might be

648
00:22:05,039 --> 00:22:09,520
and then evaluate all of this into their

649
00:22:07,919 --> 00:22:12,640
chain of evidence and look for the

650
00:22:09,520 --> 00:22:14,640
coherence across all of these linkages

651
00:22:12,640 --> 00:22:15,840
in their chain of evidence

652
00:22:14,640 --> 00:22:18,080
as they look at the evidence they

653
00:22:15,840 --> 00:22:20,880
actually look very carefully at these

654
00:22:18,080 --> 00:22:22,559
six principles um these questions are

655
00:22:20,880 --> 00:22:24,000
are critical to the way

656
00:22:22,559 --> 00:22:26,559
the task force thinks about it hence the

657
00:22:24,000 --> 00:22:28,080
name critical appraisal questions

658
00:22:26,559 --> 00:22:29,919
first do the studies have appropriate

659
00:22:28,080 --> 00:22:31,600
research designed to answer those key

660
00:22:29,919 --> 00:22:33,600
questions to what extent are the

661
00:22:31,600 --> 00:22:35,360
existing studies of sufficient qualities

662
00:22:33,600 --> 00:22:37,440
but looking specifically for example at

663
00:22:35,360 --> 00:22:38,559
their internal validity to what extent

664
00:22:37,440 --> 00:22:40,880
are the results of the studies

665
00:22:38,559 --> 00:22:43,039
generalizable again to the u.s primary

666
00:22:40,880 --> 00:22:44,799
care population of interest for the

667
00:22:43,039 --> 00:22:46,080
intervention in the particular setting

668
00:22:44,799 --> 00:22:47,760
or situation

669
00:22:46,080 --> 00:22:49,360
how many and how large are the studies

670
00:22:47,760 --> 00:22:51,440
that address these questions how

671
00:22:49,360 --> 00:22:53,679
consistent are the results and are there

672
00:22:51,440 --> 00:22:55,440
additional factors that i'll assist the

673
00:22:53,679 --> 00:22:58,080
task force in drawing conclusions so for

674
00:22:55,440 --> 00:22:59,360
example a biologic model and then across

675
00:22:58,080 --> 00:23:01,039
all of these critical appraisal

676
00:22:59,360 --> 00:23:03,120
questions they look at each individual

677
00:23:01,039 --> 00:23:05,280
key question in that analytic framework

678
00:23:03,120 --> 00:23:07,520
and then look for coherence about how it

679
00:23:05,280 --> 00:23:10,480
all hangs together is there consistency

680
00:23:07,520 --> 00:23:11,919
across those key questions

681
00:23:10,480 --> 00:23:13,360
in addition to

682
00:23:11,919 --> 00:23:16,000
those critical appraisal questions they

683
00:23:13,360 --> 00:23:17,280
also look very closely at applicability

684
00:23:16,000 --> 00:23:19,840
and that's the extent to which the

685
00:23:17,280 --> 00:23:22,720
results are generalizable to the general

686
00:23:19,840 --> 00:23:24,640
asymptomatic primary care population or

687
00:23:22,720 --> 00:23:26,640
a specific primary care population of

688
00:23:24,640 --> 00:23:28,559
interest and frequently studies evaluate

689
00:23:26,640 --> 00:23:30,720
different populations or providers or

690
00:23:28,559 --> 00:23:32,240
settings than the asymptomatic primary

691
00:23:30,720 --> 00:23:34,159
care population and that's where

692
00:23:32,240 --> 00:23:36,640
challenges can come in and create some

693
00:23:34,159 --> 00:23:38,960
of those evidence gaps so specifically

694
00:23:36,640 --> 00:23:40,240
looking at u.s populations and then when

695
00:23:38,960 --> 00:23:42,159
you're looking at interventions the

696
00:23:40,240 --> 00:23:44,640
acceptability and feasibility and the

697
00:23:42,159 --> 00:23:47,120
availability of those interventions

698
00:23:44,640 --> 00:23:49,120
in primary care and the task force has

699
00:23:47,120 --> 00:23:50,799
done an amazing job so documenting their

700
00:23:49,120 --> 00:23:52,880
processes you can find them all if you

701
00:23:50,799 --> 00:23:53,919
go to their website u.s preventive

702
00:23:52,880 --> 00:23:55,520
services

703
00:23:53,919 --> 00:23:58,159
taskforce.org

704
00:23:55,520 --> 00:24:00,240
you can find their um their methods

705
00:23:58,159 --> 00:24:02,400
manual and uh and lots of great

706
00:24:00,240 --> 00:24:04,960
information there about how

707
00:24:02,400 --> 00:24:06,640
the process works for them

708
00:24:04,960 --> 00:24:08,240
so from our perspective in the office of

709
00:24:06,640 --> 00:24:10,000
disease prevention we collaborate with

710
00:24:08,240 --> 00:24:11,919
art very closely and really trying to

711
00:24:10,000 --> 00:24:13,440
disseminate some of these evidence gaps

712
00:24:11,919 --> 00:24:15,760
in clinical prevention and you can find

713
00:24:13,440 --> 00:24:17,200
those evidence gaps in several places

714
00:24:15,760 --> 00:24:18,960
through the task force's work you can

715
00:24:17,200 --> 00:24:20,720
find them in the evidence reports that

716
00:24:18,960 --> 00:24:22,880
are generated by arc

717
00:24:20,720 --> 00:24:24,640
and in the epcs you can find those

718
00:24:22,880 --> 00:24:26,320
evidence gaps in the recommendation

719
00:24:24,640 --> 00:24:29,520
statements themselves in the research

720
00:24:26,320 --> 00:24:31,440
needs and gaps section um they also the

721
00:24:29,520 --> 00:24:33,360
task force produces an annual report to

722
00:24:31,440 --> 00:24:35,840
congress and it highlights many of the

723
00:24:33,360 --> 00:24:37,600
high priority research gaps there

724
00:24:35,840 --> 00:24:39,279
we at our website at odp we have a

725
00:24:37,600 --> 00:24:41,039
complete listing of all the i statements

726
00:24:39,279 --> 00:24:42,559
in their course to finding research

727
00:24:41,039 --> 00:24:44,520
needs and gaps

728
00:24:42,559 --> 00:24:46,320
and you can find it at

729
00:24:44,520 --> 00:24:48,799
prevention.nih.gov and just look for the

730
00:24:46,320 --> 00:24:51,039
research priorities link there

731
00:24:48,799 --> 00:24:53,600
in addition we take those i statements

732
00:24:51,039 --> 00:24:55,200
and those needs and gaps and then we we

733
00:24:53,600 --> 00:24:57,440
communicate them with all the nih

734
00:24:55,200 --> 00:25:00,240
institutes and centers through an annual

735
00:24:57,440 --> 00:25:02,080
um uh i statement reporting survey uh

736
00:25:00,240 --> 00:25:04,400
around april of every year we send out

737
00:25:02,080 --> 00:25:07,360
that survey to all the ics with some

738
00:25:04,400 --> 00:25:09,760
pre-populated um i statements that are

739
00:25:07,360 --> 00:25:11,440
relevant to their work their mission and

740
00:25:09,760 --> 00:25:14,880
try to understand what's being done in

741
00:25:11,440 --> 00:25:16,559
terms of grants uh foas contracts and

742
00:25:14,880 --> 00:25:18,480
other types of research or research

743
00:25:16,559 --> 00:25:19,679
activities that could inform some of

744
00:25:18,480 --> 00:25:22,000
those gaps

745
00:25:19,679 --> 00:25:26,400
and then our office along with arc often

746
00:25:22,000 --> 00:25:26,400
convened meetings around those nih uh

747
00:25:26,720 --> 00:25:30,880
some of that nih work um and to present

748
00:25:28,960 --> 00:25:32,720
some of those evidence gaps at national

749
00:25:30,880 --> 00:25:34,640
meetings and we're specifically

750
00:25:32,720 --> 00:25:36,320
interested in in new and innovative

751
00:25:34,640 --> 00:25:38,000
research

752
00:25:36,320 --> 00:25:40,960
and as dr compton mentioned he

753
00:25:38,000 --> 00:25:42,960
highlighted the um b and i statements

754
00:25:40,960 --> 00:25:44,720
around screening for drug use there's

755
00:25:42,960 --> 00:25:46,960
also an eye statement around primary

756
00:25:44,720 --> 00:25:48,320
care space interventions in children

757
00:25:46,960 --> 00:25:50,559
adolescents and young adults this is

758
00:25:48,320 --> 00:25:52,320
just an example of what the i statement

759
00:25:50,559 --> 00:25:54,320
looks like and i won't read it to you

760
00:25:52,320 --> 00:25:56,400
but you can get a sense for for how

761
00:25:54,320 --> 00:25:58,799
detailed it is around the populations

762
00:25:56,400 --> 00:26:01,279
the balance of benefits and harms

763
00:25:58,799 --> 00:26:03,440
and what even some when the potential

764
00:26:01,279 --> 00:26:04,559
interventions are and then here the list

765
00:26:03,440 --> 00:26:06,240
of research

766
00:26:04,559 --> 00:26:07,279
needs and gaps that came out of it you

767
00:26:06,240 --> 00:26:09,279
can see they're they're fairly

768
00:26:07,279 --> 00:26:10,960
substantial and some can be more

769
00:26:09,279 --> 00:26:12,400
specific than others

770
00:26:10,960 --> 00:26:14,240
around cannabis prevention or

771
00:26:12,400 --> 00:26:16,880
standardizing outcome measurement

772
00:26:14,240 --> 00:26:18,559
looking for replication um other types

773
00:26:16,880 --> 00:26:19,600
of things as well as other evidence

774
00:26:18,559 --> 00:26:22,840
around

775
00:26:19,600 --> 00:26:26,000
prevention i think in technology-based

776
00:26:22,840 --> 00:26:28,080
interventions so it's it's wise that you

777
00:26:26,000 --> 00:26:29,279
uh re re-title the session to some

778
00:26:28,080 --> 00:26:30,480
degree challenges and opportunities

779
00:26:29,279 --> 00:26:31,919
because that's exactly where they lie

780
00:26:30,480 --> 00:26:33,360
here with the task force and some of

781
00:26:31,919 --> 00:26:34,799
these challenges or the evidence gaps

782
00:26:33,360 --> 00:26:36,320
are often described in multiple places

783
00:26:34,799 --> 00:26:37,760
either in the epc report or in the

784
00:26:36,320 --> 00:26:39,360
recommendation statement some of the

785
00:26:37,760 --> 00:26:41,360
descriptions are non-specific so it's

786
00:26:39,360 --> 00:26:43,279
hard to know what what types of high

787
00:26:41,360 --> 00:26:44,720
quality studies which would we need but

788
00:26:43,279 --> 00:26:46,799
there's some real opportunities here for

789
00:26:44,720 --> 00:26:48,640
linking some of these gaps um to the

790
00:26:46,799 --> 00:26:50,799
task force criteria for study inclusion

791
00:26:48,640 --> 00:26:51,600
to sort of create a more coordinated

792
00:26:50,799 --> 00:26:53,679
effort

793
00:26:51,600 --> 00:26:55,360
to describe and communicate some of

794
00:26:53,679 --> 00:26:57,120
those evidence gaps with the ultimate

795
00:26:55,360 --> 00:26:59,520
goal of really reducing the number of i

796
00:26:57,120 --> 00:27:01,600
statements improving the certainty

797
00:26:59,520 --> 00:27:03,520
that clinicians have in delivering care

798
00:27:01,600 --> 00:27:06,080
and really i think nih has a critical

799
00:27:03,520 --> 00:27:07,840
role in disseminating and addressing

800
00:27:06,080 --> 00:27:10,640
those evidence gaps and while this

801
00:27:07,840 --> 00:27:12,720
doesn't speak to uh i haven't spoken to

802
00:27:10,640 --> 00:27:15,279
elect use of electronic health records

803
00:27:12,720 --> 00:27:17,200
you do get a sense of if you had a study

804
00:27:15,279 --> 00:27:19,440
where electronic health record data was

805
00:27:17,200 --> 00:27:21,120
included it would still have to be held

806
00:27:19,440 --> 00:27:23,600
to sort of a very similar standard as

807
00:27:21,120 --> 00:27:26,880
the task force worked through their um

808
00:27:23,600 --> 00:27:28,640
uh reports uh through the epc to base

809
00:27:26,880 --> 00:27:30,080
their recommendation statements so with

810
00:27:28,640 --> 00:27:32,159
that i'll just leave you with my contact

811
00:27:30,080 --> 00:27:33,919
information and i'm gonna stop sharing

812
00:27:32,159 --> 00:27:35,440
and uh thanks for the time and look

813
00:27:33,919 --> 00:27:38,399
forward to the um

814
00:27:35,440 --> 00:27:40,240
uh the conversation

815
00:27:38,399 --> 00:27:41,919
uh but um

816
00:27:40,240 --> 00:27:43,919
my name is mario turon

817
00:27:41,919 --> 00:27:48,240
i am a medical officer at physician

818
00:27:43,919 --> 00:27:50,080
informaticis uh within arc at uh within

819
00:27:48,240 --> 00:27:51,440
the division of healthcare

820
00:27:50,080 --> 00:27:53,120
research

821
00:27:51,440 --> 00:27:55,200
within the center for evidence and

822
00:27:53,120 --> 00:27:56,080
practice improvement and all that outfit

823
00:27:55,200 --> 00:27:58,080
sue

824
00:27:56,080 --> 00:28:03,000
and i'm going to take uh somewhat of a

825
00:27:58,080 --> 00:28:03,000
step back and share with you

826
00:28:03,919 --> 00:28:09,520
a greater view and sense of efforts to

827
00:28:05,919 --> 00:28:12,240
enhance ehr-based clinical decision

828
00:28:09,520 --> 00:28:14,960
support interventions to improve

829
00:28:12,240 --> 00:28:17,840
substance abuse prevention

830
00:28:14,960 --> 00:28:17,840
i have no

831
00:28:18,240 --> 00:28:24,240
uh relevant uh disclosures

832
00:28:21,039 --> 00:28:24,240
you can go to the next slide

833
00:28:24,720 --> 00:28:29,200
so our background within the division of

834
00:28:26,799 --> 00:28:32,080
healthcare research our mission is to

835
00:28:29,200 --> 00:28:33,520
determine how the various components of

836
00:28:32,080 --> 00:28:35,840
ever evolving digital healthcare

837
00:28:33,520 --> 00:28:38,000
ecosystem can best come together to

838
00:28:35,840 --> 00:28:39,600
positively affect healthcare delivery

839
00:28:38,000 --> 00:28:40,799
and create value for patients and their

840
00:28:39,600 --> 00:28:42,960
families

841
00:28:40,799 --> 00:28:44,559
our vision is that every patient and

842
00:28:42,960 --> 00:28:47,120
care team

843
00:28:44,559 --> 00:28:48,640
will have ready access to all applicable

844
00:28:47,120 --> 00:28:51,279
data and knowledge

845
00:28:48,640 --> 00:28:53,600
mediated by advanced analytics and

846
00:28:51,279 --> 00:28:57,039
understandable visualizations to address

847
00:28:53,600 --> 00:28:59,120
a patient's health and health care

848
00:28:57,039 --> 00:29:01,600
we are after the ultimate data

849
00:28:59,120 --> 00:29:03,679
and knowledge liquid liquidity

850
00:29:01,600 --> 00:29:04,880
uh at the point of care and it's this

851
00:29:03,679 --> 00:29:07,679
last point

852
00:29:04,880 --> 00:29:09,760
uh of focus and one of the ways we are

853
00:29:07,679 --> 00:29:12,320
trying to achieve this point of focus is

854
00:29:09,760 --> 00:29:15,120
through clinical decision support i can

855
00:29:12,320 --> 00:29:15,120
go to the next slide

856
00:29:18,880 --> 00:29:21,840
so

857
00:29:19,679 --> 00:29:23,440
what is clinical decision support and to

858
00:29:21,840 --> 00:29:24,559
give you a brief overview and a broad

859
00:29:23,440 --> 00:29:26,640
view is

860
00:29:24,559 --> 00:29:28,880
based on what we know as uh the five

861
00:29:26,640 --> 00:29:31,039
rights that clinical decision support or

862
00:29:28,880 --> 00:29:33,200
cds should deliver the right information

863
00:29:31,039 --> 00:29:36,000
to the right people in the right formats

864
00:29:33,200 --> 00:29:38,640
the right channels at the right times

865
00:29:36,000 --> 00:29:41,120
although it's commonly associated

866
00:29:38,640 --> 00:29:45,039
i'll make clear cds is not

867
00:29:41,120 --> 00:29:46,799
just alerts or on-screen pop-ups

868
00:29:45,039 --> 00:29:48,880
nor is it only for

869
00:29:46,799 --> 00:29:50,040
physicians or those at the

870
00:29:48,880 --> 00:29:51,279
point of care making

871
00:29:50,040 --> 00:29:53,440
[Music]

872
00:29:51,279 --> 00:29:55,360
health care decisions

873
00:29:53,440 --> 00:29:57,520
clinical decision support is a process

874
00:29:55,360 --> 00:30:00,159
in a set of tools that are technology

875
00:29:57,520 --> 00:30:03,039
enabled but in the end can support a

876
00:30:00,159 --> 00:30:05,919
person's capacities and hopefully

877
00:30:03,039 --> 00:30:08,880
improve the quality of care

878
00:30:05,919 --> 00:30:12,000
next slide

879
00:30:08,880 --> 00:30:13,600
our ongoing initiative since 2016 aims

880
00:30:12,000 --> 00:30:16,320
at getting the evidence and support and

881
00:30:13,600 --> 00:30:18,399
practice to both patients and clinicians

882
00:30:16,320 --> 00:30:20,559
through clinical decision support that

883
00:30:18,399 --> 00:30:23,120
is more shareable standards based and

884
00:30:20,559 --> 00:30:24,720
publicly available i'll go over some of

885
00:30:23,120 --> 00:30:27,039
our efforts and how

886
00:30:24,720 --> 00:30:28,320
cds and the ehr can be leveraged and

887
00:30:27,039 --> 00:30:31,279
utilized

888
00:30:28,320 --> 00:30:33,600
for substance abuse prevention

889
00:30:31,279 --> 00:30:36,159
next slide

890
00:30:33,600 --> 00:30:38,559
so due to its high penetration most but

891
00:30:36,159 --> 00:30:41,600
not all clinical settings the ehr can be

892
00:30:38,559 --> 00:30:43,840
viewed as an avenue to get cds into the

893
00:30:41,600 --> 00:30:46,240
hands of those that need it most

894
00:30:43,840 --> 00:30:48,559
this figure helps illustrate an overview

895
00:30:46,240 --> 00:30:50,399
of the cds ehr ecosystem and the

896
00:30:48,559 --> 00:30:52,559
interplay where we can see the flow of

897
00:30:50,399 --> 00:30:54,960
evidence data and knowledge and their

898
00:30:52,559 --> 00:30:56,799
sources into cds and eventually into the

899
00:30:54,960 --> 00:30:58,720
ehr

900
00:30:56,799 --> 00:31:00,640
it emphasizes the multiple channels that

901
00:30:58,720 --> 00:31:02,880
are needed to be connected not only to

902
00:31:00,640 --> 00:31:06,640
get cds into the ehr but to make sure

903
00:31:02,880 --> 00:31:08,559
that cds is relevant and trustworthy

904
00:31:06,640 --> 00:31:11,039
however due to these multiple channels

905
00:31:08,559 --> 00:31:13,840
and sources of information evidence

906
00:31:11,039 --> 00:31:16,399
exchange between them can be complex and

907
00:31:13,840 --> 00:31:16,399
difficult

908
00:31:16,480 --> 00:31:20,159
efforts

909
00:31:18,080 --> 00:31:21,200
efforts to create cds are sometimes

910
00:31:20,159 --> 00:31:23,600
siloed

911
00:31:21,200 --> 00:31:26,480
as we can see with these uh

912
00:31:23,600 --> 00:31:29,039
somewhat clinical silos at the bottom

913
00:31:26,480 --> 00:31:31,519
and only compatible for those who create

914
00:31:29,039 --> 00:31:34,159
it leading to duplicate efforts because

915
00:31:31,519 --> 00:31:35,919
of an inability to share

916
00:31:34,159 --> 00:31:37,679
cds efficiently

917
00:31:35,919 --> 00:31:41,279
so to combat these

918
00:31:37,679 --> 00:31:43,840
duplicative efforts and promote uh

919
00:31:41,279 --> 00:31:45,039
free exchange interoperability standards

920
00:31:43,840 --> 00:31:47,440
have

921
00:31:45,039 --> 00:31:50,399
are being promoted and have been created

922
00:31:47,440 --> 00:31:52,880
however they're not yet widespread

923
00:31:50,399 --> 00:31:54,720
amongst all the hr vendors and

924
00:31:52,880 --> 00:31:56,399
organizations

925
00:31:54,720 --> 00:31:59,760
uh the main exchange standard you should

926
00:31:56,399 --> 00:32:01,840
be aware of uh if not already is the

927
00:31:59,760 --> 00:32:04,399
fast healthcare intraoperatively

928
00:32:01,840 --> 00:32:06,000
resources standard also known as fire

929
00:32:04,399 --> 00:32:08,720
it is a standard for exchanging health

930
00:32:06,000 --> 00:32:10,880
information electronically and specifies

931
00:32:08,720 --> 00:32:12,880
the content of the data exchange between

932
00:32:10,880 --> 00:32:14,640
health care applications and how the

933
00:32:12,880 --> 00:32:17,360
exchange is implemented

934
00:32:14,640 --> 00:32:19,600
and managed i won't go into detail on uh

935
00:32:17,360 --> 00:32:22,159
the background of this but just to make

936
00:32:19,600 --> 00:32:24,399
those in the audience aware that

937
00:32:22,159 --> 00:32:26,320
when implementing clinical decision

938
00:32:24,399 --> 00:32:28,559
report into the ehr

939
00:32:26,320 --> 00:32:30,640
we're trying to move our efforts uh to

940
00:32:28,559 --> 00:32:33,360
include these standards to make them ehr

941
00:32:30,640 --> 00:32:36,080
agnostic and available to everyone

942
00:32:33,360 --> 00:32:36,080
uh next slide

943
00:32:37,200 --> 00:32:40,480
and um

944
00:32:38,640 --> 00:32:42,799
so these standards and their promotion

945
00:32:40,480 --> 00:32:44,720
and use are critical in our work

946
00:32:42,799 --> 00:32:46,720
surrounding cds and believe it is an

947
00:32:44,720 --> 00:32:49,519
important component in the cds

948
00:32:46,720 --> 00:32:51,760
ecosystem so to support these efforts in

949
00:32:49,519 --> 00:32:54,720
partnership with the mitre corporation

950
00:32:51,760 --> 00:32:57,840
we have created cds connect

951
00:32:54,720 --> 00:32:57,840
go to the next slide

952
00:32:59,279 --> 00:33:04,640
so this is a publicly available and free

953
00:33:02,000 --> 00:33:07,039
open source tool that not only is a

954
00:33:04,640 --> 00:33:10,399
repository for clinical decision support

955
00:33:07,039 --> 00:33:13,039
tools but also is an authoring tool to

956
00:33:10,399 --> 00:33:14,000
allow the creation of clinical decision

957
00:33:13,039 --> 00:33:16,080
support

958
00:33:14,000 --> 00:33:18,000
using interoperable standards it also

959
00:33:16,080 --> 00:33:20,720
has the capabilities to test your

960
00:33:18,000 --> 00:33:23,120
clinical decision support with uh

961
00:33:20,720 --> 00:33:26,159
synthetic patient data as well and

962
00:33:23,120 --> 00:33:29,360
this concept of operation shows how from

963
00:33:26,159 --> 00:33:31,840
going left to right how you have cds

964
00:33:29,360 --> 00:33:33,200
contributors authors

965
00:33:31,840 --> 00:33:35,720
uh can

966
00:33:33,200 --> 00:33:39,279
contribute to cds connect where it will

967
00:33:35,720 --> 00:33:41,440
create um interoperable clinical

968
00:33:39,279 --> 00:33:43,519
decision support after

969
00:33:41,440 --> 00:33:46,159
review with our partners at mitre to

970
00:33:43,519 --> 00:33:49,880
make sure that your cds is clinically

971
00:33:46,159 --> 00:33:49,880
relevant and safe

972
00:33:51,000 --> 00:33:54,109
[Music]

