1
00:00:12,346 --> 00:00:15,249
Good morning.

2
00:00:15,249 --> 00:00:16,583
I'm going to try to do two
things

3
00:00:16,583 --> 00:00:19,062
in this talk, which I hope
don't come across as

4
00:00:19,062 --> 00:00:19,853
contradictory.

5
00:00:19,853 --> 00:00:21,785
If they do, please let me know
afterwards

6
00:00:21,785 --> 00:00:23,056
and I'll try not to do it.

7
00:00:23,524 --> 00:00:26,161
Or I'll try to avoid
the apparent contradiction in

8
00:00:26,161 --> 00:00:26,793
the future.

9
00:00:27,060 --> 00:00:30,063
So the two things first are.

10
00:00:31,398 --> 00:00:33,834
To argue that doing risk benefit

11
00:00:33,834 --> 00:00:35,986
assessment is obviously just
central

12
00:00:35,986 --> 00:00:38,138
to the ethics of clinical
research,

13
00:00:39,273 --> 00:00:41,875
but it's not done very often.

14
00:00:41,875 --> 00:00:43,510
It's not done systematically.

15
00:00:43,510 --> 00:00:45,712
I don't think it's done very
well.

16
00:00:45,712 --> 00:00:48,455
And so the first thing I want to
do

17
00:00:48,455 --> 00:00:49,082
is just

18
00:00:49,650 --> 00:00:52,263
present a framework for doing
systematic

19
00:00:52,263 --> 00:00:54,288
assessments of research risks,

20
00:00:54,288 --> 00:00:57,491
and then make the argument
that it's important to do this

21
00:00:57,491 --> 00:01:00,494
systematically and come.

22
00:01:01,094 --> 00:01:03,448
This is anecdotal,
but people I can just I'm

23
00:01:03,448 --> 00:01:04,197
convinced now

24
00:01:04,531 --> 00:01:05,666
that's how I measure where I got
here.

25
00:01:05,666 --> 00:01:06,867
So I've been doing this for 30
years.

26
00:01:06,867 --> 00:01:08,232
People just aren't doing this
and

27
00:01:08,232 --> 00:01:09,970
they're not doing it in a
systematic way.

28
00:01:09,970 --> 00:01:11,204
So that's the first part of it.

29
00:01:12,172 --> 00:01:13,740
Sorry, we're not doing
systematically.

30
00:01:13,740 --> 00:01:16,743
We need to do it systematically.

31
00:01:17,010 --> 00:01:19,039
So for people just care about
the basics

32
00:01:19,039 --> 00:01:20,814
of this, that framework is
enough.

33
00:01:20,814 --> 00:01:23,817
You can just not pay attention
to the rest of the talk.

34
00:01:23,817 --> 00:01:26,472
What the second part of the talk
is going to be, which I hope

35
00:01:26,472 --> 00:01:26,820
doesn't

36
00:01:26,820 --> 00:01:29,000
seem in tension with the first
part,

37
00:01:29,000 --> 00:01:30,090
is to think about

38
00:01:30,090 --> 00:01:33,860
some of the different challenges
complexities

39
00:01:34,227 --> 00:01:36,132
that arise when you start
thinking about

40
00:01:36,132 --> 00:01:37,798
how to do risk benefit
assessment.

41
00:01:37,798 --> 00:01:37,965
Right?

42
00:01:37,965 --> 00:01:41,001
So I think the basics
are pretty straightforward,

43
00:01:41,001 --> 00:01:43,591
but I think there are also
a lot of really interesting

44
00:01:43,591 --> 00:01:44,071
questions

45
00:01:44,071 --> 00:01:46,457
that arise once
you start thinking about this

46
00:01:46,457 --> 00:01:47,040
carefully.

47
00:01:47,040 --> 00:01:49,510
And these are questions
a lot of which haven't been

48
00:01:49,510 --> 00:01:50,043
addressed.

49
00:01:50,677 --> 00:01:53,609
So people
now know the Belmont Reports of

50
00:01:53,609 --> 00:01:54,181
Belmont

51
00:01:54,181 --> 00:01:57,295
Report is one of,
if not the sort of founding

52
00:01:57,295 --> 00:01:57,918
document

53
00:01:58,185 --> 00:02:00,481
for research ethics,
particularly research

54
00:02:00,481 --> 00:02:02,122
ethics, in the United States.