973
00:33:55,519 --> 00:33:58,159
you know some nice background music

974
00:33:57,039 --> 00:34:00,399
there nice

975
00:33:58,159 --> 00:34:02,320
um and then using uh interoperable

976
00:34:00,399 --> 00:34:05,519
standards to get it into the point of

977
00:34:02,320 --> 00:34:07,360
care at the ehr or uh with patients and

978
00:34:05,519 --> 00:34:10,399
cells

979
00:34:07,360 --> 00:34:13,200
uh can you go to the next slide

980
00:34:10,399 --> 00:34:15,359
so to showcase how we leverage uh cds

981
00:34:13,200 --> 00:34:17,040
and the ehr into play

982
00:34:15,359 --> 00:34:19,280
to support substance abuse prevention

983
00:34:17,040 --> 00:34:20,960
and primary primary care i'm going to

984
00:34:19,280 --> 00:34:23,440
share some of our previous and ongoing

985
00:34:20,960 --> 00:34:25,520
efforts in chronic pain and opioid

986
00:34:23,440 --> 00:34:27,359
management although not specific to

987
00:34:25,520 --> 00:34:29,599
substance abuse prevention there are

988
00:34:27,359 --> 00:34:31,359
many parallels to both and highlight

989
00:34:29,599 --> 00:34:32,639
avenues that can be utilized for

990
00:34:31,359 --> 00:34:34,879
prevention

991
00:34:32,639 --> 00:34:37,200
these include patient patient apps uh

992
00:34:34,879 --> 00:34:39,679
the boston shared decision making and

993
00:34:37,200 --> 00:34:43,200
clinician-facing apps that utilize

994
00:34:39,679 --> 00:34:44,960
ehr data uh at the point of care uh the

995
00:34:43,200 --> 00:34:47,359
first example i'd like to highlight is

996
00:34:44,960 --> 00:34:49,599
uh the tapering and patient reported

997
00:34:47,359 --> 00:34:51,599
outcomes for chronic pain management app

998
00:34:49,599 --> 00:34:54,800
or known as the taper

999
00:34:51,599 --> 00:34:57,520
in a contracted effort with medstar and

1000
00:34:54,800 --> 00:34:58,640
also in collaboration with input from

1001
00:34:57,520 --> 00:35:01,680
chronic

1002
00:34:58,640 --> 00:35:04,240
pain patients caregivers

1003
00:35:01,680 --> 00:35:06,880
developers and pain specialists a

1004
00:35:04,240 --> 00:35:09,440
patient-basing and provider-facing

1005
00:35:06,880 --> 00:35:11,440
standards-based app was developed and is

1006
00:35:09,440 --> 00:35:12,560
currently finishing out its uh piloted

1007
00:35:11,440 --> 00:35:15,720
roll-out

1008
00:35:12,560 --> 00:35:15,720
next slide

1009
00:35:16,480 --> 00:35:20,480
so to walk you through this at an

1010
00:35:18,240 --> 00:35:22,560
initial visit a patient is introduced to

1011
00:35:20,480 --> 00:35:24,160
uh the taper app

1012
00:35:22,560 --> 00:35:26,240
and

1013
00:35:24,160 --> 00:35:29,200
uh they're able to track their pain and

1014
00:35:26,240 --> 00:35:31,440
daily function uh with the with the help

1015
00:35:29,200 --> 00:35:34,320
of the app the clinician on the back end

1016
00:35:31,440 --> 00:35:36,240
is able to see the patient's inputs and

1017
00:35:34,320 --> 00:35:38,320
either between visits or during a

1018
00:35:36,240 --> 00:35:40,640
follow-up able to review patient data

1019
00:35:38,320 --> 00:35:42,960
and work together with the patient uh to

1020
00:35:40,640 --> 00:35:44,960
create a future plan for

1021
00:35:42,960 --> 00:35:47,520
opioid tampering

1022
00:35:44,960 --> 00:35:49,280
uh next slide

1023
00:35:47,520 --> 00:35:50,320
the patient home screen welcome patient

1024
00:35:49,280 --> 00:35:51,760
and gives a

1025
00:35:50,320 --> 00:35:53,440
an overview

1026
00:35:51,760 --> 00:35:56,800
of the patient's current therapies and

1027
00:35:53,440 --> 00:35:56,800
regimen next slide

1028
00:35:57,680 --> 00:36:01,760
it allows journaling activities by the

1029
00:35:59,680 --> 00:36:04,400
patient and allows them to mark their

1030
00:36:01,760 --> 00:36:06,640
pain scores

1031
00:36:04,400 --> 00:36:10,400
next slide

1032
00:36:06,640 --> 00:36:12,400
utilizes it utilizes promised

1033
00:36:10,400 --> 00:36:13,520
questionnaire to further capture patient

1034
00:36:12,400 --> 00:36:15,599
input

1035
00:36:13,520 --> 00:36:17,520
and outcomes

1036
00:36:15,599 --> 00:36:20,000
next slide

1037
00:36:17,520 --> 00:36:21,200
and on the provider side the provider is

1038
00:36:20,000 --> 00:36:23,599
able to

1039
00:36:21,200 --> 00:36:26,240
view paper history

1040
00:36:23,599 --> 00:36:27,680
their plan summaries and also see what

1041
00:36:26,240 --> 00:36:28,800
the patient

1042
00:36:27,680 --> 00:36:30,000
reported

1043
00:36:28,800 --> 00:36:32,560
as well

1044
00:36:30,000 --> 00:36:32,560
next slide

1045
00:36:34,000 --> 00:36:36,880
so

1046
00:36:34,960 --> 00:36:39,920
on the ehr

1047
00:36:36,880 --> 00:36:41,680
data screen this relies a review of ehr

1048
00:36:39,920 --> 00:36:44,079
link patient data that is current with

1049
00:36:41,680 --> 00:36:45,599
their medications and also linked with

1050
00:36:44,079 --> 00:36:48,079
pdmp data

1051
00:36:45,599 --> 00:36:48,079
next slide

1052
00:36:48,320 --> 00:36:52,960
it allows then for a

1053
00:36:50,560 --> 00:36:55,119
taper plan to be created

1054
00:36:52,960 --> 00:36:56,800
that gives information on you can see on

1055
00:36:55,119 --> 00:36:59,119
the right hand side with your

1056
00:36:56,800 --> 00:37:00,640
recommendation and the mme reduction

1057
00:36:59,119 --> 00:37:01,680
percentage

1058
00:37:00,640 --> 00:37:04,240
along with

1059
00:37:01,680 --> 00:37:06,720
some recommendation prompts if needed

1060
00:37:04,240 --> 00:37:06,720
next slide

1061
00:37:07,520 --> 00:37:12,160
it also gives information on

1062
00:37:09,760 --> 00:37:14,480
alternative opioid pain medications and

1063
00:37:12,160 --> 00:37:16,160
other therapies that can be tried for

1064
00:37:14,480 --> 00:37:17,920
the patient as well as

1065
00:37:16,160 --> 00:37:18,960
medications for withdrawal symptoms as

1066
00:37:17,920 --> 00:37:22,119
well

1067
00:37:18,960 --> 00:37:22,119
next slide

1068
00:37:22,160 --> 00:37:26,320
our second application in partnership

1069
00:37:24,480 --> 00:37:29,200
with rti international created an

1070
00:37:26,320 --> 00:37:31,280
ehr-linked patient and provider facing

1071
00:37:29,200 --> 00:37:33,680
pain management app

1072
00:37:31,280 --> 00:37:36,640
to facilitate shared decision making

1073
00:37:33,680 --> 00:37:39,599
using a shared model approach

1074
00:37:36,640 --> 00:37:41,440
on the patient side called my pain

1075
00:37:39,599 --> 00:37:43,920
this app intakes patient reported

1076
00:37:41,440 --> 00:37:45,040
outcomes measures provides patients

1077
00:37:43,920 --> 00:37:47,920
information with their current

1078
00:37:45,040 --> 00:37:50,800
medication and alternatives

1079
00:37:47,920 --> 00:37:52,079
seeks input about their pain goals

1080
00:37:50,800 --> 00:37:53,839
as well

1081
00:37:52,079 --> 00:37:55,440
via questionnaire and then also free

1082
00:37:53,839 --> 00:37:57,200
text as well

1083
00:37:55,440 --> 00:37:59,200
on the provider side

1084
00:37:57,200 --> 00:38:01,760
called pain manager providers have

1085
00:37:59,200 --> 00:38:04,400
access to patients ehr linked histories

1086
00:38:01,760 --> 00:38:06,320
and data pdmp data

1087
00:38:04,400 --> 00:38:08,320
and results of

1088
00:38:06,320 --> 00:38:10,240
the patient's my pain questionnaire and

1089
00:38:08,320 --> 00:38:12,079
input to facilitate shared decision

1090
00:38:10,240 --> 00:38:14,000
making from the patient's pain

1091
00:38:12,079 --> 00:38:15,760
management or for the patient's pain

1092
00:38:14,000 --> 00:38:16,839
management

1093
00:38:15,760 --> 00:38:21,920
next

1094
00:38:16,839 --> 00:38:24,720
slide so to give you a quick review of

1095
00:38:21,920 --> 00:38:27,760
what the patient sees

1096
00:38:24,720 --> 00:38:29,920
and the questions asked you'll have

1097
00:38:27,760 --> 00:38:31,520
self-described goals

1098
00:38:29,920 --> 00:38:33,119
for the patient if you can click through

1099
00:38:31,520 --> 00:38:34,640
it will come up with some

1100
00:38:33,119 --> 00:38:36,960
um

1101
00:38:34,640 --> 00:38:38,880
next slides

1102
00:38:36,960 --> 00:38:41,920
and showing you what the patient

1103
00:38:38,880 --> 00:38:43,760
is able to input and review and will go

1104
00:38:41,920 --> 00:38:46,560
directly to

1105
00:38:43,760 --> 00:38:48,160
the physician-facing portion

1106
00:38:46,560 --> 00:38:49,280
my pain

1107
00:38:48,160 --> 00:38:51,440
manager

1108
00:38:49,280 --> 00:38:54,560
next slide

1109
00:38:51,440 --> 00:38:54,560
so on the provider side

1110
00:38:54,640 --> 00:38:59,760
their view is linked with ehr

1111
00:38:57,280 --> 00:39:02,000
data to embed pertinent

1112
00:38:59,760 --> 00:39:03,920
information such as past medical history

1113
00:39:02,000 --> 00:39:05,440
and other comorbidities when using

1114
00:39:03,920 --> 00:39:07,920
opioids

1115
00:39:05,440 --> 00:39:09,760
next slide

1116
00:39:07,920 --> 00:39:13,680
it shows current treatments an mme

1117
00:39:09,760 --> 00:39:15,280
calculator and also incorporates recent

1118
00:39:13,680 --> 00:39:18,400
patient stuff reported treatments as

1119
00:39:15,280 --> 00:39:18,400
well next slide

1120
00:39:18,880 --> 00:39:22,320
it also allows for review of recent ehr

1121
00:39:21,119 --> 00:39:25,359
linked data

1122
00:39:22,320 --> 00:39:28,079
from recent talk screens within the ehr

1123
00:39:25,359 --> 00:39:30,320
and also allows you to review patient

1124
00:39:28,079 --> 00:39:32,000
goals to further facilitate shared

1125
00:39:30,320 --> 00:39:32,960
decision making

1126
00:39:32,000 --> 00:39:35,680
either

1127
00:39:32,960 --> 00:39:38,960
via communication with the patient in my

1128
00:39:35,680 --> 00:39:42,160
pain or at the next clinic visit

1129
00:39:38,960 --> 00:39:44,960
next slide

1130
00:39:42,160 --> 00:39:47,839
so the next part of this work with my

1131
00:39:44,960 --> 00:39:50,000
pain pain managers to scale it and we

1132
00:39:47,839 --> 00:39:52,320
have awarded a grant to dr christopher

1133
00:39:50,000 --> 00:39:54,480
harley from the university of florida uh

1134
00:39:52,320 --> 00:39:56,880
for the opportunity to do so this work

1135
00:39:54,480 --> 00:39:59,280
will create a tailored implementation

1136
00:39:56,880 --> 00:40:00,960
strategy in order to facilitate

1137
00:39:59,280 --> 00:40:03,599
increased clinical decision support

1138
00:40:00,960 --> 00:40:06,160
adoption mainly of my pain and pain

1139
00:40:03,599 --> 00:40:08,160
manager in order to increase shared

1140
00:40:06,160 --> 00:40:09,119
decision making

1141
00:40:08,160 --> 00:40:12,000
and

1142
00:40:09,119 --> 00:40:14,240
hopefully improve outcomes improve pain

1143
00:40:12,000 --> 00:40:16,640
and physical function scores

1144
00:40:14,240 --> 00:40:18,720
uh next slide

1145
00:40:16,640 --> 00:40:21,359
uh additional work that we're funding to

1146
00:40:18,720 --> 00:40:23,359
also help facilitate getting data at the

1147
00:40:21,359 --> 00:40:27,520
point of uh delivery

1148
00:40:23,359 --> 00:40:30,160
is uh efforts uh by dr daniel

1149
00:40:27,520 --> 00:40:33,680
hartung out of ohsu

1150
00:40:30,160 --> 00:40:36,640
to facilitate ehr pdmp integration

1151
00:40:33,680 --> 00:40:38,720
uh this work hopes to determine how pmpa

1152
00:40:36,640 --> 00:40:40,560
integration facilitates provider use of

1153
00:40:38,720 --> 00:40:43,599
the pdmp program

1154
00:40:40,560 --> 00:40:45,760
and evaluates

1155
00:40:43,599 --> 00:40:48,319
these efforts on controlled setups and

1156
00:40:45,760 --> 00:40:50,800
prescribing

1157
00:40:48,319 --> 00:40:50,800
next slide

1158
00:40:50,880 --> 00:40:55,760
so although i have highlighted some

1159
00:40:53,599 --> 00:40:57,760
successes of our efforts uh there are

1160
00:40:55,760 --> 00:40:58,960
still many challenges and potential

1161
00:40:57,760 --> 00:40:59,839
opportunities

1162
00:40:58,960 --> 00:41:02,720
um

1163
00:40:59,839 --> 00:41:05,839
for not only arcs future focus but uh

1164
00:41:02,720 --> 00:41:08,000
nih as well and we hope that you will

1165
00:41:05,839 --> 00:41:10,240
think about these and moving forward uh

1166
00:41:08,000 --> 00:41:12,079
throughout the the remaining time in

1167
00:41:10,240 --> 00:41:14,640
these uh in these presentations and

1168
00:41:12,079 --> 00:41:16,079
workshops on the technical side just a

1169
00:41:14,640 --> 00:41:18,160
brief overview there are multiple

1170
00:41:16,079 --> 00:41:20,880
challenges remaining in regard to

1171
00:41:18,160 --> 00:41:23,359
ehr data interoperability

1172
00:41:20,880 --> 00:41:26,079
the use of ehr

1173
00:41:23,359 --> 00:41:28,000
data standards by vendors

1174
00:41:26,079 --> 00:41:30,960
a lack of a seamless integrated health

1175
00:41:28,000 --> 00:41:34,400
it infrastructure which ultimately

1176
00:41:30,960 --> 00:41:36,240
limits the use and scalability of cds

1177
00:41:34,400 --> 00:41:38,319
we are also still trying to show the

1178
00:41:36,240 --> 00:41:40,160
return of investment of clinical

1179
00:41:38,319 --> 00:41:43,280
decision support and there's ongoing

1180
00:41:40,160 --> 00:41:46,000
efforts to define clinical trials

1181
00:41:43,280 --> 00:41:47,200
showing how cds can improve the quality

1182
00:41:46,000 --> 00:41:49,200
of care

1183
00:41:47,200 --> 00:41:51,440
additionally how do we effectively

1184
00:41:49,200 --> 00:41:54,480
engage our patients their families and

1185
00:41:51,440 --> 00:41:56,160
caregivers in these technologies all

1186
00:41:54,480 --> 00:41:57,599
while trying not to create further

1187
00:41:56,160 --> 00:42:00,000
divides and

1188
00:41:57,599 --> 00:42:02,319
inequalities

1189
00:42:00,000 --> 00:42:04,240
also and especially for this topic

1190
00:42:02,319 --> 00:42:06,880
focused creating and disseminating the

1191
00:42:04,240 --> 00:42:09,520
appropriate evidence-based content for

1192
00:42:06,880 --> 00:42:12,480
the prevention of substance use i hope

1193
00:42:09,520 --> 00:42:15,359
in sharing some of these

1194
00:42:12,480 --> 00:42:17,040
arcs efforts and in ces can inspire and

1195
00:42:15,359 --> 00:42:19,200
provide some new ideas for those in the

1196
00:42:17,040 --> 00:42:20,720
audience i encourage everyone to reach

1197
00:42:19,200 --> 00:42:23,200
out and connect with us at our because

1198
00:42:20,720 --> 00:42:24,960
we need diverse feedback and ideas to

1199
00:42:23,200 --> 00:42:26,319
combat some of these challenges i've

1200
00:42:24,960 --> 00:42:27,760
shared

1201
00:42:26,319 --> 00:42:29,359
lastly

1202
00:42:27,760 --> 00:42:31,760
for those who are interested in

1203
00:42:29,359 --> 00:42:33,200
advancing cds efforts or creating it arc

1204
00:42:31,760 --> 00:42:35,599
has an interest in receiving health

1205
00:42:33,200 --> 00:42:37,200
services research grants

1206
00:42:35,599 --> 00:42:38,640
that propose innovative and

1207
00:42:37,200 --> 00:42:40,400
evidence-based interventions that

1208
00:42:38,640 --> 00:42:42,640
advance the nation's goal of achieving

1209
00:42:40,400 --> 00:42:44,480
equity in delivery of health care

1210
00:42:42,640 --> 00:42:47,040
services including reducing health

1211
00:42:44,480 --> 00:42:49,200
disparities in quality of care patient

1212
00:42:47,040 --> 00:42:52,240
safety health care utilization and

1213
00:42:49,200 --> 00:42:55,440
access and ultimately health outcomes uh

1214
00:42:52,240 --> 00:42:55,440
here are some of

1215
00:42:55,680 --> 00:42:59,760
information that can link us to some of

1216
00:42:57,920 --> 00:43:02,480
our efforts and also my contact

1217
00:42:59,760 --> 00:43:04,079
information so i look forward to the

1218
00:43:02,480 --> 00:43:06,079
discussion and any questions you might

1219
00:43:04,079 --> 00:43:08,960
have and hopefully you'll reach out and

1220
00:43:06,079 --> 00:43:08,960
partner with us

1221
00:43:10,400 --> 00:43:15,119
thanks so much so great as sarah

1222
00:43:13,040 --> 00:43:17,040
introduced me i'm brian jensen i'm a

1223
00:43:15,119 --> 00:43:18,720
practicing pediatrician i'm also a

1224
00:43:17,040 --> 00:43:20,240
researcher and a couple different hats i

1225
00:43:18,720 --> 00:43:21,440
have at the children's hospital of

1226
00:43:20,240 --> 00:43:22,560
philadelphia

1227
00:43:21,440 --> 00:43:24,160
uh we're gonna do something a little

1228
00:43:22,560 --> 00:43:25,520
different from the previous two speakers

1229
00:43:24,160 --> 00:43:26,560
we're gonna focus on learning through

1230
00:43:25,520 --> 00:43:28,319
failure

1231
00:43:26,560 --> 00:43:30,240
and really how to best

1232
00:43:28,319 --> 00:43:32,079
not leverage the ehr and leverage the

1233
00:43:30,240 --> 00:43:34,480
ehr for adolescent substance use

1234
00:43:32,079 --> 00:43:37,200
prevention and treatment

1235
00:43:34,480 --> 00:43:38,960
uh first i'm mainly my main focus has

1236
00:43:37,200 --> 00:43:40,560
started in the tobacco

1237
00:43:38,960 --> 00:43:42,079
control space and has moved into

1238
00:43:40,560 --> 00:43:43,680
substance use disorders other substance

1239
00:43:42,079 --> 00:43:45,520
disorders so first i have no relevant

1240
00:43:43,680 --> 00:43:47,040
conflict of interest to disclose of any

1241
00:43:45,520 --> 00:43:48,079
ownership in tobacco or e-cigarette

1242
00:43:47,040 --> 00:43:51,359
companies i also don't have any

1243
00:43:48,079 --> 00:43:54,000
ownership and any ehr companies or

1244
00:43:51,359 --> 00:43:55,440
similar health i.t companies

1245
00:43:54,000 --> 00:43:57,680
mainly what i'm going to talk about

1246
00:43:55,440 --> 00:43:59,760
briefly in about 10 15 minutes is

1247
00:43:57,680 --> 00:44:01,200
describe the lessons learned regarding

1248
00:43:59,760 --> 00:44:02,960
adolescent e-cigarette screening

1249
00:44:01,200 --> 00:44:06,480
connecting a treatment through a large

1250
00:44:02,960 --> 00:44:06,480
pediatric primary care health system

1251
00:44:06,560 --> 00:44:10,000
so

1252
00:44:07,359 --> 00:44:12,400
where this comes from first so there's

1253
00:44:10,000 --> 00:44:13,520
mario is giving a great talk about ahrq

1254
00:44:12,400 --> 00:44:15,520
and their different priorities and

1255
00:44:13,520 --> 00:44:18,000
opportunities just a little reminder

1256
00:44:15,520 --> 00:44:20,319
about how kind of the history of ehrs

1257
00:44:18,000 --> 00:44:21,280
and then how pediatric ehrs are

1258
00:44:20,319 --> 00:44:23,920
different

1259
00:44:21,280 --> 00:44:25,200
so ehrs were really built for adult

1260
00:44:23,920 --> 00:44:27,200
patients

1261
00:44:25,200 --> 00:44:28,960
uh pediatrics for anyone in that space

1262
00:44:27,200 --> 00:44:30,560
clinically or doing research knows or if

1263
00:44:28,960 --> 00:44:32,319
they have children or if you're a child

1264
00:44:30,560 --> 00:44:33,599
yourself at some point pediatrics

1265
00:44:32,319 --> 00:44:34,960
involves unique shifting development on

1266
00:44:33,599 --> 00:44:37,040
the aids of the children parents what we

1267
00:44:34,960 --> 00:44:38,720
say is there's an in for most in the

1268
00:44:37,040 --> 00:44:40,800
adult space there's a dyadic

1269
00:44:38,720 --> 00:44:42,880
relationship between the physician

1270
00:44:40,800 --> 00:44:44,880
provider and their patient in our space

1271
00:44:42,880 --> 00:44:46,319
in pediatrics it's triatic and many

1272
00:44:44,880 --> 00:44:48,240
times when i'm showing you just on this

1273
00:44:46,319 --> 00:44:49,680
classic picture yes i might take care i

1274
00:44:48,240 --> 00:44:51,599
take care of the child but in many ways

1275
00:44:49,680 --> 00:44:53,520
i'm also taking care of the mother so

1276
00:44:51,599 --> 00:44:55,359
there's a triatic relationship

1277
00:44:53,520 --> 00:44:58,880
there have been key priorities addressed

1278
00:44:55,359 --> 00:45:00,880
for improvement to help better adapt the

1279
00:44:58,880 --> 00:45:03,200
again adult focus dhr for the pediatric

1280
00:45:00,880 --> 00:45:04,400
space there are consensus guidelines i

1281
00:45:03,200 --> 00:45:05,680
have a reference to one of our papers

1282
00:45:04,400 --> 00:45:07,200
that recently highlighted this from the

1283
00:45:05,680 --> 00:45:08,640
american account of pediatrics and the

1284
00:45:07,200 --> 00:45:09,760
office of national coordinator for

1285
00:45:08,640 --> 00:45:11,599
health i.t

1286
00:45:09,760 --> 00:45:14,160
but the two main things for today for

1287
00:45:11,599 --> 00:45:16,560
this audience to know how ehrs and

1288
00:45:14,160 --> 00:45:18,640
pediatric space are different than the

1289
00:45:16,560 --> 00:45:20,720
adult is there needs to be better

1290
00:45:18,640 --> 00:45:23,200
systems to document all guardians and

1291
00:45:20,720 --> 00:45:24,640
connect families ehr is built for one

1292
00:45:23,200 --> 00:45:26,319
patient and the provider but in

1293
00:45:24,640 --> 00:45:28,079
pediatrics you might have multiple

1294
00:45:26,319 --> 00:45:30,000
different parents guardians caregivers

1295
00:45:28,079 --> 00:45:31,920
who need to access the ehr

1296
00:45:30,000 --> 00:45:34,480
and then you need to come up with ways

1297
00:45:31,920 --> 00:45:36,480
to connect the families together the

1298
00:45:34,480 --> 00:45:37,599
ehrs is currently built can't tell

1299
00:45:36,480 --> 00:45:39,599
siblings

1300
00:45:37,599 --> 00:45:41,520
can't connect siblings to one another

1301
00:45:39,599 --> 00:45:43,040
the other big issue is we need better

1302
00:45:41,520 --> 00:45:45,040
segmented access to information

1303
00:45:43,040 --> 00:45:46,240
communication channels because of that

1304
00:45:45,040 --> 00:45:49,520
shifting relationship i mentioned

1305
00:45:46,240 --> 00:45:50,720
between parent child and pediatrician

1306
00:45:49,520 --> 00:45:52,640
what happens over time is there's

1307
00:45:50,720 --> 00:45:54,480
increase in adolescent autonomy and we

1308
00:45:52,640 --> 00:45:56,160
need channels that can both communicate

1309
00:45:54,480 --> 00:45:57,760
back and forth with the teenager but

1310
00:45:56,160 --> 00:45:59,280
also separately with the parent to

1311
00:45:57,760 --> 00:46:01,440
support their relationship but also

1312
00:45:59,280 --> 00:46:02,720
protect adolescent confidentiality

1313
00:46:01,440 --> 00:46:04,079
that in and of itself is a whole talk

1314
00:46:02,720 --> 00:46:06,720
but that's just high level how the

1315
00:46:04,079 --> 00:46:08,079
unique needs of pediatric ehr

1316
00:46:06,720 --> 00:46:09,760
so what i'm going to talk more about is

1317
00:46:08,079 --> 00:46:11,920
an effort to focus on prevention and

1318
00:46:09,760 --> 00:46:13,599
treatment of tobacco use in particular

1319
00:46:11,920 --> 00:46:14,960
e-cigarette use

1320
00:46:13,599 --> 00:46:16,240
so what lessons could be learned from

1321
00:46:14,960 --> 00:46:17,760
cigarettes well

1322
00:46:16,240 --> 00:46:20,240
teen cigarette smokers want to quit

1323
00:46:17,760 --> 00:46:22,000
that's pretty clear few teens seek help

1324
00:46:20,240 --> 00:46:24,400
there are limited to essentially no

1325
00:46:22,000 --> 00:46:26,880
evidence-based treatment options

1326
00:46:24,400 --> 00:46:28,640
for adolescent cigarette use and even

1327
00:46:26,880 --> 00:46:30,800
fewer options for adolescent e-cigarette

1328
00:46:28,640 --> 00:46:32,640
use but pediatricians are uniquely

1329
00:46:30,800 --> 00:46:35,119
positioned to intervene because the vast

1330
00:46:32,640 --> 00:46:36,560
majority of children in the united

1331
00:46:35,119 --> 00:46:38,800
states the vast majority of teenagers

1332
00:46:36,560 --> 00:46:40,079
come to a pediatric provider

1333
00:46:38,800 --> 00:46:42,000
there's limited data on how to treat

1334
00:46:40,079 --> 00:46:43,920
teens for e-cigarette addiction though a

1335
00:46:42,000 --> 00:46:46,000
large percentage of teens based on most

1336
00:46:43,920 --> 00:46:47,440
recent surveys want to quit and there

1337
00:46:46,000 --> 00:46:48,720
have been some new targeted treatment

1338
00:46:47,440 --> 00:46:50,720
options and i'll highlight one that we

1339
00:46:48,720 --> 00:46:52,720
connected with

1340
00:46:50,720 --> 00:46:55,119
so we created a study that we were

1341
00:46:52,720 --> 00:46:56,640
calling our chopteen be smoke-free

1342
00:46:55,119 --> 00:46:58,640
cessation study

1343
00:46:56,640 --> 00:47:00,640
and we brought together a whole bunch of

1344
00:46:58,640 --> 00:47:02,079
innovative approaches that combined

1345
00:47:00,640 --> 00:47:03,760
interesting technology and electronic

1346
00:47:02,079 --> 00:47:05,520
health records we had a three-arm

1347
00:47:03,760 --> 00:47:07,440
pragmatic randomized trial to help teens

1348
00:47:05,520 --> 00:47:09,119
engage in treatment and quit we

1349
00:47:07,440 --> 00:47:10,319
incorporate behavioral economic concepts

1350
00:47:09,119 --> 00:47:12,240
that's beyond the scope of this talk

1351
00:47:10,319 --> 00:47:14,480
what we're focused on between us a

1352
00:47:12,240 --> 00:47:16,560
seemingly small recruitment goal with 60

1353
00:47:14,480 --> 00:47:18,400
participants i'll say seemingly small

1354
00:47:16,560 --> 00:47:20,720
and i'll show you why uh over the next

1355
00:47:18,400 --> 00:47:22,640
couple slides and we leverage three

1356
00:47:20,720 --> 00:47:24,640
critical components we have a large

1357
00:47:22,640 --> 00:47:27,920
pediatric health system

1358
00:47:24,640 --> 00:47:30,480
we were going to enroll and identify

1359
00:47:27,920 --> 00:47:32,720
children teenagers via electronic health

1360
00:47:30,480 --> 00:47:34,720
record-based prompts and then engage

1361
00:47:32,720 --> 00:47:36,640
them in treatment through a really novel

1362
00:47:34,720 --> 00:47:38,480
text messaging program and then we're

1363
00:47:36,640 --> 00:47:40,079
going to connect them to a novel

1364
00:47:38,480 --> 00:47:41,680
treatment program that was organized

1365
00:47:40,079 --> 00:47:43,359
through our state

1366
00:47:41,680 --> 00:47:44,960
quit line our state tobacco treatment

1367
00:47:43,359 --> 00:47:46,640
program

1368
00:47:44,960 --> 00:47:48,240
i mentioned we have a large primary care

1369
00:47:46,640 --> 00:47:49,280
system so

1370
00:47:48,240 --> 00:47:51,839
as sarah mentioned on the medical

1371
00:47:49,280 --> 00:47:53,359
director in our large primary care

1372
00:47:51,839 --> 00:47:55,520
health system that's based in 31

1373
00:47:53,359 --> 00:47:57,359
practices across pennsylvania new jersey

1374
00:47:55,520 --> 00:47:59,839
we have approximately 300 000 patients

1375
00:47:57,359 --> 00:48:01,920
about 80 000 teenagers and we have an

1376
00:47:59,839 --> 00:48:04,240
integrated electronic health record we

1377
00:48:01,920 --> 00:48:06,079
use epic and that system has been in use

1378
00:48:04,240 --> 00:48:07,920
for more than 20 years so we've been at

1379
00:48:06,079 --> 00:48:10,480
kind of the tip of the sphere in terms

1380
00:48:07,920 --> 00:48:12,400
of innovation efforts within the ehr our

1381
00:48:10,480 --> 00:48:15,119
demographics are for the most part

1382
00:48:12,400 --> 00:48:17,280
representative of the larger national

1383
00:48:15,119 --> 00:48:19,119
population just highlighted here on this

1384
00:48:17,280 --> 00:48:20,960
slide

1385
00:48:19,119 --> 00:48:23,440
and we leveraged relevant to this talk

1386
00:48:20,960 --> 00:48:25,359
some really advanced ehr techniques uh

1387
00:48:23,440 --> 00:48:27,040
we had screen that was based embedded

1388
00:48:25,359 --> 00:48:28,720
within clinical practice

1389
00:48:27,040 --> 00:48:30,480
what this what we did was a provider

1390
00:48:28,720 --> 00:48:33,440
prompt within the confidential

1391
00:48:30,480 --> 00:48:35,839
adolescent visit documentation to then

1392
00:48:33,440 --> 00:48:37,200
to ask about tobacco use and e-cigarette

1393
00:48:35,839 --> 00:48:38,800
use what i'm showing you on the

1394
00:48:37,200 --> 00:48:41,040
right-hand side of this screen is the

1395
00:48:38,800 --> 00:48:42,240
documentation that are prompts for the

1396
00:48:41,040 --> 00:48:44,240
physician pediatrician nurse

1397
00:48:42,240 --> 00:48:45,599
practitioner in clinical practice to go

1398
00:48:44,240 --> 00:48:47,839
all through these things during the

1399
00:48:45,599 --> 00:48:49,520
confidential visit portion of the visit

1400
00:48:47,839 --> 00:48:51,280
if the team were to be positive for

1401
00:48:49,520 --> 00:48:53,040
e-cigarette use any tobacco use

1402
00:48:51,280 --> 00:48:54,800
cigarette use there were additional

1403
00:48:53,040 --> 00:48:57,599
problems

1404
00:48:54,800 --> 00:48:59,680
the there was scripts on assessing

1405
00:48:57,599 --> 00:49:01,680
interest in quitting or engaging in

1406
00:48:59,680 --> 00:49:03,040
treatment from the pediatrician and then

1407
00:49:01,680 --> 00:49:04,480
permission from the study for the study

1408
00:49:03,040 --> 00:49:06,319
team to follow up the team via text

1409
00:49:04,480 --> 00:49:07,839
messaging this has been an approach that

1410
00:49:06,319 --> 00:49:10,319
we've been leveraging for more than 10

1411
00:49:07,839 --> 00:49:12,640
years it works really well to identify

1412
00:49:10,319 --> 00:49:14,559
individuals who could uh participate in

1413
00:49:12,640 --> 00:49:16,720
the study and then connect them to the

1414
00:49:14,559 --> 00:49:18,079
study team so this has been a well-worn

1415
00:49:16,720 --> 00:49:20,079
approach

1416
00:49:18,079 --> 00:49:21,680
we also leveraged

1417
00:49:20,079 --> 00:49:23,359
text messaging which is really the best

1418
00:49:21,680 --> 00:49:24,880
way to engage teams

1419
00:49:23,359 --> 00:49:26,720
we have a program based through the

1420
00:49:24,880 --> 00:49:29,200
university of pennsylvania that's a

1421
00:49:26,720 --> 00:49:30,559
research platform that

1422
00:49:29,200 --> 00:49:32,720
combines a whole bunch of different

1423
00:49:30,559 --> 00:49:34,640
techniques into a text messaging

1424
00:49:32,720 --> 00:49:37,520
platform to immediately engage

1425
00:49:34,640 --> 00:49:39,040
individuals in a study or in treatment

1426
00:49:37,520 --> 00:49:40,720
so a novel thing that we were connect

1427
00:49:39,040 --> 00:49:41,760
that we were able to immediately engage

1428
00:49:40,720 --> 00:49:43,599
teens in

1429
00:49:41,760 --> 00:49:45,280
and then we were collaborating with our

1430
00:49:43,599 --> 00:49:46,960
partner national jewish health they're

1431
00:49:45,280 --> 00:49:49,760
the servicer for the pennsylvania quit

1432
00:49:46,960 --> 00:49:52,559
line they had at the time a novel

1433
00:49:49,760 --> 00:49:54,319
program this was back in 2018 2019 where

1434
00:49:52,559 --> 00:49:56,240
they created both counseling over the

1435
00:49:54,319 --> 00:49:58,640
phone but a program that really

1436
00:49:56,240 --> 00:50:00,480
a text messaging based program and a

1437
00:49:58,640 --> 00:50:02,480
portal-based program to really help

1438
00:50:00,480 --> 00:50:05,599
teens who were vaping using other

1439
00:50:02,480 --> 00:50:07,119
tobacco products engaging quitting

1440
00:50:05,599 --> 00:50:09,440
and what happened here here is the

1441
00:50:07,119 --> 00:50:11,599
emphasis on learning through failure

1442
00:50:09,440 --> 00:50:14,240
unsurprisingly to anyone

1443
00:50:11,599 --> 00:50:16,720
on this meeting right now helping engage

1444
00:50:14,240 --> 00:50:18,800
teens in treatment is incredibly hard

1445
00:50:16,720 --> 00:50:20,960
but we were surprised about how much of

1446
00:50:18,800 --> 00:50:23,200
a flop this was it was a failed

1447
00:50:20,960 --> 00:50:25,760
feasibility study we screened more than

1448
00:50:23,200 --> 00:50:27,520
80 000 teams in the span of one year for

1449
00:50:25,760 --> 00:50:29,040
vaping that's a success

1450
00:50:27,520 --> 00:50:30,240
but only one percent of teens reported

1451
00:50:29,040 --> 00:50:32,559
current use

1452
00:50:30,240 --> 00:50:34,640
so either our teens are

1453
00:50:32,559 --> 00:50:36,000
perfect or we were missing something

1454
00:50:34,640 --> 00:50:38,079
because at the time

1455
00:50:36,000 --> 00:50:40,960
the school-based normative surveys were

1456
00:50:38,079 --> 00:50:42,480
suggesting 20 to 25 percent of teens

1457
00:50:40,960 --> 00:50:44,319
were current users

1458
00:50:42,480 --> 00:50:47,280
so we were missing something

1459
00:50:44,319 --> 00:50:50,000
uh of the roughly 5 000 teens that did

1460
00:50:47,280 --> 00:50:51,920
ultimately report use only 52 were

1461
00:50:50,000 --> 00:50:53,359
referred to our study another big flop

1462
00:50:51,920 --> 00:50:55,520
because we had a much better track

1463
00:50:53,359 --> 00:50:57,599
record of using engaging teens

1464
00:50:55,520 --> 00:50:59,200
in other research studies and we closed

1465
00:50:57,599 --> 00:51:01,280
the study early because we were only

1466
00:50:59,200 --> 00:51:02,480
able to enroll 16 individuals across two

1467
00:51:01,280 --> 00:51:04,800
years

1468
00:51:02,480 --> 00:51:06,880
so we published this luckily uh nice to

1469
00:51:04,800 --> 00:51:08,960
the journal to acknowledge that this was

1470
00:51:06,880 --> 00:51:11,520
a failed study but what we really

1471
00:51:08,960 --> 00:51:13,440
identified was the most immediate issue

1472
00:51:11,520 --> 00:51:15,920
before we even get to treatment

1473
00:51:13,440 --> 00:51:18,000
was improved systems to identify teens

1474
00:51:15,920 --> 00:51:20,079
at risk in other words improved clinical

1475
00:51:18,000 --> 00:51:22,400
based screening which has been a theme

1476
00:51:20,079 --> 00:51:24,400
mentioned at the start of this whole

1477
00:51:22,400 --> 00:51:26,079
session today

1478
00:51:24,400 --> 00:51:27,760
so what were our key lessons learned if

1479
00:51:26,079 --> 00:51:29,599
there's one slide

1480
00:51:27,760 --> 00:51:30,640
to remember and a couple key points it's

1481
00:51:29,599 --> 00:51:32,800
this one

1482
00:51:30,640 --> 00:51:34,720
first if you're going to do anything to

1483
00:51:32,800 --> 00:51:36,000
identify health risk you need to create

1484
00:51:34,720 --> 00:51:37,920
the right environment to maximize

1485
00:51:36,000 --> 00:51:40,319
disclosure i'm also an innovation

1486
00:51:37,920 --> 00:51:42,160
researcher so what we did was um we had

1487
00:51:40,319 --> 00:51:43,680
another study live in various practices

1488
00:51:42,160 --> 00:51:45,680
and i was walking around engaging with

1489
00:51:43,680 --> 00:51:47,200
our pediatricians

1490
00:51:45,680 --> 00:51:48,559
ostensibly for this other study but i

1491
00:51:47,200 --> 00:51:50,720
was really there to also observe what

1492
00:51:48,559 --> 00:51:52,960
was happening with screening and we

1493
00:51:50,720 --> 00:51:54,160
found that um

1494
00:51:52,960 --> 00:51:55,680
even though we had trained the

1495
00:51:54,160 --> 00:51:56,880
pediatricians and the nurse

1496
00:51:55,680 --> 00:51:58,960
practitioners and the nurses and the

1497
00:51:56,880 --> 00:52:00,720
best ways to ask certain questions we

1498
00:51:58,960 --> 00:52:02,480
often heard a nurse and i was supposed

1499
00:52:00,720 --> 00:52:04,800
to be the pediatrician asking the

1500
00:52:02,480 --> 00:52:05,920
question of the team next to the parent

1501
00:52:04,800 --> 00:52:07,200
and the parent was supposed to be

1502
00:52:05,920 --> 00:52:09,040
outside of the room

1503
00:52:07,200 --> 00:52:10,559
uh you don't you don't they do you don't

1504
00:52:09,040 --> 00:52:12,960
use any tobaccos and that's not the

1505
00:52:10,559 --> 00:52:14,800
evidence-based way of screening so you

1506
00:52:12,960 --> 00:52:15,760
really want to make sure that in this

1507
00:52:14,800 --> 00:52:17,839
environment you're not doing

1508
00:52:15,760 --> 00:52:19,520
face-to-face based screening and not

1509
00:52:17,839 --> 00:52:21,119
screening teens for health risks with

1510
00:52:19,520 --> 00:52:22,880
the parent right next to them that

1511
00:52:21,119 --> 00:52:24,240
really hasn't created the hasn't

1512
00:52:22,880 --> 00:52:25,359
maximized the right environment for

1513
00:52:24,240 --> 00:52:27,680
disclosure

1514
00:52:25,359 --> 00:52:29,760
now the other big thing related to ehr's

1515
00:52:27,680 --> 00:52:32,960
work is you really have to understand

1516
00:52:29,760 --> 00:52:35,839
the clinical workflow problem before you

1517
00:52:32,960 --> 00:52:37,599
build the correct ehr based solution

1518
00:52:35,839 --> 00:52:40,240
so for those who might be thinking oh if

1519
00:52:37,599 --> 00:52:42,480
only we could use ehr documentation to

1520
00:52:40,240 --> 00:52:44,160
identify something uh to then build a

1521
00:52:42,480 --> 00:52:45,760
better research platform or build a

1522
00:52:44,160 --> 00:52:47,680
better intervention

1523
00:52:45,760 --> 00:52:48,800
ehr documentation

1524
00:52:47,680 --> 00:52:50,559
is flawed

1525
00:52:48,800 --> 00:52:52,000
we often talk about

1526
00:52:50,559 --> 00:52:55,119
when you hear kind of terms like big

1527
00:52:52,000 --> 00:52:57,200
data big data is really found data in

1528
00:52:55,119 --> 00:52:59,040
many ways it's if you have bad data in

1529
00:52:57,200 --> 00:53:01,359
you get bad data out so what we learned

1530
00:52:59,040 --> 00:53:03,359
was yes it looked like the teens were

1531
00:53:01,359 --> 00:53:04,720
getting screened for e-cigarette use but

1532
00:53:03,359 --> 00:53:06,880
they weren't really weren't being

1533
00:53:04,720 --> 00:53:08,800
screened in the best possible way

1534
00:53:06,880 --> 00:53:10,960
so for the next couple slides

1535
00:53:08,800 --> 00:53:12,720
i'll highlight how we can do better and

1536
00:53:10,960 --> 00:53:14,240
we started to do better which is this is

1537
00:53:12,720 --> 00:53:15,599
really the direction one needs to go if

1538
00:53:14,240 --> 00:53:16,960
you're going to focus on pediatric

1539
00:53:15,599 --> 00:53:20,160
focused

1540
00:53:16,960 --> 00:53:22,319
substance use prevention and treatment

1541
00:53:20,160 --> 00:53:25,200
so pretty much all humans but especially

1542
00:53:22,319 --> 00:53:27,359
teenagers prefer electronic screening

1543
00:53:25,200 --> 00:53:29,040
directly of them as the best way to

1544
00:53:27,359 --> 00:53:30,400
identify health risk it can be

1545
00:53:29,040 --> 00:53:32,160
administered on a tablet or other mobile

1546
00:53:30,400 --> 00:53:33,599
devices and that's been identified

1547
00:53:32,160 --> 00:53:34,960
through multiple studies small studies

1548
00:53:33,599 --> 00:53:36,559
compared to either paper screen or

1549
00:53:34,960 --> 00:53:38,800
general questions that are asked by the

1550
00:53:36,559 --> 00:53:40,640
clinician face to face during the visit

1551
00:53:38,800 --> 00:53:42,880
and this approach can increase validity

1552
00:53:40,640 --> 00:53:44,400
increases the true positives

1553
00:53:42,880 --> 00:53:45,680
one can increase disclosure of health

1554
00:53:44,400 --> 00:53:47,040
risks

1555
00:53:45,680 --> 00:53:48,400
and then especially if you combine that

1556
00:53:47,040 --> 00:53:50,640
with private confidential time with the

1557
00:53:48,400 --> 00:53:51,920
pediatricians and here's the value add

1558
00:53:50,640 --> 00:53:55,680
you don't necessarily have to just focus

1559
00:53:51,920 --> 00:53:57,280
on reimbursement but giving time back to

1560
00:53:55,680 --> 00:53:59,200
physicians is such an important area

1561
00:53:57,280 --> 00:54:00,800
where you can get hidden value

1562
00:53:59,200 --> 00:54:02,240
this approach actually takes less time

1563
00:54:00,800 --> 00:54:03,599
than the current face-to-face approach

1564
00:54:02,240 --> 00:54:05,520
because most satellite social history is

1565
00:54:03,599 --> 00:54:08,240
screening this negative so you can speed

1566
00:54:05,520 --> 00:54:10,079
up the visit and focus on other issues

1567
00:54:08,240 --> 00:54:11,680
saves about 5-10 minutes of face to face

1568
00:54:10,079 --> 00:54:13,359
time and then you can more easily

1569
00:54:11,680 --> 00:54:15,599
identify the positives for in-depth

1570
00:54:13,359 --> 00:54:17,440
discussion this last section is how we

1571
00:54:15,599 --> 00:54:19,920
really pitched to our pediatricians

1572
00:54:17,440 --> 00:54:22,319
across our 31 practices why we needed to

1573
00:54:19,920 --> 00:54:23,599
move to an electronic screen approach

1574
00:54:22,319 --> 00:54:24,640
because the other way i didn't want to

1575
00:54:23,599 --> 00:54:26,000
say direct with them but they weren't

1576
00:54:24,640 --> 00:54:27,200
doing the right things to maximize

1577
00:54:26,000 --> 00:54:28,800
disclosure

1578
00:54:27,200 --> 00:54:30,480
so what we moved to

1579
00:54:28,800 --> 00:54:31,520
was a pre-visit adolescent health

1580
00:54:30,480 --> 00:54:33,680
questionnaire i'm showing on the

1581
00:54:31,520 --> 00:54:35,920
left-hand side we have eight domains

1582
00:54:33,680 --> 00:54:37,760
it's a 30-question questionnaire that's

1583
00:54:35,920 --> 00:54:39,839
administered in a confidential manner on

1584
00:54:37,760 --> 00:54:41,760
a tablet when the teenager walks into

1585
00:54:39,839 --> 00:54:43,440
the office what often happens is the

1586
00:54:41,760 --> 00:54:45,599
teenager picks up the tablet and walks

1587
00:54:43,440 --> 00:54:47,119
around the waiting area or gets shuttled

1588
00:54:45,599 --> 00:54:48,640
back shuffled back to the room and

1589
00:54:47,119 --> 00:54:50,799
completes the tablet while they're

1590
00:54:48,640 --> 00:54:52,400
waiting for the pediatrician i mentioned

1591
00:54:50,799 --> 00:54:54,480
30 questions i'm just highlighting this

1592
00:54:52,400 --> 00:54:56,720
is actually showing this audience

1593
00:54:54,480 --> 00:54:58,799
focused on substance use disorders

1594
00:54:56,720 --> 00:55:01,920
uh the craft questionnaire incorporating

1595
00:54:58,799 --> 00:55:03,760
this directly in into this questionnaire

1596
00:55:01,920 --> 00:55:05,760
and what we focused on was building this

1597
00:55:03,760 --> 00:55:07,040
into the correct workflow so i'm showing

1598
00:55:05,760 --> 00:55:08,720
you here is in our development

1599
00:55:07,040 --> 00:55:10,480
environment again with epic our main

1600
00:55:08,720 --> 00:55:11,760
electronic health record really

1601
00:55:10,480 --> 00:55:13,599
integrating this into the physician

1602
00:55:11,760 --> 00:55:15,520
workflow you don't have to just focus on

1603
00:55:13,599 --> 00:55:17,520
reimbursement again making lives easier

1604
00:55:15,520 --> 00:55:19,599
for physicians and nurses is a great

1605
00:55:17,520 --> 00:55:21,440
value add and so this is showing you the

1606
00:55:19,599 --> 00:55:22,960
questionnaire on the left-hand side and

1607
00:55:21,440 --> 00:55:24,960
how we automatically pull the

1608
00:55:22,960 --> 00:55:26,960
documentation into the notes for the

1609
00:55:24,960 --> 00:55:28,640
pediatricians so if there's no issue

1610
00:55:26,960 --> 00:55:30,319
they can automatically say oh no issues

1611
00:55:28,640 --> 00:55:32,319
discuss priorities and that there were

1612
00:55:30,319 --> 00:55:34,400
issues we supported the documentation

1613
00:55:32,319 --> 00:55:36,079
for that

1614
00:55:34,400 --> 00:55:37,839
it's been a huge operational success

1615
00:55:36,079 --> 00:55:41,440
this questionnaire is now live in all 31

1616
00:55:37,839 --> 00:55:43,680
of our practices as of january 2022

1617
00:55:41,440 --> 00:55:45,599
i mentioned we have 80 000 teenagers the

1618
00:55:43,680 --> 00:55:48,640
questionnaire has been completed 31 000

1619
00:55:45,599 --> 00:55:49,920
times across our 31 primary care sites

1620
00:55:48,640 --> 00:55:51,760
lower numbers right now because we just

1621
00:55:49,920 --> 00:55:53,760
went live in our practices we have a

1622
00:55:51,760 --> 00:55:55,920
high completion rate for the individual

1623
00:55:53,760 --> 00:55:57,440
visit about 92 percent of the time the

1624
00:55:55,920 --> 00:55:58,559
questionnaire is completed

1625
00:55:57,440 --> 00:55:59,760
and there's no real meaningful

1626
00:55:58,559 --> 00:56:01,760
difference in completion by practice

1627
00:55:59,760 --> 00:56:03,520
setting vast majority of teens prefer to

1628
00:56:01,760 --> 00:56:04,799
answer these questionnaires questions

1629
00:56:03,520 --> 00:56:06,160
electronically we embedded a

1630
00:56:04,799 --> 00:56:08,160
questionnaire into the questionnaire to

1631
00:56:06,160 --> 00:56:10,240
confirm that and on average just takes

1632
00:56:08,160 --> 00:56:12,480
five to ten minutes now here's the real

1633
00:56:10,240 --> 00:56:14,000
success again learning through failure

1634
00:56:12,480 --> 00:56:16,319
prior to moving live which i'm going to

1635
00:56:14,000 --> 00:56:18,400
show you in the next slide to this

1636
00:56:16,319 --> 00:56:19,680
electronic approach when you looked at

1637
00:56:18,400 --> 00:56:21,839
health risks that were supposed to be

1638
00:56:19,680 --> 00:56:23,680
captured across the board we found about

1639
00:56:21,839 --> 00:56:26,079
one to two percent of teens reported

1640
00:56:23,680 --> 00:56:28,640
that they were using tobacco using

1641
00:56:26,079 --> 00:56:30,079
marijuana using alcohol and even only

1642
00:56:28,640 --> 00:56:32,160
three to four percent were reporting

1643
00:56:30,079 --> 00:56:33,839
sexual activity which also is not

1644
00:56:32,160 --> 00:56:36,000
aligned with the school-based norm of

1645
00:56:33,839 --> 00:56:38,000
surveys nor our own anecdotal experience

1646
00:56:36,000 --> 00:56:40,000
as pediatricians

1647
00:56:38,000 --> 00:56:41,760
moving to electronic screening it's a

1648
00:56:40,000 --> 00:56:43,599
pre-post model this is kind of a messy

1649
00:56:41,760 --> 00:56:45,599
innovation study but it's embedded now

1650
00:56:43,599 --> 00:56:46,880
in the large health system dramatically

1651
00:56:45,599 --> 00:56:48,880
improves disclosure i'm just showing you

1652
00:56:46,880 --> 00:56:50,079
some of the high level uh unsurprisingly

1653
00:56:48,880 --> 00:56:51,040
the vast majority of teens don't eat the

1654
00:56:50,079 --> 00:56:52,240
right amount of fruits and veggies i

1655
00:56:51,040 --> 00:56:53,200
know this is beyond substance use but

1656
00:56:52,240 --> 00:56:54,400
just highlighting how there are a whole

1657
00:56:53,200 --> 00:56:56,400
bunch of other health risks we talk

1658
00:56:54,400 --> 00:56:58,880
about a lot of teenagers have a gun in

1659
00:56:56,400 --> 00:57:00,799
the home that they know about most many

1660
00:56:58,880 --> 00:57:02,720
teens don't wear their seatbelt but of

1661
00:57:00,799 --> 00:57:04,079
the interests of this audience so

1662
00:57:02,720 --> 00:57:06,000
previous implementation we're talking

1663
00:57:04,079 --> 00:57:08,319
about one percent of teens disclosing

1664
00:57:06,000 --> 00:57:10,240
substance use uh now we're up to nine

1665
00:57:08,319 --> 00:57:11,920
percent reporting alcohol nine percent

1666
00:57:10,240 --> 00:57:14,240
reporting marijuana use eight percent

1667
00:57:11,920 --> 00:57:15,359
regarding tobacco product use and um

1668
00:57:14,240 --> 00:57:16,880
another area that we're focused on

1669
00:57:15,359 --> 00:57:18,960
sexual activity 23 percent of

1670
00:57:16,880 --> 00:57:20,799
adolescents recruit sexual activity and

1671
00:57:18,960 --> 00:57:23,040
something else is we also include um

1672
00:57:20,799 --> 00:57:24,960
questions around sexual identity 14 of

1673
00:57:23,040 --> 00:57:26,640
our teens i'm proud of this next one 14

1674
00:57:24,960 --> 00:57:29,200
of our teams i feel comfortable

1675
00:57:26,640 --> 00:57:31,760
disclosing that they're they identify as

1676
00:57:29,200 --> 00:57:33,680
lgbtq on finally on the right hand side

1677
00:57:31,760 --> 00:57:35,119
i'm just showing the craft score to show

1678
00:57:33,680 --> 00:57:37,119
you where we can move next we can do

1679
00:57:35,119 --> 00:57:38,720
profiling obtained or kind of i should

1680
00:57:37,119 --> 00:57:40,960
say stratification of teams that are at

1681
00:57:38,720 --> 00:57:41,760
risk and then target interventions for

1682
00:57:40,960 --> 00:57:43,440
them

1683
00:57:41,760 --> 00:57:45,040
the final thing we're focused on now is

1684
00:57:43,440 --> 00:57:46,160
really decision support to address the

1685
00:57:45,040 --> 00:57:48,640
needs

1686
00:57:46,160 --> 00:57:50,400
based on issues identified through our

1687
00:57:48,640 --> 00:57:52,400
electronic questionnaire approach this

1688
00:57:50,400 --> 00:57:54,799
is a busy slide but it's showing you

1689
00:57:52,400 --> 00:57:56,640
prompts towards the physicians around

1690
00:57:54,799 --> 00:57:59,359
sexual

1691
00:57:56,640 --> 00:58:01,520
sti-based screening prevention and

1692
00:57:59,359 --> 00:58:03,520
treatment meaning that if a teen

1693
00:58:01,520 --> 00:58:04,720
identifies as being sexually active but

1694
00:58:03,520 --> 00:58:05,920
our system shows they've never been

1695
00:58:04,720 --> 00:58:07,839
screened or haven't recently been

1696
00:58:05,920 --> 00:58:09,760
appropriately screened for stis this

1697
00:58:07,839 --> 00:58:11,599
then is one of our ways of getting this

1698
00:58:09,760 --> 00:58:15,040
in front of a physician to order

1699
00:58:11,599 --> 00:58:16,400
appropriate labs and get those labs done

1700
00:58:15,040 --> 00:58:18,400
with that

1701
00:58:16,400 --> 00:58:20,559
i'll close and i think we'll open up for

1702
00:58:18,400 --> 00:58:20,559
q

1703
00:58:20,839 --> 00:58:26,240
a thank you ryan um

1704
00:58:23,920 --> 00:58:27,760
and thanks to each of you that was great

1705
00:58:26,240 --> 00:58:28,480
um

1706
00:58:27,760 --> 00:58:29,520
we

1707
00:58:28,480 --> 00:58:32,559
um

1708
00:58:29,520 --> 00:58:34,559
i'd love for anyone who has questions to

1709
00:58:32,559 --> 00:58:37,200
um to put them in the chat

1710
00:58:34,559 --> 00:58:38,559
um go ahead and ask and just as we're

1711
00:58:37,200 --> 00:58:41,280
we're kind of assembling and others are

1712
00:58:38,559 --> 00:58:42,720
thinking of questions um brian there was

1713
00:58:41,280 --> 00:58:45,200
a question that i know we talked about

1714
00:58:42,720 --> 00:58:46,000
and you and you kind of alluded to but

1715
00:58:45,200 --> 00:58:47,920
um

1716
00:58:46,000 --> 00:58:50,240
a question about how close the school

1717
00:58:47,920 --> 00:58:52,559
reported data populations align with

1718
00:58:50,240 --> 00:58:55,520
those um in this study i know it was way

1719
00:58:52,559 --> 00:58:56,640
off for your initial study um but what

1720
00:58:55,520 --> 00:58:58,880
um

1721
00:58:56,640 --> 00:59:00,240
how close did you get to

1722
00:58:58,880 --> 00:59:03,040
kind of

1723
00:59:00,240 --> 00:59:04,799
believing that those screening data were

1724
00:59:03,040 --> 00:59:07,119
aligned with what we're

1725
00:59:04,799 --> 00:59:08,640
seeing in population-based studies

1726
00:59:07,119 --> 00:59:10,400
yeah i'll launch that in two ways first

1727
00:59:08,640 --> 00:59:12,000
as the preface is often thinking about

1728
00:59:10,400 --> 00:59:13,359
the alternative and recognizing that

1729
00:59:12,000 --> 00:59:14,799
when you move things from a research

1730
00:59:13,359 --> 00:59:15,680
enterprise into clinic enterprise it's

1731
00:59:14,799 --> 00:59:17,599
messy

1732
00:59:15,680 --> 00:59:19,040
and so i would say though it's closer

1733
00:59:17,599 --> 00:59:21,040
when you look at our regional

1734
00:59:19,040 --> 00:59:23,119
school-based surveys

1735
00:59:21,040 --> 00:59:24,960
e-cigarette use is more like in tobacco

1736
00:59:23,119 --> 00:59:26,960
use overall more like in the

1737
00:59:24,960 --> 00:59:28,799
mid-teen range at this point which is

1738
00:59:26,960 --> 00:59:30,960
probably still another report

1739
00:59:28,799 --> 00:59:32,799
similarly for alcohol marijuana the

1740
00:59:30,960 --> 00:59:35,520
sexual activity though is interesting uh

1741
00:59:32,799 --> 00:59:37,040
is more and more aligned um so i would

1742
00:59:35,520 --> 00:59:38,880
say we've moved in the right direction

1743
00:59:37,040 --> 00:59:41,119
dramatically the final thing i think

1744
00:59:38,880 --> 00:59:44,319
about even if it doesn't perfectly align

1745
00:59:41,119 --> 00:59:46,160
with anonymous school-based surveys you

1746
00:59:44,319 --> 00:59:48,799
at least now have a population of teens

1747
00:59:46,160 --> 00:59:50,640
that are identifying that they use which

1748
00:59:48,799 --> 00:59:52,640
is probably a t a group of teens that

1749
00:59:50,640 --> 00:59:54,000
are more amenable to engaging in some

1750
00:59:52,640 --> 00:59:58,440
sort of treatment

1751
00:59:54,000 --> 00:59:58,440
or prevention-based kind of exercise

1752
01:00:00,400 --> 01:00:06,240
so to to back up a little bit so we

1753
01:00:03,760 --> 01:00:08,160
most of the folks on this call um are

1754
01:00:06,240 --> 01:00:10,240
researchers and we have three we have

1755
01:00:08,160 --> 01:00:12,720
three clinicians um

1756
01:00:10,240 --> 01:00:15,520
on our panel um so

1757
01:00:12,720 --> 01:00:16,799
as we're thinking about research um

1758
01:00:15,520 --> 01:00:19,119
you know

1759
01:00:16,799 --> 01:00:20,960
we sort of i i i think of and we have

1760
01:00:19,119 --> 01:00:22,720
another question as well about about

1761
01:00:20,960 --> 01:00:24,960
parents within a pediatric population

1762
01:00:22,720 --> 01:00:26,839
but there's kind of three users of these

1763
01:00:24,960 --> 01:00:29,359
data potentially there's there's the

1764
01:00:26,839 --> 01:00:32,319
patient um or the patient in the family

1765
01:00:29,359 --> 01:00:33,839
the caregiver um there's the clinician

1766
01:00:32,319 --> 01:00:36,559
and then there's the

1767
01:00:33,839 --> 01:00:37,920
then there's a potential the researcher

1768
01:00:36,559 --> 01:00:39,119
there's other masters right there's

1769
01:00:37,920 --> 01:00:40,319
there's uh

1770
01:00:39,119 --> 01:00:42,240
there's

1771
01:00:40,319 --> 01:00:43,680
the health system and there's billing

1772
01:00:42,240 --> 01:00:46,640
and there's quality right there's

1773
01:00:43,680 --> 01:00:48,559
there's a lot of uses for these data um

1774
01:00:46,640 --> 01:00:52,000
but how you know putting on your

1775
01:00:48,559 --> 01:00:54,640
clinician hats how um

1776
01:00:52,000 --> 01:00:56,319
how can we sort of think about

1777
01:00:54,640 --> 01:00:58,960
the questions that we're asking as

1778
01:00:56,319 --> 01:01:01,040
researchers um and make sure that what

1779
01:00:58,960 --> 01:01:01,920
we're asking is

1780
01:01:01,040 --> 01:01:03,839
um

1781
01:01:01,920 --> 01:01:07,280
and what we what we're hoping to see in

1782
01:01:03,839 --> 01:01:10,559
ehr data is is feasible and useful for

1783
01:01:07,280 --> 01:01:12,559
clinicians and and patients as well um

1784
01:01:10,559 --> 01:01:14,640
so i think i think you know

1785
01:01:12,559 --> 01:01:17,359
one of our goals is to increase the use

1786
01:01:14,640 --> 01:01:18,960
of ehr data and research um but we don't

1787
01:01:17,359 --> 01:01:21,280
want to do that um

1788
01:01:18,960 --> 01:01:22,720
to the detriment of of clinical

1789
01:01:21,280 --> 01:01:25,359
encounters to

1790
01:01:22,720 --> 01:01:26,880
you know to the detriment of clinicians

1791
01:01:25,359 --> 01:01:28,880
um and the work that they need to get

1792
01:01:26,880 --> 01:01:30,640
done and patient patient health outcomes

1793
01:01:28,880 --> 01:01:32,240
obviously

1794
01:01:30,640 --> 01:01:35,520
and that's kind of a big question does

1795
01:01:32,240 --> 01:01:35,520
anyone have any thoughts on it

1796
01:01:36,240 --> 01:01:39,920
well i can i can feel that but i'm going

1797
01:01:37,839 --> 01:01:41,920
to defer to mario and bob who are

1798
01:01:39,920 --> 01:01:43,440
smarter than me but i i think the um

1799
01:01:41,920 --> 01:01:45,520
what you need to do on the research side

1800
01:01:43,440 --> 01:01:46,559
so part of our role is i

1801
01:01:45,520 --> 01:01:48,400
you need to find people who multiple

1802
01:01:46,559 --> 01:01:50,000
hats who can bridge the gap so i think

1803
01:01:48,400 --> 01:01:51,839
if you're thinking you know clinical

1804
01:01:50,000 --> 01:01:53,200
based research one make sure that you're

1805
01:01:51,839 --> 01:01:55,200
falling in love with problem not

1806
01:01:53,200 --> 01:01:56,960
solutions and really helping discover

1807
01:01:55,200 --> 01:01:58,000
your problem by talking to your clinical

1808
01:01:56,960 --> 01:02:00,000
partners

1809
01:01:58,000 --> 01:02:01,920
and i think the other thing on the ehr

1810
01:02:00,000 --> 01:02:04,000
side is you need to find good partners

1811
01:02:01,920 --> 01:02:05,920
who are clinical information like mar

1812
01:02:04,000 --> 01:02:07,200
and i who can help you you can challenge

1813
01:02:05,920 --> 01:02:08,880
them but also can help you think about

1814
01:02:07,200 --> 01:02:11,520
what's feasible and what's not it's

1815
01:02:08,880 --> 01:02:12,880
gonna identify well you might wanna

1816
01:02:11,520 --> 01:02:14,240
focus on that but we've done that that's

1817
01:02:12,880 --> 01:02:16,240
not the best approach here's another

1818
01:02:14,240 --> 01:02:17,760
suggestion so i think it's really just

1819
01:02:16,240 --> 01:02:20,160
to summarize is finding those right

1820
01:02:17,760 --> 01:02:22,480
clinical partners uh to talk with and

1821
01:02:20,160 --> 01:02:23,680
really looking them as partners not that

1822
01:02:22,480 --> 01:02:25,520
you're going to tell them how to do

1823
01:02:23,680 --> 01:02:27,280
something so you can really fall in love

1824
01:02:25,520 --> 01:02:28,400
with the problem not your potential

1825
01:02:27,280 --> 01:02:31,119
solution

1826
01:02:28,400 --> 01:02:31,119
if that makes sense

1827
01:02:32,480 --> 01:02:36,240
yeah to add on to what brian just said

1828
01:02:35,440 --> 01:02:39,039
um

1829
01:02:36,240 --> 01:02:41,599
and that partnering part is uh

1830
01:02:39,039 --> 01:02:43,359
collaborating with patients uh

1831
01:02:41,599 --> 01:02:45,440
caregivers and clinicians actually

1832
01:02:43,359 --> 01:02:47,920
getting their input on

1833
01:02:45,440 --> 01:02:50,880
what they would like to see the outcomes

1834
01:02:47,920 --> 01:02:53,359
are what is important to them and

1835
01:02:50,880 --> 01:02:56,799
working with them to try to execute that

1836
01:02:53,359 --> 01:02:58,400
it's uh it's almost like you're

1837
01:02:56,799 --> 01:02:59,680
trying to like the

1838
01:02:58,400 --> 01:03:01,839
example that comes to mind is like the

1839
01:02:59,680 --> 01:03:03,200
delorean it looks cool it's pretty

1840
01:03:01,839 --> 01:03:05,440
outside of

1841
01:03:03,200 --> 01:03:08,720
time travel you know how much was it

1842
01:03:05,440 --> 01:03:10,559
utilized and how much was it uptake

1843
01:03:08,720 --> 01:03:12,799
from the general population so if you

1844
01:03:10,559 --> 01:03:14,000
can work with your stakeholders

1845
01:03:12,799 --> 01:03:17,039
to

1846
01:03:14,000 --> 01:03:18,720
create common goals and create uh

1847
01:03:17,039 --> 01:03:20,799
get what their

1848
01:03:18,720 --> 01:03:22,960
uh outcomes would uh would be what

1849
01:03:20,799 --> 01:03:24,559
they'd like to see then uh i believe you

1850
01:03:22,960 --> 01:03:26,960
get uh

1851
01:03:24,559 --> 01:03:29,039
greater access to uh to ehr and then

1852
01:03:26,960 --> 01:03:31,280
also

1853
01:03:29,039 --> 01:03:33,760
uh input and uh

1854
01:03:31,280 --> 01:03:35,760
and serviceability from uh from patients

1855
01:03:33,760 --> 01:03:38,559
and those providing uh

1856
01:03:35,760 --> 01:03:38,559
uh support to them

1857
01:03:39,839 --> 01:03:43,599
the only thing i'll add and i love the

1858
01:03:41,680 --> 01:03:45,680
way you deframed is multiple uses for

1859
01:03:43,599 --> 01:03:47,920
the ehr sarah i mean and that's part of

1860
01:03:45,680 --> 01:03:49,200
the problem too um you know is and we've

1861
01:03:47,920 --> 01:03:50,799
really had to focus at least in my

1862
01:03:49,200 --> 01:03:53,039
experience is the usability piece it's

1863
01:03:50,799 --> 01:03:54,799
got to be something that's easily usable

1864
01:03:53,039 --> 01:03:56,960
by the clinician otherwise you know the

1865
01:03:54,799 --> 01:03:58,319
stories we hear about burnout and people

1866
01:03:56,960 --> 01:04:00,240
have spending more screen time than

1867
01:03:58,319 --> 01:04:01,680
patient time i think what will continue

1868
01:04:00,240 --> 01:04:03,119
to be to be a problem and then as i

1869
01:04:01,680 --> 01:04:05,119
think further about it you know we think

1870
01:04:03,119 --> 01:04:06,640
about ehrs for individual patient care

1871
01:04:05,119 --> 01:04:09,039
right so the information in there has to

1872
01:04:06,640 --> 01:04:10,640
be adequate to deliver patient care

1873
01:04:09,039 --> 01:04:11,760
we've explored whether it's useful for

1874
01:04:10,640 --> 01:04:13,119
quality improvement can you get a

1875
01:04:11,760 --> 01:04:14,720
feedback loop in there enough so you can

1876
01:04:13,119 --> 01:04:16,720
see how well you're doing and i think

1877
01:04:14,720 --> 01:04:18,240
that's not quite ready yet i mean i

1878
01:04:16,720 --> 01:04:19,680
think it's getting better but not quite

1879
01:04:18,240 --> 01:04:21,280
ready then the third one is for research

1880
01:04:19,680 --> 01:04:22,799
can you actually generate data that will

1881
01:04:21,280 --> 01:04:24,880
give you generalizable knowledge that

1882
01:04:22,799 --> 01:04:26,559
you can apply to other populations i

1883
01:04:24,880 --> 01:04:28,400
think all three of those they're not at

1884
01:04:26,559 --> 01:04:30,799
odds with each other but but man you got

1885
01:04:28,400 --> 01:04:32,480
to sort of work to to get them all lined

1886
01:04:30,799 --> 01:04:33,680
up and not put more burden on the

1887
01:04:32,480 --> 01:04:36,240
clinician i think if you really want to

1888
01:04:33,680 --> 01:04:36,240
make this work

1889
01:04:38,880 --> 01:04:44,480
one other another question um

1890
01:04:42,480 --> 01:04:46,880
at what level do you think clinical

1891
01:04:44,480 --> 01:04:48,880
decision support is best scaled up after

1892
01:04:46,880 --> 01:04:50,319
testing at the health system level a

1893
01:04:48,880 --> 01:04:54,160
state level

1894
01:04:50,319 --> 01:04:56,799
um ehr vendor level um you know how much

1895
01:04:54,160 --> 01:04:59,680
do we need to know about what's working

1896
01:04:56,799 --> 01:05:01,760
um and with whom um and what sort of

1897
01:04:59,680 --> 01:05:03,680
what ehr

1898
01:05:01,760 --> 01:05:05,920
what health systems or what populations

1899
01:05:03,680 --> 01:05:09,760
or what clinician types

1900
01:05:05,920 --> 01:05:11,520
before we can start to to deploy this

1901
01:05:09,760 --> 01:05:13,280
more broadly and therefore kind of you

1902
01:05:11,520 --> 01:05:15,280
know start thinking about scaling up

1903
01:05:13,280 --> 01:05:16,960
some of these primary care based

1904
01:05:15,280 --> 01:05:20,319
interventions that are supported by

1905
01:05:16,960 --> 01:05:20,319
clinical decision support

1906
01:05:21,440 --> 01:05:25,680
i can jump in so this is uh actually one

1907
01:05:23,680 --> 01:05:28,000
of the areas of focus of one of our

1908
01:05:25,680 --> 01:05:30,079
funding opportunities and i highlighted

1909
01:05:28,000 --> 01:05:32,079
uh dr christopher harley's work uh

1910
01:05:30,079 --> 01:05:33,839
university of florida so

1911
01:05:32,079 --> 01:05:35,359
there really isn't a

1912
01:05:33,839 --> 01:05:37,280
right answer

1913
01:05:35,359 --> 01:05:39,599
of how you scale up and who you scale up

1914
01:05:37,280 --> 01:05:41,680
to do you start you know at the state

1915
01:05:39,599 --> 01:05:43,599
level uh do you work your

1916
01:05:41,680 --> 01:05:46,400
work your way down or do you start just

1917
01:05:43,599 --> 01:05:48,240
by trying to go across the street to the

1918
01:05:46,400 --> 01:05:50,880
primary care clinic

1919
01:05:48,240 --> 01:05:53,280
that's affiliated with uh your hospital

1920
01:05:50,880 --> 01:05:55,119
uh so there is no right answer the i

1921
01:05:53,280 --> 01:05:57,200
think uh one of the main tenants that

1922
01:05:55,119 --> 01:05:59,839
we're trying to get is one once you

1923
01:05:57,200 --> 01:06:00,799
piloted your clinical decision support

1924
01:05:59,839 --> 01:06:03,760
shows

1925
01:06:00,799 --> 01:06:06,000
show that it works in your local setting

1926
01:06:03,760 --> 01:06:08,319
has some benefit or even if you're

1927
01:06:06,000 --> 01:06:12,400
unsure if it has benefit that's another

1928
01:06:08,319 --> 01:06:14,160
opportunity for scaling is partner with

1929
01:06:12,400 --> 01:06:17,440
other collaborators your other

1930
01:06:14,160 --> 01:06:19,839
stakeholders who want to scale this

1931
01:06:17,440 --> 01:06:22,079
as researchers you know given time

1932
01:06:19,839 --> 01:06:24,000
effort budget you know do what is

1933
01:06:22,079 --> 01:06:26,160
allowed in what you're you're capable

1934
01:06:24,000 --> 01:06:27,280
and feel comfortable executing and then

1935
01:06:26,160 --> 01:06:29,680
from there

1936
01:06:27,280 --> 01:06:31,359
you know reporting on your outcomes

1937
01:06:29,680 --> 01:06:33,119
reporting on your results and then

1938
01:06:31,359 --> 01:06:36,000
sharing that information disseminating

1939
01:06:33,119 --> 01:06:37,440
it with not only uh peers but ourselves

1940
01:06:36,000 --> 01:06:39,039
so that we can get a sense of what

1941
01:06:37,440 --> 01:06:41,359
worked what didn't work and then hope

1942
01:06:39,039 --> 01:06:43,760
you can share lessons learned and then

1943
01:06:41,359 --> 01:06:47,839
from there someone can pivot on

1944
01:06:43,760 --> 01:06:47,839
on on your work so

1945
01:06:51,520 --> 01:06:54,640
great um thanks

1946
01:06:53,440 --> 01:06:56,720
um

1947
01:06:54,640 --> 01:06:59,200
maybe we might just have a question for

1948
01:06:56,720 --> 01:07:01,039
or time for one more um

1949
01:06:59,200 --> 01:07:02,960
so um

1950
01:07:01,039 --> 01:07:04,240
as we are um

1951
01:07:02,960 --> 01:07:06,160
there's been a lot of questions and i

1952
01:07:04,240 --> 01:07:06,960
think we're all still thinking about

1953
01:07:06,160 --> 01:07:10,319
this

1954
01:07:06,960 --> 01:07:12,319
um about the the relationship or how

1955
01:07:10,319 --> 01:07:14,160
we're using these tools and virtual

1956
01:07:12,319 --> 01:07:17,760
versus

1957
01:07:14,160 --> 01:07:19,359
in-person um in person settings so um

1958
01:07:17,760 --> 01:07:21,119
are there any is there anything that we

1959
01:07:19,359 --> 01:07:22,559
need to do differently and and brian

1960
01:07:21,119 --> 01:07:25,119
there's you know kind of some specific

1961
01:07:22,559 --> 01:07:26,640
questions for you about whether or not

1962
01:07:25,119 --> 01:07:29,039
your screening

1963
01:07:26,640 --> 01:07:31,039
was done in person or only during

1964
01:07:29,039 --> 01:07:32,240
in-person visits or if it's something

1965
01:07:31,039 --> 01:07:34,000
that can be

1966
01:07:32,240 --> 01:07:35,119
deployed in virtual care settings as

1967
01:07:34,000 --> 01:07:37,440
well

1968
01:07:35,119 --> 01:07:39,119
yeah it's a lot of great questions um

1969
01:07:37,440 --> 01:07:41,920
let me feel that in a couple of ways so

1970
01:07:39,119 --> 01:07:43,359
what we've um one of the big issues is

1971
01:07:41,920 --> 01:07:44,720
as we mentioned as i mentioned and the

1972
01:07:43,359 --> 01:07:46,640
comments are speaking to is the

1973
01:07:44,720 --> 01:07:48,640
confidentiality issue of how do you get

1974
01:07:46,640 --> 01:07:49,839
information from the teen electronically

1975
01:07:48,640 --> 01:07:51,760
ensuring

1976
01:07:49,839 --> 01:07:52,960
especially if it's remote and ensuring

1977
01:07:51,760 --> 01:07:54,960
it's not the parent who's completing

1978
01:07:52,960 --> 01:07:56,640
that content so we did our own pilot

1979
01:07:54,960 --> 01:07:59,119
studies to show that we couldn't rely on

1980
01:07:56,640 --> 01:08:01,280
sending this content ahead of time

1981
01:07:59,119 --> 01:08:02,640
because um and the stanford-based group

1982
01:08:01,280 --> 01:08:04,160
also identified this when they did

1983
01:08:02,640 --> 01:08:06,000
natural language processing to look at

1984
01:08:04,160 --> 01:08:07,920
the language coming back from supposedly

1985
01:08:06,000 --> 01:08:09,520
the teenager it was the parent who was

1986
01:08:07,920 --> 01:08:11,599
completing the content which is

1987
01:08:09,520 --> 01:08:12,720
frustrating because it's like

1988
01:08:11,599 --> 01:08:14,319
you know you think you're committed to

1989
01:08:12,720 --> 01:08:16,080
the teen it says team but it's actually

1990
01:08:14,319 --> 01:08:17,440
the mom dad responding back so there's

1991
01:08:16,080 --> 01:08:19,679
an intellectual dishonesty there that

1992
01:08:17,440 --> 01:08:21,040
you have to navigate um as a physician

1993
01:08:19,679 --> 01:08:23,359
as a health system

1994
01:08:21,040 --> 01:08:25,440
so what we did first is make sure we

1995
01:08:23,359 --> 01:08:27,279
have this completed when the individual

1996
01:08:25,440 --> 01:08:28,480
team comes into the office that was our

1997
01:08:27,279 --> 01:08:29,759
first approach because that was the only

1998
01:08:28,480 --> 01:08:31,600
way you could verify it was the team

1999
01:08:29,759 --> 01:08:33,520
actually completing this

2000
01:08:31,600 --> 01:08:35,520
now what we've done with an acceleration

2001
01:08:33,520 --> 01:08:37,520
because of pandemic and better services

2002
01:08:35,520 --> 01:08:39,839
for remote services better opportunities

2003
01:08:37,520 --> 01:08:41,679
and technology remote services

2004
01:08:39,839 --> 01:08:43,839
we've figured out

2005
01:08:41,679 --> 01:08:45,279
the proxy systems and making sure that

2006
01:08:43,839 --> 01:08:46,799
it's actually the team who's completing

2007
01:08:45,279 --> 01:08:48,640
the content and once we did that you

2008
01:08:46,799 --> 01:08:50,000
could verify that then we could send out

2009
01:08:48,640 --> 01:08:51,759
and open up opportunities to do

2010
01:08:50,000 --> 01:08:53,759
telehealth well visits where we're

2011
01:08:51,759 --> 01:08:54,880
actually kind of screening and capturing

2012
01:08:53,759 --> 01:08:57,120
this data

2013
01:08:54,880 --> 01:08:58,960
and i really um for those interested in

2014
01:08:57,120 --> 01:09:01,679
this is and i'm playing off of what uh

2015
01:08:58,960 --> 01:09:02,799
wilson had commented on it's not

2016
01:09:01,679 --> 01:09:04,400
one of the other reasons why you want to

2017
01:09:02,799 --> 01:09:05,679
find your collaborators ask them what

2018
01:09:04,400 --> 01:09:07,440
are the problems their facing so again

2019
01:09:05,679 --> 01:09:09,600
fall in love with problems not solutions

2020
01:09:07,440 --> 01:09:11,040
and bob had mentioned usability is so

2021
01:09:09,600 --> 01:09:13,120
important making sure you're building

2022
01:09:11,040 --> 01:09:15,040
solutions or we're helping other people

2023
01:09:13,120 --> 01:09:16,319
identify the solutions to really fit and

2024
01:09:15,040 --> 01:09:18,000
meet the needs

2025
01:09:16,319 --> 01:09:20,719
when we heard people saying that you

2026
01:09:18,000 --> 01:09:22,080
know talking to teens takes too long

2027
01:09:20,719 --> 01:09:23,679
we were like okay we're on to something

2028
01:09:22,080 --> 01:09:25,040
we can we can come up with we can do

2029
01:09:23,679 --> 01:09:26,560
time saving there and just get the right

2030
01:09:25,040 --> 01:09:27,920
information to them at the right time

2031
01:09:26,560 --> 01:09:30,880
which tomorrow's talking about the five

2032
01:09:27,920 --> 01:09:32,719
tenets of effective um ehr based

2033
01:09:30,880 --> 01:09:34,319
solutions and cds

2034
01:09:32,719 --> 01:09:35,759
so there it was through discussions that

2035
01:09:34,319 --> 01:09:36,799
we learned more about this problem so i

2036
01:09:35,759 --> 01:09:38,640
know i just answered kind of three

2037
01:09:36,799 --> 01:09:40,080
questions in the comments at once but

2038
01:09:38,640 --> 01:09:42,239
i'm happy to talk further anyway i left

2039
01:09:40,080 --> 01:09:45,600
my email at any time more about some of

2040
01:09:42,239 --> 01:09:45,600
these issues and i'll stop talking

2041
01:09:46,159 --> 01:09:50,239
thank you brian um and i know that um i

2042
01:09:48,960 --> 01:09:52,719
know that we all have a lot more to say

2043
01:09:50,239 --> 01:09:55,600
but we're gonna take a very quick five

2044
01:09:52,719 --> 01:09:56,400
minute break um grab some water um thank

2045
01:09:55,600 --> 01:09:58,960
you

2046
01:09:56,400 --> 01:10:00,880
bob mario and brian for um

2047
01:09:58,960 --> 01:10:03,280
for your um

2048
01:10:00,880 --> 01:10:06,320
for your contributions um and for your

2049
01:10:03,280 --> 01:10:08,640
um your time today and um yes feel free

2050
01:10:06,320 --> 01:10:10,239
to continue to ask questions in the chat

2051
01:10:08,640 --> 01:10:11,520
um and anything that we didn't get to we

2052
01:10:10,239 --> 01:10:14,320
will try to

2053
01:10:11,520 --> 01:10:16,480
um we'll try to get into our um

2054
01:10:14,320 --> 01:10:18,719
our summary um of our

2055
01:10:16,480 --> 01:10:20,000
of our um meeting that we'll be putting

2056
01:10:18,719 --> 01:10:21,840
together

2057
01:10:20,000 --> 01:10:23,280
um okay so we've got five minute break

2058
01:10:21,840 --> 01:10:24,640
um

2059
01:10:23,280 --> 01:10:26,159
we have a

2060
01:10:24,640 --> 01:10:29,600
uh

2061
01:10:26,159 --> 01:10:31,760
timer up on the screen so everyone can

2062
01:10:29,600 --> 01:10:33,920
run get some get some water stretch your

2063
01:10:31,760 --> 01:10:36,880
legs and we'll meet you back here in

2064
01:10:33,920 --> 01:10:36,880
five minutes thank you

2065
01:10:37,360 --> 01:10:40,880
like to welcome everybody back from

2066
01:10:38,800 --> 01:10:42,640
break the serene water picture has

2067
01:10:40,880 --> 01:10:44,880
disappeared

2068
01:10:42,640 --> 01:10:46,320
i will take over from there and welcome

2069
01:10:44,880 --> 01:10:48,320
you back to our second panel

2070
01:10:46,320 --> 01:10:49,840
transitioning from our first

2071
01:10:48,320 --> 01:10:51,920
we're now going to focus on examples of

2072
01:10:49,840 --> 01:10:54,400
research that have successfully used ehr

2073
01:10:51,920 --> 01:10:56,640
data in mental health pediatrics health

2074
01:10:54,400 --> 01:10:59,600
disparities substance use and veterans

2075
01:10:56,640 --> 01:11:01,360
health research so um as sarah did i'll

2076
01:10:59,600 --> 01:11:03,520
go ahead and introduce all the speakers

2077
01:11:01,360 --> 01:11:05,280
at the start and then you know feel free

2078
01:11:03,520 --> 01:11:06,880
to start dumping questions in the chat

2079
01:11:05,280 --> 01:11:08,400
and we'll have a opportunity for

2080
01:11:06,880 --> 01:11:10,080
discussion at the end

2081
01:11:08,400 --> 01:11:12,320
so our first speaker will be dr greg

2082
01:11:10,080 --> 01:11:14,400
simon dr simon is a psychiatrist and a

2083
01:11:12,320 --> 01:11:16,080
senior investigator at kaiser permanente

2084
01:11:14,400 --> 01:11:17,520
in washington

2085
01:11:16,080 --> 01:11:19,280
he leads the mental health research

2086
01:11:17,520 --> 01:11:21,280
network a consortium of research centers

2087
01:11:19,280 --> 01:11:22,880
affiliated with 13 large health systems

2088
01:11:21,280 --> 01:11:24,960
across the united states conducting

2089
01:11:22,880 --> 01:11:26,560
practical mental health research

2090
01:11:24,960 --> 01:11:28,560
next we'll hear from dr dina chisholm

2091
01:11:26,560 --> 01:11:30,159
who's a health services epidemiologist

2092
01:11:28,560 --> 01:11:32,159
and the director of the center for child

2093
01:11:30,159 --> 01:11:34,159
health equity and outcomes research at

2094
01:11:32,159 --> 01:11:36,239
abigail wexner research institute at

2095
01:11:34,159 --> 01:11:37,920
nationwide children's hospital

2096
01:11:36,239 --> 01:11:39,840
she's also the vice president of health

2097
01:11:37,920 --> 01:11:41,679
services research in the abigail wexner

2098
01:11:39,840 --> 01:11:43,520
research institute and a professor of

2099
01:11:41,679 --> 01:11:45,840
pediatrics at ohio state's college of

2100
01:11:43,520 --> 01:11:47,679
medicine and public health

2101
01:11:45,840 --> 01:11:49,920
following dr chisholm will be joined by

2102
01:11:47,679 --> 01:11:51,360
dr stacie sterling a research scientist

2103
01:11:49,920 --> 01:11:53,600
with kaiser permanente northern

2104
01:11:51,360 --> 01:11:54,880
california she's co-director for the

2105
01:11:53,600 --> 01:11:56,239
center for addiction and mental health

2106
01:11:54,880 --> 01:11:57,440
research and kaiser's division of

2107
01:11:56,239 --> 01:11:59,360
research

2108
01:11:57,440 --> 01:12:01,360
and finally we'll conclude this past

2109
01:11:59,360 --> 01:12:03,600
presentations with dr elizabeth

2110
01:12:01,360 --> 01:12:05,840
dr oliva is an investigator at the va

2111
01:12:03,600 --> 01:12:08,320
center for innovation to implementation

2112
01:12:05,840 --> 01:12:10,159
at the va palo alto healthcare system

2113
01:12:08,320 --> 01:12:12,640
and a senior evaluator for the va

2114
01:12:10,159 --> 01:12:15,280
program evaluation and research center

2115
01:12:12,640 --> 01:12:16,880
she is currently the va national opioid

2116
01:12:15,280 --> 01:12:19,040
overdose education and naloxone

2117
01:12:16,880 --> 01:12:21,440
distribution coordinator

2118
01:12:19,040 --> 01:12:23,360
so with that i will turn it over to dr

2119
01:12:21,440 --> 01:12:27,120
simon

2120
01:12:23,360 --> 01:12:32,320
thanks very much can you hear me okay

2121
01:12:27,120 --> 01:12:32,320
okay and i will share my screen here

2122
01:12:32,640 --> 01:12:35,640
and

2123
01:12:36,880 --> 01:12:43,199
taking a second let's see

2124
01:12:40,159 --> 01:12:45,600
i think we are okay and then

2125
01:12:43,199 --> 01:12:47,600
are you you know i will do the high risk

2126
01:12:45,600 --> 01:12:49,280
maneuver or switching to presentation

2127
01:12:47,600 --> 01:12:51,600
mode are you seeing the correct version

2128
01:12:49,280 --> 01:12:52,480
now or the other version

2129
01:12:51,600 --> 01:12:54,000
correct

2130
01:12:52,480 --> 01:12:56,640
okay good all right then i think we're

2131
01:12:54,000 --> 01:12:59,440
all good so thanks very much um i'm

2132
01:12:56,640 --> 01:13:01,280
going to uh talk today about uh using

2133
01:12:59,440 --> 01:13:02,880
what i call the sort of data exhaust of

2134
01:13:01,280 --> 01:13:04,320
health care i'm actually gonna you'll

2135
01:13:02,880 --> 01:13:05,760
i'll echo a few of the things that i

2136
01:13:04,320 --> 01:13:08,880
heard brian jensen say just a few

2137
01:13:05,760 --> 01:13:10,960
minutes ago um i'm talking about some of

2138
01:13:08,880 --> 01:13:12,320
our work using those sorts of data to

2139
01:13:10,960 --> 01:13:14,640
assess the mental health impact of

2140
01:13:12,320 --> 01:13:17,360
covid19 not because that's the primary

2141
01:13:14,640 --> 01:13:19,199
uh topic of this webinar but because i

2142
01:13:17,360 --> 01:13:21,199
think there's some nice cautionary tales

2143
01:13:19,199 --> 01:13:23,360
we can tell this was work done by our

2144
01:13:21,199 --> 01:13:24,800
mental health research network

2145
01:13:23,360 --> 01:13:26,159
which involves several large health

2146
01:13:24,800 --> 01:13:27,440
systems but i want to especially

2147
01:13:26,159 --> 01:13:28,800
acknowledge some of the people more

2148
01:13:27,440 --> 01:13:31,120
involved in the work i'll talk about

2149
01:13:28,800 --> 01:13:33,120
here rebecca rossum at health partners

2150
01:13:31,120 --> 01:13:35,520
brian amadani and henry ford and my

2151
01:13:33,120 --> 01:13:38,159
colleagues chris stewart and rob penfold

2152
01:13:35,520 --> 01:13:41,040
here at kp washington

2153
01:13:38,159 --> 01:13:42,880
so my main message is you know when you

2154
01:13:41,040 --> 01:13:44,800
ask about can you trust the data that

2155
01:13:42,880 --> 01:13:46,400
come out of electronic health records or

2156
01:13:44,800 --> 01:13:48,480
sometimes health insurance claims or

2157
01:13:46,400 --> 01:13:50,159
pharmacy dispensing data to measure

2158
01:13:48,480 --> 01:13:52,640
things about people's health the answer

2159
01:13:50,159 --> 01:13:54,880
is it depends um you often hear sort of

2160
01:13:52,640 --> 01:13:56,719
broad over generalizations

2161
01:13:54,880 --> 01:13:58,400
that you know big data will solve all

2162
01:13:56,719 --> 01:14:00,080
problems and a large sample size

2163
01:13:58,400 --> 01:14:01,920
overcomes all issues with measurement

2164
01:14:00,080 --> 01:14:03,520
error and on the other hand you'll hear

2165
01:14:01,920 --> 01:14:05,600
data from health records are garbage and

2166
01:14:03,520 --> 01:14:07,199
nonsense and you can't learn anything

2167
01:14:05,600 --> 01:14:09,040
and the truth is it's somewhere in

2168
01:14:07,199 --> 01:14:10,320
between but it's more complicated

2169
01:14:09,040 --> 01:14:12,800
because i think you need to get very

2170
01:14:10,320 --> 01:14:14,480
specific about that so along the lines

2171
01:14:12,800 --> 01:14:16,400
of getting specific i'm going to put in

2172
01:14:14,480 --> 01:14:18,480
a plug here for a paper that i wrote

2173
01:14:16,400 --> 01:14:20,960
along with some other people this was

2174
01:14:18,480 --> 01:14:23,199
came out of a national academy's

2175
01:14:20,960 --> 01:14:25,120
series of workshops which was focused

2176
01:14:23,199 --> 01:14:26,640
more on evaluating medical treatments

2177
01:14:25,120 --> 01:14:29,040
but i think the points we make in this

2178
01:14:26,640 --> 01:14:30,880
paper are relevant very relevant to what

2179
01:14:29,040 --> 01:14:32,960
we're talking about today when we ask

2180
01:14:30,880 --> 01:14:35,679
when can we trust so-called real world

2181
01:14:32,960 --> 01:14:37,520
data or the data exhausted healthcare to

2182
01:14:35,679 --> 01:14:38,800
accurately measure things about people's

2183
01:14:37,520 --> 01:14:41,360
health

2184
01:14:38,800 --> 01:14:43,040
and our point was that there is a clear

2185
01:14:41,360 --> 01:14:45,520
sort of you could say chain of custody

2186
01:14:43,040 --> 01:14:47,520
or series of steps if we begin with a

2187
01:14:45,520 --> 01:14:49,280
health state that a person experiences

2188
01:14:47,520 --> 01:14:51,040
and at the end of the day we have data

2189
01:14:49,280 --> 01:14:53,760
in a data warehouse we need to think

2190
01:14:51,040 --> 01:14:55,600
about each of the steps along the way

2191
01:14:53,760 --> 01:14:57,120
whether that person who experiences that

2192
01:14:55,600 --> 01:14:59,199
health state would actually present to

2193
01:14:57,120 --> 01:15:01,360
the health care system how that health

2194
01:14:59,199 --> 01:15:03,600
state would be diagnosed or recognized

2195
01:15:01,360 --> 01:15:05,760
by a health care provider how it would

2196
01:15:03,600 --> 01:15:07,679
be recorded and how that recording would

2197
01:15:05,760 --> 01:15:09,920
be influenced by the local sort of

2198
01:15:07,679 --> 01:15:12,000
characteristics the social cultural

2199
01:15:09,920 --> 01:15:13,600
economic and technical characteristics

2200
01:15:12,000 --> 01:15:15,600
of the practice setting

2201
01:15:13,600 --> 01:15:17,280
how those data would be extracted and

2202
01:15:15,600 --> 01:15:20,000
harmonized before they were put into

2203
01:15:17,280 --> 01:15:21,600
some research data warehouse or data set

2204
01:15:20,000 --> 01:15:23,360
so we need to consider each of those

2205
01:15:21,600 --> 01:15:25,120
steps and i'm going to illustrate that

2206
01:15:23,360 --> 01:15:27,120
process by talking about some of our

2207
01:15:25,120 --> 01:15:29,440
work on looking at the mental health

2208
01:15:27,120 --> 01:15:30,560
impacts of covert 19.

2209
01:15:29,440 --> 01:15:32,320
so

2210
01:15:30,560 --> 01:15:34,560
there were three big questions that we

2211
01:15:32,320 --> 01:15:37,920
were thinking about back sort of in mid

2212
01:15:34,560 --> 01:15:39,840
to late 2020 how did the covet 19 uh

2213
01:15:37,920 --> 01:15:42,800
pandemic affect overall levels of

2214
01:15:39,840 --> 01:15:44,800
psychological distress how did it affect

2215
01:15:42,800 --> 01:15:47,040
more in in more concerning the sort of

2216
01:15:44,800 --> 01:15:49,280
rate of mental health crises or suicide

2217
01:15:47,040 --> 01:15:51,120
attempts and then finally how did it

2218
01:15:49,280 --> 01:15:52,880
affect suicide deaths or suicide

2219
01:15:51,120 --> 01:15:55,679
mortality and i'm going to talk about

2220
01:15:52,880 --> 01:15:57,360
those actually in reverse order now

2221
01:15:55,679 --> 01:15:59,360
if we go back to those early days of the

2222
01:15:57,360 --> 01:16:01,520
pandemic some of you may remember there

2223
01:15:59,360 --> 01:16:04,000
were some very early alarms about the

2224
01:16:01,520 --> 01:16:06,080
idea that the covet 19 pandemic might

2225
01:16:04,000 --> 01:16:07,520
produce sort of an epidemic of suicide

2226
01:16:06,080 --> 01:16:09,520
this was a prominent paper that was

2227
01:16:07,520 --> 01:16:11,120
published you could see april 10th so

2228
01:16:09,520 --> 01:16:12,880
just a few weeks after the sort of

2229
01:16:11,120 --> 01:16:15,040
pandemic began to really break out

2230
01:16:12,880 --> 01:16:16,960
across the u.s and it was thought to be

2231
01:16:15,040 --> 01:16:19,040
a potential perfect storm

2232
01:16:16,960 --> 01:16:21,199
and as you may remember

2233
01:16:19,040 --> 01:16:22,880
it even got fairly political you know

2234
01:16:21,199 --> 01:16:25,679
this is from that medical journal the

2235
01:16:22,880 --> 01:16:28,000
wall street journal an editorial talking

2236
01:16:25,679 --> 01:16:29,679
about how so-called lockdowns or public

2237
01:16:28,000 --> 01:16:32,320
health measures to reduce the spread of

2238
01:16:29,679 --> 01:16:33,840
covert 19 might lead to an epidemic of

2239
01:16:32,320 --> 01:16:35,760
suicide

2240
01:16:33,840 --> 01:16:37,440
if you looked at the literature about

2241
01:16:35,760 --> 01:16:39,440
various kinds of sort of natural

2242
01:16:37,440 --> 01:16:42,239
disasters or other crises and how they

2243
01:16:39,440 --> 01:16:44,480
affected suicide mortality it actually

2244
01:16:42,239 --> 01:16:46,320
is mixed historically there's even some

2245
01:16:44,480 --> 01:16:48,320
evidence that sort of things like

2246
01:16:46,320 --> 01:16:50,800
earthquakes and floods things that may

2247
01:16:48,320 --> 01:16:52,480
affect a community may sometimes

2248
01:16:50,800 --> 01:16:54,239
actually be followed by decreased

2249
01:16:52,480 --> 01:16:56,239
especially in the short term suicide

2250
01:16:54,239 --> 01:16:58,480
mortality possibly because people feel

2251
01:16:56,239 --> 01:17:00,400
more connected to each other so when we

2252
01:16:58,480 --> 01:17:02,719
were interested in assessing this we

2253
01:17:00,400 --> 01:17:04,560
said what data can we get and that

2254
01:17:02,719 --> 01:17:06,320
naturally those would be data that our

2255
01:17:04,560 --> 01:17:08,159
health systems often

2256
01:17:06,320 --> 01:17:09,920
get from state health departments

2257
01:17:08,159 --> 01:17:11,920
because we've been doing ongoing work

2258
01:17:09,920 --> 01:17:14,560
looking at uh suicide mortality in these

2259
01:17:11,920 --> 01:17:16,480
health systems

2260
01:17:14,560 --> 01:17:18,640
so as many of you may know the sort of

2261
01:17:16,480 --> 01:17:20,320
so-called final mortality data that

2262
01:17:18,640 --> 01:17:21,920
passed from state health departments to

2263
01:17:20,320 --> 01:17:24,239
the center for disease control and come

2264
01:17:21,920 --> 01:17:26,400
back via the national death index are

2265
01:17:24,239 --> 01:17:28,880
typically delayed by up to 20 months

2266
01:17:26,400 --> 01:17:31,440
that means that 2020 data would not be

2267
01:17:28,880 --> 01:17:33,600
available until the fall of 2021 and we

2268
01:17:31,440 --> 01:17:35,040
would hope not to wait that long

2269
01:17:33,600 --> 01:17:37,120
several of the health systems in our

2270
01:17:35,040 --> 01:17:38,800
network had established relationships

2271
01:17:37,120 --> 01:17:40,320
with state departments of health to get

2272
01:17:38,800 --> 01:17:42,159
access to interim data which were

2273
01:17:40,320 --> 01:17:44,560
typically produced quarterly or even

2274
01:17:42,159 --> 01:17:46,800
monthly but we had some questions about

2275
01:17:44,560 --> 01:17:48,960
the timeliness or accuracy of those data

2276
01:17:46,800 --> 01:17:51,520
especially during the pandemic

2277
01:17:48,960 --> 01:17:53,120
so here we are in fall of 2020 and we're

2278
01:17:51,520 --> 01:17:54,960
looking at data from three healthcare

2279
01:17:53,120 --> 01:17:57,440
systems kaiser permanente washington

2280
01:17:54,960 --> 01:17:59,360
henry ford and health partners

2281
01:17:57,440 --> 01:18:01,520
and looking at what we see about the

2282
01:17:59,360 --> 01:18:03,280
number of suicide deaths per month in

2283
01:18:01,520 --> 01:18:04,560
each of those health care systems using

2284
01:18:03,280 --> 01:18:06,159
these data that each of these health

2285
01:18:04,560 --> 01:18:08,159
care systems receives from their

2286
01:18:06,159 --> 01:18:10,719
respective states from washington from

2287
01:18:08,159 --> 01:18:12,320
michigan and from minnesota

2288
01:18:10,719 --> 01:18:14,400
what you see here when you look at when

2289
01:18:12,320 --> 01:18:17,520
the pandemic was really taking off their

2290
01:18:14,400 --> 01:18:19,679
march april and may um certainly uh you

2291
01:18:17,520 --> 01:18:21,840
know you you don't see any evidence of

2292
01:18:19,679 --> 01:18:24,640
an increase it's possible that you see a

2293
01:18:21,840 --> 01:18:26,400
decrease in suicide mortality but also

2294
01:18:24,640 --> 01:18:28,080
what you see there is the data for the

2295
01:18:26,400 --> 01:18:30,560
most recent few months in each of those

2296
01:18:28,080 --> 01:18:31,760
places may really not be credible

2297
01:18:30,560 --> 01:18:33,600
certainly if we look at kaiser

2298
01:18:31,760 --> 01:18:35,199
permanente washington data for june of

2299
01:18:33,600 --> 01:18:37,360
2020 we say

2300
01:18:35,199 --> 01:18:39,120
we must have really incomplete capture

2301
01:18:37,360 --> 01:18:41,840
or if we look at henry ford data from

2302
01:18:39,120 --> 01:18:43,520
october we must have incomplete capture

2303
01:18:41,840 --> 01:18:46,080
the question is how believable are the

2304
01:18:43,520 --> 01:18:49,280
data from a few months back

2305
01:18:46,080 --> 01:18:51,280
so the next step we do is we look at

2306
01:18:49,280 --> 01:18:53,040
at kp washington what happens when we

2307
01:18:51,280 --> 01:18:54,400
look at the data from second quarter and

2308
01:18:53,040 --> 01:18:56,239
then the data from third quarter and

2309
01:18:54,400 --> 01:18:58,320
match them up with each other

2310
01:18:56,239 --> 01:19:00,239
and what we see here is and this is to

2311
01:18:58,320 --> 01:19:02,000
me a very interesting story i won't go

2312
01:19:00,239 --> 01:19:03,840
into the details of all the numbers

2313
01:19:02,000 --> 01:19:06,239
because that would take too long

2314
01:19:03,840 --> 01:19:08,400
but what we see here is that

2315
01:19:06,239 --> 01:19:09,440
when we look at the quarter three data

2316
01:19:08,400 --> 01:19:11,440
and and

2317
01:19:09,440 --> 01:19:13,600
cross check those against what we got

2318
01:19:11,440 --> 01:19:16,159
sort of preliminarily in quarter two we

2319
01:19:13,600 --> 01:19:18,080
see that the data on suicide mortality

2320
01:19:16,159 --> 01:19:20,480
look like they may not be complete for

2321
01:19:18,080 --> 01:19:22,640
three or four months back interestingly

2322
01:19:20,480 --> 01:19:24,560
the data on covet 19 deaths look like

2323
01:19:22,640 --> 01:19:26,320
they were really up to date and if you

2324
01:19:24,560 --> 01:19:29,040
look those total mortality numbers

2325
01:19:26,320 --> 01:19:31,040
hardly changed so we see here a very

2326
01:19:29,040 --> 01:19:33,760
important message that data regarding

2327
01:19:31,040 --> 01:19:35,679
suicide mortality especially seemed to

2328
01:19:33,760 --> 01:19:37,760
be delayed by a few months and really

2329
01:19:35,679 --> 01:19:40,480
are not trustworthy for the most recent

2330
01:19:37,760 --> 01:19:43,120
few months at least in 2020

2331
01:19:40,480 --> 01:19:46,000
and then final piece of that picture

2332
01:19:43,120 --> 01:19:48,960
these are data looking at final uh 2020

2333
01:19:46,000 --> 01:19:51,920
data that arrived in 2021 and how those

2334
01:19:48,960 --> 01:19:54,320
compared to fourth quarter 2020 data and

2335
01:19:51,920 --> 01:19:57,040
this is sort of with these uh uh what

2336
01:19:54,320 --> 01:19:59,199
the lines show is what proportion of

2337
01:19:57,040 --> 01:20:01,199
those deaths in the final data were

2338
01:19:59,199 --> 01:20:02,960
actually recorded in the interim data

2339
01:20:01,199 --> 01:20:05,440
for the fourth quarter by the sort of

2340
01:20:02,960 --> 01:20:07,679
broad category of cause of death

2341
01:20:05,440 --> 01:20:09,840
what we see is for most medical causes

2342
01:20:07,679 --> 01:20:12,239
of death they're relatively up to date

2343
01:20:09,840 --> 01:20:14,560
that means those lines stay near 100

2344
01:20:12,239 --> 01:20:15,840
almost out to the end of the year

2345
01:20:14,560 --> 01:20:17,840
deaths that were said to be due to

2346
01:20:15,840 --> 01:20:19,840
undetermined intent or maybe of interest

2347
01:20:17,840 --> 01:20:21,280
to this group especially accidental

2348
01:20:19,840 --> 01:20:22,960
poisonings which would include deaths

2349
01:20:21,280 --> 01:20:24,639
related to opioid overdose that was

2350
01:20:22,960 --> 01:20:26,880
considered to be accidental were

2351
01:20:24,639 --> 01:20:28,480
actually quite a bit delayed and suicide

2352
01:20:26,880 --> 01:20:30,159
or self-harm deaths are somewhere in

2353
01:20:28,480 --> 01:20:32,080
between

2354
01:20:30,159 --> 01:20:34,000
so what we see here is we certainly

2355
01:20:32,080 --> 01:20:36,159
found when we looked at those data no

2356
01:20:34,000 --> 01:20:37,840
evidence of increase we saw a potential

2357
01:20:36,159 --> 01:20:39,920
signal of decrease which we thought was

2358
01:20:37,840 --> 01:20:42,239
really not plausible when we dug into

2359
01:20:39,920 --> 01:20:43,920
the data we said we really don't have we

2360
01:20:42,239 --> 01:20:46,400
can't trust these up to the minute data

2361
01:20:43,920 --> 01:20:47,600
we need to wait and the final version of

2362
01:20:46,400 --> 01:20:48,960
the story

2363
01:20:47,600 --> 01:20:51,760
you know when this is a paper that was

2364
01:20:48,960 --> 01:20:53,440
published uh later in 2020 looking at

2365
01:20:51,760 --> 01:20:54,880
all across you know the data that were

2366
01:20:53,440 --> 01:20:57,199
available different countries you know

2367
01:20:54,880 --> 01:20:58,880
what suggests is the final story about

2368
01:20:57,199 --> 01:21:01,280
suicide mortality data during the

2369
01:20:58,880 --> 01:21:04,239
pandemic was it probably did not change

2370
01:21:01,280 --> 01:21:05,920
much at all anywhere um so going back to

2371
01:21:04,239 --> 01:21:08,159
that little picture i showed you you

2372
01:21:05,920 --> 01:21:10,480
know what we're saying here is when we

2373
01:21:08,159 --> 01:21:12,880
look at this step in the data generating

2374
01:21:10,480 --> 01:21:15,120
process how data were extracted and

2375
01:21:12,880 --> 01:21:16,880
harmonized especially those data from

2376
01:21:15,120 --> 01:21:19,120
health departments during the pandemic

2377
01:21:16,880 --> 01:21:21,440
we found that there was a sort of you

2378
01:21:19,120 --> 01:21:23,280
could say a change in the process or in

2379
01:21:21,440 --> 01:21:25,280
some ways a breakdown in the process

2380
01:21:23,280 --> 01:21:27,679
here that it was important for us to

2381
01:21:25,280 --> 01:21:30,639
recognize or we might have been misled

2382
01:21:27,679 --> 01:21:32,239
about what was really going on

2383
01:21:30,639 --> 01:21:33,920
so on to that second question the

2384
01:21:32,239 --> 01:21:35,440
effects of the pandemic on mental health

2385
01:21:33,920 --> 01:21:37,360
crises

2386
01:21:35,440 --> 01:21:40,239
this was an early report that got a fair

2387
01:21:37,360 --> 01:21:42,800
bit of attention um from the mmwr

2388
01:21:40,239 --> 01:21:44,960
looking at the proportion of emergency

2389
01:21:42,800 --> 01:21:47,440
department visits by adolescents that

2390
01:21:44,960 --> 01:21:49,199
had mental health diagnoses and this was

2391
01:21:47,440 --> 01:21:50,560
a report saying that the proportion of

2392
01:21:49,199 --> 01:21:52,400
emergency department visits with mental

2393
01:21:50,560 --> 01:21:54,159
health diagnoses had really bumped up

2394
01:21:52,400 --> 01:21:56,320
you can see this in this graph at the

2395
01:21:54,159 --> 01:21:58,080
bottom these are weeks during 2020 so

2396
01:21:56,320 --> 01:21:59,679
that's right about march april when the

2397
01:21:58,080 --> 01:22:01,440
pandemic took off

2398
01:21:59,679 --> 01:22:03,120
and a big bump in the proportion that

2399
01:22:01,440 --> 01:22:05,199
appeared to be due to mental health

2400
01:22:03,120 --> 01:22:06,159
which was sustained throughout the year

2401
01:22:05,199 --> 01:22:08,000
um

2402
01:22:06,159 --> 01:22:10,159
when we took a look at the data from our

2403
01:22:08,000 --> 01:22:12,800
health systems what we saw looking not

2404
01:22:10,159 --> 01:22:14,960
at the proportion but at the absolute uh

2405
01:22:12,800 --> 01:22:16,880
you know the the population rate the

2406
01:22:14,960 --> 01:22:19,120
population rate of mental health

2407
01:22:16,880 --> 01:22:21,679
emergency department visits we actually

2408
01:22:19,120 --> 01:22:25,280
see a big drop there you can see in

2409
01:22:21,679 --> 01:22:26,800
march april of 2020 some recovery

2410
01:22:25,280 --> 01:22:28,639
in some health systems up to the

2411
01:22:26,800 --> 01:22:30,880
previous level and some to a somewhat

2412
01:22:28,639 --> 01:22:33,120
below the previous level

2413
01:22:30,880 --> 01:22:35,520
so it seems to us there was a general

2414
01:22:33,120 --> 01:22:36,960
drop in emergency department utilization

2415
01:22:35,520 --> 01:22:39,040
there was a drop in mental health

2416
01:22:36,960 --> 01:22:40,400
utilization and the question in our

2417
01:22:39,040 --> 01:22:42,080
minds was

2418
01:22:40,400 --> 01:22:45,120
how can we say

2419
01:22:42,080 --> 01:22:48,960
whether mental health visits decreased

2420
01:22:45,120 --> 01:22:50,719
more or less than would be expected

2421
01:22:48,960 --> 01:22:53,679
and what we started to search around for

2422
01:22:50,719 --> 01:22:55,360
was what's the comparison condition

2423
01:22:53,679 --> 01:22:57,360
what's a type of visitation to the

2424
01:22:55,360 --> 01:22:59,199
emergency department that would not have

2425
01:22:57,360 --> 01:23:01,040
been affected by the pandemic that we

2426
01:22:59,199 --> 01:23:03,199
could use for sort of a difference in

2427
01:23:01,040 --> 01:23:06,000
difference design and to say what's our

2428
01:23:03,199 --> 01:23:07,440
sort of control or comparison group the

2429
01:23:06,000 --> 01:23:10,239
challenge was we really couldn't come up

2430
01:23:07,440 --> 01:23:12,080
with any when we looked we saw that uh

2431
01:23:10,239 --> 01:23:14,000
all visits to emergency departments

2432
01:23:12,080 --> 01:23:16,159
dramatically decreased during spring

2433
01:23:14,000 --> 01:23:18,320
2020 if we looked at visits for

2434
01:23:16,159 --> 01:23:19,600
accidents those dramatically decreased

2435
01:23:18,320 --> 01:23:21,120
some of that may have been people were

2436
01:23:19,600 --> 01:23:22,320
avoiding the emergency room or some of

2437
01:23:21,120 --> 01:23:24,560
that may have been people were driving

2438
01:23:22,320 --> 01:23:25,840
their cars less or going out less

2439
01:23:24,560 --> 01:23:27,440
and as many of you know even if we

2440
01:23:25,840 --> 01:23:29,920
looked at so-called serious medical

2441
01:23:27,440 --> 01:23:31,600
emergencies strokes and heart attacks

2442
01:23:29,920 --> 01:23:34,320
we don't think the pandemic probably

2443
01:23:31,600 --> 01:23:35,679
much decreased the true rate of strokes

2444
01:23:34,320 --> 01:23:37,760
and heart attacks but emergency

2445
01:23:35,679 --> 01:23:40,560
department visits decreased

2446
01:23:37,760 --> 01:23:42,560
so our conclusion was we really don't

2447
01:23:40,560 --> 01:23:44,480
think we can use data from emergency

2448
01:23:42,560 --> 01:23:47,120
department visits to say anything about

2449
01:23:44,480 --> 01:23:49,120
the true rate of mental health crises

2450
01:23:47,120 --> 01:23:50,960
that is the the health care data exhaust

2451
01:23:49,120 --> 01:23:52,159
don't really tell us about the true

2452
01:23:50,960 --> 01:23:53,920
health state

2453
01:23:52,159 --> 01:23:55,600
and this is a follow-up report from the

2454
01:23:53,920 --> 01:23:57,199
people at the cdc

2455
01:23:55,600 --> 01:23:58,560
showing essentially the same picture we

2456
01:23:57,199 --> 01:24:00,400
saw that if you look at emergency

2457
01:23:58,560 --> 01:24:03,760
department visits with various specific

2458
01:24:00,400 --> 01:24:05,679
mental health diagnoses throughout 2020

2459
01:24:03,760 --> 01:24:08,080
big drop in the spring

2460
01:24:05,679 --> 01:24:10,000
partial to full recovery but possibly

2461
01:24:08,080 --> 01:24:12,400
continuing at a somewhat lower rate

2462
01:24:10,000 --> 01:24:14,719
possibly reflecting or to my mind more

2463
01:24:12,400 --> 01:24:17,679
likely reflecting still avoidance of

2464
01:24:14,719 --> 01:24:19,520
emergency departments during the sort of

2465
01:24:17,679 --> 01:24:22,639
increases in the pandemic that happened

2466
01:24:19,520 --> 01:24:24,400
later in 2020

2467
01:24:22,639 --> 01:24:26,000
so in this case we look back here and

2468
01:24:24,400 --> 01:24:28,000
say this is where we think there was a

2469
01:24:26,000 --> 01:24:30,239
problem or an issue with our measurement

2470
01:24:28,000 --> 01:24:31,760
system that people were less likely to

2471
01:24:30,239 --> 01:24:34,159
present to the health care system

2472
01:24:31,760 --> 01:24:35,520
regardless so we really cannot use data

2473
01:24:34,159 --> 01:24:38,000
regarding emergency department

2474
01:24:35,520 --> 01:24:39,440
visitations there on the right hand side

2475
01:24:38,000 --> 01:24:41,440
to tell us anything about the true

2476
01:24:39,440 --> 01:24:43,040
health state or the actual rate of

2477
01:24:41,440 --> 01:24:44,880
mental health crises that people

2478
01:24:43,040 --> 01:24:46,560
experienced

2479
01:24:44,880 --> 01:24:47,520
the third question

2480
01:24:46,560 --> 01:24:50,400
about

2481
01:24:47,520 --> 01:24:52,000
overall rates of general psychological

2482
01:24:50,400 --> 01:24:53,280
distress because there's been a lot of

2483
01:24:52,000 --> 01:24:55,360
concern about the effects of the

2484
01:24:53,280 --> 01:24:56,800
pandemic on on sort of it's the general

2485
01:24:55,360 --> 01:24:58,719
mental health state of the american

2486
01:24:56,800 --> 01:24:59,920
public and maybe more especially in

2487
01:24:58,719 --> 01:25:01,120
teens

2488
01:24:59,920 --> 01:25:02,880
what happens if we look at various

2489
01:25:01,120 --> 01:25:05,120
indicators across our health systems

2490
01:25:02,880 --> 01:25:06,960
about mental health conditions

2491
01:25:05,120 --> 01:25:10,239
interestingly we see here rates of

2492
01:25:06,960 --> 01:25:11,760
depression diagnoses during 19 2019

2493
01:25:10,239 --> 01:25:13,760
through 2020.

2494
01:25:11,760 --> 01:25:15,199
what you see here is there is a dip

2495
01:25:13,760 --> 01:25:17,520
actually in rates of depression

2496
01:25:15,199 --> 01:25:19,440
diagnosis in spring of 2020 which

2497
01:25:17,520 --> 01:25:21,520
probably to my mind reflects more people

2498
01:25:19,440 --> 01:25:23,360
avoiding outpatient visits

2499
01:25:21,520 --> 01:25:25,440
interestingly that dip you know is

2500
01:25:23,360 --> 01:25:27,679
relatively small compared to the spike

2501
01:25:25,440 --> 01:25:29,520
at the end of 2019

2502
01:25:27,679 --> 01:25:31,440
once again here you need to be familiar

2503
01:25:29,520 --> 01:25:32,719
with sort of the the technical and to be

2504
01:25:31,440 --> 01:25:34,719
honest with you the financial

2505
01:25:32,719 --> 01:25:36,239
environment of health care which is you

2506
01:25:34,719 --> 01:25:38,320
typically see a big spike in the

2507
01:25:36,239 --> 01:25:40,159
recording of chronic condition diagnoses

2508
01:25:38,320 --> 01:25:42,000
and at the end of the year to maximize

2509
01:25:40,159 --> 01:25:44,560
medicare reimbursement we can talk more

2510
01:25:42,000 --> 01:25:46,000
about that if people want to

2511
01:25:44,560 --> 01:25:47,600
we look here at antidepressant

2512
01:25:46,000 --> 01:25:49,360
prescriptions and say could we use

2513
01:25:47,600 --> 01:25:51,920
antidepressant prescriptions as some

2514
01:25:49,360 --> 01:25:54,080
indicator we do see a blip in filling

2515
01:25:51,920 --> 01:25:56,560
antidepressant prescriptions right in

2516
01:25:54,080 --> 01:25:58,159
the spring there of 2020 but it turns

2517
01:25:56,560 --> 01:26:00,400
out what this is about when we look more

2518
01:25:58,159 --> 01:26:02,159
deeply is as you may remember in the

2519
01:26:00,400 --> 01:26:03,840
early weeks of the pandemic people were

2520
01:26:02,159 --> 01:26:05,760
concerned that pharmacies would close

2521
01:26:03,840 --> 01:26:07,440
down so people were actually instructed

2522
01:26:05,760 --> 01:26:09,920
to fill an extra month so they have a

2523
01:26:07,440 --> 01:26:11,760
supply on hand so this was sort of

2524
01:26:09,920 --> 01:26:13,520
hoarding maybe the wrong term to use

2525
01:26:11,760 --> 01:26:15,199
because that has a negative connotation

2526
01:26:13,520 --> 01:26:17,280
but this maybe have been appropriate

2527
01:26:15,199 --> 01:26:19,040
medication stockpiling and actually

2528
01:26:17,280 --> 01:26:20,480
following some fairly common advice

2529
01:26:19,040 --> 01:26:22,320
about what to do about your chronic

2530
01:26:20,480 --> 01:26:24,080
medications

2531
01:26:22,320 --> 01:26:26,320
if we look at psychotherapy visits we

2532
01:26:24,080 --> 01:26:28,480
see that those dropped

2533
01:26:26,320 --> 01:26:30,639
and had a partial recovery but the

2534
01:26:28,480 --> 01:26:32,639
question is is this really a disruption

2535
01:26:30,639 --> 01:26:34,639
in care people not getting care or is

2536
01:26:32,639 --> 01:26:36,639
this really a coding problem

2537
01:26:34,639 --> 01:26:39,040
when we dig deeper into this this is

2538
01:26:36,639 --> 01:26:41,520
data from kp washington only what we see

2539
01:26:39,040 --> 01:26:43,440
is you know a pretty dramatic decrease

2540
01:26:41,520 --> 01:26:46,320
in in-person psychotherapy visits there

2541
01:26:43,440 --> 01:26:48,159
in march april of 2020 pretty rapidly

2542
01:26:46,320 --> 01:26:49,679
replaced by telephone and video

2543
01:26:48,159 --> 01:26:52,080
encounters and actually going up to a

2544
01:26:49,679 --> 01:26:53,760
somewhat higher level

2545
01:26:52,080 --> 01:26:55,679
when we look across all the health

2546
01:26:53,760 --> 01:26:57,840
systems in our network we find that the

2547
01:26:55,679 --> 01:26:59,760
overall rates of mental health

2548
01:26:57,840 --> 01:27:02,639
utilization when we consider all visit

2549
01:26:59,760 --> 01:27:04,960
types remain fairly stable but one

2550
01:27:02,639 --> 01:27:07,040
interesting part of the story here is

2551
01:27:04,960 --> 01:27:08,800
that when we're monitoring visit rates

2552
01:27:07,040 --> 01:27:10,400
as an indicator of people's mental

2553
01:27:08,800 --> 01:27:11,520
health condition

2554
01:27:10,400 --> 01:27:12,560
it's important to recognize they're

2555
01:27:11,520 --> 01:27:13,920
important they're things we're not

2556
01:27:12,560 --> 01:27:16,400
measuring

2557
01:27:13,920 --> 01:27:17,920
if we um look take a cast a broader view

2558
01:27:16,400 --> 01:27:19,679
and say how often were these health

2559
01:27:17,920 --> 01:27:21,600
systems receiving new requests for

2560
01:27:19,679 --> 01:27:23,360
mental health services people saying i

2561
01:27:21,600 --> 01:27:25,600
need to see a doctor about mental health

2562
01:27:23,360 --> 01:27:27,600
problems i need to see a therapist

2563
01:27:25,600 --> 01:27:30,000
those actually increased quite a bit

2564
01:27:27,600 --> 01:27:32,080
during 2020 increased by 50 and in some

2565
01:27:30,000 --> 01:27:34,159
cases almost 100

2566
01:27:32,080 --> 01:27:35,280
we saw increased wait times and as many

2567
01:27:34,159 --> 01:27:36,960
of you know there's sort of been a

2568
01:27:35,280 --> 01:27:38,960
crisis in access to mental health

2569
01:27:36,960 --> 01:27:40,960
services across the country which

2570
01:27:38,960 --> 01:27:42,159
persists now even two years into the

2571
01:27:40,960 --> 01:27:45,679
pandemic

2572
01:27:42,159 --> 01:27:48,800
and my my point here is that if we had

2573
01:27:45,679 --> 01:27:51,440
been using use of mental health care as

2574
01:27:48,800 --> 01:27:53,840
a as a valid indicator of mental health

2575
01:27:51,440 --> 01:27:56,080
need what we would be ignoring is the

2576
01:27:53,840 --> 01:27:58,719
fact that the supply of mental health

2577
01:27:56,080 --> 01:28:00,400
visits is actually relatively fixed

2578
01:27:58,719 --> 01:28:02,159
even at the health system level

2579
01:28:00,400 --> 01:28:04,239
certainly say in one region of kaiser

2580
01:28:02,159 --> 01:28:05,840
permanente but even in the broader sense

2581
01:28:04,239 --> 01:28:07,679
at the national level there are only so

2582
01:28:05,840 --> 01:28:09,600
many providers they're only working so

2583
01:28:07,679 --> 01:28:11,360
many hours a week there are only so many

2584
01:28:09,600 --> 01:28:14,239
mental health visits that can possibly

2585
01:28:11,360 --> 01:28:17,440
occur so if a system is at or even above

2586
01:28:14,239 --> 01:28:19,920
capacity utilization is really not an

2587
01:28:17,440 --> 01:28:22,159
adequate measure of need or demand so

2588
01:28:19,920 --> 01:28:24,080
this is another example where

2589
01:28:22,159 --> 01:28:26,880
the presentation to the health care

2590
01:28:24,080 --> 01:28:28,960
system was a bottleneck and prevented us

2591
01:28:26,880 --> 01:28:30,880
we were just using data on visit rates

2592
01:28:28,960 --> 01:28:32,880
from observing an increase in need for

2593
01:28:30,880 --> 01:28:34,719
mental health services so we might have

2594
01:28:32,880 --> 01:28:36,719
been misled

2595
01:28:34,719 --> 01:28:38,560
so those are just three examples we may

2596
01:28:36,719 --> 01:28:41,120
have time to talk in the question period

2597
01:28:38,560 --> 01:28:43,920
more about those but you know as a as a

2598
01:28:41,120 --> 01:28:46,239
last word you know when when people try

2599
01:28:43,920 --> 01:28:47,920
to oversimplify you know you either can

2600
01:28:46,239 --> 01:28:49,600
or can't trust data that come from

2601
01:28:47,920 --> 01:28:51,199
electronic health records or insurance

2602
01:28:49,600 --> 01:28:53,280
claims to me that's that's

2603
01:28:51,199 --> 01:28:55,280
oversimplified the question is not can

2604
01:28:53,280 --> 01:28:57,840
we trust real world data the question is

2605
01:28:55,280 --> 01:28:59,600
when can we trust real world data and my

2606
01:28:57,840 --> 01:29:02,480
answer is it depends and it depends on

2607
01:28:59,600 --> 01:29:04,800
some really specific things so my advice

2608
01:29:02,480 --> 01:29:07,040
is when you're thinking about using you

2609
01:29:04,800 --> 01:29:08,159
know data from some data warehouse to to

2610
01:29:07,040 --> 01:29:10,000
actually give you an accurate

2611
01:29:08,159 --> 01:29:11,520
representation of a health state you

2612
01:29:10,000 --> 01:29:12,719
need to ask yourself

2613
01:29:11,520 --> 01:29:15,120
questions about each of those

2614
01:29:12,719 --> 01:29:17,199
intermediate steps along the way

2615
01:29:15,120 --> 01:29:19,199
so i will stop there so we can get on to

2616
01:29:17,199 --> 01:29:21,920
our other presentations and i'll look in

2617
01:29:19,199 --> 01:29:24,000
the chat for your questions thanks

2618
01:29:21,920 --> 01:29:27,040
thanks greg that was great um i have a

2619
01:29:24,000 --> 01:29:29,440
lot of questions about them for the end

2620
01:29:27,040 --> 01:29:31,840
and now i'll turn it over to dr deana

2621
01:29:29,440 --> 01:29:31,840
chisholm

2622
01:29:34,080 --> 01:29:38,880
i'm going to talk about

2623
01:29:36,960 --> 01:29:40,080
electronic medical record used in

2624
01:29:38,880 --> 01:29:42,639
research from a little different

2625
01:29:40,080 --> 01:29:45,920
perspective um i

2626
01:29:42,639 --> 01:29:49,120
do a lot of work in the space of health

2627
01:29:45,920 --> 01:29:52,800
equity and within that space think a lot

2628
01:29:49,120 --> 01:29:55,600
about the interactions between social

2629
01:29:52,800 --> 01:29:57,199
factors and health factors including

2630
01:29:55,600 --> 01:29:58,800
health services utilization and

2631
01:29:57,199 --> 01:29:59,679
including outcomes including substance

2632
01:29:58,800 --> 01:30:00,800
use

2633
01:29:59,679 --> 01:30:02,400
so what i

2634
01:30:00,800 --> 01:30:04,480
want to talk a little bit about is some

2635
01:30:02,400 --> 01:30:06,480
work that has tried to

2636
01:30:04,480 --> 01:30:09,440
look at how we get all of those pieces

2637
01:30:06,480 --> 01:30:11,120
of information inside and

2638
01:30:09,440 --> 01:30:12,719
and ehr

2639
01:30:11,120 --> 01:30:14,960
so

2640
01:30:12,719 --> 01:30:16,639
starting with that um

2641
01:30:14,960 --> 01:30:19,679
we at nationwide children's hospital

2642
01:30:16,639 --> 01:30:21,440
decided around 2018 that we needed to

2643
01:30:19,679 --> 01:30:24,400
think really hard about integrating

2644
01:30:21,440 --> 01:30:26,800
social determinants of health into our

2645
01:30:24,400 --> 01:30:28,320
electronic health record we use epic

2646
01:30:26,800 --> 01:30:30,719
which is of course one of the more

2647
01:30:28,320 --> 01:30:33,840
commonly used systems um and at the time

2648
01:30:30,719 --> 01:30:35,280
we did it epic did not have a module for

2649
01:30:33,840 --> 01:30:36,800
social determinants of health that was

2650
01:30:35,280 --> 01:30:38,400
broadly in use

2651
01:30:36,800 --> 01:30:40,639
so we designed one

2652
01:30:38,400 --> 01:30:42,880
put together a team that included

2653
01:30:40,639 --> 01:30:44,960
clinicians and social workers and

2654
01:30:42,880 --> 01:30:47,040
members of the community and researchers

2655
01:30:44,960 --> 01:30:49,040
and pulled questions out of a bunch of

2656
01:30:47,040 --> 01:30:50,639
different tools because nobody loved any

2657
01:30:49,040 --> 01:30:51,760
one tool that existed on the market at

2658
01:30:50,639 --> 01:30:54,000
the time

2659
01:30:51,760 --> 01:30:54,880
and came up with basically a four item

2660
01:30:54,000 --> 01:30:56,960
screen

2661
01:30:54,880 --> 01:30:58,960
which focused on

2662
01:30:56,960 --> 01:31:01,199
housing food

2663
01:30:58,960 --> 01:31:03,440
transportation and utilities four

2664
01:31:01,199 --> 01:31:05,760
questions one on each of those topics

2665
01:31:03,440 --> 01:31:08,239
and then if you answered yes on any of

2666
01:31:05,760 --> 01:31:10,719
those topics you got a

2667
01:31:08,239 --> 01:31:13,120
follow-on question that basically asked

2668
01:31:10,719 --> 01:31:14,880
is this a quest problem today so the

2669
01:31:13,120 --> 01:31:17,199
questions were the general questions

2670
01:31:14,880 --> 01:31:19,280
looked over 12 month period but then if

2671
01:31:17,199 --> 01:31:21,840
you said yes over the 12 month period

2672
01:31:19,280 --> 01:31:24,159
you got a detailed question that says is

2673
01:31:21,840 --> 01:31:26,159
this a problem right now and that

2674
01:31:24,159 --> 01:31:29,520
question really determined how we

2675
01:31:26,159 --> 01:31:30,880
responded to the data there was a talk

2676
01:31:29,520 --> 01:31:33,760
about you know how do you collect this

2677
01:31:30,880 --> 01:31:36,639
data this was collected uh at the time

2678
01:31:33,760 --> 01:31:39,600
primarily in uh outpatient ambulatory

2679
01:31:36,639 --> 01:31:41,840
clinic settings primary care um and done

2680
01:31:39,600 --> 01:31:43,679
on tablets

2681
01:31:41,840 --> 01:31:45,600
that either the parent or the teen could

2682
01:31:43,679 --> 01:31:47,199
do because some of these questions like

2683
01:31:45,600 --> 01:31:49,840
you know are you worried about being

2684
01:31:47,199 --> 01:31:51,840
evicted your teen may or may not know is

2685
01:31:49,840 --> 01:31:55,040
actually a rationale to have parents

2686
01:31:51,840 --> 01:31:56,400
collect some of this information

2687
01:31:55,040 --> 01:31:59,360
so

2688
01:31:56,400 --> 01:32:02,960
when you look at what the rates we were

2689
01:31:59,360 --> 01:32:06,239
getting of these uh questions were

2690
01:32:02,960 --> 01:32:08,960
we saw that about 12 percent said they

2691
01:32:06,239 --> 01:32:10,719
had within the past 12 months

2692
01:32:08,960 --> 01:32:12,639
at least one of those problems and you

2693
01:32:10,719 --> 01:32:14,080
can see that food insecurity was the

2694
01:32:12,639 --> 01:32:16,080
most commonly

2695
01:32:14,080 --> 01:32:17,679
uh espoused need

2696
01:32:16,080 --> 01:32:19,199
and then of those who said they had a

2697
01:32:17,679 --> 01:32:21,199
general need what percentage of them

2698
01:32:19,199 --> 01:32:23,040
said they had a need today

2699
01:32:21,199 --> 01:32:24,880
about 28

2700
01:32:23,040 --> 01:32:26,480
and again you can see that food was the

2701
01:32:24,880 --> 01:32:29,280
one that was the most

2702
01:32:26,480 --> 01:32:31,040
likely to have an immediate need this is

2703
01:32:29,280 --> 01:32:33,679
data from june

2704
01:32:31,040 --> 01:32:35,840
2018 to 2019. we have the

2705
01:32:33,679 --> 01:32:37,040
um more recent data but i pulled this

2706
01:32:35,840 --> 01:32:38,560
out of a paper that i'm going to share

2707
01:32:37,040 --> 01:32:42,480
some other numbers from so i want to

2708
01:32:38,560 --> 01:32:44,480
stay consistent in the period

2709
01:32:42,480 --> 01:32:47,360
so when i said the response to the

2710
01:32:44,480 --> 01:32:50,080
urgency question determined what we did

2711
01:32:47,360 --> 01:32:52,480
i say that because if they said yes to

2712
01:32:50,080 --> 01:32:54,400
an urgent to a general need so i've had

2713
01:32:52,480 --> 01:32:57,280
a need in the past 12 months they got a

2714
01:32:54,400 --> 01:32:59,360
research a resource sheet that was

2715
01:32:57,280 --> 01:33:01,600
generated directly out of the ehr

2716
01:32:59,360 --> 01:33:04,000
directly in the after visit summary

2717
01:33:01,600 --> 01:33:06,159
process so anybody who had yes checked

2718
01:33:04,000 --> 01:33:08,159
on any of those four questions got a

2719
01:33:06,159 --> 01:33:10,239
resource needs sheep printed out that

2720
01:33:08,159 --> 01:33:12,080
was specific to

2721
01:33:10,239 --> 01:33:14,400
the part of town they lived in or the

2722
01:33:12,080 --> 01:33:15,840
city that they lived in and to their

2723
01:33:14,400 --> 01:33:18,239
language

2724
01:33:15,840 --> 01:33:19,280
for the four languages that we generated

2725
01:33:18,239 --> 01:33:21,120
these in

2726
01:33:19,280 --> 01:33:23,199
we also were able to do translations for

2727
01:33:21,120 --> 01:33:26,239
other languages but we picked top four

2728
01:33:23,199 --> 01:33:27,679
which was more than 80 of our population

2729
01:33:26,239 --> 01:33:29,120
to

2730
01:33:27,679 --> 01:33:31,360
give these needs to

2731
01:33:29,120 --> 01:33:32,800
if they said they had an urgent need

2732
01:33:31,360 --> 01:33:34,880
they got the resource sheet but they

2733
01:33:32,800 --> 01:33:38,320
also got a referral to

2734
01:33:34,880 --> 01:33:39,520
social work to have somebody either talk

2735
01:33:38,320 --> 01:33:41,679
to them on the phone while they were

2736
01:33:39,520 --> 01:33:43,520
there in that visit or if they couldn't

2737
01:33:41,679 --> 01:33:44,719
talk to them on the phone it's also like

2738
01:33:43,520 --> 01:33:46,880
physically there to come and talk to

2739
01:33:44,719 --> 01:33:48,560
them physically there or to schedule a

2740
01:33:46,880 --> 01:33:50,880
time to have a follow-up call within the

2741
01:33:48,560 --> 01:33:52,239
next three days about their needs so

2742
01:33:50,880 --> 01:33:53,360
this got

2743
01:33:52,239 --> 01:33:56,400
really

2744
01:33:53,360 --> 01:33:58,719
um embedded in the clinical

2745
01:33:56,400 --> 01:34:01,040
processes but of course i'm a health

2746
01:33:58,719 --> 01:34:02,320
services researcher so as much as i was

2747
01:34:01,040 --> 01:34:04,719
excited about it being part of our

2748
01:34:02,320 --> 01:34:06,080
clinical processes i wanted to see how

2749
01:34:04,719 --> 01:34:08,320
it could help us understand what was

2750
01:34:06,080 --> 01:34:12,239
going on in our

2751
01:34:08,320 --> 01:34:14,080
population from a analytic perspective

2752
01:34:12,239 --> 01:34:16,400
so one of the papers that we did out of

2753
01:34:14,080 --> 01:34:18,320
this work was looking at social needs

2754
01:34:16,400 --> 01:34:20,080
data and looking at how it was

2755
01:34:18,320 --> 01:34:22,239
associated with

2756
01:34:20,080 --> 01:34:23,040
use of healthcare

2757
01:34:22,239 --> 01:34:24,639
so

2758
01:34:23,040 --> 01:34:26,800
over that time period that i presented

2759
01:34:24,639 --> 01:34:30,480
those numbers from at the beginning that

2760
01:34:26,800 --> 01:34:31,360
was just under 60 000 screens

2761
01:34:30,480 --> 01:34:33,679
and

2762
01:34:31,360 --> 01:34:35,440
that represented the fact that even

2763
01:34:33,679 --> 01:34:37,520
though that was a little over 12 months

2764
01:34:35,440 --> 01:34:39,120
there was a roll in so the first couple

2765
01:34:37,520 --> 01:34:41,199
of months it was just one or two clinics

2766
01:34:39,120 --> 01:34:42,719
and then three or four clinics so this

2767
01:34:41,199 --> 01:34:45,440
is a little smaller than the number of

2768
01:34:42,719 --> 01:34:48,400
patients with a period but again wanted

2769
01:34:45,440 --> 01:34:50,800
to be consistent with the data i shared

2770
01:34:48,400 --> 01:34:52,159
so we removed the patients who were out

2771
01:34:50,800 --> 01:34:53,280
of state because we figured if they

2772
01:34:52,159 --> 01:34:54,800
needed more care they were probably

2773
01:34:53,280 --> 01:34:56,400
going to get it at home

2774
01:34:54,800 --> 01:34:58,719
we removed ones where we didn't have

2775
01:34:56,400 --> 01:35:00,719
enough demographic information we

2776
01:34:58,719 --> 01:35:03,199
removed foster care because

2777
01:35:00,719 --> 01:35:05,199
it's unclear whether their social needs

2778
01:35:03,199 --> 01:35:06,880
were associated with their primary

2779
01:35:05,199 --> 01:35:09,440
parent or their foster parent and how

2780
01:35:06,880 --> 01:35:10,800
that might be changing frequently and

2781
01:35:09,440 --> 01:35:13,199
then we did

2782
01:35:10,800 --> 01:35:15,360
uh removed a few for just trying to make

2783
01:35:13,199 --> 01:35:17,520
sure we didn't incidentally identify any

2784
01:35:15,360 --> 01:35:18,480
patients based on their pair so we ended

2785
01:35:17,520 --> 01:35:20,080
up with about

2786
01:35:18,480 --> 01:35:22,639
56 000

2787
01:35:20,080 --> 01:35:24,239
youth in our analyses

2788
01:35:22,639 --> 01:35:25,760
and what we saw was if we said we

2789
01:35:24,239 --> 01:35:29,280
started with the index visit where we

2790
01:35:25,760 --> 01:35:31,760
did a screener we did that once a year

2791
01:35:29,280 --> 01:35:34,880
and we looked at their healthcare

2792
01:35:31,760 --> 01:35:37,520
utilization in the six months following

2793
01:35:34,880 --> 01:35:40,080
that index visit what we saw was the

2794
01:35:37,520 --> 01:35:42,639
youth with social needs were

2795
01:35:40,080 --> 01:35:44,480
more likely to have an ed visit in six

2796
01:35:42,639 --> 01:35:47,840
months following that visit and more

2797
01:35:44,480 --> 01:35:50,400
likely to be admitted inpatient in the

2798
01:35:47,840 --> 01:35:52,320
six months post that visit so that tells

2799
01:35:50,400 --> 01:35:54,320
us that collecting that information

2800
01:35:52,320 --> 01:35:57,360
while valuable for

2801
01:35:54,320 --> 01:35:59,280
um supporting uh

2802
01:35:57,360 --> 01:36:02,719
social assistance

2803
01:35:59,280 --> 01:36:05,600
uh is also a predictor for understanding

2804
01:36:02,719 --> 01:36:07,840
populations that need more support in

2805
01:36:05,600 --> 01:36:10,080
more ways to reduce healthcare

2806
01:36:07,840 --> 01:36:11,840
utilization that is not the most

2807
01:36:10,080 --> 01:36:14,000
efficient or effective way to provide

2808
01:36:11,840 --> 01:36:16,880
care

2809
01:36:14,000 --> 01:36:18,960
similarly when we looked at well child

2810
01:36:16,880 --> 01:36:20,880
visits what we saw was those with a

2811
01:36:18,960 --> 01:36:23,440
social need were significantly less

2812
01:36:20,880 --> 01:36:26,000
likely to

2813
01:36:23,440 --> 01:36:28,159
have a well child visit in the 12 month

2814
01:36:26,000 --> 01:36:30,000
in the six months following now since

2815
01:36:28,159 --> 01:36:31,520
most of these were done at primary care

2816
01:36:30,000 --> 01:36:33,360
visits you could say well they probably

2817
01:36:31,520 --> 01:36:34,880
didn't need another visit since this is

2818
01:36:33,360 --> 01:36:36,639
six months later

2819
01:36:34,880 --> 01:36:39,119
but we believe that these are children

2820
01:36:36,639 --> 01:36:40,560
who have needs so it shouldn't be a

2821
01:36:39,119 --> 01:36:41,600
surprise that they would come back in

2822
01:36:40,560 --> 01:36:43,920
six months

2823
01:36:41,600 --> 01:36:46,239
what we saw though is that that number

2824
01:36:43,920 --> 01:36:47,840
is lower which means we need to be doing

2825
01:36:46,239 --> 01:36:49,840
continuing to do a better job of

2826
01:36:47,840 --> 01:36:51,760
identifying needs at those index visits

2827
01:36:49,840 --> 01:36:54,400
because we're going to be less likely to

2828
01:36:51,760 --> 01:36:56,320
see them quickly

2829
01:36:54,400 --> 01:36:57,600
so this helped us to start and think

2830
01:36:56,320 --> 01:36:58,800
about

2831
01:36:57,600 --> 01:37:01,440
why

2832
01:36:58,800 --> 01:37:03,520
embedding social factors in the

2833
01:37:01,440 --> 01:37:06,320
electronic health record with the

2834
01:37:03,520 --> 01:37:08,000
healthcare utilization data has value it

2835
01:37:06,320 --> 01:37:11,600
has value for

2836
01:37:08,000 --> 01:37:13,760
care delivery and it has value for

2837
01:37:11,600 --> 01:37:16,080
intervention development and it has

2838
01:37:13,760 --> 01:37:19,199
value for predictive health services

2839
01:37:16,080 --> 01:37:22,320
research type stuff

2840
01:37:19,199 --> 01:37:25,040
now we have since uh

2841
01:37:22,320 --> 01:37:28,320
switched to using the epic foundation

2842
01:37:25,040 --> 01:37:31,360
screening questions uh these are

2843
01:37:28,320 --> 01:37:32,960
um embedded in the standard epic module

2844
01:37:31,360 --> 01:37:34,639
they're different than our questions we

2845
01:37:32,960 --> 01:37:36,159
don't necessarily love them as much as

2846
01:37:34,639 --> 01:37:39,280
we loved our questions that's because we

2847
01:37:36,159 --> 01:37:42,400
didn't lovingly create them ourselves um

2848
01:37:39,280 --> 01:37:44,400
the value of them though is that many

2849
01:37:42,400 --> 01:37:46,800
organizations that are doing this type

2850
01:37:44,400 --> 01:37:48,719
of screening are going to be using these

2851
01:37:46,800 --> 01:37:50,880
specific sets of questions so now we get

2852
01:37:48,719 --> 01:37:53,520
a chance to not just be stuck doing

2853
01:37:50,880 --> 01:37:55,199
single site emr studies

2854
01:37:53,520 --> 01:37:58,239
but really being able to look at

2855
01:37:55,199 --> 01:38:00,480
harmonization across data that will

2856
01:37:58,239 --> 01:38:02,960
allow us to do

2857
01:38:00,480 --> 01:38:06,239
larger studies that will allow us to

2858
01:38:02,960 --> 01:38:09,159
either identify rarer behaviors

2859
01:38:06,239 --> 01:38:11,600
or to be able to do better analysis in

2860
01:38:09,159 --> 01:38:13,840
subpopulations that we don't have large

2861
01:38:11,600 --> 01:38:15,920
numbers of in any single site which is

2862
01:38:13,840 --> 01:38:18,480
really important to an equity researcher

2863
01:38:15,920 --> 01:38:20,960
who hates the fact that we frequently

2864
01:38:18,480 --> 01:38:23,280
can't look at specific racial ethnic

2865
01:38:20,960 --> 01:38:24,320
populations because no one site has a

2866
01:38:23,280 --> 01:38:25,920
sufficient

2867
01:38:24,320 --> 01:38:27,280
sample size

2868
01:38:25,920 --> 01:38:31,199
another thing that's interesting about

2869
01:38:27,280 --> 01:38:34,239
this is that this built into epic

2870
01:38:31,199 --> 01:38:36,639
um and every time a clinician opens the

2871
01:38:34,239 --> 01:38:38,960
landing page for a patient their social

2872
01:38:36,639 --> 01:38:41,199
determinants data appears

2873
01:38:38,960 --> 01:38:44,400
and they see this big circle that says

2874
01:38:41,199 --> 01:38:46,960
what their social needs are so if a

2875
01:38:44,400 --> 01:38:48,480
child has serious social needs lots of

2876
01:38:46,960 --> 01:38:49,920
social needs are going to see a big red

2877
01:38:48,480 --> 01:38:51,360
circle

2878
01:38:49,920 --> 01:38:52,719
what does that mean

2879
01:38:51,360 --> 01:38:54,719
does that mean the doc's going to spend

2880
01:38:52,719 --> 01:38:56,159
some extra time with that patient does

2881
01:38:54,719 --> 01:38:57,920
that mean they're going to make sure

2882
01:38:56,159 --> 01:38:59,920
there's a social worker available when

2883
01:38:57,920 --> 01:39:01,920
that patient comes in doesn't mean

2884
01:38:59,920 --> 01:39:04,239
they're going to

2885
01:39:01,920 --> 01:39:06,000
sigh and be annoyed that this is going

2886
01:39:04,239 --> 01:39:08,000
to be a longer visit than they want are

2887
01:39:06,000 --> 01:39:09,600
they going to start with negative

2888
01:39:08,000 --> 01:39:12,159
expectations about this patient and

2889
01:39:09,600 --> 01:39:14,320
their needs there's two sides to that

2890
01:39:12,159 --> 01:39:16,480
coin and it's a researchable question

2891
01:39:14,320 --> 01:39:18,159
that um we'll certainly need to look at

2892
01:39:16,480 --> 01:39:19,679
in the future

2893
01:39:18,159 --> 01:39:22,239
so i want to talk real quickly about

2894
01:39:19,679 --> 01:39:23,679
another social determinant that is less

2895
01:39:22,239 --> 01:39:25,600
likely to be asked about you know those

2896
01:39:23,679 --> 01:39:28,000
for transportation housing food

2897
01:39:25,600 --> 01:39:30,800
utilities are pretty standard types of

2898
01:39:28,000 --> 01:39:30,800
social determinants

2899
01:39:31,040 --> 01:39:36,560
the

2900
01:39:32,639 --> 01:39:38,480
issue of mass incarceration of high

2901
01:39:36,560 --> 01:39:40,320
levels of justice involvement in

2902
01:39:38,480 --> 01:39:42,239
children particularly children in racial

2903
01:39:40,320 --> 01:39:43,920
ethnic minority populations

2904
01:39:42,239 --> 01:39:45,760
has

2905
01:39:43,920 --> 01:39:47,360
a meaningful impact on health but it's

2906
01:39:45,760 --> 01:39:49,840
really hard to study because nobody has

2907
01:39:47,360 --> 01:39:52,639
the numbers so what we wanted to see is

2908
01:39:49,840 --> 01:39:54,400
can we figure this out using an emr

2909
01:39:52,639 --> 01:39:56,000
so what we wanted to know was how many

2910
01:39:54,400 --> 01:39:58,080
children seen in our institution have a

2911
01:39:56,000 --> 01:40:00,320
history of an incarcerated parent is

2912
01:39:58,080 --> 01:40:02,480
that number changing over time

2913
01:40:00,320 --> 01:40:04,320
and what health challenges do these

2914
01:40:02,480 --> 01:40:06,840
youth face

2915
01:40:04,320 --> 01:40:09,520
so what we did

2916
01:40:06,840 --> 01:40:12,239
was use the

2917
01:40:09,520 --> 01:40:13,840
emr but not using the coded fields that

2918
01:40:12,239 --> 01:40:15,760
we created because we hadn't decided

2919
01:40:13,840 --> 01:40:18,239
that that was one of our questions but

2920
01:40:15,760 --> 01:40:21,360
using natural language processing

2921
01:40:18,239 --> 01:40:25,360
through the clinical notes to identify

2922
01:40:21,360 --> 01:40:27,040
key words that might be associated with

2923
01:40:25,360 --> 01:40:29,199
experiences with the justice system

2924
01:40:27,040 --> 01:40:32,239
whether self or whether an incarcerated

2925
01:40:29,199 --> 01:40:35,440
parent or a guardian

2926
01:40:32,239 --> 01:40:37,679
so we developed a set of keywords and

2927
01:40:35,440 --> 01:40:41,199
used a natural language processor to run

2928
01:40:37,679 --> 01:40:43,360
through our epic system between 2006 and

2929
01:40:41,199 --> 01:40:46,159
2020

2930
01:40:43,360 --> 01:40:49,679
and what you can see is just looking at

2931
01:40:46,159 --> 01:40:51,920
the probability of a word associated

2932
01:40:49,679 --> 01:40:55,600
with incarceration showing up in our

2933
01:40:51,920 --> 01:40:57,840
records it has shot up dramatically um

2934
01:40:55,600 --> 01:40:59,360
over the past 15 years

2935
01:40:57,840 --> 01:41:01,360
so you know that's something that we

2936
01:40:59,360 --> 01:41:02,719
wouldn't have otherwise known

2937
01:41:01,360 --> 01:41:04,159
maybe there might be some anecdotal

2938
01:41:02,719 --> 01:41:06,239
thoughts but this is some data that says

2939
01:41:04,159 --> 01:41:08,000
hey there's really something going on

2940
01:41:06,239 --> 01:41:09,600
here does that mean the docs are more

2941
01:41:08,000 --> 01:41:10,800
likely to be talking about that today

2942
01:41:09,600 --> 01:41:13,679
than they were

2943
01:41:10,800 --> 01:41:16,480
10 15 years ago maybe so we can't really

2944
01:41:13,679 --> 01:41:18,159
tease that out but we can tease out that

2945
01:41:16,480 --> 01:41:20,719
there's more than a little bit of it

2946
01:41:18,159 --> 01:41:20,719
going on

2947
01:41:20,880 --> 01:41:27,840
when we linked that data from the nlp

2948
01:41:23,760 --> 01:41:30,239
to the data in the medical record

2949
01:41:27,840 --> 01:41:32,639
for utilization what we saw kind of

2950
01:41:30,239 --> 01:41:34,239
dropped our jaws a little bit

2951
01:41:32,639 --> 01:41:35,199
and i know this might be a little small

2952
01:41:34,239 --> 01:41:37,440
to see

2953
01:41:35,199 --> 01:41:40,080
but when you look at for example

2954
01:41:37,440 --> 01:41:41,679
youth with alcohol disorder related

2955
01:41:40,080 --> 01:41:43,840
diagnoses

2956
01:41:41,679 --> 01:41:46,480
within the

2957
01:41:43,840 --> 01:41:50,400
children who had a correctional word in

2958
01:41:46,480 --> 01:41:52,080
their ehr about 0.37 percent had

2959
01:41:50,400 --> 01:41:55,920
alcohol disorder

2960
01:41:52,080 --> 01:41:58,800
across all the patients in that

2961
01:41:55,920 --> 01:42:00,639
set of records it was about 0.02

2962
01:41:58,800 --> 01:42:02,560
and to understand what that means is

2963
01:42:00,639 --> 01:42:04,639
that 43

2964
01:42:02,560 --> 01:42:06,800
of the children over that time period

2965
01:42:04,639 --> 01:42:08,159
that had an alcohol disorder related

2966
01:42:06,800 --> 01:42:10,719
diagnosis

2967
01:42:08,159 --> 01:42:11,760
in their record had a correctional

2968
01:42:10,719 --> 01:42:12,840
keyword

2969
01:42:11,760 --> 01:42:15,840
in their

2970
01:42:12,840 --> 01:42:18,000
diagnosis so that 42 if you look at

2971
01:42:15,840 --> 01:42:20,880
cannabis 66

2972
01:42:18,000 --> 01:42:22,960
if you look at broader substance abuse

2973
01:42:20,880 --> 01:42:26,239
diagnosis 54

2974
01:42:22,960 --> 01:42:27,440
so this small group of

2975
01:42:26,239 --> 01:42:29,679
children

2976
01:42:27,440 --> 01:42:32,000
making up a huge percentage of our youth

2977
01:42:29,679 --> 01:42:33,440
who had substance abuse

2978
01:42:32,000 --> 01:42:36,000
issues

2979
01:42:33,440 --> 01:42:37,679
um and i haven't highlighted and read

2980
01:42:36,000 --> 01:42:39,360
post-traumatic stress because seeing

2981
01:42:37,679 --> 01:42:41,040
your parent arrested having a parent

2982
01:42:39,360 --> 01:42:43,040
taken out of your house being arrested

2983
01:42:41,040 --> 01:42:44,239
yourself is a traumatic experience

2984
01:42:43,040 --> 01:42:46,159
um so

2985
01:42:44,239 --> 01:42:48,480
not surprisingly

2986
01:42:46,159 --> 01:42:50,480
um you know almost 50 percent of our

2987
01:42:48,480 --> 01:42:53,520
youth with post-traumatic stress

2988
01:42:50,480 --> 01:42:56,719
diagnosis were also um had a

2989
01:42:53,520 --> 01:42:58,239
correctional keyword in their record

2990
01:42:56,719 --> 01:42:59,600
now you might say well what does that

2991
01:42:58,239 --> 01:43:01,440
mean correctional keyword probably

2992
01:42:59,600 --> 01:43:03,040
doesn't mean anything at all

2993
01:43:01,440 --> 01:43:05,920
we did a thousand records we went

2994
01:43:03,040 --> 01:43:09,119
through them read them by hand

2995
01:43:05,920 --> 01:43:11,440
and saw that the misclassification

2996
01:43:09,119 --> 01:43:15,199
was around 17

2997
01:43:11,440 --> 01:43:17,280
so of that 1000 about 17

2998
01:43:15,199 --> 01:43:18,880
about 173

2999
01:43:17,280 --> 01:43:20,800
did not have any evidence so these are

3000
01:43:18,880 --> 01:43:22,239
people who might have said something

3001
01:43:20,800 --> 01:43:25,600
like they were afraid of being

3002
01:43:22,239 --> 01:43:27,840
incarcerated a few of them were

3003
01:43:25,600 --> 01:43:29,760
like a car incarcerated appendix or

3004
01:43:27,840 --> 01:43:31,280
something like that so

3005
01:43:29,760 --> 01:43:33,520
those got

3006
01:43:31,280 --> 01:43:35,920
pulled out but even just using keywords

3007
01:43:33,520 --> 01:43:38,800
we were able to have

3008
01:43:35,920 --> 01:43:41,119
about 83

3009
01:43:38,800 --> 01:43:43,760
or 87 rather

3010
01:43:41,119 --> 01:43:45,119
validity we took the next level

3011
01:43:43,760 --> 01:43:48,400
and we took

3012
01:43:45,119 --> 01:43:50,800
7 500 records and did all of them coded

3013
01:43:48,400 --> 01:43:53,040
out by reading them by hand and used

3014
01:43:50,800 --> 01:43:56,320
those to train a

3015
01:43:53,040 --> 01:43:59,440
uh machine learning predictive model uh

3016
01:43:56,320 --> 01:44:02,400
to identify um and what you can see

3017
01:43:59,440 --> 01:44:03,360
on here is that we went from our model

3018
01:44:02,400 --> 01:44:06,400
having

3019
01:44:03,360 --> 01:44:08,480
this level of predictiveness using this

3020
01:44:06,400 --> 01:44:10,880
approach similar to an roc curve i won't

3021
01:44:08,480 --> 01:44:12,880
go into a bunch of detail about it uh

3022
01:44:10,880 --> 01:44:16,000
but moving up to

3023
01:44:12,880 --> 01:44:18,639
a really high level of predictive value

3024
01:44:16,000 --> 01:44:20,800
so this is a way without even training

3025
01:44:18,639 --> 01:44:22,239
everybody on doing something new

3026
01:44:20,800 --> 01:44:23,440
without adding questions and going

3027
01:44:22,239 --> 01:44:24,960
through the whole process that we go

3028
01:44:23,440 --> 01:44:27,119
through to add questions to say we can

3029
01:44:24,960 --> 01:44:28,560
use the data that's already there in a

3030
01:44:27,119 --> 01:44:31,440
way that predicts

3031
01:44:28,560 --> 01:44:33,360
a outcome that predicts health

3032
01:44:31,440 --> 01:44:34,719
and helps us to identify people who we

3033
01:44:33,360 --> 01:44:37,199
would want to

3034
01:44:34,719 --> 01:44:39,760
catch early and intervene with before it

3035
01:44:37,199 --> 01:44:42,159
gets to that point

3036
01:44:39,760 --> 01:44:44,159
so i'd say that the

3037
01:44:42,159 --> 01:44:46,880
electronic medical record has to be part

3038
01:44:44,159 --> 01:44:49,600
of how we think about not just

3039
01:44:46,880 --> 01:44:51,840
outcomes research or

3040
01:44:49,600 --> 01:44:53,520
utilization research but really the

3041
01:44:51,840 --> 01:44:55,119
kinds of social research that allow us

3042
01:44:53,520 --> 01:44:56,560
to think about health at a population

3043
01:44:55,119 --> 01:44:58,560
level

3044
01:44:56,560 --> 01:45:00,719
ehrs are increasingly being used to

3045
01:44:58,560 --> 01:45:03,119
capture social determinants of health we

3046
01:45:00,719 --> 01:45:05,920
believe that that data can be used to

3047
01:45:03,119 --> 01:45:08,480
guide care and to study predictors and

3048
01:45:05,920 --> 01:45:10,560
outcomes and that even when that data is

3049
01:45:08,480 --> 01:45:12,639
not collected in coded fields natural

3050
01:45:10,560 --> 01:45:15,280
language processing can help to find

3051
01:45:12,639 --> 01:45:16,560
information that matters in text-based

3052
01:45:15,280 --> 01:45:19,040
notes

3053
01:45:16,560 --> 01:45:21,679
so i will close with that just want to

3054
01:45:19,040 --> 01:45:22,719
thank an amazing group of colleagues who

3055
01:45:21,679 --> 01:45:26,000
have been

3056
01:45:22,719 --> 01:45:28,320
uh part of this work um specifically

3057
01:45:26,000 --> 01:45:30,320
shout out to samantha buck the paper

3058
01:45:28,320 --> 01:45:33,760
that we that i showed at the end with

3059
01:45:30,320 --> 01:45:34,560
the roc curve was published today

3060
01:45:33,760 --> 01:45:36,320
and

3061
01:45:34,560 --> 01:45:38,400
she's the first author on that a former

3062
01:45:36,320 --> 01:45:41,440
postdoc at nationwide children's now at

3063
01:45:38,400 --> 01:45:43,520
university cinematic um and also to rose

3064
01:45:41,440 --> 01:45:45,840
hardy a former postdoc who was the first

3065
01:45:43,520 --> 01:45:49,440
author on the utilization paper

3066
01:45:45,840 --> 01:45:49,440
so i will stop there thank you

3067
01:45:49,920 --> 01:45:53,440
great thank you and congratulations on

3068
01:45:51,920 --> 01:45:56,760
that publication i look forward to

3069
01:45:53,440 --> 01:45:56,760
reading it

3070
01:45:58,400 --> 01:46:02,880
oop dr cece sperling will be our next

3071
01:46:01,440 --> 01:46:05,360
speaker

3072
01:46:02,880 --> 01:46:05,360
i'm sorry

3073
01:46:05,840 --> 01:46:10,080
great okay

3074
01:46:07,440 --> 01:46:12,400
uh well thank you and i'm gonna share

3075
01:46:10,080 --> 01:46:15,280
some of the you know a few examples

3076
01:46:12,400 --> 01:46:18,159
of how we have used the ehr at kaiser

3077
01:46:15,280 --> 01:46:21,040
permanente northern california to

3078
01:46:18,159 --> 01:46:23,360
create tools to help further the

3079
01:46:21,040 --> 01:46:26,159
integration of prevention and early

3080
01:46:23,360 --> 01:46:28,560
intervention efforts and then some of

3081
01:46:26,159 --> 01:46:31,199
the ways that we have been able to study

3082
01:46:28,560 --> 01:46:31,199
these efforts

3083
01:46:32,080 --> 01:46:35,840
so a little bit of context this is

3084
01:46:33,920 --> 01:46:37,920
kaiser permanente northern california

3085
01:46:35,840 --> 01:46:41,840
and we're a large integrated healthcare

3086
01:46:37,920 --> 01:46:44,480
delivery system we have many members

3087
01:46:41,840 --> 01:46:47,360
4.5 million about a third of the

3088
01:46:44,480 --> 01:46:49,600
northern california region's population

3089
01:46:47,360 --> 01:46:52,400
it's a very diverse membership

3090
01:46:49,600 --> 01:46:54,960
in many different ways

3091
01:46:52,400 --> 01:46:57,199
it's an integrated system so we have

3092
01:46:54,960 --> 01:46:59,199
carved in mental health psychiatry and

3093
01:46:57,199 --> 01:47:02,719
specialty addiction medicine

3094
01:46:59,199 --> 01:47:03,920
treatment programs um and i think uh

3095
01:47:02,719 --> 01:47:06,639
important

3096
01:47:03,920 --> 01:47:10,960
to our conversation today a very mature

3097
01:47:06,639 --> 01:47:12,880
robust ehr and embedded research with

3098
01:47:10,960 --> 01:47:15,600
robust database

3099
01:47:12,880 --> 01:47:15,600
infrastructure

3100
01:47:17,119 --> 01:47:20,320
so the first project that i'm going to

3101
01:47:18,880 --> 01:47:22,320
be talking about today is one where

3102
01:47:20,320 --> 01:47:24,480
we've used machine learning and

3103
01:47:22,320 --> 01:47:26,719
electronic health record data to develop

3104
01:47:24,480 --> 01:47:30,320
a predictive model of adolescent

3105
01:47:26,719 --> 01:47:30,320
substance use development

3106
01:47:30,639 --> 01:47:34,800
and this was a study funded by the

3107
01:47:32,400 --> 01:47:37,280
conrad and hilton foundations adolescent

3108
01:47:34,800 --> 01:47:39,760
early intervention initiative and our

3109
01:47:37,280 --> 01:47:41,760
aim was to develop and validate a

3110
01:47:39,760 --> 01:47:44,080
predictive model of adolescent substance

3111
01:47:41,760 --> 01:47:47,360
use problem development and we

3112
01:47:44,080 --> 01:47:48,320
collaborated with teams from three other

3113
01:47:47,360 --> 01:47:50,719
very

3114
01:47:48,320 --> 01:47:52,960
geographically diverse health care

3115
01:47:50,719 --> 01:47:54,719
systems you see those here kaiser

3116
01:47:52,960 --> 01:47:56,960
permanente hawaii

3117
01:47:54,719 --> 01:47:58,800
kaiser permanente i am

3118
01:47:56,960 --> 01:48:01,440
northern california of course geisinger

3119
01:47:58,800 --> 01:48:03,679
health system in rural pennsylvania and

3120
01:48:01,440 --> 01:48:05,280
henry ford health system in metropolitan

3121
01:48:03,679 --> 01:48:08,400
detroit

3122
01:48:05,280 --> 01:48:10,000
and we identified over 40 000 birth

3123
01:48:08,400 --> 01:48:13,199
cohorts where we had both the

3124
01:48:10,000 --> 01:48:15,679
adolescents and their mothers data and

3125
01:48:13,199 --> 01:48:18,679
who had continuous membership since

3126
01:48:15,679 --> 01:48:18,679
birth

3127
01:48:19,520 --> 01:48:24,000
and working with a variety of clinical

3128
01:48:22,000 --> 01:48:25,840
experts in our systems such as

3129
01:48:24,000 --> 01:48:28,239
developmental pediatricians and

3130
01:48:25,840 --> 01:48:30,880
psychologists we identified a large

3131
01:48:28,239 --> 01:48:33,840
group of potential candidate variables

3132
01:48:30,880 --> 01:48:36,480
to look at

3133
01:48:33,840 --> 01:48:39,360
we used cox regression models for

3134
01:48:36,480 --> 01:48:41,520
survival analyses and we built the

3135
01:48:39,360 --> 01:48:44,480
models using kaiser permanente northern

3136
01:48:41,520 --> 01:48:47,119
california data and then evaluated them

3137
01:48:44,480 --> 01:48:49,520
using data from the other health systems

3138
01:48:47,119 --> 01:48:52,239
and we created two sets of models a

3139
01:48:49,520 --> 01:48:54,159
baseline model that examined factors

3140
01:48:52,239 --> 01:48:55,760
occurring in childhood prior to

3141
01:48:54,159 --> 01:48:56,880
adolescent

3142
01:48:55,760 --> 01:48:59,440
and then

3143
01:48:56,880 --> 01:49:02,000
maternal factors occurring before

3144
01:48:59,440 --> 01:49:04,239
age 12 and then we created models that

3145
01:49:02,000 --> 01:49:07,280
really took into account time varying

3146
01:49:04,239 --> 01:49:09,840
predictors occurring between 12 and 18

3147
01:49:07,280 --> 01:49:11,199
because of course exposures that may

3148
01:49:09,840 --> 01:49:12,560
influence

3149
01:49:11,199 --> 01:49:14,960
the development of substance use

3150
01:49:12,560 --> 01:49:18,400
problems you know continue to occur

3151
01:49:14,960 --> 01:49:18,400
occur throughout adelaide

3152
01:49:20,000 --> 01:49:25,360
so we created the baseline model here

3153
01:49:22,320 --> 01:49:28,080
you can see some of the child and mother

3154
01:49:25,360 --> 01:49:30,880
level factors occurring before age 12

3155
01:49:28,080 --> 01:49:34,800
that we found to increase the risk of

3156
01:49:30,880 --> 01:49:34,800
developing a substance use problem

3157
01:49:35,119 --> 01:49:39,599
and then informed by

3158
01:49:37,440 --> 01:49:41,840
that baseline

3159
01:49:39,599 --> 01:49:43,760
model we developed time varying models

3160
01:49:41,840 --> 01:49:46,000
examining predictors significant in the

3161
01:49:43,760 --> 01:49:48,560
baseline model and also factors

3162
01:49:46,000 --> 01:49:51,199
occurring throughout adolescence and the

3163
01:49:48,560 --> 01:49:53,679
scope and duration of the ehr data we

3164
01:49:51,199 --> 01:49:56,239
had from these birth cohorts really

3165
01:49:53,679 --> 01:49:59,679
allowed us to examine the chronology of

3166
01:49:56,239 --> 01:50:02,159
risk exposure and problem development so

3167
01:49:59,679 --> 01:50:05,760
we looked at having the impact of each

3168
01:50:02,159 --> 01:50:08,239
of these risk factors prior to age 12

3169
01:50:05,760 --> 01:50:09,599
after age 12 and then an incident

3170
01:50:08,239 --> 01:50:12,159
diagnosis

3171
01:50:09,599 --> 01:50:16,320
only in the three months prior to

3172
01:50:12,159 --> 01:50:16,320
their substance use problem diagnosis

3173
01:50:17,760 --> 01:50:23,760
and we created categories of exposures

3174
01:50:21,119 --> 01:50:26,880
based on when they occurred and whether

3175
01:50:23,760 --> 01:50:28,719
they persisted over time so early

3176
01:50:26,880 --> 01:50:29,760
early persistent

3177
01:50:28,719 --> 01:50:32,320
late

3178
01:50:29,760 --> 01:50:34,800
early persistent and recent and late and

3179
01:50:32,320 --> 01:50:34,800
recent

3180
01:50:36,400 --> 01:50:40,639
and we identified a number of

3181
01:50:38,320 --> 01:50:42,239
significant predictors at both the child

3182
01:50:40,639 --> 01:50:44,239
and maternal level

3183
01:50:42,239 --> 01:50:45,119
um looking at these

3184
01:50:44,239 --> 01:50:47,199
as

3185
01:50:45,119 --> 01:50:48,159
researchers or clinicians in this field

3186
01:50:47,199 --> 01:50:50,639
they

3187
01:50:48,159 --> 01:50:53,280
may seem intuitive but it's been nice to

3188
01:50:50,639 --> 01:50:56,080
be able to add some empirical support to

3189
01:50:53,280 --> 01:50:58,800
what we see clinically and anecdotally

3190
01:50:56,080 --> 01:51:01,360
and hopefully provide some evidence that

3191
01:50:58,800 --> 01:51:05,400
can be useful for clinicians working on

3192
01:51:01,360 --> 01:51:05,400
the pediatric front lines

3193
01:51:06,159 --> 01:51:11,360
and we've now taken these findings from

3194
01:51:09,040 --> 01:51:13,280
this predictive modeling work and

3195
01:51:11,360 --> 01:51:15,840
working in collaboration with the team

3196
01:51:13,280 --> 01:51:18,239
from the partnership to end addiction we

3197
01:51:15,840 --> 01:51:20,800
conducted focus groups around the

3198
01:51:18,239 --> 01:51:22,960
country with groups of parents and

3199
01:51:20,800 --> 01:51:26,320
groups of pediatricians to really get

3200
01:51:22,960 --> 01:51:27,920
their perspectives on how our findings

3201
01:51:26,320 --> 01:51:30,800
from the predictive modeling work could

3202
01:51:27,920 --> 01:51:32,400
best help them with their prevention and

3203
01:51:30,800 --> 01:51:34,800
early intervention

3204
01:51:32,400 --> 01:51:38,000
with the teens that they work with or

3205
01:51:34,800 --> 01:51:38,000
that their children

3206
01:51:38,320 --> 01:51:44,320
um this work is ongoing and we're in the

3207
01:51:41,920 --> 01:51:46,480
process of dissemination and developing

3208
01:51:44,320 --> 01:51:49,760
additional tools but one of the products

3209
01:51:46,480 --> 01:51:52,719
so far has been the creation of a free

3210
01:51:49,760 --> 01:51:55,199
accessible online substance use problem

3211
01:51:52,719 --> 01:51:58,880
risk assessment tool on the partnership

3212
01:51:55,199 --> 01:52:00,960
to win addiction website

3213
01:51:58,880 --> 01:52:03,040
and the questions in the tool are

3214
01:52:00,960 --> 01:52:07,679
informed by the predictive modeling

3215
01:52:03,040 --> 01:52:07,679
study and the larger literature

3216
01:52:08,320 --> 01:52:12,880
and they're really intended to provide

3217
01:52:10,320 --> 01:52:15,040
some additional empirically supported

3218
01:52:12,880 --> 01:52:17,440
information to help parents and

3219
01:52:15,040 --> 01:52:19,440
pediatricians and other clinicians who

3220
01:52:17,440 --> 01:52:21,520
are concerned and wanting to figure out

3221
01:52:19,440 --> 01:52:24,239
if their child might be at heightened

3222
01:52:21,520 --> 01:52:26,800
risk of developing a problem

3223
01:52:24,239 --> 01:52:30,320
and this is the kind of tool that can be

3224
01:52:26,800 --> 01:52:32,000
employed in an ehr again and so for

3225
01:52:30,320 --> 01:52:34,480
example pediatricians and other

3226
01:52:32,000 --> 01:52:37,360
clinicians could embed

3227
01:52:34,480 --> 01:52:39,119
um this you know the link to this

3228
01:52:37,360 --> 01:52:41,920
risk assessment tool in their

3229
01:52:39,119 --> 01:52:45,040
patient-facing pages or push it out to

3230
01:52:41,920 --> 01:52:47,599
families at visit time or in after visit

3231
01:52:45,040 --> 01:52:47,599
summaries

3232
01:52:49,599 --> 01:52:53,760
so the second set of projects i'm going

3233
01:52:51,119 --> 01:52:55,840
to show are based on our adult and

3234
01:52:53,760 --> 01:53:00,159
adolescent screening brief intervention

3235
01:52:55,840 --> 01:53:00,159
and referral to treatment or expert work

3236
01:53:01,280 --> 01:53:05,199
so we have in kaiser permanente northern

3237
01:53:03,520 --> 01:53:06,800
california we

3238
01:53:05,199 --> 01:53:09,840
have implemented

3239
01:53:06,800 --> 01:53:11,440
a comprehensive alcohol expert

3240
01:53:09,840 --> 01:53:14,719
initiative

3241
01:53:11,440 --> 01:53:17,040
in adult primary care and this was

3242
01:53:14,719 --> 01:53:20,080
based on a cluster randomized

3243
01:53:17,040 --> 01:53:21,840
implementation trial that used the ehr a

3244
01:53:20,080 --> 01:53:24,080
completely pragmatic trial where we

3245
01:53:21,840 --> 01:53:24,880
looked at outcomes

3246
01:53:24,080 --> 01:53:27,920
in

3247
01:53:24,880 --> 01:53:28,719
using ehr data and that was a study of

3248
01:53:27,920 --> 01:53:31,119
the

3249
01:53:28,719 --> 01:53:33,199
best most effective modalities for

3250
01:53:31,119 --> 01:53:35,760
delivering espert

3251
01:53:33,199 --> 01:53:38,480
and the health system took the findings

3252
01:53:35,760 --> 01:53:40,320
from that trial and implemented a

3253
01:53:38,480 --> 01:53:42,960
region-wide

3254
01:53:40,320 --> 01:53:44,080
alcohol expert program in adult primary

3255
01:53:42,960 --> 01:53:46,000
care

3256
01:53:44,080 --> 01:53:48,000
where medical assistants do the

3257
01:53:46,000 --> 01:53:51,599
screening for unhealthy alcohol use as

3258
01:53:48,000 --> 01:53:54,480
part of the rooming process alongside

3259
01:53:51,599 --> 01:53:57,040
blood pressure tobacco and exercise and

3260
01:53:54,480 --> 01:53:59,199
then primary care physicians deliver the

3261
01:53:57,040 --> 01:54:01,440
brief advice or brief intervention and

3262
01:53:59,199 --> 01:54:03,840
referral to specialty treatment as

3263
01:54:01,440 --> 01:54:03,840
needed

3264
01:54:04,560 --> 01:54:10,080
and this has been a very successful

3265
01:54:06,719 --> 01:54:11,599
program since its inception in 2013

3266
01:54:10,080 --> 01:54:14,880
we've conducted

3267
01:54:11,599 --> 01:54:15,840
over 15 million total alcohol screenings

3268
01:54:14,880 --> 01:54:19,280
of

3269
01:54:15,840 --> 01:54:22,080
over 4.5 million unique adults we have

3270
01:54:19,280 --> 01:54:23,590
very high screening rate

3271
01:54:22,080 --> 01:54:24,719
and a very

3272
01:54:23,590 --> 01:54:26,880
[Music]

3273
01:54:24,719 --> 01:54:29,440
you know healthy brief intervention rate

3274
01:54:26,880 --> 01:54:32,639
among those people who screen positive

3275
01:54:29,440 --> 01:54:32,639
for unhealthy drinking

3276
01:54:33,920 --> 01:54:38,159
so these next few slides are screen

3277
01:54:36,080 --> 01:54:40,880
shots of some of the decision support

3278
01:54:38,159 --> 01:54:41,840
tools that we were able to get embedded

3279
01:54:40,880 --> 01:54:44,960
into

3280
01:54:41,840 --> 01:54:46,719
the ehr so this one here is what pops up

3281
01:54:44,960 --> 01:54:49,920
when the medical assistant is rooming

3282
01:54:46,719 --> 01:54:53,040
the patient to remind them to ask the

3283
01:54:49,920 --> 01:54:53,040
screening questions

3284
01:54:54,080 --> 01:54:59,679
the these are the screening questions

3285
01:54:56,000 --> 01:55:01,840
these are a modified version of the nia

3286
01:54:59,679 --> 01:55:05,280
single items

3287
01:55:01,840 --> 01:55:08,480
binge drinking question tailored to the

3288
01:55:05,280 --> 01:55:10,800
age and gender of the patient and then

3289
01:55:08,480 --> 01:55:13,760
that follow up with daily and weekly

3290
01:55:10,800 --> 01:55:18,280
frequency and then a calculated weekly

3291
01:55:13,760 --> 01:55:18,280
quantity of of drinks

3292
01:55:19,360 --> 01:55:24,159
and then based on that this is the best

3293
01:55:21,840 --> 01:55:26,800
practice alert that informs the primary

3294
01:55:24,159 --> 01:55:28,560
care clinician of the patient's

3295
01:55:26,800 --> 01:55:31,360
screening responses

3296
01:55:28,560 --> 01:55:33,679
it provides prompts for next steps so

3297
01:55:31,360 --> 01:55:36,719
suggesting you know this person

3298
01:55:33,679 --> 01:55:38,639
has endorsed unhealthy drinking

3299
01:55:36,719 --> 01:55:40,320
we suggest that you deliver a brief

3300
01:55:38,639 --> 01:55:42,239
intervention

3301
01:55:40,320 --> 01:55:44,000
and then provides them with some

3302
01:55:42,239 --> 01:55:46,480
additional dependence risk questions

3303
01:55:44,000 --> 01:55:48,320
that they can use for further assessment

3304
01:55:46,480 --> 01:55:50,159
if they want to help them

3305
01:55:48,320 --> 01:55:54,320
in their decision making about a

3306
01:55:50,159 --> 01:55:54,320
referral to specialty care

3307
01:55:56,080 --> 01:55:58,960
here in this

3308
01:55:57,599 --> 01:56:02,239
flow sheet

3309
01:55:58,960 --> 01:56:06,239
the physicians are able to see patients

3310
01:56:02,239 --> 01:56:08,560
alcohol use over time across visits

3311
01:56:06,239 --> 01:56:12,320
in order to stimulate conversations with

3312
01:56:08,560 --> 01:56:12,320
patients about how they're doing

3313
01:56:12,800 --> 01:56:18,560
and then this is a

3314
01:56:15,440 --> 01:56:21,119
referral documentation screen that

3315
01:56:18,560 --> 01:56:22,960
clinicians can use to easily refer

3316
01:56:21,119 --> 01:56:25,360
patients to specialty addiction

3317
01:56:22,960 --> 01:56:25,360
treatment

3318
01:56:26,880 --> 01:56:32,159
so because of all of these tools and

3319
01:56:29,920 --> 01:56:33,440
combined with the comprehensive medical

3320
01:56:32,159 --> 01:56:36,960
and mental health and services

3321
01:56:33,440 --> 01:56:39,599
utilization data that we have in the ehr

3322
01:56:36,960 --> 01:56:41,679
we've been able to examine you know

3323
01:56:39,599 --> 01:56:44,239
different ways of the you know the

3324
01:56:41,679 --> 01:56:46,159
effectiveness of you know different ways

3325
01:56:44,239 --> 01:56:48,080
of delivering espert

3326
01:56:46,159 --> 01:56:50,960
uh we've been able to look at the

3327
01:56:48,080 --> 01:56:54,400
effects on something like blood pressure

3328
01:56:50,960 --> 01:56:57,040
of receiving a brief intervention

3329
01:56:54,400 --> 01:57:00,080
and then we've been able to look at the

3330
01:56:57,040 --> 01:57:04,400
associations between different levels of

3331
01:57:00,080 --> 01:57:05,280
and patterns of unhealthy drinking on um

3332
01:57:04,400 --> 01:57:07,360
with

3333
01:57:05,280 --> 01:57:10,080
different medical conditions psychiatric

3334
01:57:07,360 --> 01:57:12,719
disorders and with other substance use

3335
01:57:10,080 --> 01:57:12,719
disorders

3336
01:57:17,199 --> 01:57:23,840
we are also using

3337
01:57:20,639 --> 01:57:25,599
the ehr data to study factors associated

3338
01:57:23,840 --> 01:57:28,080
with successful implementation and

3339
01:57:25,599 --> 01:57:30,000
sustainment of expert

3340
01:57:28,080 --> 01:57:31,840
and looking at the fidelity and quality

3341
01:57:30,000 --> 01:57:33,760
of brief interventions using an

3342
01:57:31,840 --> 01:57:37,239
implementation

3343
01:57:33,760 --> 01:57:37,239
science framework

3344
01:57:38,560 --> 01:57:43,280
so similar to the adult

3345
01:57:41,119 --> 01:57:46,719
alcohol screening work we've been able

3346
01:57:43,280 --> 01:57:48,800
to study expert and prevention and early

3347
01:57:46,719 --> 01:57:51,280
intervention in adolescence using the

3348
01:57:48,800 --> 01:57:54,239
tools that were already in place in the

3349
01:57:51,280 --> 01:57:56,960
systems ehr as part of

3350
01:57:54,239 --> 01:57:58,880
the well visit template so you know kids

3351
01:57:56,960 --> 01:58:01,199
fill out a comprehensive

3352
01:57:58,880 --> 01:58:04,639
screening questionnaire

3353
01:58:01,199 --> 01:58:06,560
as they're sitting and waiting for

3354
01:58:04,639 --> 01:58:09,840
for their well visit and part of the

3355
01:58:06,560 --> 01:58:13,440
those of course are alcohol and drug and

3356
01:58:09,840 --> 01:58:13,440
mood and other questions

3357
01:58:13,599 --> 01:58:21,679
uh we've also been able to we were able

3358
01:58:16,320 --> 01:58:22,639
to get the craft um embedded in the ehr

3359
01:58:21,679 --> 01:58:26,560
and

3360
01:58:22,639 --> 01:58:29,920
some other alcohol and drug questions

3361
01:58:26,560 --> 01:58:34,159
that clinicians can use to assess

3362
01:58:29,920 --> 01:58:35,920
severity and use patterns

3363
01:58:34,159 --> 01:58:39,119
and then similar to what i showed you

3364
01:58:35,920 --> 01:58:41,599
with the adults you're we're able to um

3365
01:58:39,119 --> 01:58:44,400
show in this flow sheet you know

3366
01:58:41,599 --> 01:58:47,040
kids response over time so that doctors

3367
01:58:44,400 --> 01:58:50,239
can sort of see how they're doing you

3368
01:58:47,040 --> 01:58:50,239
know from visit to visit

3369
01:58:52,000 --> 01:58:58,560
and again using these tools and other

3370
01:58:55,520 --> 01:59:01,520
ehr data we've been able to conduct

3371
01:58:58,560 --> 01:59:03,520
pragmatic trials to examine

3372
01:59:01,520 --> 01:59:05,360
you know what works in terms of

3373
01:59:03,520 --> 01:59:08,400
implementation of screening brief

3374
01:59:05,360 --> 01:59:10,800
intervention and referral to treatment

3375
01:59:08,400 --> 01:59:14,080
being able to look at specialty

3376
01:59:10,800 --> 01:59:16,719
treatment initiation and engagement

3377
01:59:14,080 --> 01:59:19,920
as a result of expert

3378
01:59:16,719 --> 01:59:21,920
able to look at patient outcomes like

3379
01:59:19,920 --> 01:59:23,440
substance use outcomes and depression

3380
01:59:21,920 --> 01:59:27,199
outcomes

3381
01:59:23,440 --> 01:59:30,239
and then uh because of the the scope and

3382
01:59:27,199 --> 01:59:33,119
the the you know the stable membership

3383
01:59:30,239 --> 01:59:37,199
we're able to look at long-term outcomes

3384
01:59:33,119 --> 01:59:39,360
from espert over time so uh looking at

3385
01:59:37,199 --> 01:59:41,840
three-year outcomes and now we're uh we

3386
01:59:39,360 --> 01:59:43,599
have a paper coming out soon looking at

3387
01:59:41,840 --> 01:59:47,199
seven-year outcomes

3388
01:59:43,599 --> 01:59:47,199
over time from esper

3389
01:59:49,119 --> 01:59:51,440
so

3390
01:59:50,080 --> 01:59:54,320
you know we've seen a number of the

3391
01:59:51,440 --> 01:59:57,520
different ways that ehr data can be used

3392
01:59:54,320 --> 01:59:59,639
for studying and delivering prevention

3393
01:59:57,520 --> 02:00:01,679
activities obviously

3394
01:59:59,639 --> 02:00:03,360
epidemiological data

3395
02:00:01,679 --> 02:00:05,760
the kind of machine learning and

3396
02:00:03,360 --> 02:00:08,480
predictive modeling that i talked about

3397
02:00:05,760 --> 02:00:10,080
pragmatic trials clinical decision

3398
02:00:08,480 --> 02:00:12,159
support tools

3399
02:00:10,080 --> 02:00:14,239
the sort of patient reported outcomes

3400
02:00:12,159 --> 02:00:16,400
and symptoms and functioning that can be

3401
02:00:14,239 --> 02:00:18,159
used both for research and for quality

3402
02:00:16,400 --> 02:00:20,800
improvement

3403
02:00:18,159 --> 02:00:22,239
as we just heard from dr chisholm we're

3404
02:00:20,800 --> 02:00:24,840
increasingly

3405
02:00:22,239 --> 02:00:27,119
assessing social determinants of health

3406
02:00:24,840 --> 02:00:29,679
uh including things like adverse

3407
02:00:27,119 --> 02:00:32,960
childhood experiences

3408
02:00:29,679 --> 02:00:35,760
food housing and transportation

3409
02:00:32,960 --> 02:00:38,239
other sorts of critical outcomes beyond

3410
02:00:35,760 --> 02:00:40,880
just substance use so medical and mental

3411
02:00:38,239 --> 02:00:42,400
health and services utilization pharmacy

3412
02:00:40,880 --> 02:00:44,560
utilization

3413
02:00:42,400 --> 02:00:47,280
pharmacotherapy

3414
02:00:44,560 --> 02:00:48,239
and then really able to look at numbers

3415
02:00:47,280 --> 02:00:51,040
and

3416
02:00:48,239 --> 02:00:53,599
having the ability to look at the reach

3417
02:00:51,040 --> 02:00:56,719
of intervention so you know in a regular

3418
02:00:53,599 --> 02:00:59,119
clinical trial looking at you know um

3419
02:00:56,719 --> 02:01:01,679
outcomes it's great when you you know

3420
02:00:59,119 --> 02:01:04,000
but you may have a sample of only

3421
02:01:01,679 --> 02:01:06,960
um you know hundreds of people but being

3422
02:01:04,000 --> 02:01:09,360
able to look at this over thousands or

3423
02:01:06,960 --> 02:01:11,199
or millions of people

3424
02:01:09,360 --> 02:01:12,960
you know is something that ehr data

3425
02:01:11,199 --> 02:01:16,000
really enables so

3426
02:01:12,960 --> 02:01:17,520
and looking at outcomes over time so i

3427
02:01:16,000 --> 02:01:20,239
will stop there

3428
02:01:17,520 --> 02:01:20,239
and thank you

3429
02:01:21,119 --> 02:01:28,480
great thank you for that and now i will

3430
02:01:23,199 --> 02:01:28,480
turn to our final presenter hello

3431
02:01:28,560 --> 02:01:33,440
um yeah so i'm excited to kind of close

3432
02:01:30,800 --> 02:01:35,840
this out and we'll try to get us

3433
02:01:33,440 --> 02:01:38,639
uh by some of our time back because i

3434
02:01:35,840 --> 02:01:41,760
would love to have some time for a q a

3435
02:01:38,639 --> 02:01:43,280
so let me just um okay i'm sharing the

3436
02:01:41,760 --> 02:01:45,599
right one

3437
02:01:43,280 --> 02:01:47,440
and we'll get started so i'm going to

3438
02:01:45,599 --> 02:01:48,239
talk uh today about

3439
02:01:47,440 --> 02:01:51,040
um

3440
02:01:48,239 --> 02:01:53,760
how we leverage the va electronic health

3441
02:01:51,040 --> 02:01:56,080
record data to improve care um and i'm

3442
02:01:53,760 --> 02:01:57,520
going to use storm as an exemplar so i

3443
02:01:56,080 --> 02:01:59,760
was super excited that in the first

3444
02:01:57,520 --> 02:02:03,440
session there were some opioid related

3445
02:01:59,760 --> 02:02:05,679
examples shared from arc or ahrq so

3446
02:02:03,440 --> 02:02:06,960
again we'll hopefully uh

3447
02:02:05,679 --> 02:02:09,040
have some nice

3448
02:02:06,960 --> 02:02:11,360
info that will help

3449
02:02:09,040 --> 02:02:13,760
piggyback on some of that so

3450
02:02:11,360 --> 02:02:15,760
in va we have a lot of data so we just

3451
02:02:13,760 --> 02:02:17,440
wanted to make it very clear

3452
02:02:15,760 --> 02:02:19,520
you know we have this is a little bit

3453
02:02:17,440 --> 02:02:21,440
older so we're up to 24 million patients

3454
02:02:19,520 --> 02:02:24,320
but the point being you can see the bees

3455
02:02:21,440 --> 02:02:27,440
on the right hand side about um how many

3456
02:02:24,320 --> 02:02:29,280
billion uh records we have uh regarding

3457
02:02:27,440 --> 02:02:32,239
all sorts of things clinical notes labs

3458
02:02:29,280 --> 02:02:35,280
and and whatnot so we have a lot of data

3459
02:02:32,239 --> 02:02:37,199
and it's organized um in regions and so

3460
02:02:35,280 --> 02:02:40,000
hopefully that just gives a quick

3461
02:02:37,199 --> 02:02:41,520
uh background on on what types of data

3462
02:02:40,000 --> 02:02:43,040
we do have and

3463
02:02:41,520 --> 02:02:45,679
the amount as

3464
02:02:43,040 --> 02:02:47,679
really the us's national only national

3465
02:02:45,679 --> 02:02:48,560
healthcare system so one thing in the

3466
02:02:47,679 --> 02:02:50,880
opioid

3467
02:02:48,560 --> 02:02:53,360
related space we did use our corporate

3468
02:02:50,880 --> 02:02:55,679
data warehouse for is looking at

3469
02:02:53,360 --> 02:02:57,440
overdose and suicide mortality among

3470
02:02:55,679 --> 02:02:59,840
patients

3471
02:02:57,440 --> 02:03:02,480
that were prescribed opioids and this

3472
02:02:59,840 --> 02:03:06,239
was an earlier analysis we did so it's a

3473
02:03:02,480 --> 02:03:08,560
little bit older data looking at fy 2013

3474
02:03:06,239 --> 02:03:10,560
overdose and suicide mortality and what

3475
02:03:08,560 --> 02:03:13,440
we found given that this is a naida

3476
02:03:10,560 --> 02:03:16,239
webinar is a good chunk of those who end

3477
02:03:13,440 --> 02:03:18,239
up dying this was this analysis was done

3478
02:03:16,239 --> 02:03:21,599
when there was a huge focus on morphine

3479
02:03:18,239 --> 02:03:23,760
equivalent daily dose um limits and what

3480
02:03:21,599 --> 02:03:25,760
we found is again you know 20 percent of

3481
02:03:23,760 --> 02:03:28,159
people people who died were above you

3482
02:03:25,760 --> 02:03:30,639
know that you know cut off that a lot of

3483
02:03:28,159 --> 02:03:32,719
people were focused on and a good 80 of

3484
02:03:30,639 --> 02:03:34,800
people who died um

3485
02:03:32,719 --> 02:03:36,800
were below that but moreover four out of

3486
02:03:34,800 --> 02:03:38,400
every five patients who died from an

3487
02:03:36,800 --> 02:03:40,480
overdose or suicide were again

3488
02:03:38,400 --> 02:03:42,080
prescribed below 90 and three out of

3489
02:03:40,480 --> 02:03:45,119
four had a mental health or stud

3490
02:03:42,080 --> 02:03:46,800
diagnosis so again really it really

3491
02:03:45,119 --> 02:03:47,840
emphasized to us that we need to

3492
02:03:46,800 --> 02:03:51,119
consider

3493
02:03:47,840 --> 02:03:52,719
um a number of other factors given that

3494
02:03:51,119 --> 02:03:54,880
those at greatest risk of overdose or

3495
02:03:52,719 --> 02:03:56,320
suicide do we know tend to be complete

3496
02:03:54,880 --> 02:03:58,639
complex patients with multiple

3497
02:03:56,320 --> 02:04:00,560
comorbidities and again you know that

3498
02:03:58,639 --> 02:04:02,560
opioid that small little red dot even if

3499
02:04:00,560 --> 02:04:04,480
you take that away risk for adverse

3500
02:04:02,560 --> 02:04:06,239
events related to pain mental health

3501
02:04:04,480 --> 02:04:08,639
medical conditions are all really

3502
02:04:06,239 --> 02:04:11,360
important and so really thinking through

3503
02:04:08,639 --> 02:04:13,119
how can we support patient-centered risk

3504
02:04:11,360 --> 02:04:15,280
mitigation addressing all the

3505
02:04:13,119 --> 02:04:18,000
co-morbidities we know to be associated

3506
02:04:15,280 --> 02:04:19,920
with adverse outcomes like overdose and

3507
02:04:18,000 --> 02:04:22,400
suicide among patients prescribed

3508
02:04:19,920 --> 02:04:25,119
opioids so that's where storm enters in

3509
02:04:22,400 --> 02:04:26,079
so what is storm it's a

3510
02:04:25,119 --> 02:04:29,119
a

3511
02:04:26,079 --> 02:04:31,679
we leverage the ehr data um and we

3512
02:04:29,119 --> 02:04:35,119
predict risk of overdose or suicide from

3513
02:04:31,679 --> 02:04:37,520
that data and we look at prediction in

3514
02:04:35,119 --> 02:04:39,199
the next year of an overdose or suicide

3515
02:04:37,520 --> 02:04:41,920
event and it generates a patient

3516
02:04:39,199 --> 02:04:45,119
specific score and so these

3517
02:04:41,920 --> 02:04:47,840
parameters are applied on real-time data

3518
02:04:45,119 --> 02:04:50,159
and updated nightly so every night these

3519
02:04:47,840 --> 02:04:53,199
data are updated to create individual

3520
02:04:50,159 --> 02:04:55,840
risk estimates for overdose or suicide

3521
02:04:53,199 --> 02:04:57,760
among all patients in the va so not just

3522
02:04:55,840 --> 02:04:59,599
those prescribed opioids but we also

3523
02:04:57,760 --> 02:05:01,199
have what we call a hypothetical score

3524
02:04:59,599 --> 02:05:04,000
so if a patient was going to be

3525
02:05:01,199 --> 02:05:04,960
introduced um or started on opioids what

3526
02:05:04,000 --> 02:05:06,960
their

3527
02:05:04,960 --> 02:05:09,920
risk would be and so this has been

3528
02:05:06,960 --> 02:05:11,679
translated into other ehrs we've worked

3529
02:05:09,920 --> 02:05:15,280
on getting the department of defense a

3530
02:05:11,679 --> 02:05:16,880
tri-storm um example as well as uh we're

3531
02:05:15,280 --> 02:05:19,280
currently working with cerner to

3532
02:05:16,880 --> 02:05:21,599
integrate that into the ehr as well and

3533
02:05:19,280 --> 02:05:23,520
so some of the key cohorts are those

3534
02:05:21,599 --> 02:05:25,360
patients prescribed opioids again

3535
02:05:23,520 --> 02:05:27,360
actively prescribed and we have some

3536
02:05:25,360 --> 02:05:29,679
data that we've um done showing that

3537
02:05:27,360 --> 02:05:31,520
recent discontinuation also puts people

3538
02:05:29,679 --> 02:05:33,679
at high risk particularly those who've

3539
02:05:31,520 --> 02:05:35,599
been on for longer periods of time so we

3540
02:05:33,679 --> 02:05:37,360
do include that as a cohort and storm

3541
02:05:35,599 --> 02:05:39,520
which i'll show you a little bit later

3542
02:05:37,360 --> 02:05:41,599
and again we do include patients with an

3543
02:05:39,520 --> 02:05:43,440
opiate use disorder just in general on

3544
02:05:41,599 --> 02:05:45,119
the site because we do know those are

3545
02:05:43,440 --> 02:05:47,280
very high risk patients and again as i

3546
02:05:45,119 --> 02:05:49,520
mentioned those considering opioid

3547
02:05:47,280 --> 02:05:51,199
therapy as well as patients with an

3548
02:05:49,520 --> 02:05:53,199
overdose we have some very compelling

3549
02:05:51,199 --> 02:05:55,280
data that uh you know patients who

3550
02:05:53,199 --> 02:05:56,560
overdose are obviously at a higher risk

3551
02:05:55,280 --> 02:05:58,000
for um

3552
02:05:56,560 --> 02:06:00,239
overdose or suicide in the following

3553
02:05:58,000 --> 02:06:02,239
year so we also track those patients so

3554
02:06:00,239 --> 02:06:04,079
we have a lot of information um on how

3555
02:06:02,239 --> 02:06:05,599
we developed our risk model and we are

3556
02:06:04,079 --> 02:06:09,280
actually doing a randomized program

3557
02:06:05,599 --> 02:06:11,440
evaluation of actually mandating um

3558
02:06:09,280 --> 02:06:14,079
interdisciplinary reviews of these

3559
02:06:11,440 --> 02:06:16,159
patients um who are identified as very

3560
02:06:14,079 --> 02:06:19,520
high risk in storm which i'll talk about

3561
02:06:16,159 --> 02:06:21,360
and so ahrq has noted this as a an

3562
02:06:19,520 --> 02:06:22,880
innovation and posted on their patient

3563
02:06:21,360 --> 02:06:25,679
safety network so you can also learn

3564
02:06:22,880 --> 02:06:28,239
more about that um via the web that

3565
02:06:25,679 --> 02:06:31,360
website so the big thing with regards to

3566
02:06:28,239 --> 02:06:33,440
our predictive model and we have over uh

3567
02:06:31,360 --> 02:06:35,760
you know 50 variables but i'm just kind

3568
02:06:33,440 --> 02:06:38,719
of showing you some examples of some of

3569
02:06:35,760 --> 02:06:40,560
the key domains and again we have again

3570
02:06:38,719 --> 02:06:42,400
medical comorbidities we know increased

3571
02:06:40,560 --> 02:06:44,400
risk as well as psychiatric substance

3572
02:06:42,400 --> 02:06:46,320
use and healthcare utilization again the

3573
02:06:44,400 --> 02:06:48,480
best predictor of future behaviors past

3574
02:06:46,320 --> 02:06:50,960
behavior so we do see very very high

3575
02:06:48,480 --> 02:06:53,440
odds ratios for those events in there uh

3576
02:06:50,960 --> 02:06:54,719
interestingly again with the focus

3577
02:06:53,440 --> 02:06:58,320
previously on

3578
02:06:54,719 --> 02:07:00,880
medd levels 120 medd would increase

3579
02:06:58,320 --> 02:07:02,960
modeled risk by about as much as a ptsd

3580
02:07:00,880 --> 02:07:05,119
or aud diagnosis

3581
02:07:02,960 --> 02:07:06,800
so really you know letting folks know

3582
02:07:05,119 --> 02:07:09,360
that as much effort as you're putting

3583
02:07:06,800 --> 02:07:10,639
into addressing mdd we should be

3584
02:07:09,360 --> 02:07:12,480
thinking about putting those efforts

3585
02:07:10,639 --> 02:07:15,280
into addressing these mental health and

3586
02:07:12,480 --> 02:07:18,719
substance use disorder comorbidities

3587
02:07:15,280 --> 02:07:21,520
moreover what our paper did also show is

3588
02:07:18,719 --> 02:07:24,159
that there are high odds ratios for

3589
02:07:21,520 --> 02:07:26,159
other evidence-based sedating pain meds

3590
02:07:24,159 --> 02:07:27,840
so there has been a lot of focus on uh

3591
02:07:26,159 --> 02:07:30,159
benzodiazepines or sedatives for

3592
02:07:27,840 --> 02:07:31,520
instance and you can see actually if you

3593
02:07:30,159 --> 02:07:33,119
have um

3594
02:07:31,520 --> 02:07:35,520
an opioid sedative

3595
02:07:33,119 --> 02:07:37,920
your increa your risk increases you know

3596
02:07:35,520 --> 02:07:39,679
1.4 times but if you have an opioid plus

3597
02:07:37,920 --> 02:07:42,400
one of these other evidence-based but

3598
02:07:39,679 --> 02:07:45,360
sedating pain meds tcas snris

3599
02:07:42,400 --> 02:07:48,000
anticonvulsants you see a gradient right

3600
02:07:45,360 --> 02:07:49,920
with the more you have on the greater

3601
02:07:48,000 --> 02:07:52,960
the risk for an overdose or suicide

3602
02:07:49,920 --> 02:07:55,199
event even much so more than the

3603
02:07:52,960 --> 02:07:57,920
sedatives so just really orienting the

3604
02:07:55,199 --> 02:07:59,760
field and uh and folks and providers to

3605
02:07:57,920 --> 02:08:02,480
this you know it's not just about benzos

3606
02:07:59,760 --> 02:08:05,119
it's about other sedating meds as well

3607
02:08:02,480 --> 02:08:08,560
so this is what storm looks like um we

3608
02:08:05,119 --> 02:08:10,800
have again the big thing from a provider

3609
02:08:08,560 --> 02:08:13,119
perspective right is and we're

3610
02:08:10,800 --> 02:08:14,639
incorporating again the predictive model

3611
02:08:13,119 --> 02:08:17,040
you have the predictive model risk on

3612
02:08:14,639 --> 02:08:19,520
the left hand side so we classify them

3613
02:08:17,040 --> 02:08:22,000
as very high high medium or low risk so

3614
02:08:19,520 --> 02:08:25,360
it gives you um that in the red their

3615
02:08:22,000 --> 02:08:26,960
risk level um strata and then we also

3616
02:08:25,360 --> 02:08:28,639
have resort for those who are of you or

3617
02:08:26,960 --> 02:08:31,040
who are in this space resource another

3618
02:08:28,639 --> 02:08:32,880
way of um stratifying patients so we

3619
02:08:31,040 --> 02:08:35,360
have that their risk scores and resort

3620
02:08:32,880 --> 02:08:37,199
also embedded in there um and then you

3621
02:08:35,360 --> 02:08:39,599
have what factors contribute to my

3622
02:08:37,199 --> 02:08:41,679
patient's risk so of those 50 or so

3623
02:08:39,599 --> 02:08:43,520
variables what are the ones that

3624
02:08:41,679 --> 02:08:44,880
contribute to this patient's risk and it

3625
02:08:43,520 --> 02:08:47,840
lists them so you have mental health

3626
02:08:44,880 --> 02:08:49,040
medical comorbidities adverse events

3627
02:08:47,840 --> 02:08:50,880
moreover what are the relevant

3628
02:08:49,040 --> 02:08:52,320
medications this patient's taking that

3629
02:08:50,880 --> 02:08:54,560
may also be

3630
02:08:52,320 --> 02:08:55,760
putting them at risk and the biggest

3631
02:08:54,560 --> 02:08:58,719
thing again where the rubber meets the

3632
02:08:55,760 --> 02:09:01,599
road is clinically how to do i better

3633
02:08:58,719 --> 02:09:03,440
manage my patients risks so these risk

3634
02:09:01,599 --> 02:09:05,920
mitigation strategies are based on

3635
02:09:03,440 --> 02:09:07,520
clinical practice guidelines and so we

3636
02:09:05,920 --> 02:09:09,840
integrate those clinical practice

3637
02:09:07,520 --> 02:09:12,800
guideline recommended interventions and

3638
02:09:09,840 --> 02:09:14,560
treatments on like within this system

3639
02:09:12,800 --> 02:09:16,639
and we include the dates in which any of

3640
02:09:14,560 --> 02:09:18,560
them were done and then on you see on

3641
02:09:16,639 --> 02:09:20,639
the right hand side also when they have

3642
02:09:18,560 --> 02:09:22,000
recent they've had recent and upcoming

3643
02:09:20,639 --> 02:09:24,320
appointments and who their care

3644
02:09:22,000 --> 02:09:27,199
providers are interestingly we color

3645
02:09:24,320 --> 02:09:28,560
code all of this so that um

3646
02:09:27,199 --> 02:09:31,199
things done at one facility are a

3647
02:09:28,560 --> 02:09:33,440
certain color so this also gives a sense

3648
02:09:31,199 --> 02:09:35,040
of how much potentially uncoordinated

3649
02:09:33,440 --> 02:09:36,480
uncoordinated care might be happening if

3650
02:09:35,040 --> 02:09:38,880
a patient's for instance being seen at

3651
02:09:36,480 --> 02:09:40,480
multiple facilities it really emphasizes

3652
02:09:38,880 --> 02:09:42,639
the need to um get them that care

3653
02:09:40,480 --> 02:09:44,159
coordination so again one of the big

3654
02:09:42,639 --> 02:09:45,679
things i love as well is that we also

3655
02:09:44,159 --> 02:09:46,639
include non-pharmacological pain

3656
02:09:45,679 --> 02:09:48,079
treatments

3657
02:09:46,639 --> 02:09:49,599
and when

3658
02:09:48,079 --> 02:09:50,560
patients have received them as well so

3659
02:09:49,599 --> 02:09:53,920
again

3660
02:09:50,560 --> 02:09:55,280
it really in one single click gives

3661
02:09:53,920 --> 02:09:59,040
providers

3662
02:09:55,280 --> 02:10:01,840
a real-time view of what their patient's

3663
02:09:59,040 --> 02:10:03,760
risk is and what puts them at risk and

3664
02:10:01,840 --> 02:10:05,599
what they can do about it which again is

3665
02:10:03,760 --> 02:10:07,119
actionable information

3666
02:10:05,599 --> 02:10:10,239
so we have a randomized program

3667
02:10:07,119 --> 02:10:12,000
evaluation um that again and this has

3668
02:10:10,239 --> 02:10:13,840
been embedded in a decision support

3669
02:10:12,000 --> 02:10:16,159
system to encourage guideline-based risk

3670
02:10:13,840 --> 02:10:18,000
mitigation and so there was a policy

3671
02:10:16,159 --> 02:10:20,480
that we released to national uh

3672
02:10:18,000 --> 02:10:22,239
randomized program evaluation and it

3673
02:10:20,480 --> 02:10:24,079
required interdisciplinary team review

3674
02:10:22,239 --> 02:10:26,400
of patients again estimated at a very

3675
02:10:24,079 --> 02:10:29,679
high risk and we randomized facilities

3676
02:10:26,400 --> 02:10:31,679
um to having an action planning uh

3677
02:10:29,679 --> 02:10:34,239
element or not having it and then we

3678
02:10:31,679 --> 02:10:37,679
monitored it uh via metrics so we

3679
02:10:34,239 --> 02:10:39,599
expanded uh storm from one percent to

3680
02:10:37,679 --> 02:10:41,679
the top five percent as part of a step

3681
02:10:39,599 --> 02:10:43,199
wedge design and i'll kind of show you

3682
02:10:41,679 --> 02:10:45,280
what that looks like but this was a

3683
02:10:43,199 --> 02:10:46,400
policy that we put out

3684
02:10:45,280 --> 02:10:48,400
to

3685
02:10:46,400 --> 02:10:50,239
have folks oriented to the fact that

3686
02:10:48,400 --> 02:10:52,800
very high-risk patients needed to have a

3687
02:10:50,239 --> 02:10:56,400
risk review completed and that we would

3688
02:10:52,800 --> 02:10:59,280
be monitoring and if they didn't uh meet

3689
02:10:56,400 --> 02:11:02,239
uh the uh meet the metrics that they

3690
02:10:59,280 --> 02:11:03,760
would be um given some action planning

3691
02:11:02,239 --> 02:11:05,920
so this is a trial design this is what

3692
02:11:03,760 --> 02:11:07,840
the stepwidge design looked like again

3693
02:11:05,920 --> 02:11:09,599
we expanded from the top one percent to

3694
02:11:07,840 --> 02:11:12,239
the top five percent

3695
02:11:09,599 --> 02:11:14,560
um and the biggest thing for us that was

3696
02:11:12,239 --> 02:11:16,480
really exciting was the fact that when

3697
02:11:14,560 --> 02:11:18,719
you actually expanded um

3698
02:11:16,480 --> 02:11:20,719
interdisciplinary reviews from the top

3699
02:11:18,719 --> 02:11:24,320
one uh to five percent in the step wedge

3700
02:11:20,719 --> 02:11:26,719
design we found a 22 percent reduction

3701
02:11:24,320 --> 02:11:28,880
in all cause mortality in the subsequent

3702
02:11:26,719 --> 02:11:32,560
four months approximately 180 lives

3703
02:11:28,880 --> 02:11:35,360
saved again these are um you know rates

3704
02:11:32,560 --> 02:11:37,040
on par with what you see in medication

3705
02:11:35,360 --> 02:11:38,960
related interventions

3706
02:11:37,040 --> 02:11:40,239
moreover we think that the mechanism was

3707
02:11:38,960 --> 02:11:42,880
these case reviews these

3708
02:11:40,239 --> 02:11:45,280
interdisciplinary reviews um in that uh

3709
02:11:42,880 --> 02:11:47,040
folks were more five times more likely

3710
02:11:45,280 --> 02:11:49,199
um to receive a case review when they

3711
02:11:47,040 --> 02:11:51,119
were listed as very high risk on storm

3712
02:11:49,199 --> 02:11:54,480
than uh than those

3713
02:11:51,119 --> 02:11:56,320
controls so again um we think that this

3714
02:11:54,480 --> 02:11:58,400
may have actually underestimated the

3715
02:11:56,320 --> 02:12:00,639
full benefits given that only 30 of

3716
02:11:58,400 --> 02:12:02,000
those um in the top one to five percent

3717
02:12:00,639 --> 02:12:03,920
received a full case review but we're

3718
02:12:02,000 --> 02:12:06,800
still seeing these significant um

3719
02:12:03,920 --> 02:12:07,599
decreases in mortality based on that

3720
02:12:06,800 --> 02:12:08,960
so

3721
02:12:07,599 --> 02:12:10,800
tauronologis is the first study to

3722
02:12:08,960 --> 02:12:13,119
evaluate the impact of a predictive

3723
02:12:10,800 --> 02:12:16,079
model targeted prevention program in

3724
02:12:13,119 --> 02:12:18,800
reducing adverse outcomes on high-risk

3725
02:12:16,079 --> 02:12:21,760
patients prescribed opioids specifically

3726
02:12:18,800 --> 02:12:23,280
by providing a case review and this is a

3727
02:12:21,760 --> 02:12:26,000
definitely a way we have leveraged the

3728
02:12:23,280 --> 02:12:27,679
ehr to combine risk identification with

3729
02:12:26,000 --> 02:12:29,360
predictive modeling

3730
02:12:27,679 --> 02:12:31,440
and interdisciplinary case review to

3731
02:12:29,360 --> 02:12:33,520
improve patient safety this again is

3732
02:12:31,440 --> 02:12:35,440
just one of many ways va leverages the

3733
02:12:33,520 --> 02:12:37,119
ehr we have lots of different other

3734
02:12:35,440 --> 02:12:38,000
examples but just wanted to use one for

3735
02:12:37,119 --> 02:12:39,199
now

3736
02:12:38,000 --> 02:12:41,199
and we do

3737
02:12:39,199 --> 02:12:43,119
suggest that this shows that combining

3738
02:12:41,199 --> 02:12:45,599
data systems like our corporate data

3739
02:12:43,119 --> 02:12:47,440
warehouse with policy enables

3740
02:12:45,599 --> 02:12:49,360
implementation designs that do a lot

3741
02:12:47,440 --> 02:12:51,920
allow for a strong evaluation of

3742
02:12:49,360 --> 02:12:55,199
outcomes so um here are some references

3743
02:12:51,920 --> 02:12:57,599
and i will stop sharing and pause uh for

3744
02:12:55,199 --> 02:12:59,679
for questions

3745
02:12:57,599 --> 02:13:01,599
all right thank you i think we have

3746
02:12:59,679 --> 02:13:03,280
about five minutes available for

3747
02:13:01,599 --> 02:13:05,199
questions um before we're going to go on

3748
02:13:03,280 --> 02:13:06,400
another brief break and i know that

3749
02:13:05,199 --> 02:13:08,639
there's been some that have been

3750
02:13:06,400 --> 02:13:10,239
floating up in the chat so i invite our

3751
02:13:08,639 --> 02:13:12,239
discussions especially the questions

3752
02:13:10,239 --> 02:13:14,079
that are a little bit more clarifying

3753
02:13:12,239 --> 02:13:15,280
to peek in the chat and see

3754
02:13:14,079 --> 02:13:17,440
um

3755
02:13:15,280 --> 02:13:19,520
i guess i'll ask a sort of a broad

3756
02:13:17,440 --> 02:13:20,960
question that can go in a couple

3757
02:13:19,520 --> 02:13:23,040
different directions so i'll put all the

3758
02:13:20,960 --> 02:13:24,880
directions out and let our presenters

3759
02:13:23,040 --> 02:13:26,480
decide which way to go in it but one

3760
02:13:24,880 --> 02:13:29,280
thing that was touched on a little bit

3761
02:13:26,480 --> 02:13:32,159
when we think about the ehr data

3762
02:13:29,280 --> 02:13:34,639
is the free text data versus like what's

3763
02:13:32,159 --> 02:13:36,880
in a field and a click box

3764
02:13:34,639 --> 02:13:38,800
um and in listening to everybody talk i

3765
02:13:36,880 --> 02:13:41,040
can sort of understand how a research in

3766
02:13:38,800 --> 02:13:42,800
a research design you can use natural

3767
02:13:41,040 --> 02:13:44,320
language processing to pull out some of

3768
02:13:42,800 --> 02:13:46,400
that free text data and then integrate

3769
02:13:44,320 --> 02:13:48,239
it in an algorithm to either predict you

3770
02:13:46,400 --> 02:13:50,320
or combine with some of the field date

3771
02:13:48,239 --> 02:13:51,599
the kind of field

3772
02:13:50,320 --> 02:13:53,280
i don't know what to call it but the

3773
02:13:51,599 --> 02:13:54,639
click box kind of data

3774
02:13:53,280 --> 02:13:56,639
i'm wondering one like how that

3775
02:13:54,639 --> 02:13:58,960
translates into practice if you want to

3776
02:13:56,639 --> 02:14:00,320
build algorithms that are predictive

3777
02:13:58,960 --> 02:14:02,560
um two

3778
02:14:00,320 --> 02:14:04,639
i'm wondering your thoughts on the issue

3779
02:14:02,560 --> 02:14:06,560
of provider variability

3780
02:14:04,639 --> 02:14:08,320
and how that factors into your research

3781
02:14:06,560 --> 02:14:10,000
you know in terms of providers who are

3782
02:14:08,320 --> 02:14:11,440
more likely to click boxes versus

3783
02:14:10,000 --> 02:14:12,800
providers more likely to put all their

3784
02:14:11,440 --> 02:14:15,119
notes in a free text and how you

3785
02:14:12,800 --> 02:14:16,800
potentially change that behavior

3786
02:14:15,119 --> 02:14:19,280
and then the third direction i could go

3787
02:14:16,800 --> 02:14:21,599
in with this question is you know we all

3788
02:14:19,280 --> 02:14:23,760
have our pet area and our priority area

3789
02:14:21,599 --> 02:14:25,679
where we'd love to get that question in

3790
02:14:23,760 --> 02:14:27,679
you know or those items in an electronic

3791
02:14:25,679 --> 02:14:30,000
health record and how do we do that

3792
02:14:27,679 --> 02:14:31,679
without contributing to physician burden

3793
02:14:30,000 --> 02:14:34,000
and adding to the hours that physicians

3794
02:14:31,679 --> 02:14:37,000
spend just documenting things in health

3795
02:14:34,000 --> 02:14:37,000
records

3796
02:14:40,800 --> 02:14:43,679
um

3797
02:14:41,840 --> 02:14:45,840
this is jody trafton i can try to take

3798
02:14:43,679 --> 02:14:48,079
that a little bit for va i mean most of

3799
02:14:45,840 --> 02:14:49,840
our work to date has been using

3800
02:14:48,079 --> 02:14:51,280
structured data because it's so much

3801
02:14:49,840 --> 02:14:54,480
easier to use and so much more

3802
02:14:51,280 --> 02:14:57,040
accessible um but we are uh right now

3803
02:14:54,480 --> 02:14:59,599
starting to implement um

3804
02:14:57,040 --> 02:15:01,920
nlp algorithms for so free text note

3805
02:14:59,599 --> 02:15:04,400
searches in our nightly updated decision

3806
02:15:01,920 --> 02:15:07,040
support in clinical practice um it took

3807
02:15:04,400 --> 02:15:09,599
a lot of technical work to

3808
02:15:07,040 --> 02:15:11,440
to set up um high performance computing

3809
02:15:09,599 --> 02:15:13,920
capabilities that

3810
02:15:11,440 --> 02:15:15,760
that allow us to go through we get 1.74

3811
02:15:13,920 --> 02:15:18,159
million notes a day

3812
02:15:15,760 --> 02:15:19,760
um which is just a lot of text to go

3813
02:15:18,159 --> 02:15:21,599
through but i think some of the other

3814
02:15:19,760 --> 02:15:23,599
pieces that that we struggle with a

3815
02:15:21,599 --> 02:15:24,560
little bit are um

3816
02:15:23,599 --> 02:15:26,320
just

3817
02:15:24,560 --> 02:15:28,400
getting

3818
02:15:26,320 --> 02:15:30,239
it would be so much easier to implement

3819
02:15:28,400 --> 02:15:32,560
things from research into practice if

3820
02:15:30,239 --> 02:15:34,480
the researchers would work with

3821
02:15:32,560 --> 02:15:36,480
your clinical uh

3822
02:15:34,480 --> 02:15:38,159
with your your informatics folks from

3823
02:15:36,480 --> 02:15:39,599
the start because it's really hard to

3824
02:15:38,159 --> 02:15:41,840
take somebody's

3825
02:15:39,599 --> 02:15:43,679
pre-built research algorithm and

3826
02:15:41,840 --> 02:15:46,239
implement it kind of have to go back and

3827
02:15:43,679 --> 02:15:47,760
start over again not completely but but

3828
02:15:46,239 --> 02:15:49,440
you have to do a lot of additional

3829
02:15:47,760 --> 02:15:51,840
revalidation that you wouldn't have to

3830
02:15:49,440 --> 02:15:53,679
do necessarily if if it was done as a

3831
02:15:51,840 --> 02:15:55,599
partnership so i would just highly

3832
02:15:53,679 --> 02:15:57,119
encourage people to work

3833
02:15:55,599 --> 02:15:58,880
you know within and

3834
02:15:57,119 --> 02:16:01,679
with your informatics specialists in

3835
02:15:58,880 --> 02:16:01,679
your health care system

3836
02:16:02,320 --> 02:16:08,639
i would add on to that um

3837
02:16:05,440 --> 02:16:12,719
i think the structure field data

3838
02:16:08,639 --> 02:16:15,840
is much more well suited to

3839
02:16:12,719 --> 02:16:18,159
uh clinical decision support um to

3840
02:16:15,840 --> 02:16:20,880
making decisions about you know brief

3841
02:16:18,159 --> 02:16:24,320
intervention or referrals during a visit

3842
02:16:20,880 --> 02:16:26,480
i think the nlp and the algorithms that

3843
02:16:24,320 --> 02:16:28,800
you can build with that are more

3844
02:16:26,480 --> 02:16:31,599
back-end things that you might do as

3845
02:16:28,800 --> 02:16:34,000
predictors outside of a visit that you

3846
02:16:31,599 --> 02:16:35,920
might use for doing targeted outreach

3847
02:16:34,000 --> 02:16:38,319
and that sort of thing just because the

3848
02:16:35,920 --> 02:16:40,160
real timeness of it just doesn't work

3849
02:16:38,319 --> 02:16:42,559
quite as well so i think that they both

3850
02:16:40,160 --> 02:16:45,200
have strengths and weaknesses i think

3851
02:16:42,559 --> 02:16:47,840
they both have a lot of embedded biases

3852
02:16:45,200 --> 02:16:49,359
um i think that we know you know even

3853
02:16:47,840 --> 02:16:50,319
our work with children of incarcerated

3854
02:16:49,359 --> 02:16:52,639
parents

3855
02:16:50,319 --> 02:16:54,639
we are pretty sure that docs are talking

3856
02:16:52,639 --> 02:16:56,160
about incarcerated parents with some

3857
02:16:54,639 --> 02:16:57,840
types of patients more than they're

3858
02:16:56,160 --> 02:16:59,599
talking about incarcerated parents with

3859
02:16:57,840 --> 02:17:02,319
other types of patients therefore we're

3860
02:16:59,599 --> 02:17:03,599
more likely to recognize it um in

3861
02:17:02,319 --> 02:17:05,359
patients of color who they're more

3862
02:17:03,599 --> 02:17:06,399
likely to be asking that question or

3863
02:17:05,359 --> 02:17:08,800
have been patients or they're more

3864
02:17:06,399 --> 02:17:10,000
likely to be having that discussion so

3865
02:17:08,800 --> 02:17:12,240
we sort of

3866
02:17:10,000 --> 02:17:12,240
um

3867
02:17:12,960 --> 02:17:19,200
reinforce existing biases by using data

3868
02:17:16,960 --> 02:17:21,120
that was biased when we collected it so

3869
02:17:19,200 --> 02:17:23,200
you know part of it is just recognizing

3870
02:17:21,120 --> 02:17:26,240
that at the time and starting to figure

3871
02:17:23,200 --> 02:17:28,160
out how you can standardize but we know

3872
02:17:26,240 --> 02:17:29,840
even when you have a quick box the

3873
02:17:28,160 --> 02:17:31,840
decision about whether or not to ask the

3874
02:17:29,840 --> 02:17:33,920
question versus whether or not to guess

3875
02:17:31,840 --> 02:17:35,760
whether to say you haven't your parents

3876
02:17:33,920 --> 02:17:37,519
haven't been in jail yet have they

3877
02:17:35,760 --> 02:17:39,920
versus um

3878
02:17:37,519 --> 02:17:41,920
are your parents in jail um

3879
02:17:39,920 --> 02:17:44,399
get you different responses so we know

3880
02:17:41,920 --> 02:17:46,240
that we get biased results whether we go

3881
02:17:44,399 --> 02:17:49,359
with the natural language processes or

3882
02:17:46,240 --> 02:17:51,359
whether we go with the structure fields

3883
02:17:49,359 --> 02:17:54,319
so a lot of that is just training and a

3884
02:17:51,359 --> 02:17:56,880
lot of that is

3885
02:17:54,319 --> 02:17:58,000
you know figuring out how we best

3886
02:17:56,880 --> 02:18:00,960
validate

3887
02:17:58,000 --> 02:18:02,559
um and how we best weigh the value of

3888
02:18:00,960 --> 02:18:05,200
that data when we're making decisions

3889
02:18:02,559 --> 02:18:05,200
based on it

3890
02:18:08,080 --> 02:18:11,439
yeah i was gonna say the only thing i'd

3891
02:18:09,920 --> 02:18:13,599
add you know the low hanging fruit here

3892
02:18:11,439 --> 02:18:15,040
i think is you know when people and this

3893
02:18:13,599 --> 02:18:16,960
is often happens you know when people

3894
02:18:15,040 --> 02:18:18,719
are recording discrete information as

3895
02:18:16,960 --> 02:18:20,160
text you know to which you want to say

3896
02:18:18,719 --> 02:18:22,240
don't do that

3897
02:18:20,160 --> 02:18:24,479
you know record discrete information as

3898
02:18:22,240 --> 02:18:26,399
discrete data there's a lot about health

3899
02:18:24,479 --> 02:18:27,519
and health care that is not discrete you

3900
02:18:26,399 --> 02:18:28,960
know there's a lot about health and

3901
02:18:27,519 --> 02:18:31,040
health care that's complicated and

3902
02:18:28,960 --> 02:18:33,200
continuous and cannot be reduced to

3903
02:18:31,040 --> 02:18:34,559
discrete data but there's still plenty

3904
02:18:33,200 --> 02:18:36,399
of health care where people are

3905
02:18:34,559 --> 02:18:38,559
recording things that are discrete data

3906
02:18:36,399 --> 02:18:40,559
by typing them in as text and that's

3907
02:18:38,559 --> 02:18:42,080
just a loss for everybody it's a loss in

3908
02:18:40,559 --> 02:18:44,479
terms of efficient practice for the

3909
02:18:42,080 --> 02:18:46,240
clinicians it's a loss in terms of you

3910
02:18:44,479 --> 02:18:47,840
know being able to improve care it's a

3911
02:18:46,240 --> 02:18:49,280
loss for research it's just a loss all

3912
02:18:47,840 --> 02:18:51,519
around so that's the low-hanging one i

3913
02:18:49,280 --> 02:18:51,519
think

3914
02:18:51,599 --> 02:18:55,200
i guess there's a follow-up thought to

3915
02:18:53,200 --> 02:18:57,040
something that jody said you know

3916
02:18:55,200 --> 02:18:58,639
our ehr is the kind of thing where you

3917
02:18:57,040 --> 02:18:59,760
know you've seen one ehr you've seen one

3918
02:18:58,639 --> 02:19:02,160
ehr

3919
02:18:59,760 --> 02:19:04,240
you know how generalizable

3920
02:19:02,160 --> 02:19:06,319
are some of these algorithms that you

3921
02:19:04,240 --> 02:19:08,800
all are working on so taking it from one

3922
02:19:06,319 --> 02:19:11,439
healthcare system to moving it out to

3923
02:19:08,800 --> 02:19:12,880
another one or nationally or broadly you

3924
02:19:11,439 --> 02:19:14,719
know how much work would have to be done

3925
02:19:12,880 --> 02:19:16,240
in translation for something developed

3926
02:19:14,719 --> 02:19:17,519
in one healthcare system to move it out

3927
02:19:16,240 --> 02:19:20,240
to another

3928
02:19:17,519 --> 02:19:22,479
there's a lot of translation so we have

3929
02:19:20,240 --> 02:19:25,280
for example this for this storm model we

3930
02:19:22,479 --> 02:19:29,359
have translated to the dod medical

3931
02:19:25,280 --> 02:19:32,240
record system um as well as to cerner

3932
02:19:29,359 --> 02:19:33,280
um and and i will say um

3933
02:19:32,240 --> 02:19:35,760
we

3934
02:19:33,280 --> 02:19:37,840
i think it's not just a medical record

3935
02:19:35,760 --> 02:19:40,719
system issue but also

3936
02:19:37,840 --> 02:19:42,000
you need you can't just it depends on

3937
02:19:40,719 --> 02:19:44,639
the patient populations that you're

3938
02:19:42,000 --> 02:19:47,040
seeing right so you can't take a model

3939
02:19:44,639 --> 02:19:48,640
or a decision support platform that you

3940
02:19:47,040 --> 02:19:50,720
built within

3941
02:19:48,640 --> 02:19:52,640
for use in one population and just

3942
02:19:50,720 --> 02:19:55,600
assume that it's going to work just fine

3943
02:19:52,640 --> 02:19:57,920
on another population a lot of the

3944
02:19:55,600 --> 02:20:00,800
that you're working on

3945
02:19:57,920 --> 02:20:02,399
most of your predictors are based upon

3946
02:20:00,800 --> 02:20:06,240
the way that the healthcare

3947
02:20:02,399 --> 02:20:08,080
record is used and in practice and if

3948
02:20:06,240 --> 02:20:09,920
you have differences in that or

3949
02:20:08,080 --> 02:20:11,680
differences in the populations being

3950
02:20:09,920 --> 02:20:13,359
treated the patterns that you're going

3951
02:20:11,680 --> 02:20:15,200
to see in the data could be very very

3952
02:20:13,359 --> 02:20:17,120
different and the interpretations can be

3953
02:20:15,200 --> 02:20:20,880
very very different as well so it's

3954
02:20:17,120 --> 02:20:24,000
really important i think to revalidate

3955
02:20:20,880 --> 02:20:25,920
and and ideally build your platforms and

3956
02:20:24,000 --> 02:20:28,640
adapt them with the groups that you're

3957
02:20:25,920 --> 02:20:31,120
in right i would i would not recommend

3958
02:20:28,640 --> 02:20:33,280
just kind of plug and playing like we we

3959
02:20:31,120 --> 02:20:35,280
worry a lot people are constantly trying

3960
02:20:33,280 --> 02:20:37,840
to sell va um

3961
02:20:35,280 --> 02:20:40,000
outside built systems that they claim

3962
02:20:37,840 --> 02:20:41,600
will do all sorts of things and they

3963
02:20:40,000 --> 02:20:44,080
kind of terrify us because they're black

3964
02:20:41,600 --> 02:20:45,760
boxes we have no idea way to validate

3965
02:20:44,080 --> 02:20:48,160
them we have no idea how well they're

3966
02:20:45,760 --> 02:20:49,760
working um and

3967
02:20:48,160 --> 02:20:52,479
and we have no way to fix them if we

3968
02:20:49,760 --> 02:20:53,200
find problems with them right so i i do

3969
02:20:52,479 --> 02:20:55,760
think

3970
02:20:53,200 --> 02:20:57,680
it's really important to consider these

3971
02:20:55,760 --> 02:21:00,080
tools part of

3972
02:20:57,680 --> 02:21:02,319
the health system that you're working in

3973
02:21:00,080 --> 02:21:04,399
and adapt it just the same way that you

3974
02:21:02,319 --> 02:21:06,479
would any clinical practice pattern

3975
02:21:04,399 --> 02:21:09,120
within that system

3976
02:21:06,479 --> 02:21:10,560
i would just add on to that

3977
02:21:09,120 --> 02:21:14,000
i agree with everything you said

3978
02:21:10,560 --> 02:21:14,960
particularly in the clinical delivery

3979
02:21:14,000 --> 02:21:19,040
setting

3980
02:21:14,960 --> 02:21:21,359
for the research setting i think we can

3981
02:21:19,040 --> 02:21:23,280
find ways of harmonizing data across

3982
02:21:21,359 --> 02:21:24,399
systems to be able to do larger scale

3983
02:21:23,280 --> 02:21:26,399
research

3984
02:21:24,399 --> 02:21:29,520
i think groups like peedsnet which i'm

3985
02:21:26,399 --> 02:21:32,319
engaged with is doing that work it is

3986
02:21:29,520 --> 02:21:35,040
very expensive it's very time consuming

3987
02:21:32,319 --> 02:21:36,560
and it's not always easy to get federal

3988
02:21:35,040 --> 02:21:38,560
grants that want to fund that kind of

3989
02:21:36,560 --> 02:21:41,040
work so i'll put in a plug

3990
02:21:38,560 --> 02:21:43,520
for funding those harmonization

3991
02:21:41,040 --> 02:21:46,000
processes that allow us to do

3992
02:21:43,520 --> 02:21:48,319
large-scale harmonization across

3993
02:21:46,000 --> 02:21:50,720
uh systems i mean even if everybody uses

3994
02:21:48,319 --> 02:21:52,560
epic they still don't match so there's

3995
02:21:50,720 --> 02:21:54,640
still a lot of work that has to be done

3996
02:21:52,560 --> 02:21:56,240
and we can't just do things under one

3997
02:21:54,640 --> 02:21:59,280
vendor so

3998
02:21:56,240 --> 02:22:00,640
i we have to do the work but it is work

3999
02:21:59,280 --> 02:22:03,200
and it's time consuming and it's

4000
02:22:00,640 --> 02:22:03,200
expensive

4001
02:22:04,640 --> 02:22:09,040
okay well thank you for a really

4002
02:22:06,960 --> 02:22:10,880
interesting and thought-provoking

4003
02:22:09,040 --> 02:22:12,800
panel there are several questions in the

4004
02:22:10,880 --> 02:22:14,720
chat that we didn't really have time for

4005
02:22:12,800 --> 02:22:15,920
because our presentations just took up a

4006
02:22:14,720 --> 02:22:18,160
lot of time which was great because they

4007
02:22:15,920 --> 02:22:19,600
were so interesting um so feel free to

4008
02:22:18,160 --> 02:22:21,600
kind of continue discussion in the chat

4009
02:22:19,600 --> 02:22:24,880
we're going to go on a five minute break

4010
02:22:21,600 --> 02:22:24,880
and we'll see you all back in a few

4011
02:22:25,040 --> 02:22:30,880
welcome back everyone um

4012
02:22:27,680 --> 02:22:30,880
thanks for hanging with us

4013
02:22:31,040 --> 02:22:34,399
and we have

4014
02:22:32,640 --> 02:22:35,680
our last session

4015
02:22:34,399 --> 02:22:38,160
starting now

4016
02:22:35,680 --> 02:22:40,000
and we're going to be hearing from two

4017
02:22:38,160 --> 02:22:42,800
nida projects that are using the ehr

4018
02:22:40,000 --> 02:22:44,160
data to meet their aims so

4019
02:22:42,800 --> 02:22:45,760
our the first project that we're going

4020
02:22:44,160 --> 02:22:48,080
to hear from is one of our heel

4021
02:22:45,760 --> 02:22:51,359
prevention projects and we have both dr

4022
02:22:48,080 --> 02:22:54,319
amy ewell and dr vinod rao with us to

4023
02:22:51,359 --> 02:22:57,040
discuss their project dr amy yule is a

4024
02:22:54,319 --> 02:22:58,800
board certified in adult child and

4025
02:22:57,040 --> 02:23:00,160
addiction psychiatry

4026
02:22:58,800 --> 02:23:02,160
she is the director of adolescent

4027
02:23:00,160 --> 02:23:03,920
addiction psychiatry at boston medical

4028
02:23:02,160 --> 02:23:05,920
center and an assistant professor at the

4029
02:23:03,920 --> 02:23:08,080
brown university school of boston

4030
02:23:05,920 --> 02:23:10,080
university school of medicine sorry

4031
02:23:08,080 --> 02:23:12,479
and dr rao is an addiction psychiatrist

4032
02:23:10,080 --> 02:23:14,800
the medical director of massachusetts uh

4033
02:23:12,479 --> 02:23:16,319
general hospital's west end clinic

4034
02:23:14,800 --> 02:23:18,960
and is also an instructor at harvard

4035
02:23:16,319 --> 02:23:21,520
medical school he also works as a

4036
02:23:18,960 --> 02:23:23,040
physician informaticist for

4037
02:23:21,520 --> 02:23:24,399
mgh

4038
02:23:23,040 --> 02:23:27,680
and then um

4039
02:23:24,399 --> 02:23:30,640
after doctors yule and rao we'll turn

4040
02:23:27,680 --> 02:23:34,240
the um i'll turn it over to

4041
02:23:30,640 --> 02:23:36,160
dr sarah beale dr dr beal is a dev

4042
02:23:34,240 --> 02:23:37,200
developmental psychologist at cincinnati

4043
02:23:36,160 --> 02:23:40,479
children's

4044
02:23:37,200 --> 02:23:42,319
um scientific director of the child welf

4045
02:23:40,479 --> 02:23:43,439
child welfare research lab and associate

4046
02:23:42,319 --> 02:23:46,479
professor at the university of

4047
02:23:43,439 --> 02:23:47,920
cincinnati department of pediatrics

4048
02:23:46,479 --> 02:23:50,240
so thank you

4049
02:23:47,920 --> 02:23:52,399
to both of your teams for joining us and

4050
02:23:50,240 --> 02:23:55,439
i will turn it over

4051
02:23:52,399 --> 02:23:55,439
to amy and vinod

4052
02:23:56,880 --> 02:24:00,319
thanks sarah

4053
02:23:58,319 --> 02:24:01,920
so i'm excited to be here today to talk

4054
02:24:00,319 --> 02:24:04,080
about how we're using data from the

4055
02:24:01,920 --> 02:24:05,840
electronic health record for a project

4056
02:24:04,080 --> 02:24:08,000
treat and identify psychopathology and

4057
02:24:05,840 --> 02:24:09,040
substance use disorders and as sarah

4058
02:24:08,000 --> 02:24:10,800
noted this is part of the heal

4059
02:24:09,040 --> 02:24:14,880
prevention cooperative and i'm the mpi

4060
02:24:10,800 --> 02:24:16,399
along with dr timothy williams next line

4061
02:24:14,880 --> 02:24:18,880
and so as we go on to discuss how we're

4062
02:24:16,399 --> 02:24:20,319
using data from the ehr for our project

4063
02:24:18,880 --> 02:24:22,640
we wanted to first describe our study

4064
02:24:20,319 --> 02:24:24,080
aims and the project itself and to start

4065
02:24:22,640 --> 02:24:26,080
with um i wanted to note that our

4066
02:24:24,080 --> 02:24:28,000
projects taking place within outpatient

4067
02:24:26,080 --> 02:24:30,160
behavioral health within the outpatient

4068
02:24:28,000 --> 02:24:31,359
behavioral setting and we chose the

4069
02:24:30,160 --> 02:24:33,120
setting because because we know that

4070
02:24:31,359 --> 02:24:34,560
mental health conditions are associated

4071
02:24:33,120 --> 02:24:36,880
with increased risk for early substance

4072
02:24:34,560 --> 02:24:40,800
use including opiate misuse and

4073
02:24:36,880 --> 02:24:40,800
substance use disorders next slide

4074
02:24:40,880 --> 02:24:44,479
in our first study aim was to implement

4075
02:24:42,640 --> 02:24:46,319
measurement-based care in the behavioral

4076
02:24:44,479 --> 02:24:47,680
health setting using patient-reported

4077
02:24:46,319 --> 02:24:49,760
outcome measures that are linked to the

4078
02:24:47,680 --> 02:24:50,960
electronic health record and our goal

4079
02:24:49,760 --> 02:24:52,640
with this was really to support

4080
02:24:50,960 --> 02:24:55,120
systematic screening and monitoring of

4081
02:24:52,640 --> 02:24:57,439
substance use and psychopathology with a

4082
02:24:55,120 --> 02:24:59,680
focus on depression and anxiety in the

4083
02:24:57,439 --> 02:25:01,600
behavioral health setting

4084
02:24:59,680 --> 02:25:03,439
as part of this we really focused on

4085
02:25:01,600 --> 02:25:05,040
four clinics at mass general hospital

4086
02:25:03,439 --> 02:25:06,319
and then two clinics at boston medical

4087
02:25:05,040 --> 02:25:08,560
center

4088
02:25:06,319 --> 02:25:10,880
within these four clinics at mgh two of

4089
02:25:08,560 --> 02:25:12,880
these are a child and adult behavioral

4090
02:25:10,880 --> 02:25:14,479
health clinic and then two are

4091
02:25:12,880 --> 02:25:16,000
substance use disorder clinics that are

4092
02:25:14,479 --> 02:25:17,920
within the department of psychiatry

4093
02:25:16,000 --> 02:25:20,080
focused on care for youth and then

4094
02:25:17,920 --> 02:25:22,319
another one on care for adults next

4095
02:25:20,080 --> 02:25:22,319
slide

4096
02:25:22,840 --> 02:25:26,399
next

4097
02:25:24,640 --> 02:25:28,080
and then our second study aim is really

4098
02:25:26,399 --> 02:25:30,240
to be assessing the impact of treatment

4099
02:25:28,080 --> 02:25:32,000
of psychopathology on the development of

4100
02:25:30,240 --> 02:25:34,240
an opiate use disorder substance use

4101
02:25:32,000 --> 02:25:36,800
disorder and nicotine use disorder in

4102
02:25:34,240 --> 02:25:38,479
young people 16 to 30 years of age who

4103
02:25:36,800 --> 02:25:40,960
are engaged in behavioral health

4104
02:25:38,479 --> 02:25:42,640
treatment and we're really doing this

4105
02:25:40,960 --> 02:25:45,200
through two different ways one is a

4106
02:25:42,640 --> 02:25:46,800
de-identified pull of longitudinal data

4107
02:25:45,200 --> 02:25:49,760
from the ehr within our larger

4108
02:25:46,800 --> 02:25:51,600
healthcare system of 16 to 30 year olds

4109
02:25:49,760 --> 02:25:53,200
and then secondly we're doing

4110
02:25:51,600 --> 02:25:55,520
we're recruiting from the six clinics

4111
02:25:53,200 --> 02:25:57,359
that i mentioned before um patients to

4112
02:25:55,520 --> 02:25:59,680
participate in a naturalistic

4113
02:25:57,359 --> 02:26:01,359
longitudinal study that involves self

4114
02:25:59,680 --> 02:26:04,080
self assessments that are done every six

4115
02:26:01,359 --> 02:26:06,399
months and then data pulled from the ehr

4116
02:26:04,080 --> 02:26:08,720
next slide

4117
02:26:06,399 --> 02:26:10,240
and as we think about our the second aim

4118
02:26:08,720 --> 02:26:11,920
you know really evaluating the impact of

4119
02:26:10,240 --> 02:26:13,680
treating psychopathology on the

4120
02:26:11,920 --> 02:26:15,040
development of an opiate use disorder

4121
02:26:13,680 --> 02:26:16,720
there's a lot of information that's

4122
02:26:15,040 --> 02:26:18,479
within the hr that can help answer our

4123
02:26:16,720 --> 02:26:20,960
research questions and we've talked

4124
02:26:18,479 --> 02:26:23,120
about these types of data um kind of

4125
02:26:20,960 --> 02:26:25,280
earlier on in the presentations and and

4126
02:26:23,120 --> 02:26:27,200
so i'm going to really focus um for my

4127
02:26:25,280 --> 02:26:29,120
small section a little bit on patient

4128
02:26:27,200 --> 02:26:30,479
reported outcome measures which are one

4129
02:26:29,120 --> 02:26:32,720
way that we can really get at severity

4130
02:26:30,479 --> 02:26:34,160
of illness which may may be a really

4131
02:26:32,720 --> 02:26:36,240
important factor as we think about the

4132
02:26:34,160 --> 02:26:38,800
impact of treatment on someone's risk of

4133
02:26:36,240 --> 02:26:40,880
developing an opiate use disorder um and

4134
02:26:38,800 --> 02:26:42,479
as we noted before you know the free

4135
02:26:40,880 --> 02:26:44,240
text data that's available in the

4136
02:26:42,479 --> 02:26:46,960
clinical note there's often a lot of

4137
02:26:44,240 --> 02:26:49,600
rich data in the clinical note um but

4138
02:26:46,960 --> 02:26:51,680
it's hard to um to make that systematic

4139
02:26:49,600 --> 02:26:53,439
or to quantify it and that's where

4140
02:26:51,680 --> 02:26:56,560
future reported outcome measures have a

4141
02:26:53,439 --> 02:26:58,399
lot of promise next slide

4142
02:26:56,560 --> 02:27:00,399
so that being said these reported

4143
02:26:58,399 --> 02:27:01,520
outcome measures don't end up in the ehr

4144
02:27:00,399 --> 02:27:03,200
magically

4145
02:27:01,520 --> 02:27:05,040
it's really they're part of a process of

4146
02:27:03,200 --> 02:27:06,640
measurement-based care and the idea with

4147
02:27:05,040 --> 02:27:08,000
measurement based here is that we're

4148
02:27:06,640 --> 02:27:10,000
routinely

4149
02:27:08,000 --> 02:27:11,840
and systematically administering a

4150
02:27:10,000 --> 02:27:13,920
patient-reported outcome as your tool to

4151
02:27:11,840 --> 02:27:16,080
assess patient symptoms and that the

4152
02:27:13,920 --> 02:27:18,080
patients receiving the tool completing

4153
02:27:16,080 --> 02:27:19,520
the tool and then the information from

4154
02:27:18,080 --> 02:27:20,640
this questionnaire is getting into the

4155
02:27:19,520 --> 02:27:22,560
hr

4156
02:27:20,640 --> 02:27:24,880
and then within the hr the provider has

4157
02:27:22,560 --> 02:27:27,760
to find this information and then review

4158
02:27:24,880 --> 02:27:29,359
it and then review it with the patient

4159
02:27:27,760 --> 02:27:31,600
it's really this process of the provider

4160
02:27:29,359 --> 02:27:34,080
and patient reviewing the data together

4161
02:27:31,600 --> 02:27:35,600
that has an impact on care and they can

4162
02:27:34,080 --> 02:27:36,960
use this data to think about the

4163
02:27:35,600 --> 02:27:38,960
treatment plan

4164
02:27:36,960 --> 02:27:40,399
and next steps within treatment

4165
02:27:38,960 --> 02:27:42,640
and what we really know from previous

4166
02:27:40,399 --> 02:27:43,920
research is that to be successful in

4167
02:27:42,640 --> 02:27:46,000
collecting patient-reported outcome

4168
02:27:43,920 --> 02:27:47,760
measures over time

4169
02:27:46,000 --> 02:27:49,680
this really has to be integrated into

4170
02:27:47,760 --> 02:27:51,760
clinical care in a meaningful way for

4171
02:27:49,680 --> 02:27:53,840
both patients and for

4172
02:27:51,760 --> 02:27:55,520
for the clinician so if the patient

4173
02:27:53,840 --> 02:27:57,280
completes the outcome measure and the

4174
02:27:55,520 --> 02:27:59,760
clinician ever reviews it they're going

4175
02:27:57,280 --> 02:28:01,359
to stop completing this outcome measure

4176
02:27:59,760 --> 02:28:04,080
and if the patient the provider can't

4177
02:28:01,359 --> 02:28:05,840
find the information within the ehr or

4178
02:28:04,080 --> 02:28:07,439
you know doesn't view the information to

4179
02:28:05,840 --> 02:28:08,960
be helpful for the the clinical

4180
02:28:07,439 --> 02:28:12,800
encounter they're going to also stop

4181
02:28:08,960 --> 02:28:13,920
using it as well so next slide

4182
02:28:12,800 --> 02:28:15,280
so as we think about kind of

4183
02:28:13,920 --> 02:28:16,640
measurement-based care implementation

4184
02:28:15,280 --> 02:28:18,560
which we've been working on over the

4185
02:28:16,640 --> 02:28:19,920
past um two to three years there you

4186
02:28:18,560 --> 02:28:22,160
know are many opportunities and

4187
02:28:19,920 --> 02:28:23,840
challenges and um and so as we think

4188
02:28:22,160 --> 02:28:25,760
about the behavioral health setting you

4189
02:28:23,840 --> 02:28:27,439
know one challenge is that we don't have

4190
02:28:25,760 --> 02:28:28,960
ancillary supports you know we don't

4191
02:28:27,439 --> 02:28:30,640
have a medical assistant who's checking

4192
02:28:28,960 --> 02:28:32,479
in patients and getting vital signs and

4193
02:28:30,640 --> 02:28:34,560
also looking at you know did a patient

4194
02:28:32,479 --> 02:28:37,359
complete this outcome measure um

4195
02:28:34,560 --> 02:28:38,720
completely and and kind of um correctly

4196
02:28:37,359 --> 02:28:40,640
and then you know helping to put the

4197
02:28:38,720 --> 02:28:42,479
information into the ehr

4198
02:28:40,640 --> 02:28:45,120
and so at mgh with a larger

4199
02:28:42,479 --> 02:28:46,880
infrastructure support pre-covered um

4200
02:28:45,120 --> 02:28:48,319
did have patients completing these

4201
02:28:46,880 --> 02:28:50,880
questionnaires on tablets that were

4202
02:28:48,319 --> 02:28:53,280
automatically synced to the ehr

4203
02:28:50,880 --> 02:28:55,040
and then with kovid had this opportunity

4204
02:28:53,280 --> 02:28:57,920
and challenge where for everything

4205
02:28:55,040 --> 02:28:59,920
switched to being done over telehealth

4206
02:28:57,920 --> 02:29:01,760
and so at mgh most of these visits were

4207
02:28:59,920 --> 02:29:03,920
done through a patient portal um and

4208
02:29:01,760 --> 02:29:06,000
there was because these um patient

4209
02:29:03,920 --> 02:29:08,000
report outcome measures were already um

4210
02:29:06,000 --> 02:29:09,840
built into the patient portal it was

4211
02:29:08,000 --> 02:29:11,359
fairly seamless to transition to

4212
02:29:09,840 --> 02:29:13,040
completing these questionnaires as part

4213
02:29:11,359 --> 02:29:15,040
of telehealth visits

4214
02:29:13,040 --> 02:29:16,800
um that was not the case at boston

4215
02:29:15,040 --> 02:29:18,880
medical center so boston medical center

4216
02:29:16,800 --> 02:29:20,399
is only two miles away from mgh but it's

4217
02:29:18,880 --> 02:29:22,479
a very different patient population

4218
02:29:20,399 --> 02:29:24,880
where we have a much larger

4219
02:29:22,479 --> 02:29:26,880
medicaid population um with different

4220
02:29:24,880 --> 02:29:29,120
infusions with um that are often

4221
02:29:26,880 --> 02:29:30,880
multi-level and so as we think about

4222
02:29:29,120 --> 02:29:32,399
patient population this is also really

4223
02:29:30,880 --> 02:29:34,319
important and particularly as we think

4224
02:29:32,399 --> 02:29:35,760
about administering administering

4225
02:29:34,319 --> 02:29:38,080
patient-reported outcome measures

4226
02:29:35,760 --> 02:29:40,640
through patient portals um

4227
02:29:38,080 --> 02:29:42,560
since we saw as we switched to doing

4228
02:29:40,640 --> 02:29:45,040
telehealth within behavioral health at

4229
02:29:42,560 --> 02:29:47,280
both institutions we saw significantly

4230
02:29:45,040 --> 02:29:49,200
less engagement with the patient portal

4231
02:29:47,280 --> 02:29:51,920
among behavioral health patients at

4232
02:29:49,200 --> 02:29:53,920
boston medical center so only 65 of

4233
02:29:51,920 --> 02:29:55,280
patients um you know kind of at the

4234
02:29:53,920 --> 02:29:57,920
beginning of the pandemic and then kind

4235
02:29:55,280 --> 02:30:00,640
of throughout within behavioral health

4236
02:29:57,920 --> 02:30:02,080
have an active patient portal

4237
02:30:00,640 --> 02:30:04,479
the other factor is thinking about

4238
02:30:02,080 --> 02:30:06,000
language are these measures um

4239
02:30:04,479 --> 02:30:07,439
translated into different languages and

4240
02:30:06,000 --> 02:30:09,520
available in different languages through

4241
02:30:07,439 --> 02:30:11,040
the ehr and then the patient portals

4242
02:30:09,520 --> 02:30:12,240
themselves are often

4243
02:30:11,040 --> 02:30:14,000
generally

4244
02:30:12,240 --> 02:30:15,840
more routinely available in english so

4245
02:30:14,000 --> 02:30:18,319
they're not they're not all translated

4246
02:30:15,840 --> 02:30:19,840
or easily translated um and so i think

4247
02:30:18,319 --> 02:30:22,319
another thing to be thinking about as we

4248
02:30:19,840 --> 02:30:25,359
think about equitable screening um is

4249
02:30:22,319 --> 02:30:27,280
that um the kind of administering these

4250
02:30:25,359 --> 02:30:29,120
problems through the patient portal and

4251
02:30:27,280 --> 02:30:31,280
this came up within the the master

4252
02:30:29,120 --> 02:30:33,600
general hospital system the larger

4253
02:30:31,280 --> 02:30:35,520
system where they saw that when they

4254
02:30:33,600 --> 02:30:37,520
looked at race and ethnicity there were

4255
02:30:35,520 --> 02:30:39,359
no differences in patient reported

4256
02:30:37,520 --> 02:30:41,760
outcome completion when done in the

4257
02:30:39,359 --> 02:30:43,359
office on these tablets but when they

4258
02:30:41,760 --> 02:30:45,200
switched to doing these patient reported

4259
02:30:43,359 --> 02:30:47,200
outcome measures across the whole system

4260
02:30:45,200 --> 02:30:50,080
in different departments um through the

4261
02:30:47,200 --> 02:30:52,399
patient portal saw a very steep decline

4262
02:30:50,080 --> 02:30:55,280
in um completion rates by individuals

4263
02:30:52,399 --> 02:30:57,200
who were black or hispanic latino um so

4264
02:30:55,280 --> 02:30:59,520
again thinking about equity is really

4265
02:30:57,200 --> 02:31:01,120
important as we're thinking about

4266
02:30:59,520 --> 02:31:03,040
implementing measurement-based care and

4267
02:31:01,120 --> 02:31:05,680
collecting this type of really important

4268
02:31:03,040 --> 02:31:07,359
data as we think about data available to

4269
02:31:05,680 --> 02:31:08,640
us through the ehr

4270
02:31:07,359 --> 02:31:10,080
and then lastly this has been talked

4271
02:31:08,640 --> 02:31:12,160
about quite a bit is just really

4272
02:31:10,080 --> 02:31:15,120
thinking about the ehr build so these

4273
02:31:12,160 --> 02:31:16,319
are not all the same ehr systems um i

4274
02:31:15,120 --> 02:31:18,240
think it's kind of like thinking about

4275
02:31:16,319 --> 02:31:19,840
like we all have a ford um and there are

4276
02:31:18,240 --> 02:31:21,600
many different types of models of fords

4277
02:31:19,840 --> 02:31:23,439
or different types of cars

4278
02:31:21,600 --> 02:31:26,080
and then also across systems when you're

4279
02:31:23,439 --> 02:31:28,240
outside of these really large systems um

4280
02:31:26,080 --> 02:31:30,080
the it infrastructure support is is

4281
02:31:28,240 --> 02:31:31,760
quite variable um and this can be a

4282
02:31:30,080 --> 02:31:34,560
really important factor to be thinking

4283
02:31:31,760 --> 02:31:36,319
about when we're trying to integrate um

4284
02:31:34,560 --> 02:31:37,600
these type of things into clinical care

4285
02:31:36,319 --> 02:31:40,160
into research

4286
02:31:37,600 --> 02:31:42,399
so next time i'm going to switch over to

4287
02:31:40,160 --> 02:31:45,280
now to dr rao

4288
02:31:42,399 --> 02:31:46,640
all right um sorry about that i will be

4289
02:31:45,280 --> 02:31:48,240
so um

4290
02:31:46,640 --> 02:31:51,200
thank you dr yul i will be discussing

4291
02:31:48,240 --> 02:31:53,520
some of the practical aspects of ehr

4292
02:31:51,200 --> 02:31:55,600
study design both generally and in the

4293
02:31:53,520 --> 02:31:57,359
process and the process in the context

4294
02:31:55,600 --> 02:31:58,880
of some of the projects that we have

4295
02:31:57,359 --> 02:32:00,000
been working on

4296
02:31:58,880 --> 02:32:02,000
so

4297
02:32:00,000 --> 02:32:03,680
um as you've heard um there are a couple

4298
02:32:02,000 --> 02:32:06,399
of different places where information

4299
02:32:03,680 --> 02:32:08,880
gets put into the ehr often times by

4300
02:32:06,399 --> 02:32:11,439
administrators non-clinical staff

4301
02:32:08,880 --> 02:32:12,880
um through

4302
02:32:11,439 --> 02:32:14,399
various opportunities to put information

4303
02:32:12,880 --> 02:32:16,000
into coded fields

4304
02:32:14,399 --> 02:32:18,080
and

4305
02:32:16,000 --> 02:32:21,840
in the context of measurement-based care

4306
02:32:18,080 --> 02:32:23,840
of patients to indirectly through these

4307
02:32:21,840 --> 02:32:25,680
questionnaires and then clinicians do

4308
02:32:23,840 --> 02:32:26,960
through these clinical notes and

4309
02:32:25,680 --> 02:32:29,040
what clinicians are interested in is

4310
02:32:26,960 --> 02:32:31,040
getting their work done accurately and

4311
02:32:29,040 --> 02:32:33,200
quickly um so they'll write these notes

4312
02:32:31,040 --> 02:32:36,800
for regulatory compliance for insurance

4313
02:32:33,200 --> 02:32:39,680
for colleagues in the era of open notes

4314
02:32:36,800 --> 02:32:39,930
for patients and what this means is that

4315
02:32:39,680 --> 02:32:42,160
we

4316
02:32:39,930 --> 02:32:44,000
[Music]

4317
02:32:42,160 --> 02:32:47,920
the clinical documentation does not

4318
02:32:44,000 --> 02:32:49,840
yield high fidelity data sets

4319
02:32:47,920 --> 02:32:52,000
and that's something that we need to

4320
02:32:49,840 --> 02:32:54,399
live with and keep in mind as we're

4321
02:32:52,000 --> 02:32:55,359
designing these studies so the variables

4322
02:32:54,399 --> 02:32:56,350
that we're

4323
02:32:55,359 --> 02:32:57,920
using

4324
02:32:56,350 --> 02:32:59,120
[Music]

4325
02:32:57,920 --> 02:33:01,280
need to

4326
02:32:59,120 --> 02:33:02,800
act to be appropriate

4327
02:33:01,280 --> 02:33:04,720
we'll need to actually capture the

4328
02:33:02,800 --> 02:33:07,120
concept of interest they have to be

4329
02:33:04,720 --> 02:33:10,000
available and reliable in order to

4330
02:33:07,120 --> 02:33:12,720
effectively build a model um to

4331
02:33:10,000 --> 02:33:15,120
assess or follow the clinical question

4332
02:33:12,720 --> 02:33:16,479
so these use so the features of a useful

4333
02:33:15,120 --> 02:33:17,439
variable

4334
02:33:16,479 --> 02:33:18,240
um

4335
02:33:17,439 --> 02:33:20,399
in

4336
02:33:18,240 --> 02:33:23,200
our preliminary experiences

4337
02:33:20,399 --> 02:33:25,359
living in coded fields

4338
02:33:23,200 --> 02:33:26,960
the value should be ideally discrete or

4339
02:33:25,359 --> 02:33:29,359
numerical

4340
02:33:26,960 --> 02:33:32,000
reliably updated by

4341
02:33:29,359 --> 02:33:33,200
clinicians or patients as soon as they

4342
02:33:32,000 --> 02:33:35,200
change

4343
02:33:33,200 --> 02:33:39,520
and should be carefully and accurately

4344
02:33:35,200 --> 02:33:39,520
described by everyone that's doing this

4345
02:33:41,600 --> 02:33:46,319
and so as a result of this not all

4346
02:33:44,000 --> 02:33:49,359
variables end up being encoded equally

4347
02:33:46,319 --> 02:33:51,760
some of them are pretty reliable age sex

4348
02:33:49,359 --> 02:33:53,439
is pretty straightforward race you would

4349
02:33:51,760 --> 02:33:55,439
think would be and often is but it

4350
02:33:53,439 --> 02:33:58,880
depends on the institution whether

4351
02:33:55,439 --> 02:34:00,800
um how accurately that might be captured

4352
02:33:58,880 --> 02:34:02,960
some employment information billing

4353
02:34:00,800 --> 02:34:05,040
codes visit diagnosis these are there

4354
02:34:02,960 --> 02:34:06,000
but may not map onto the concept you

4355
02:34:05,040 --> 02:34:07,680
care about

4356
02:34:06,000 --> 02:34:10,319
ed visits and hospitalizations are

4357
02:34:07,680 --> 02:34:12,960
reliable within that system and

4358
02:34:10,319 --> 02:34:15,359
insurance information is often accurate

4359
02:34:12,960 --> 02:34:17,760
other kinds of data are

4360
02:34:15,359 --> 02:34:19,760
are possible to

4361
02:34:17,760 --> 02:34:21,359
extract but require a little bit of

4362
02:34:19,760 --> 02:34:22,960
computation a little bit of work in

4363
02:34:21,359 --> 02:34:24,560
order to turn it into a useful variable

4364
02:34:22,960 --> 02:34:26,000
for your analysis

4365
02:34:24,560 --> 02:34:27,439
prompts questionnaires we'll see an

4366
02:34:26,000 --> 02:34:28,560
example in a moment

4367
02:34:27,439 --> 02:34:29,840
of

4368
02:34:28,560 --> 02:34:31,520
some of the work that needs to be done

4369
02:34:29,840 --> 02:34:34,640
in order to make use of questionnaire

4370
02:34:31,520 --> 02:34:36,000
data similarly toxicology screens sounds

4371
02:34:34,640 --> 02:34:38,080
straightforward but there are a lot of

4372
02:34:36,000 --> 02:34:39,280
talk screens you need to think about how

4373
02:34:38,080 --> 02:34:40,800
each one gets

4374
02:34:39,280 --> 02:34:42,960
interpreted

4375
02:34:40,800 --> 02:34:44,479
similarly with medications

4376
02:34:42,960 --> 02:34:47,359
it's not just the medication it's the

4377
02:34:44,479 --> 02:34:49,600
strength it's the um dosing instruction

4378
02:34:47,359 --> 02:34:52,399
it's when it was prescribed

4379
02:34:49,600 --> 02:34:54,000
um overdoses can be

4380
02:34:52,399 --> 02:34:56,160
can be extracted from the record but

4381
02:34:54,000 --> 02:34:57,680
maybe available in different places and

4382
02:34:56,160 --> 02:34:59,359
the comorbidities there may need to be

4383
02:34:57,680 --> 02:35:01,680
lumping of information

4384
02:34:59,359 --> 02:35:03,200
um in order to get uh

4385
02:35:01,680 --> 02:35:05,359
really diverse

4386
02:35:03,200 --> 02:35:07,120
diagnosis into something more tractable

4387
02:35:05,359 --> 02:35:08,960
for analysis

4388
02:35:07,120 --> 02:35:11,200
and then some information is

4389
02:35:08,960 --> 02:35:13,520
challenging education marital status

4390
02:35:11,200 --> 02:35:14,560
employment homelessness these are

4391
02:35:13,520 --> 02:35:16,880
pretty

4392
02:35:14,560 --> 02:35:19,760
basic concepts which are not necessarily

4393
02:35:16,880 --> 02:35:20,880
reliably updated or stored in coded

4394
02:35:19,760 --> 02:35:22,560
fields

4395
02:35:20,880 --> 02:35:24,080
remission status is extremely important

4396
02:35:22,560 --> 02:35:26,560
for substance use care

4397
02:35:24,080 --> 02:35:28,240
um and yet maybe embedded within the

4398
02:35:26,560 --> 02:35:30,560
note it may not always be reliably

4399
02:35:28,240 --> 02:35:32,000
updated in the physics diagnosis

4400
02:35:30,560 --> 02:35:34,240
similarly that some of the types of the

4401
02:35:32,000 --> 02:35:36,800
interventions especially psychotherapies

4402
02:35:34,240 --> 02:35:38,560
are not um reliably encoded in coded

4403
02:35:36,800 --> 02:35:41,200
fields

4404
02:35:38,560 --> 02:35:44,319
so with that said let's um walk through

4405
02:35:41,200 --> 02:35:46,160
what the process of um data in getting

4406
02:35:44,319 --> 02:35:48,080
putting data into the system and turning

4407
02:35:46,160 --> 02:35:48,880
it into analysis can look like

4408
02:35:48,080 --> 02:35:51,040
so

4409
02:35:48,880 --> 02:35:53,439
in this case dr young and i both exist

4410
02:35:51,040 --> 02:35:55,520
within epic ecosystems so when a

4411
02:35:53,439 --> 02:35:57,840
clinician or administrator puts

4412
02:35:55,520 --> 02:35:58,880
input into epic in real time it goes

4413
02:35:57,840 --> 02:36:01,439
into a

4414
02:35:58,880 --> 02:36:04,479
server called the epic calls chronicles

4415
02:36:01,439 --> 02:36:05,920
and at least in our system every night a

4416
02:36:04,479 --> 02:36:09,040
large amount of this

4417
02:36:05,920 --> 02:36:10,960
gets exported to a relational database

4418
02:36:09,040 --> 02:36:13,680
relational database called enterpr what

4419
02:36:10,960 --> 02:36:15,280
we call enterprise data warehouse and it

4420
02:36:13,680 --> 02:36:16,720
was sort of

4421
02:36:15,280 --> 02:36:19,200
inappropriate to represent this as a

4422
02:36:16,720 --> 02:36:21,120
spreadsheet but in reality it's a whole

4423
02:36:19,200 --> 02:36:22,160
bunch of tables

4424
02:36:21,120 --> 02:36:24,399
in

4425
02:36:22,160 --> 02:36:26,160
this fictional example there is

4426
02:36:24,399 --> 02:36:28,560
patient information

4427
02:36:26,160 --> 02:36:30,560
in one table medication information

4428
02:36:28,560 --> 02:36:31,600
another table appointment information

4429
02:36:30,560 --> 02:36:33,600
the third

4430
02:36:31,600 --> 02:36:35,520
and what allows people to connect

4431
02:36:33,600 --> 02:36:37,760
information across these tables is the

4432
02:36:35,520 --> 02:36:40,000
existence of these common columns

4433
02:36:37,760 --> 02:36:42,160
and because of that you might be able to

4434
02:36:40,000 --> 02:36:43,359
see that oh this patient martin

4435
02:36:42,160 --> 02:36:45,520
odergaard

4436
02:36:43,359 --> 02:36:47,680
had a

4437
02:36:45,520 --> 02:36:49,520
medication of listening to 10 milligrams

4438
02:36:47,680 --> 02:36:52,720
at some point and had an appointment on

4439
02:36:49,520 --> 02:36:54,359
january 11 2019.

4440
02:36:52,720 --> 02:36:56,399
now that's the drastic

4441
02:36:54,359 --> 02:36:58,560
oversimplification um

4442
02:36:56,399 --> 02:37:01,040
each institution may have

4443
02:36:58,560 --> 02:37:02,479
variants of this kind of data model and

4444
02:37:01,040 --> 02:37:04,960
this data model

4445
02:37:02,479 --> 02:37:07,280
may have not three but rather dozens if

4446
02:37:04,960 --> 02:37:10,720
not hundreds of tables and you need to

4447
02:37:07,280 --> 02:37:10,720
have a map to navigate these

4448
02:37:10,880 --> 02:37:14,880
in order to extract the data um from a

4449
02:37:13,280 --> 02:37:16,319
general relational database the most

4450
02:37:14,880 --> 02:37:18,960
common tool is something called

4451
02:37:16,319 --> 02:37:22,000
structured querying language

4452
02:37:18,960 --> 02:37:23,520
or institutions provide tools that

4453
02:37:22,000 --> 02:37:25,200
allow you to

4454
02:37:23,520 --> 02:37:26,800
extract this data

4455
02:37:25,200 --> 02:37:28,960
after navigating that then you can use

4456
02:37:26,800 --> 02:37:32,319
your analysis tool of choice

4457
02:37:28,960 --> 02:37:34,240
our status python to do your statistics

4458
02:37:32,319 --> 02:37:36,399
or visualizations

4459
02:37:34,240 --> 02:37:38,399
so let's look at an example

4460
02:37:36,399 --> 02:37:41,600
we were interested in

4461
02:37:38,399 --> 02:37:43,840
the clinical question does treating adhd

4462
02:37:41,600 --> 02:37:45,520
with medications improve retention in a

4463
02:37:43,840 --> 02:37:47,600
dual diagnosis clinic

4464
02:37:45,520 --> 02:37:50,560
let's talk about how we concept how we

4465
02:37:47,600 --> 02:37:53,760
operationalize some of these concepts

4466
02:37:50,560 --> 02:37:55,680
so the notion of a clinic we capture as

4467
02:37:53,760 --> 02:37:57,200
the appointment department id and it's

4468
02:37:55,680 --> 02:37:58,880
pretty straightforward to filter based

4469
02:37:57,200 --> 02:38:01,920
on

4470
02:37:58,880 --> 02:38:03,120
happy to jump in here and what uh vinod

4471
02:38:01,920 --> 02:38:04,640
was saying is

4472
02:38:03,120 --> 02:38:06,000
uh as we were

4473
02:38:04,640 --> 02:38:08,080
lining up to

4474
02:38:06,000 --> 02:38:09,600
just look at this issue this sort of

4475
02:38:08,080 --> 02:38:10,800
straightforward issue

4476
02:38:09,600 --> 02:38:13,040
we realized that there are certain

4477
02:38:10,800 --> 02:38:14,800
things that were relatively easy and

4478
02:38:13,040 --> 02:38:16,720
predictable and reliable

4479
02:38:14,800 --> 02:38:18,319
and this was a project that we did very

4480
02:38:16,720 --> 02:38:20,080
early because we hadn't gotten other

4481
02:38:18,319 --> 02:38:21,840
information yet because of the

4482
02:38:20,080 --> 02:38:22,960
implementation of the prompts that amy

4483
02:38:21,840 --> 02:38:24,640
discussed

4484
02:38:22,960 --> 02:38:26,960
and here you can see

4485
02:38:24,640 --> 02:38:30,479
the things i'm not going to repeat what

4486
02:38:26,960 --> 02:38:32,080
benad was saying but the bottom line is

4487
02:38:30,479 --> 02:38:34,080
we in here

4488
02:38:32,080 --> 02:38:35,600
looked at a thing such as medication

4489
02:38:34,080 --> 02:38:37,680
which was very important because we were

4490
02:38:35,600 --> 02:38:40,720
looking at the question of medication

4491
02:38:37,680 --> 02:38:42,479
for adhd and then we looked at the time

4492
02:38:40,720 --> 02:38:45,120
covariate about when did you get the

4493
02:38:42,479 --> 02:38:47,120
medication which became more complicated

4494
02:38:45,120 --> 02:38:49,120
we looked at medication

4495
02:38:47,120 --> 02:38:50,240
yes no and then we looked at medication

4496
02:38:49,120 --> 02:38:52,319
type and

4497
02:38:50,240 --> 02:38:54,000
we you know had to filter it by

4498
02:38:52,319 --> 02:38:56,399
stimulant yes no

4499
02:38:54,000 --> 02:38:57,680
or nonstimulant which was almost always

4500
02:38:56,399 --> 02:38:59,840
because this is an adult clinic

4501
02:38:57,680 --> 02:39:01,280
atomoxetine just to give people people

4502
02:38:59,840 --> 02:39:03,120
have clinical side

4503
02:39:01,280 --> 02:39:05,439
um and then the retention

4504
02:39:03,120 --> 02:39:07,200
uh was trying to we were modeling

4505
02:39:05,439 --> 02:39:08,720
survival curves to take a look at this

4506
02:39:07,200 --> 02:39:11,760
and we were looking at when they dropped

4507
02:39:08,720 --> 02:39:13,760
out and then had to uh looked at the

4508
02:39:11,760 --> 02:39:15,520
best of your ability and one of the

4509
02:39:13,760 --> 02:39:17,439
issues you run into if somebody drops

4510
02:39:15,520 --> 02:39:19,040
out and then comes back there's all

4511
02:39:17,439 --> 02:39:20,720
these rules you have to define when

4512
02:39:19,040 --> 02:39:24,760
you're coming up with this

4513
02:39:20,720 --> 02:39:24,760
um if you can go to the next

4514
02:39:26,720 --> 02:39:33,520
sorry about that so um the the

4515
02:39:30,479 --> 02:39:36,479
end result is when one goes through the

4516
02:39:33,520 --> 02:39:38,479
steps of um filtering the data as such

4517
02:39:36,479 --> 02:39:40,960
we can produce these survival analyses

4518
02:39:38,479 --> 02:39:43,120
um looking at when do patients

4519
02:39:40,960 --> 02:39:45,359
um drop out of the clinic

4520
02:39:43,120 --> 02:39:47,520
and as a function of whether they've

4521
02:39:45,359 --> 02:39:49,520
been prescribed medications to treat

4522
02:39:47,520 --> 02:39:51,200
adhd the

4523
02:39:49,520 --> 02:39:52,720
blue curve

4524
02:39:51,200 --> 02:39:53,760
indicates that the patient should have

4525
02:39:52,720 --> 02:39:56,240
had

4526
02:39:53,760 --> 02:39:57,840
medications for adhd and found that

4527
02:39:56,240 --> 02:39:59,200
there is uh that will be higher

4528
02:39:57,840 --> 02:40:01,520
retention and treatment

4529
02:39:59,200 --> 02:40:03,920
um for those patients unlike those who

4530
02:40:01,520 --> 02:40:05,680
do not receive um this data now note

4531
02:40:03,920 --> 02:40:06,880
this isn't really big data this is a

4532
02:40:05,680 --> 02:40:09,520
relatively

4533
02:40:06,880 --> 02:40:11,680
small sample size and one is able to

4534
02:40:09,520 --> 02:40:15,200
extract these

4535
02:40:11,680 --> 02:40:18,680
notable signals from this data

4536
02:40:15,200 --> 02:40:18,680
next slide please

4537
02:40:20,880 --> 02:40:24,880
so um

4538
02:40:23,120 --> 02:40:26,319
the

4539
02:40:24,880 --> 02:40:28,880
the next um

4540
02:40:26,319 --> 02:40:30,720
thing that i wanted to discuss was how

4541
02:40:28,880 --> 02:40:32,160
might um

4542
02:40:30,720 --> 02:40:35,680
measurement based care enriched data

4543
02:40:32,160 --> 02:40:37,439
from the ehr so as part of this we were

4544
02:40:35,680 --> 02:40:39,120
interested in asking the question are

4545
02:40:37,439 --> 02:40:41,520
people with depression more likely to

4546
02:40:39,120 --> 02:40:43,840
develop a substance use disorder

4547
02:40:41,520 --> 02:40:45,840
and we took a first crack at this um

4548
02:40:43,840 --> 02:40:49,760
which is sort of iconically

4549
02:40:45,840 --> 02:40:52,000
depicted on the right but our um

4550
02:40:49,760 --> 02:40:53,920
sense was that 23 of patients with

4551
02:40:52,000 --> 02:40:55,920
depression subsequently developed a

4552
02:40:53,920 --> 02:40:56,960
substance use disorder

4553
02:40:55,920 --> 02:40:58,880
and

4554
02:40:56,960 --> 02:41:01,920
while that was interesting these

4555
02:40:58,880 --> 02:41:04,160
diagnoses were based on basic diagnosis

4556
02:41:01,920 --> 02:41:05,359
but measurement based care allows us to

4557
02:41:04,160 --> 02:41:06,720
have a lot of information from

4558
02:41:05,359 --> 02:41:09,920
questionnaires

4559
02:41:06,720 --> 02:41:13,279
so how can that be incorporated into the

4560
02:41:09,920 --> 02:41:14,960
analysis next slide please

4561
02:41:13,279 --> 02:41:17,120
so here's what some of the challenges

4562
02:41:14,960 --> 02:41:19,120
come up when we try to do that um here's

4563
02:41:17,120 --> 02:41:22,160
a screenshot of um some of the

4564
02:41:19,120 --> 02:41:24,080
individual question questions that are

4565
02:41:22,160 --> 02:41:25,840
asked to the patient so if you look at

4566
02:41:24,080 --> 02:41:28,319
the third one how often do you have a

4567
02:41:25,840 --> 02:41:29,760
drink containing alcohol

4568
02:41:28,319 --> 02:41:31,600
many of you in these in this audience

4569
02:41:29,760 --> 02:41:34,640
will answer something non-zero just two

4570
02:41:31,600 --> 02:41:37,520
to three times per week um is not

4571
02:41:34,640 --> 02:41:39,520
necessarily indicative of a problem

4572
02:41:37,520 --> 02:41:41,200
however look at the first question how

4573
02:41:39,520 --> 02:41:42,960
many times in the past year have you

4574
02:41:41,200 --> 02:41:46,000
used an illegal drug or prescription

4575
02:41:42,960 --> 02:41:48,000
medication for non-medical reasons

4576
02:41:46,000 --> 02:41:49,840
answers that are different from xero may

4577
02:41:48,000 --> 02:41:51,600
actually uh

4578
02:41:49,840 --> 02:41:53,760
be recently interpreted as a problem and

4579
02:41:51,600 --> 02:41:55,600
should be considered as such

4580
02:41:53,760 --> 02:41:57,760
each question needs to have its own

4581
02:41:55,600 --> 02:41:59,359
threshold scent and we need sort of

4582
02:41:57,760 --> 02:42:01,520
clinical expertise in order to develop a

4583
02:41:59,359 --> 02:42:04,479
mapping in order to appropriately use

4584
02:42:01,520 --> 02:42:06,960
these questionnaire data um and use it

4585
02:42:04,479 --> 02:42:08,399
to help define an outcome that you want

4586
02:42:06,960 --> 02:42:10,640
to use in your

4587
02:42:08,399 --> 02:42:13,040
further analyses so this is some of what

4588
02:42:10,640 --> 02:42:14,399
needs to be worked out when using

4589
02:42:13,040 --> 02:42:16,479
problems data

4590
02:42:14,399 --> 02:42:18,800
with medical record

4591
02:42:16,479 --> 02:42:20,640
next slide please

4592
02:42:18,800 --> 02:42:22,960
so to end with opportunities and

4593
02:42:20,640 --> 02:42:24,720
challenges with these ehr studies they

4594
02:42:22,960 --> 02:42:26,640
have a because they're

4595
02:42:24,720 --> 02:42:27,840
fundamentally retrospective

4596
02:42:26,640 --> 02:42:28,720
and

4597
02:42:27,840 --> 02:42:30,560
they're

4598
02:42:28,720 --> 02:42:33,200
it's a real opportunity to pick up

4599
02:42:30,560 --> 02:42:36,160
subtle signals that are present

4600
02:42:33,200 --> 02:42:40,080
that can then be used to this

4601
02:42:36,160 --> 02:42:41,680
to define testable hypotheses which

4602
02:42:40,080 --> 02:42:43,040
future

4603
02:42:41,680 --> 02:42:44,640
randomized trials might be able to

4604
02:42:43,040 --> 02:42:46,319
better get at

4605
02:42:44,640 --> 02:42:48,800
next

4606
02:42:46,319 --> 02:42:50,560
and as we've talked about um there is a

4607
02:42:48,800 --> 02:42:52,479
real potential for integrating across

4608
02:42:50,560 --> 02:42:54,880
health care systems we've seen examples

4609
02:42:52,479 --> 02:42:56,000
of that in today's sessions what they

4610
02:42:54,880 --> 02:42:58,160
probably haven't gone through with all

4611
02:42:56,000 --> 02:43:01,120
the challenges they've dealt with um

4612
02:42:58,160 --> 02:43:04,240
merging across different ehrs even uh um

4613
02:43:01,120 --> 02:43:05,920
dr yule and i um are we both using epic

4614
02:43:04,240 --> 02:43:07,840
slightly different flavors of it and

4615
02:43:05,920 --> 02:43:08,960
there are challenges to navigate with

4616
02:43:07,840 --> 02:43:10,640
that

4617
02:43:08,960 --> 02:43:12,399
and one thing that needs to be kept in

4618
02:43:10,640 --> 02:43:14,880
mind um

4619
02:43:12,399 --> 02:43:15,920
with new nih rules regarding

4620
02:43:14,880 --> 02:43:18,160
um

4621
02:43:15,920 --> 02:43:19,040
sharing of data sets any information

4622
02:43:18,160 --> 02:43:20,800
that

4623
02:43:19,040 --> 02:43:23,520
you extract for

4624
02:43:20,800 --> 02:43:25,760
for an electronic health record-based

4625
02:43:23,520 --> 02:43:26,640
data set you need to be comfortable with

4626
02:43:25,760 --> 02:43:27,760
sharing

4627
02:43:26,640 --> 02:43:29,120
that information so that's just

4628
02:43:27,760 --> 02:43:32,160
something needs to be kept in mind

4629
02:43:29,120 --> 02:43:32,160
during study design

4630
02:43:32,800 --> 02:43:36,560
we've already seen examples of how

4631
02:43:34,640 --> 02:43:37,840
natural language processing can be used

4632
02:43:36,560 --> 02:43:40,640
to

4633
02:43:37,840 --> 02:43:41,760
enrich the available out measures um

4634
02:43:40,640 --> 02:43:44,000
that are

4635
02:43:41,760 --> 02:43:46,240
available in some situations that can be

4636
02:43:44,000 --> 02:43:47,600
done really well it's going to be a work

4637
02:43:46,240 --> 02:43:50,160
in progress to develop this we can

4638
02:43:47,600 --> 02:43:51,520
capture more ideas and then of course

4639
02:43:50,160 --> 02:43:54,000
things are happening in real time we've

4640
02:43:51,520 --> 02:43:55,840
seen examples today of how

4641
02:43:54,000 --> 02:43:57,600
information about the pandemic has been

4642
02:43:55,840 --> 02:43:58,800
captured and

4643
02:43:57,600 --> 02:44:01,680
identified

4644
02:43:58,800 --> 02:44:04,160
and as changes in policy change toxins

4645
02:44:01,680 --> 02:44:06,640
or environmental stressors

4646
02:44:04,160 --> 02:44:08,319
emerge these can better be tracked

4647
02:44:06,640 --> 02:44:11,040
and we can intervene better for our

4648
02:44:08,319 --> 02:44:13,439
patients so with that thank you for the

4649
02:44:11,040 --> 02:44:16,880
tolerating the technology interruptions

4650
02:44:13,439 --> 02:44:16,880
and we'll go on to the next talk

4651
02:44:18,560 --> 02:44:22,640
no problem

4652
02:44:19,920 --> 02:44:24,240
um sarah

4653
02:44:22,640 --> 02:44:26,960
are you

4654
02:44:24,240 --> 02:44:26,960
able to share

4655
02:44:27,359 --> 02:44:31,040
yes can everyone see

4656
02:44:29,200 --> 02:44:34,399
yes perfect

4657
02:44:31,040 --> 02:44:35,920
so i um want to thank everyone it is a

4658
02:44:34,399 --> 02:44:37,120
blessing that i get to go last if

4659
02:44:35,920 --> 02:44:38,560
there's a lot that i'm actually not

4660
02:44:37,120 --> 02:44:40,880
going to cover that i didn't initially

4661
02:44:38,560 --> 02:44:42,240
plan but i'm going to give the talk that

4662
02:44:40,880 --> 02:44:44,399
i want to give now

4663
02:44:42,240 --> 02:44:46,160
um having heard all of your talks which

4664
02:44:44,399 --> 02:44:47,840
really lay the foundation for some of

4665
02:44:46,160 --> 02:44:50,560
the work that's been done here and thank

4666
02:44:47,840 --> 02:44:51,840
you to nida for the continued support of

4667
02:44:50,560 --> 02:44:52,800
the research that i'm talking through

4668
02:44:51,840 --> 02:44:53,520
today

4669
02:44:52,800 --> 02:44:54,560
um

4670
02:44:53,520 --> 02:44:56,240
so

4671
02:44:54,560 --> 02:44:58,640
very briefly

4672
02:44:56,240 --> 02:44:59,680
um we know that substance use prevention

4673
02:44:58,640 --> 02:45:02,800
efforts

4674
02:44:59,680 --> 02:45:04,800
in adolescence typically rely on systems

4675
02:45:02,800 --> 02:45:07,359
so this can be the family system it's

4676
02:45:04,800 --> 02:45:08,720
also peers that are influencing

4677
02:45:07,359 --> 02:45:10,479
the experience that young people have in

4678
02:45:08,720 --> 02:45:12,160
their exposure to substance use as well

4679
02:45:10,479 --> 02:45:14,319
as schools

4680
02:45:12,160 --> 02:45:17,439
social norms and policies and all of

4681
02:45:14,319 --> 02:45:19,840
these factors um introduce opportunities

4682
02:45:17,439 --> 02:45:21,600
um for threat when we uh have young

4683
02:45:19,840 --> 02:45:24,080
people who are experiencing other social

4684
02:45:21,600 --> 02:45:25,760
vulnerability as dr chisholm laid out

4685
02:45:24,080 --> 02:45:28,000
and there's an opportunity as was

4686
02:45:25,760 --> 02:45:29,760
highlighted by dr jensen to also

4687
02:45:28,000 --> 02:45:31,760
consider the health care system as an

4688
02:45:29,760 --> 02:45:33,600
opportunity for prevention

4689
02:45:31,760 --> 02:45:35,200
um we know and i'm not going to go

4690
02:45:33,600 --> 02:45:37,040
through this in detail but we know that

4691
02:45:35,200 --> 02:45:38,720
families are relying on pediatricians

4692
02:45:37,040 --> 02:45:40,240
and are coming to pediatricians that

4693
02:45:38,720 --> 02:45:42,160
screening is normal we talked about

4694
02:45:40,240 --> 02:45:43,760
confidentiality earlier today as well

4695
02:45:42,160 --> 02:45:46,160
which is important

4696
02:45:43,760 --> 02:45:48,640
um and that access to healthcare systems

4697
02:45:46,160 --> 02:45:51,279
is at least intended to be universal

4698
02:45:48,640 --> 02:45:52,880
but that there is opportunity to explore

4699
02:45:51,279 --> 02:45:56,240
whether

4700
02:45:52,880 --> 02:45:58,960
young people can actually benefit from

4701
02:45:56,240 --> 02:46:01,279
prevention and substance use screening

4702
02:45:58,960 --> 02:46:03,040
when we look across the broad spectrum

4703
02:46:01,279 --> 02:46:06,160
of where young people are seen and not

4704
02:46:03,040 --> 02:46:08,479
necessarily only in primary care

4705
02:46:06,160 --> 02:46:10,720
but specifically related to foster care

4706
02:46:08,479 --> 02:46:12,319
which is the population that i work with

4707
02:46:10,720 --> 02:46:14,000
there's a whole lot of social system

4708
02:46:12,319 --> 02:46:16,560
disruption that's happening that might

4709
02:46:14,000 --> 02:46:18,720
actually exacerbate risk so we know

4710
02:46:16,560 --> 02:46:21,040
there are over 400 000 young people in

4711
02:46:18,720 --> 02:46:23,600
foster care on any given day here in the

4712
02:46:21,040 --> 02:46:25,439
united states about 40 of those are ages

4713
02:46:23,600 --> 02:46:27,120
10 and older which indicates that they

4714
02:46:25,439 --> 02:46:30,240
are vulnerable for

4715
02:46:27,120 --> 02:46:32,160
voluntarily initiating substance use and

4716
02:46:30,240 --> 02:46:34,960
that there are disparities with respect

4717
02:46:32,160 --> 02:46:37,040
to race ethnicity and poverty on who

4718
02:46:34,960 --> 02:46:38,880
comes into foster care such that we have

4719
02:46:37,040 --> 02:46:40,960
an over-representation of young people

4720
02:46:38,880 --> 02:46:42,800
who have experienced poverty and young

4721
02:46:40,960 --> 02:46:45,840
people of color and foster care compared

4722
02:46:42,800 --> 02:46:47,439
to the general population

4723
02:46:45,840 --> 02:46:49,439
that those that are coming into foster

4724
02:46:47,439 --> 02:46:50,960
care are placed outside of their family

4725
02:46:49,439 --> 02:46:52,800
home which means that they are

4726
02:46:50,960 --> 02:46:54,720
frequently disrupted from things like

4727
02:46:52,800 --> 02:46:56,399
school healthcare and social settings

4728
02:46:54,720 --> 02:46:58,560
which means a lot of the opportunities

4729
02:46:56,399 --> 02:47:00,479
for prevention that we tend to think of

4730
02:46:58,560 --> 02:47:01,760
as being effective for prevention

4731
02:47:00,479 --> 02:47:03,920
delivery

4732
02:47:01,760 --> 02:47:05,680
for these young people are disrupted and

4733
02:47:03,920 --> 02:47:07,920
even though we have an increased

4734
02:47:05,680 --> 02:47:08,800
emphasis here in the united states now

4735
02:47:07,920 --> 02:47:10,720
on

4736
02:47:08,800 --> 02:47:12,399
kinship caregiving with the goal that

4737
02:47:10,720 --> 02:47:13,840
young people might come into foster care

4738
02:47:12,399 --> 02:47:15,600
but they won't have as much social

4739
02:47:13,840 --> 02:47:18,479
disruption we're still seeing a lot of

4740
02:47:15,600 --> 02:47:21,200
changes in school and healthcare

4741
02:47:18,479 --> 02:47:23,359
this contributes to exacerbated risk for

4742
02:47:21,200 --> 02:47:26,319
substance use for adolescents and adults

4743
02:47:23,359 --> 02:47:28,160
who have experienced foster care on top

4744
02:47:26,319 --> 02:47:30,560
of mechanisms we already know are

4745
02:47:28,160 --> 02:47:32,080
contributing to increased risk

4746
02:47:30,560 --> 02:47:34,640
and that's things like maltreatment and

4747
02:47:32,080 --> 02:47:35,920
parental substance use and i'm going to

4748
02:47:34,640 --> 02:47:37,520
specifically highlight parental

4749
02:47:35,920 --> 02:47:39,200
substance use here because you'll see it

4750
02:47:37,520 --> 02:47:41,120
repeated later on

4751
02:47:39,200 --> 02:47:43,279
so the first of two studies that have

4752
02:47:41,120 --> 02:47:45,520
been funded by nida

4753
02:47:43,279 --> 02:47:47,760
were what i call careful which is where

4754
02:47:45,520 --> 02:47:49,600
we linked child welfare

4755
02:47:47,760 --> 02:47:51,920
records with electronic health records

4756
02:47:49,600 --> 02:47:53,359
data for young people in hamilton county

4757
02:47:51,920 --> 02:47:55,760
which is the county that cincinnati

4758
02:47:53,359 --> 02:47:58,000
children sits in um

4759
02:47:55,760 --> 02:47:59,680
with their ehr data

4760
02:47:58,000 --> 02:48:02,960
if they were in foster care on at least

4761
02:47:59,680 --> 02:48:04,800
one day between 2012 and 2017 and they

4762
02:48:02,960 --> 02:48:06,720
were ages 10 or older

4763
02:48:04,800 --> 02:48:09,279
we were successful at linking more than

4764
02:48:06,720 --> 02:48:10,880
99 of those data so there were three

4765
02:48:09,279 --> 02:48:12,720
young people in our whole sample that we

4766
02:48:10,880 --> 02:48:14,319
weren't able to link records for and i'm

4767
02:48:12,720 --> 02:48:15,680
happy to have a separate conversation

4768
02:48:14,319 --> 02:48:17,840
about how

4769
02:48:15,680 --> 02:48:19,399
that process occurred but we ended up

4770
02:48:17,840 --> 02:48:22,080
with a cohort of

4771
02:48:19,399 --> 02:48:24,880
2787 young people who were represented

4772
02:48:22,080 --> 02:48:26,160
in our ehr at cincinnati children's

4773
02:48:24,880 --> 02:48:28,240
and then were able to extract a

4774
02:48:26,160 --> 02:48:31,200
comparison sample who was matched

4775
02:48:28,240 --> 02:48:33,600
one-to-one on gender race and ethnicity

4776
02:48:31,200 --> 02:48:36,479
date of birth within six months medicaid

4777
02:48:33,600 --> 02:48:38,399
status um and received primary care here

4778
02:48:36,479 --> 02:48:40,319
at cincinnati children's and we're never

4779
02:48:38,399 --> 02:48:42,080
in foster care so that we really were

4780
02:48:40,319 --> 02:48:45,279
looking at differences in

4781
02:48:42,080 --> 02:48:47,680
foster care involvement on screening and

4782
02:48:45,279 --> 02:48:49,040
referral practices and the goal of this

4783
02:48:47,680 --> 02:48:50,880
was to understand differences in

4784
02:48:49,040 --> 02:48:52,720
substance use screening and when our

4785
02:48:50,880 --> 02:48:55,120
health care system is identifying these

4786
02:48:52,720 --> 02:48:56,800
young people as being at risk so that we

4787
02:48:55,120 --> 02:48:58,640
could start seeing if the health care

4788
02:48:56,800 --> 02:48:59,680
system is an appropriate target for

4789
02:48:58,640 --> 02:49:01,760
healthcare

4790
02:48:59,680 --> 02:49:03,520
for prevention delivery

4791
02:49:01,760 --> 02:49:05,680
and also to identify disparities in

4792
02:49:03,520 --> 02:49:08,399
healthcare access that might have an

4793
02:49:05,680 --> 02:49:10,240
impact on substance use prevention

4794
02:49:08,399 --> 02:49:12,479
not surprisingly based on the data that

4795
02:49:10,240 --> 02:49:14,640
we saw earlier today

4796
02:49:12,479 --> 02:49:16,800
we when we pulled the structured ehr

4797
02:49:14,640 --> 02:49:18,240
data discovered that

4798
02:49:16,800 --> 02:49:21,120
only four percent of our clinical

4799
02:49:18,240 --> 02:49:23,600
encounters contain structured data about

4800
02:49:21,120 --> 02:49:26,399
substance use which obviously is a hard

4801
02:49:23,600 --> 02:49:28,720
stop um and i appreciate the challenges

4802
02:49:26,399 --> 02:49:32,240
that dr johnson experienced with his

4803
02:49:28,720 --> 02:49:34,080
uh study as well um so for us uh because

4804
02:49:32,240 --> 02:49:36,479
we knew that there had to be more data

4805
02:49:34,080 --> 02:49:38,319
in the ehr based on conversations we had

4806
02:49:36,479 --> 02:49:40,000
with our clinical team and

4807
02:49:38,319 --> 02:49:41,840
my colleagues in adolescent medicine in

4808
02:49:40,000 --> 02:49:43,439
our foster care clinics that i knew were

4809
02:49:41,840 --> 02:49:44,800
talking to young people about substance

4810
02:49:43,439 --> 02:49:46,880
use we did a

4811
02:49:44,800 --> 02:49:48,399
quick chart review and as all of you

4812
02:49:46,880 --> 02:49:50,319
have already discovered and we've

4813
02:49:48,399 --> 02:49:51,840
covered thoroughly today um that's

4814
02:49:50,319 --> 02:49:54,080
because that screening was in

4815
02:49:51,840 --> 02:49:56,640
unstructured clinical notes

4816
02:49:54,080 --> 02:49:59,040
and with us and this is not surprising

4817
02:49:56,640 --> 02:50:00,800
given that um the screening practices

4818
02:49:59,040 --> 02:50:03,359
were also not standardized across our

4819
02:50:00,800 --> 02:50:06,000
hospital system we were seeing a lot of

4820
02:50:03,359 --> 02:50:07,439
ambiguity in those clinical notes where

4821
02:50:06,000 --> 02:50:09,120
frequently

4822
02:50:07,439 --> 02:50:10,880
the reference to substance use was

4823
02:50:09,120 --> 02:50:13,040
actually to apparent or where it would

4824
02:50:10,880 --> 02:50:14,640
be difficult for us to take an out of

4825
02:50:13,040 --> 02:50:16,479
the box

4826
02:50:14,640 --> 02:50:18,880
natural language processing algorithm

4827
02:50:16,479 --> 02:50:20,720
and actually apply it to the data that

4828
02:50:18,880 --> 02:50:23,120
we were seeing with this population so i

4829
02:50:20,720 --> 02:50:24,880
give two real examples there one patient

4830
02:50:23,120 --> 02:50:26,960
has a neutral exposure to drugs and

4831
02:50:24,880 --> 02:50:28,479
alcohol and a second she discovered he

4832
02:50:26,960 --> 02:50:30,240
began drinking alcohol and using

4833
02:50:28,479 --> 02:50:32,640
marijuana where it would have been

4834
02:50:30,240 --> 02:50:34,720
difficult to know whether she is the mom

4835
02:50:32,640 --> 02:50:37,200
or whether she is the daughter who's

4836
02:50:34,720 --> 02:50:39,760
discovered her father's

4837
02:50:37,200 --> 02:50:41,920
substance use for example so

4838
02:50:39,760 --> 02:50:43,840
led by an amazing study coordinator

4839
02:50:41,920 --> 02:50:46,880
katie nass in my lab

4840
02:50:43,840 --> 02:50:48,640
we recruited a whole host of undergrads

4841
02:50:46,880 --> 02:50:51,840
from the university of cincinnati who

4842
02:50:48,640 --> 02:50:53,840
helped us to manually code all 26 000

4843
02:50:51,840 --> 02:50:56,640
notes that were represented by these

4844
02:50:53,840 --> 02:50:58,880
data um with high reliability we have

4845
02:50:56,640 --> 02:51:00,640
kappas about 0.8

4846
02:50:58,880 --> 02:51:02,479
and so we treated these clinical notes

4847
02:51:00,640 --> 02:51:04,960
as if they were

4848
02:51:02,479 --> 02:51:05,760
transcripts from qualitative interviews

4849
02:51:04,960 --> 02:51:07,040
or

4850
02:51:05,760 --> 02:51:08,720
um

4851
02:51:07,040 --> 02:51:10,560
the clinical data that we tend to get

4852
02:51:08,720 --> 02:51:13,120
from child welfare reports where we're

4853
02:51:10,560 --> 02:51:16,240
coding for maltreatment exposure

4854
02:51:13,120 --> 02:51:18,399
to really ensure that we had accurate

4855
02:51:16,240 --> 02:51:19,840
substance use data

4856
02:51:18,399 --> 02:51:21,600
on both the

4857
02:51:19,840 --> 02:51:22,880
foster care sample and the comparison

4858
02:51:21,600 --> 02:51:24,479
sample

4859
02:51:22,880 --> 02:51:26,319
and what we found with that and again i

4860
02:51:24,479 --> 02:51:28,080
want to emphasize that this is not with

4861
02:51:26,319 --> 02:51:30,479
us changing anything about the health

4862
02:51:28,080 --> 02:51:31,680
care system this is just what was in the

4863
02:51:30,479 --> 02:51:33,439
data

4864
02:51:31,680 --> 02:51:35,680
we were actually able to identify that

4865
02:51:33,439 --> 02:51:37,439
over 70 percent of our population had

4866
02:51:35,680 --> 02:51:40,399
been screened at least once over that

4867
02:51:37,439 --> 02:51:42,640
five-year period for substance use um

4868
02:51:40,399 --> 02:51:44,960
and not surprising for all of the

4869
02:51:42,640 --> 02:51:47,359
conversations we've had today

4870
02:51:44,960 --> 02:51:49,439
there was significantly more likely that

4871
02:51:47,359 --> 02:51:51,359
the positive screening information was

4872
02:51:49,439 --> 02:51:53,279
going to be stored in the unstructured

4873
02:51:51,359 --> 02:51:56,479
compared to structured data fields even

4874
02:51:53,279 --> 02:51:58,319
though we had a place in the ehr

4875
02:51:56,479 --> 02:51:59,680
for individuals to record that

4876
02:51:58,319 --> 02:52:01,920
information

4877
02:51:59,680 --> 02:52:03,760
we also found that foster youth were

4878
02:52:01,920 --> 02:52:05,920
significantly more likely to receive

4879
02:52:03,760 --> 02:52:08,000
screening and part of that is because

4880
02:52:05,920 --> 02:52:10,000
screening had been somewhat standardized

4881
02:52:08,000 --> 02:52:11,520
in our foster care clinic

4882
02:52:10,000 --> 02:52:13,359
and so we were seeing a difference

4883
02:52:11,520 --> 02:52:15,279
between those that were in foster care

4884
02:52:13,359 --> 02:52:17,200
versus the young people that weren't in

4885
02:52:15,279 --> 02:52:19,120
foster care but were receiving primary

4886
02:52:17,200 --> 02:52:21,120
care because of differences in screening

4887
02:52:19,120 --> 02:52:22,880
across those settings

4888
02:52:21,120 --> 02:52:24,080
among those young people who were

4889
02:52:22,880 --> 02:52:26,080
screened

4890
02:52:24,080 --> 02:52:28,560
we were able to do survival analysis to

4891
02:52:26,080 --> 02:52:30,399
look at when substance use was initiated

4892
02:52:28,560 --> 02:52:31,840
and what we found was consistent with

4893
02:52:30,399 --> 02:52:33,359
the literature and that is that our

4894
02:52:31,840 --> 02:52:36,160
health care system

4895
02:52:33,359 --> 02:52:38,399
is uh effectively identifying when young

4896
02:52:36,160 --> 02:52:40,080
people initiate substance use and that

4897
02:52:38,399 --> 02:52:42,240
youth in foster care are initiating

4898
02:52:40,080 --> 02:52:45,040
substances earlier and a higher

4899
02:52:42,240 --> 02:52:47,200
frequency than the comparison sample

4900
02:52:45,040 --> 02:52:49,040
this was important because it was

4901
02:52:47,200 --> 02:52:50,960
demonstrated to our health care system

4902
02:52:49,040 --> 02:52:52,800
for the first time that actually there

4903
02:52:50,960 --> 02:52:54,319
was an opportunity for us to engage in

4904
02:52:52,800 --> 02:52:55,600
some of these prevention efforts and

4905
02:52:54,319 --> 02:52:57,920
we'll talk a little bit more about what

4906
02:52:55,600 --> 02:52:59,439
we did with that here in a minute

4907
02:52:57,920 --> 02:53:01,279
um but

4908
02:52:59,439 --> 02:53:03,680
we were also able to look at predictive

4909
02:53:01,279 --> 02:53:05,920
data and i think the work of dr chisholm

4910
02:53:03,680 --> 02:53:07,680
sets this up quite nicely where we were

4911
02:53:05,920 --> 02:53:10,560
able to look to see what the

4912
02:53:07,680 --> 02:53:13,279
true impact of child welfare involvement

4913
02:53:10,560 --> 02:53:15,920
and other factors was on

4914
02:53:13,279 --> 02:53:17,600
exacerbating risk for young people and

4915
02:53:15,920 --> 02:53:19,920
what we found there because we had

4916
02:53:17,600 --> 02:53:21,680
reliable information about foster care

4917
02:53:19,920 --> 02:53:23,680
status from our child welfare record

4918
02:53:21,680 --> 02:53:25,920
that was linked is that a history of

4919
02:53:23,680 --> 02:53:28,560
foster care significantly exacerbated

4920
02:53:25,920 --> 02:53:31,600
risk with an um

4921
02:53:28,560 --> 02:53:34,720
uh irr of around point eight

4922
02:53:31,600 --> 02:53:36,560
um that that somewhat changed over time

4923
02:53:34,720 --> 02:53:38,000
but what was more interesting for us is

4924
02:53:36,560 --> 02:53:40,319
that we found that there was a

4925
02:53:38,000 --> 02:53:42,399
protective effect of being in foster

4926
02:53:40,319 --> 02:53:43,760
care at age 13 so these are young people

4927
02:53:42,399 --> 02:53:44,800
that were screening

4928
02:53:43,760 --> 02:53:46,160
um

4929
02:53:44,800 --> 02:53:47,920
at the time of their

4930
02:53:46,160 --> 02:53:50,319
around the age of time of their 13th

4931
02:53:47,920 --> 02:53:52,399
birthday um but that seemed to be

4932
02:53:50,319 --> 02:53:54,720
protective for foster care placement but

4933
02:53:52,399 --> 02:53:57,680
that shifted over time so

4934
02:53:54,720 --> 02:54:00,800
the impact of being in foster care

4935
02:53:57,680 --> 02:54:02,240
changes with age on risk for initiating

4936
02:54:00,800 --> 02:54:04,479
substance use

4937
02:54:02,240 --> 02:54:07,840
and we saw similar interactions with age

4938
02:54:04,479 --> 02:54:08,880
for being a woman or a young person of

4939
02:54:07,840 --> 02:54:10,080
color

4940
02:54:08,880 --> 02:54:11,520
i mean that's consistent with the

4941
02:54:10,080 --> 02:54:12,800
literature

4942
02:54:11,520 --> 02:54:15,040
in general

4943
02:54:12,800 --> 02:54:16,800
we also detected significant differences

4944
02:54:15,040 --> 02:54:19,040
in health care utilization for these two

4945
02:54:16,800 --> 02:54:21,200
populations where

4946
02:54:19,040 --> 02:54:23,359
a first look

4947
02:54:21,200 --> 02:54:24,800
we found that young people in foster

4948
02:54:23,359 --> 02:54:26,960
care were

4949
02:54:24,800 --> 02:54:28,800
using more health care across a variety

4950
02:54:26,960 --> 02:54:31,439
of different settings compared to young

4951
02:54:28,800 --> 02:54:33,439
people that were not in foster care

4952
02:54:31,439 --> 02:54:35,840
but interestingly and i think this

4953
02:54:33,439 --> 02:54:37,840
echoes some of the work of others

4954
02:54:35,840 --> 02:54:39,520
once we accounted for substance use

4955
02:54:37,840 --> 02:54:40,800
those effects

4956
02:54:39,520 --> 02:54:43,439
were reversed

4957
02:54:40,800 --> 02:54:45,760
particularly for primary care

4958
02:54:43,439 --> 02:54:48,479
and so what this suggests to us is that

4959
02:54:45,760 --> 02:54:51,120
actually we have some differences in

4960
02:54:48,479 --> 02:54:52,560
access to healthcare services based on

4961
02:54:51,120 --> 02:54:54,720
whether young people are using

4962
02:54:52,560 --> 02:54:56,319
substances or not

4963
02:54:54,720 --> 02:54:58,479
where likely those young people are

4964
02:54:56,319 --> 02:55:00,240
going to be seen more often in emergency

4965
02:54:58,479 --> 02:55:02,800
care settings and are going to be less

4966
02:55:00,240 --> 02:55:06,240
likely to be seen in primary care

4967
02:55:02,800 --> 02:55:08,080
and so thinking about what we need to do

4968
02:55:06,240 --> 02:55:10,399
to make sure that we have

4969
02:55:08,080 --> 02:55:12,160
accurate data across all settings when

4970
02:55:10,399 --> 02:55:14,240
we're thinking about trying to implement

4971
02:55:12,160 --> 02:55:16,080
and standardize screening

4972
02:55:14,240 --> 02:55:17,680
for substance use and where that data is

4973
02:55:16,080 --> 02:55:18,960
going to come from is important that

4974
02:55:17,680 --> 02:55:20,160
young people are seen in lots of

4975
02:55:18,960 --> 02:55:22,080
different settings and we have to

4976
02:55:20,160 --> 02:55:23,120
account for that both in our research

4977
02:55:22,080 --> 02:55:25,279
and in what we're going to do with

4978
02:55:23,120 --> 02:55:27,520
clinical care

4979
02:55:25,279 --> 02:55:29,279
it turned out that the coded data which

4980
02:55:27,520 --> 02:55:31,120
we weren't initially setting out to

4981
02:55:29,279 --> 02:55:33,439
develop ai

4982
02:55:31,120 --> 02:55:35,600
around was actually really great as a

4983
02:55:33,439 --> 02:55:37,279
training set and so we partnered with dr

4984
02:55:35,600 --> 02:55:38,720
yiz auni who was here at cincinnati

4985
02:55:37,279 --> 02:55:40,880
children's at the time and has since

4986
02:55:38,720 --> 02:55:41,760
moved to kaiser

4987
02:55:40,880 --> 02:55:45,760
who

4988
02:55:41,760 --> 02:55:48,080
helps to develop some rule-based matches

4989
02:55:45,760 --> 02:55:49,680
matching algorithms as well as some nlp

4990
02:55:48,080 --> 02:55:52,160
and deep learning models so that we

4991
02:55:49,680 --> 02:55:54,160
could take that structured data when it

4992
02:55:52,160 --> 02:55:55,279
was present because it was there some of

4993
02:55:54,160 --> 02:55:57,279
the time

4994
02:55:55,279 --> 02:55:59,120
and the unstructured clinical narrative

4995
02:55:57,279 --> 02:55:59,840
and integrate that together

4996
02:55:59,120 --> 02:56:02,240
to

4997
02:55:59,840 --> 02:56:03,920
set up a series of classifications so

4998
02:56:02,240 --> 02:56:05,840
that young people are classified based

4999
02:56:03,920 --> 02:56:08,160
on whether or not they have received any

5000
02:56:05,840 --> 02:56:09,920
screening so had they been screened yes

5001
02:56:08,160 --> 02:56:12,479
or no and then if they had been screened

5002
02:56:09,920 --> 02:56:14,319
was that screening positive or negative

5003
02:56:12,479 --> 02:56:16,800
for each of the different substances or

5004
02:56:14,319 --> 02:56:18,720
for any substance use if it was generic

5005
02:56:16,800 --> 02:56:20,640
and then we added another category that

5006
02:56:18,720 --> 02:56:22,080
was actually family history so was there

5007
02:56:20,640 --> 02:56:23,359
an evidence of family history of

5008
02:56:22,080 --> 02:56:26,479
substance use

5009
02:56:23,359 --> 02:56:28,479
um and it turned out that uh linking

5010
02:56:26,479 --> 02:56:30,479
this structured and unstructured data

5011
02:56:28,479 --> 02:56:32,800
together across all of these different

5012
02:56:30,479 --> 02:56:35,200
sets was very important

5013
02:56:32,800 --> 02:56:37,120
because we were able to secure a samsa

5014
02:56:35,200 --> 02:56:39,520
grant to implement standardized

5015
02:56:37,120 --> 02:56:41,680
screening the craft on welcome tablets

5016
02:56:39,520 --> 02:56:42,720
and are now doing substance use

5017
02:56:41,680 --> 02:56:44,640
um

5018
02:56:42,720 --> 02:56:46,080
brief intervention and refer screening

5019
02:56:44,640 --> 02:56:47,200
brief intervention and refer all the

5020
02:56:46,080 --> 02:56:49,200
treatment

5021
02:56:47,200 --> 02:56:51,840
where now all of our young people are

5022
02:56:49,200 --> 02:56:53,439
receiving um the craft on welcome

5023
02:56:51,840 --> 02:56:55,760
tablets at the beginning of their visit

5024
02:56:53,439 --> 02:56:56,960
so that they can report themselves

5025
02:56:55,760 --> 02:56:59,120
and that screening is happening

5026
02:56:56,960 --> 02:57:01,840
universally we've just started this so

5027
02:56:59,120 --> 02:57:04,319
we are i don't have any um

5028
02:57:01,840 --> 02:57:06,160
any hard data to present but i do want

5029
02:57:04,319 --> 02:57:07,520
to share some learnings early on that

5030
02:57:06,160 --> 02:57:09,600
we've had

5031
02:57:07,520 --> 02:57:11,359
which is first that that when we

5032
02:57:09,600 --> 02:57:12,880
implemented we decided that we would

5033
02:57:11,359 --> 02:57:14,800
train our health care providers to

5034
02:57:12,880 --> 02:57:16,720
validate abstinence

5035
02:57:14,800 --> 02:57:19,359
um so those young people that reported

5036
02:57:16,720 --> 02:57:21,120
that they weren't using substances and

5037
02:57:19,359 --> 02:57:23,120
in part driven by some conversations i

5038
02:57:21,120 --> 02:57:25,120
had with dr sterling

5039
02:57:23,120 --> 02:57:26,960
and that was turn has turned out to be

5040
02:57:25,120 --> 02:57:29,040
an opportunity to confirm screening

5041
02:57:26,960 --> 02:57:30,720
results so now we have a healthcare

5042
02:57:29,040 --> 02:57:32,640
provider who says i'm so proud of you

5043
02:57:30,720 --> 02:57:34,000
for being abstinent and that young

5044
02:57:32,640 --> 02:57:36,960
person will say

5045
02:57:34,000 --> 02:57:39,359
actually i'm not um and and that's being

5046
02:57:36,960 --> 02:57:41,840
documented in our clinical notes and so

5047
02:57:39,359 --> 02:57:43,920
by integrating our standardized

5048
02:57:41,840 --> 02:57:46,080
structured data with our unstandardized

5049
02:57:43,920 --> 02:57:47,520
screening results we're able to confirm

5050
02:57:46,080 --> 02:57:49,439
that validation

5051
02:57:47,520 --> 02:57:51,200
um and report discrepancies so that

5052
02:57:49,439 --> 02:57:52,960
clinicians and researchers can use both

5053
02:57:51,200 --> 02:57:54,640
pieces of information

5054
02:57:52,960 --> 02:57:56,399
and we're finding that that's happening

5055
02:57:54,640 --> 02:57:58,160
so far and we're just two months into

5056
02:57:56,399 --> 02:58:00,080
this but in about five percent of our

5057
02:57:58,160 --> 02:58:02,080
population

5058
02:58:00,080 --> 02:58:03,279
it also allows us to ensure that our

5059
02:58:02,080 --> 02:58:06,160
behavioral health providers are

5060
02:58:03,279 --> 02:58:08,000
addressing you so we can pull that

5061
02:58:06,160 --> 02:58:09,840
pull that unstructured data and ensure

5062
02:58:08,000 --> 02:58:11,359
that even if they aren't being seen in

5063
02:58:09,840 --> 02:58:12,960
our clinic where we have standardized

5064
02:58:11,359 --> 02:58:15,040
universal screening that we have that

5065
02:58:12,960 --> 02:58:17,359
information from other places

5066
02:58:15,040 --> 02:58:20,399
and are able to continuously monitor

5067
02:58:17,359 --> 02:58:22,800
that over time so that if we have

5068
02:58:20,399 --> 02:58:24,880
an issue where a young person didn't

5069
02:58:22,800 --> 02:58:26,640
respond or skipped those questions that

5070
02:58:24,880 --> 02:58:28,160
we can catch them the next time and we

5071
02:58:26,640 --> 02:58:29,680
know whether they were screened or not

5072
02:58:28,160 --> 02:58:31,520
which has been extremely helpful as

5073
02:58:29,680 --> 02:58:32,960
we've implemented

5074
02:58:31,520 --> 02:58:35,120
expert

5075
02:58:32,960 --> 02:58:36,560
but we're also fortunate to be able to

5076
02:58:35,120 --> 02:58:38,720
continue to

5077
02:58:36,560 --> 02:58:39,920
validate the algorithm that we created

5078
02:58:38,720 --> 02:58:42,080
initially

5079
02:58:39,920 --> 02:58:43,920
so with our new knight of funding we're

5080
02:58:42,080 --> 02:58:46,479
going to be

5081
02:58:43,920 --> 02:58:48,240
pulling a broader set of notes

5082
02:58:46,479 --> 02:58:50,160
that allows us to look at both publicly

5083
02:58:48,240 --> 02:58:51,920
and privately insured young people who

5084
02:58:50,160 --> 02:58:54,080
are in urban and suburban pediatric

5085
02:58:51,920 --> 02:58:56,479
settings and begin to examine some of

5086
02:58:54,080 --> 02:58:59,520
the bias that others have raised earlier

5087
02:58:56,479 --> 02:59:01,040
today in thinking about what differences

5088
02:58:59,520 --> 02:59:02,960
we might have around screening and

5089
02:59:01,040 --> 02:59:05,040
documentation

5090
02:59:02,960 --> 02:59:06,720
for substance use based on gender race

5091
02:59:05,040 --> 02:59:08,080
and ethnicity insurance status and

5092
02:59:06,720 --> 02:59:09,680
counter sites

5093
02:59:08,080 --> 02:59:11,680
so what are people doing in different

5094
02:59:09,680 --> 02:59:14,000
settings and how important is that

5095
02:59:11,680 --> 02:59:15,600
one of our hypotheses that we'll be able

5096
02:59:14,000 --> 02:59:17,840
to test out in this grant is actually

5097
02:59:15,600 --> 02:59:19,200
that by capturing both screening

5098
02:59:17,840 --> 02:59:21,200
standardized

5099
02:59:19,200 --> 02:59:23,120
structured data and the unstandardized

5100
02:59:21,200 --> 02:59:25,040
screening results that we can

5101
02:59:23,120 --> 02:59:26,960
potentially provide a more comprehensive

5102
02:59:25,040 --> 02:59:29,359
view that's longitudinal and we can

5103
02:59:26,960 --> 02:59:31,200
reduce some of those biases because

5104
02:59:29,359 --> 02:59:33,600
young people that might have experienced

5105
02:59:31,200 --> 02:59:35,359
some targeted screening in one setting

5106
02:59:33,600 --> 02:59:36,880
may be captured in another setting of a

5107
02:59:35,359 --> 02:59:38,640
different encounter

5108
02:59:36,880 --> 02:59:40,640
where they are engaging in universal

5109
02:59:38,640 --> 02:59:42,399
screening and we can ensure a more

5110
02:59:40,640 --> 02:59:44,479
comprehensive coverage of young people

5111
02:59:42,399 --> 02:59:46,960
by that by using that information

5112
02:59:44,479 --> 02:59:48,560
longitudinally from multiple sites

5113
02:59:46,960 --> 02:59:49,600
so with that

5114
02:59:48,560 --> 02:59:51,760
um

5115
02:59:49,600 --> 02:59:53,279
i want to echo the the comments of

5116
02:59:51,760 --> 02:59:55,439
others that i think it's important for

5117
02:59:53,279 --> 02:59:57,040
us to be thinking about this both from a

5118
02:59:55,439 --> 02:59:58,800
clinical standpoint and why it's

5119
02:59:57,040 --> 03:00:01,359
important to know whether your patient

5120
02:59:58,800 --> 03:00:03,439
was screened at a previous encounter and

5121
03:00:01,359 --> 03:00:05,120
what that clinician reported or observed

5122
03:00:03,439 --> 03:00:06,960
about that screening

5123
03:00:05,120 --> 03:00:09,439
needs to be shared for clinical care

5124
03:00:06,960 --> 03:00:11,279
delivery as well as for um

5125
03:00:09,439 --> 03:00:12,880
for research purposes

5126
03:00:11,279 --> 03:00:14,880
so we need to be able to track and

5127
03:00:12,880 --> 03:00:16,560
monitor that over time but also because

5128
03:00:14,880 --> 03:00:18,319
we know that substance use is changing

5129
03:00:16,560 --> 03:00:19,840
with age

5130
03:00:18,319 --> 03:00:21,760
and that we're reliant on both the

5131
03:00:19,840 --> 03:00:24,080
health care system to properly document

5132
03:00:21,760 --> 03:00:26,640
and adolescence to accurately report so

5133
03:00:24,080 --> 03:00:28,240
we really likely do need to think about

5134
03:00:26,640 --> 03:00:29,279
integrating both the standardized

5135
03:00:28,240 --> 03:00:31,840
screening

5136
03:00:29,279 --> 03:00:33,760
and trained clinical follow-up that can

5137
03:00:31,840 --> 03:00:35,680
ensure that we're capturing

5138
03:00:33,760 --> 03:00:37,200
recanting or other issues that might be

5139
03:00:35,680 --> 03:00:38,800
happening that result resulted

5140
03:00:37,200 --> 03:00:40,800
discrepancies between structured and

5141
03:00:38,800 --> 03:00:44,960
unstructured data collection

5142
03:00:40,800 --> 03:00:45,760
so with that i am going to turn it over

5143
03:00:44,960 --> 03:00:49,600
um

5144
03:00:45,760 --> 03:00:49,600
go and stop sharing my screen

5145
03:00:50,640 --> 03:00:53,920
thank you sarah

5146
03:00:52,720 --> 03:00:55,840
there we go

5147
03:00:53,920 --> 03:00:57,279
great thank you so much that was great

5148
03:00:55,840 --> 03:00:59,840
um

5149
03:00:57,279 --> 03:01:02,000
and we just have a few few questions or

5150
03:00:59,840 --> 03:01:03,920
a few few minutes for questions um if

5151
03:01:02,000 --> 03:01:05,520
you have any additional questions please

5152
03:01:03,920 --> 03:01:08,160
go ahead and

5153
03:01:05,520 --> 03:01:10,240
pop them into the chat

5154
03:01:08,160 --> 03:01:12,560
but i wanted to pick up on

5155
03:01:10,240 --> 03:01:14,319
an issue that you just

5156
03:01:12,560 --> 03:01:16,240
were kind of um

5157
03:01:14,319 --> 03:01:18,399
you were discussing at the end sarah and

5158
03:01:16,240 --> 03:01:20,319
i think maybe um

5159
03:01:18,399 --> 03:01:21,120
amy and vinod that you probably have

5160
03:01:20,319 --> 03:01:23,680
some

5161
03:01:21,120 --> 03:01:26,240
um some nuggets for us as well thinking

5162
03:01:23,680 --> 03:01:27,920
about the these issues of sharing across

5163
03:01:26,240 --> 03:01:30,800
providers um

5164
03:01:27,920 --> 03:01:33,920
a lot of our earlier speakers are in

5165
03:01:30,800 --> 03:01:36,240
more closed systems where um kaiser or

5166
03:01:33,920 --> 03:01:38,640
the va where you know you can kind of

5167
03:01:36,240 --> 03:01:39,840
rely on the fact that um

5168
03:01:38,640 --> 03:01:41,200
different clinicians are going to be

5169
03:01:39,840 --> 03:01:43,200
able to see

5170
03:01:41,200 --> 03:01:44,720
the same ehr or at least going to be

5171
03:01:43,200 --> 03:01:46,399
able to see the same data even if they

5172
03:01:44,720 --> 03:01:48,880
have a little bit of a different view

5173
03:01:46,399 --> 03:01:50,240
within the hr um but

5174
03:01:48,880 --> 03:01:51,920
you know when we're when we're dealing

5175
03:01:50,240 --> 03:01:54,319
with sort of different patient

5176
03:01:51,920 --> 03:01:55,200
populations who may be seeing

5177
03:01:54,319 --> 03:01:57,600
um

5178
03:01:55,200 --> 03:02:00,640
seeing providers outside of um the

5179
03:01:57,600 --> 03:02:02,800
current system um you know are there any

5180
03:02:00,640 --> 03:02:05,040
lessons learned any any

5181
03:02:02,800 --> 03:02:06,800
tools that can be used that we can kind

5182
03:02:05,040 --> 03:02:07,600
of think about in making sure that we're

5183
03:02:06,800 --> 03:02:10,080
not

5184
03:02:07,600 --> 03:02:12,319
um sort of duplicating efforts and we're

5185
03:02:10,080 --> 03:02:14,960
streamlining data but also

5186
03:02:12,319 --> 03:02:17,200
obviously navigating a whole lot of

5187
03:02:14,960 --> 03:02:20,399
technological and privacy

5188
03:02:17,200 --> 03:02:21,840
um privacy concerns that that will will

5189
03:02:20,399 --> 03:02:23,760
naturally come up in those sort of

5190
03:02:21,840 --> 03:02:25,520
situations

5191
03:02:23,760 --> 03:02:27,439
i did kind of bring up one of those kind

5192
03:02:25,520 --> 03:02:29,200
of privacy concerns that hasn't come up

5193
03:02:27,439 --> 03:02:32,640
yet and that's just thinking about 42

5194
03:02:29,200 --> 03:02:34,560
cfr um and you know that different

5195
03:02:32,640 --> 03:02:37,040
institutions interpret that in different

5196
03:02:34,560 --> 03:02:38,640
ways and so um vanilla can speak to this

5197
03:02:37,040 --> 03:02:40,880
more thoughtfully but you know when we

5198
03:02:38,640 --> 03:02:42,560
have two substances disorder clinics as

5199
03:02:40,880 --> 03:02:44,960
part of our project as we're looking at

5200
03:02:42,560 --> 03:02:47,439
risk of developing an open use disorder

5201
03:02:44,960 --> 03:02:49,359
um and kind of overnight one day the

5202
03:02:47,439 --> 03:02:51,680
proms data from the substance use

5203
03:02:49,359 --> 03:02:53,520
disorder clinics disappeared um you know

5204
03:02:51,680 --> 03:02:55,040
and then knew that that data was there

5205
03:02:53,520 --> 03:02:56,720
and so new to kind of advocating kind of

5206
03:02:55,040 --> 03:02:59,279
ask for it to come back but you know

5207
03:02:56,720 --> 03:03:00,640
anything if it's a clinician you know if

5208
03:02:59,279 --> 03:03:01,840
researchers are partnering with

5209
03:03:00,640 --> 03:03:04,319
clinicians they're going to you know

5210
03:03:01,840 --> 03:03:06,479
pretty quickly pick up this issue but um

5211
03:03:04,319 --> 03:03:09,040
but it's an issue that me you know some

5212
03:03:06,479 --> 03:03:12,600
people may not be as aware of um it

5213
03:03:09,040 --> 03:03:12,600
exists so

5214
03:03:18,479 --> 03:03:23,520
all right and then another sort of along

5215
03:03:20,960 --> 03:03:25,439
those same same lines um this was a

5216
03:03:23,520 --> 03:03:26,640
question that came into the chat

5217
03:03:25,439 --> 03:03:28,399
from

5218
03:03:26,640 --> 03:03:30,560
one of our colleagues but how can we set

5219
03:03:28,399 --> 03:03:32,479
up tools or support

5220
03:03:30,560 --> 03:03:35,520
to help providers reflect and choose the

5221
03:03:32,479 --> 03:03:37,359
appropriate course and avoid automation

5222
03:03:35,520 --> 03:03:40,319
bias that can lead to

5223
03:03:37,359 --> 03:03:41,600
draconian response or bias care um and i

5224
03:03:40,319 --> 03:03:43,680
think i you know you think about this a

5225
03:03:41,600 --> 03:03:46,319
lot um you know we haven't quite said

5226
03:03:43,680 --> 03:03:47,600
you know there is a stigma related you

5227
03:03:46,319 --> 03:03:51,120
know both

5228
03:03:47,600 --> 03:03:53,200
by patients and clinicians related to um

5229
03:03:51,120 --> 03:03:56,000
related to substance use

5230
03:03:53,200 --> 03:03:58,640
also related to

5231
03:03:56,000 --> 03:03:59,600
foster care also related to mental

5232
03:03:58,640 --> 03:04:01,279
health

5233
03:03:59,600 --> 03:04:02,800
so there's there's a lot kind of that

5234
03:04:01,279 --> 03:04:04,399
we're navigating

5235
03:04:02,800 --> 03:04:07,520
but any thoughts on that of how we can

5236
03:04:04,399 --> 03:04:09,680
kind of safeguard or at least put some

5237
03:04:07,520 --> 03:04:12,560
put some safeguards in place to try to

5238
03:04:09,680 --> 03:04:12,560
minimize bias

5239
03:04:12,800 --> 03:04:15,840
um so

5240
03:04:14,000 --> 03:04:19,840
at least one strategy that we're using

5241
03:04:15,840 --> 03:04:21,680
within cincinnati children's so far

5242
03:04:19,840 --> 03:04:23,520
is that

5243
03:04:21,680 --> 03:04:26,319
we're trying to highlight that if it's

5244
03:04:23,520 --> 03:04:28,560
universal that that reduces stigma right

5245
03:04:26,319 --> 03:04:30,960
so if everybody is screened

5246
03:04:28,560 --> 03:04:32,479
um and we screen all the time

5247
03:04:30,960 --> 03:04:34,479
um then it's

5248
03:04:32,479 --> 03:04:38,240
it becomes a little bit easier to be

5249
03:04:34,479 --> 03:04:40,479
able to um advocate and and frame the

5250
03:04:38,240 --> 03:04:43,120
questions that we're asking everyone and

5251
03:04:40,479 --> 03:04:44,880
so it's it's not as biasing

5252
03:04:43,120 --> 03:04:47,680
the second thing is

5253
03:04:44,880 --> 03:04:49,439
for us to highlight that substance use

5254
03:04:47,680 --> 03:04:51,200
is somewhat normative in adolescence and

5255
03:04:49,439 --> 03:04:53,359
we know what's going to happen

5256
03:04:51,200 --> 03:04:55,200
but that we really want to set up and

5257
03:04:53,359 --> 03:04:57,040
partner with youth and families to

5258
03:04:55,200 --> 03:05:00,560
ensure that there are safeguards so that

5259
03:04:57,040 --> 03:05:02,080
they are are not moving from

5260
03:05:00,560 --> 03:05:04,800
experimental substance use to

5261
03:05:02,080 --> 03:05:06,640
problematic substance use and how we can

5262
03:05:04,800 --> 03:05:09,120
frame that in such a way that young

5263
03:05:06,640 --> 03:05:10,319
people feel comfortable disclosing their

5264
03:05:09,120 --> 03:05:12,399
use

5265
03:05:10,319 --> 03:05:14,160
and trusting providers and then the last

5266
03:05:12,399 --> 03:05:15,040
thing is that

5267
03:05:14,160 --> 03:05:16,800
i think

5268
03:05:15,040 --> 03:05:18,640
it's also important to train healthcare

5269
03:05:16,800 --> 03:05:20,240
providers to be aware that there are

5270
03:05:18,640 --> 03:05:22,560
some biases

5271
03:05:20,240 --> 03:05:24,800
even in the way that we ask and talk

5272
03:05:22,560 --> 03:05:26,640
with young people around

5273
03:05:24,800 --> 03:05:28,880
substance use and screen for that and

5274
03:05:26,640 --> 03:05:30,800
that we have to be prepared

5275
03:05:28,880 --> 03:05:33,040
to respond to it so

5276
03:05:30,800 --> 03:05:33,920
so that we're not just looking at a

5277
03:05:33,040 --> 03:05:35,840
score

5278
03:05:33,920 --> 03:05:38,000
um and a great example of that is that

5279
03:05:35,840 --> 03:05:40,000
we have a

5280
03:05:38,000 --> 03:05:42,080
non-ignorable number of young people on

5281
03:05:40,000 --> 03:05:43,359
the craft who are reporting that they

5282
03:05:42,080 --> 03:05:45,120
have been in the vehicle with an

5283
03:05:43,359 --> 03:05:47,520
intoxicated driver

5284
03:05:45,120 --> 03:05:49,600
and it's not the same risk that we think

5285
03:05:47,520 --> 03:05:51,600
of when we when that that question was

5286
03:05:49,600 --> 03:05:53,840
initially described around peers right

5287
03:05:51,600 --> 03:05:56,399
but it's actually a caregiver

5288
03:05:53,840 --> 03:05:58,080
um and so explaining to a clinician that

5289
03:05:56,399 --> 03:06:00,160
they need to have a conversation around

5290
03:05:58,080 --> 03:06:01,760
how to safely

5291
03:06:00,160 --> 03:06:03,439
navigate that when it's your parent

5292
03:06:01,760 --> 03:06:04,880
who's intoxicated is an important

5293
03:06:03,439 --> 03:06:06,720
conversation that's about child

5294
03:06:04,880 --> 03:06:08,479
protection and not about bias towards

5295
03:06:06,720 --> 03:06:09,520
that child they're spending their time

5296
03:06:08,479 --> 03:06:12,080
with

5297
03:06:09,520 --> 03:06:13,520
um and so so that all of those things

5298
03:06:12,080 --> 03:06:15,920
helping to contextualize that i think

5299
03:06:13,520 --> 03:06:19,120
motivates discussion with providers

5300
03:06:15,920 --> 03:06:19,120
that's less judgmental

5301
03:06:20,160 --> 03:06:24,640
all right so my final question um

5302
03:06:22,880 --> 03:06:27,279
is is

5303
03:06:24,640 --> 03:06:29,520
advice for us so um

5304
03:06:27,279 --> 03:06:32,560
as nida if we're thinking about how we

5305
03:06:29,520 --> 03:06:35,520
can support

5306
03:06:32,560 --> 03:06:36,960
support prevention research using ehrs

5307
03:06:35,520 --> 03:06:39,120
if we can if we're thinking about how we

5308
03:06:36,960 --> 03:06:41,120
can support

5309
03:06:39,120 --> 03:06:43,279
prevention

5310
03:06:41,120 --> 03:06:45,680
within primary care and especially

5311
03:06:43,279 --> 03:06:47,279
implementation research to promote the

5312
03:06:45,680 --> 03:06:49,200
you know the implementation of of

5313
03:06:47,279 --> 03:06:50,960
intervention screening or you know

5314
03:06:49,200 --> 03:06:52,720
interventions that we know work

5315
03:06:50,960 --> 03:06:54,479
um how

5316
03:06:52,720 --> 03:06:55,600
what are do you have any advice for us

5317
03:06:54,479 --> 03:06:57,520
and this is probably a question we

5318
03:06:55,600 --> 03:06:59,680
should have asked everybody uh but

5319
03:06:57,520 --> 03:07:00,479
others can kind of chime in too if they

5320
03:06:59,680 --> 03:07:02,560
um

5321
03:07:00,479 --> 03:07:04,319
if they still have energy but you know

5322
03:07:02,560 --> 03:07:05,760
do you have advice for us for thinking

5323
03:07:04,319 --> 03:07:08,479
through like what we should be

5324
03:07:05,760 --> 03:07:10,080
prioritizing and how we can kind of

5325
03:07:08,479 --> 03:07:12,160
you know ensure that we're asking the

5326
03:07:10,080 --> 03:07:13,680
right questions and that we're um we're

5327
03:07:12,160 --> 03:07:15,279
not also not duplicating efforts and

5328
03:07:13,680 --> 03:07:17,200
that we're we're also supporting you

5329
03:07:15,279 --> 03:07:19,200
know sarah you mentioned that you talked

5330
03:07:17,200 --> 03:07:21,520
to um stacy

5331
03:07:19,200 --> 03:07:24,880
um which is wonderful um but that we're

5332
03:07:21,520 --> 03:07:26,720
kind of helping to support um to support

5333
03:07:24,880 --> 03:07:28,880
each other and kind of asking the right

5334
03:07:26,720 --> 03:07:31,040
questions avoiding duplication and

5335
03:07:28,880 --> 03:07:33,920
getting to answers more quickly do you

5336
03:07:31,040 --> 03:07:33,920
have any advice for us

5337
03:07:36,000 --> 03:07:39,760
great question i mean i will say that

5338
03:07:37,840 --> 03:07:41,279
this has been a really helpful workshop

5339
03:07:39,760 --> 03:07:43,120
um you know that i think kind of we're

5340
03:07:41,279 --> 03:07:44,640
all have been eagerly listening to the

5341
03:07:43,120 --> 03:07:46,880
different talks and so having

5342
03:07:44,640 --> 03:07:48,880
opportunities to to get people together

5343
03:07:46,880 --> 03:07:49,840
to talk about oh screening adolescence

5344
03:07:48,880 --> 03:07:52,000
via

5345
03:07:49,840 --> 03:07:53,840
you know kind of telehealth um you know

5346
03:07:52,000 --> 03:07:55,200
i think he brian mentioned kind of oh

5347
03:07:53,840 --> 03:07:57,439
you just set up the proxies but like

5348
03:07:55,200 --> 03:07:59,600
that's actually a talk in and of itself

5349
03:07:57,439 --> 03:08:00,720
is how you set up proxies and

5350
03:07:59,600 --> 03:08:02,160
then i think also learning from

5351
03:08:00,720 --> 03:08:04,479
different systems we have children's

5352
03:08:02,160 --> 03:08:06,319
hospitals and then hospital systems that

5353
03:08:04,479 --> 03:08:09,279
treat adults and children and so that

5354
03:08:06,319 --> 03:08:11,359
there's lessons to be learned um there

5355
03:08:09,279 --> 03:08:13,439
as well so so thank you for the

5356
03:08:11,359 --> 03:08:15,279
opportunity to have us all together and

5357
03:08:13,439 --> 03:08:16,800
and to hear all the information that was

5358
03:08:15,279 --> 03:08:20,120
presented today

5359
03:08:16,800 --> 03:08:20,120
my only thoughts

5360
03:08:21,680 --> 03:08:27,359
all right well on that note um we i'm

5361
03:08:24,399 --> 03:08:29,120
going to turn it over to carlos um

5362
03:08:27,359 --> 03:08:31,840
blanco to

5363
03:08:29,120 --> 03:08:34,000
finish us off and close us off um before

5364
03:08:31,840 --> 03:08:35,680
i turn it over to carlos i want to let

5365
03:08:34,000 --> 03:08:38,640
everybody know that we did record this

5366
03:08:35,680 --> 03:08:40,960
and we'll be putting it online

5367
03:08:38,640 --> 03:08:42,399
and we also will be putting together a

5368
03:08:40,960 --> 03:08:45,040
summary

5369
03:08:42,399 --> 03:08:47,279
of the of the discussion so

5370
03:08:45,040 --> 03:08:49,840
a lot of the issues that have come up

5371
03:08:47,279 --> 03:08:51,040
today we'll be able to document and put

5372
03:08:49,840 --> 03:08:54,880
online

5373
03:08:51,040 --> 03:08:57,600
um if you'd like copies of uh of the

5374
03:08:54,880 --> 03:08:59,840
the slides you can reach out to

5375
03:08:57,600 --> 03:09:01,920
um i believe caitlyn um and caitlin i

5376
03:08:59,840 --> 03:09:04,479
don't know if you can put in the chat um

5377
03:09:01,920 --> 03:09:06,000
your email again i know you it was up up

5378
03:09:04,479 --> 03:09:08,319
a ways in the chat

5379
03:09:06,000 --> 03:09:11,120
um but we can provide those um or at

5380
03:09:08,319 --> 03:09:13,840
least most of them that have released um

5381
03:09:11,120 --> 03:09:15,920
to you all and i just want to thank the

5382
03:09:13,840 --> 03:09:17,680
um i know carlos will do this but i

5383
03:09:15,920 --> 03:09:19,760
really want to thank

5384
03:09:17,680 --> 03:09:21,439
the speakers who um

5385
03:09:19,760 --> 03:09:23,279
took time out of their

5386
03:09:21,439 --> 03:09:25,439
their week to

5387
03:09:23,279 --> 03:09:26,319
prepare and you know time today to

5388
03:09:25,439 --> 03:09:27,439
present

5389
03:09:26,319 --> 03:09:29,680
and discuss

5390
03:09:27,439 --> 03:09:31,120
their research and um their expertise

5391
03:09:29,680 --> 03:09:33,120
with us

5392
03:09:31,120 --> 03:09:34,319
carlos are you there

5393
03:09:33,120 --> 03:09:35,359
i'm here

5394
03:09:34,319 --> 03:09:37,120
yes

5395
03:09:35,359 --> 03:09:39,600
no thanks sarah i mean this was a

5396
03:09:37,120 --> 03:09:41,359
wonderful workshop as i think we all

5397
03:09:39,600 --> 03:09:43,439
expected so i want to

5398
03:09:41,359 --> 03:09:45,359
to thank everybody for participating for

5399
03:09:43,439 --> 03:09:48,000
the questions and for sharing your your

5400
03:09:45,359 --> 03:09:49,760
knowledge and and that was no questions

5401
03:09:48,000 --> 03:09:51,279
and we'll have to we'll have to chew on

5402
03:09:49,760 --> 03:09:53,359
it and see

5403
03:09:51,279 --> 03:09:56,000
where we go but certainly you have a lot

5404
03:09:53,359 --> 03:09:57,680
of you have given us a lot to think to

5405
03:09:56,000 --> 03:09:58,479
think about so thanks as sarah was

5406
03:09:57,680 --> 03:10:00,479
saying

5407
03:09:58,479 --> 03:10:03,120
to all the speakers and everybody who

5408
03:10:00,479 --> 03:10:05,040
has attended the workshop and also to

5409
03:10:03,120 --> 03:10:07,279
the contractors for setting up

5410
03:10:05,040 --> 03:10:08,880
everything so that everything went um

5411
03:10:07,279 --> 03:10:11,520
went very smoothly

5412
03:10:08,880 --> 03:10:14,640
so sarah i'll send it back to you for

5413
03:10:11,520 --> 03:10:16,319
for you to close the workshop

5414
03:10:14,640 --> 03:10:18,160
all right thank you

5415
03:10:16,319 --> 03:10:22,080
um i don't think there is anything else

5416
03:10:18,160 --> 03:10:23,680
from us so um thank you so much for

5417
03:10:22,080 --> 03:10:24,880
for joining us everyone who's been on

5418
03:10:23,680 --> 03:10:26,880
the line

5419
03:10:24,880 --> 03:10:29,279
thanks for your engagement

5420
03:10:26,880 --> 03:10:31,840
and your interest and

5421
03:10:29,279 --> 03:10:33,120
look out for some additional information

5422
03:10:31,840 --> 03:10:35,040
as our

5423
03:10:33,120 --> 03:10:35,760
as our summary and our recording come

5424
03:10:35,040 --> 03:10:39,040
out

5425
03:10:35,760 --> 03:10:40,880
and please be in touch with any

5426
03:10:39,040 --> 03:10:45,160
any thoughts or comments that you'd like

5427
03:10:40,880 --> 03:10:45,160
to provide to us thank you