55
00:02:02,522 --> 00:02:04,887
And here are some things
that they have to say with

56
00:02:04,887 --> 00:02:05,258
respect

57
00:02:05,258 --> 00:02:07,060
to risk benefit assessment.

58
00:02:07,060 --> 00:02:09,539
They say things like,
we should try to do it as

59
00:02:09,539 --> 00:02:10,330
systematically

60
00:02:10,330 --> 00:02:13,533
as we can, as clearly as we can.

61
00:02:13,834 --> 00:02:17,070
I like this last one,
which is just never done.

62
00:02:17,070 --> 00:02:22,209
Is that you shouldn't just do it
based on intuition,

63
00:02:22,409 --> 00:02:25,684
but risk benefit assessments,
they're empirical questions,

64
00:02:25,684 --> 00:02:26,079
right?

65
00:02:26,079 --> 00:02:29,216
They're questions of
if I take a certain drug,

66
00:02:29,216 --> 00:02:30,751
what are the chances
it's going to help me?

67
00:02:30,751 --> 00:02:32,586
What are the chances
it's going to hurt me?

68
00:02:32,586 --> 00:02:34,221
Well, those are very empirical
questions.

69
00:02:34,221 --> 00:02:37,824
And presumably in order
to address them, it's

70
00:02:37,824 --> 00:02:40,260
valuable to have as much of the
relevant

71
00:02:40,260 --> 00:02:41,294
data as you can.

72
00:02:41,294 --> 00:02:42,562
Belmont to

73
00:02:43,630 --> 00:02:46,833
emphasize that 50 years ago now,
and basically nobody does it.

74
00:02:47,200 --> 00:02:50,184
And so one of the things
we've been trying to argue is

75
00:02:50,184 --> 00:02:50,570
trying

76
00:02:50,570 --> 00:02:53,781
to convince people of the value
and the importance of doing

77
00:02:53,781 --> 00:02:54,107
that.

78
00:02:54,307 --> 00:02:55,709
So I'm just going to give.

79
00:02:55,709 --> 00:02:59,179
So when I give talks,
I usually say I get to tangents,

80
00:02:59,913 --> 00:03:01,394
but what I do, what I realize
now is

81
00:03:01,394 --> 00:03:02,916
if I put some of them on the
slides,

82
00:03:02,916 --> 00:03:04,651
they're not actually tangents.
And then I get more.

83
00:03:04,651 --> 00:03:06,486
So I'm sneaking in more tangents
here.

84
00:03:06,486 --> 00:03:09,056
So this is to leading into this
talk.

85
00:03:09,056 --> 00:03:12,392
So this is a paper
you can see we did 20 years ago

86
00:03:12,392 --> 00:03:12,726
now.

87
00:03:12,726 --> 00:03:15,395
And CMA was actually a post back
fellow here.

88
00:03:15,395 --> 00:03:17,408
Her first. She came through the
department

89
00:03:17,408 --> 00:03:18,031
three times.

90
00:03:18,031 --> 00:03:19,399
This was her first time.

91
00:03:19,399 --> 00:03:22,703
And this was a survey of IRB
chairpersons

92
00:03:23,036 --> 00:03:24,855
who had worked in pediatric
research,

93
00:03:24,855 --> 00:03:26,673
reviewed pediatric research
studies.

94
00:03:27,007 --> 00:03:31,462
We gave them a list of common
research interventions, like a

95
00:03:31,462 --> 00:03:32,279
blood draw

96
00:03:32,612 --> 00:03:36,159
like an allergy, skin
testing like a bone marrow

97
00:03:36,159 --> 00:03:36,750
biopsy.

98
00:03:38,351 --> 00:03:40,400
And we
just asked them, how risky are

99
00:03:40,400 --> 00:03:40,787
these?

100
00:03:40,787 --> 00:03:42,189
Are they minimal risk?

101
00:03:42,189 --> 00:03:43,623
Are they a minor increase or a
mask?

102
00:03:43,623 --> 00:03:45,021
Are they more than a minor
increase

103
00:03:45,021 --> 00:03:45,859
over a minimal risk?

104
00:03:45,859 --> 00:03:47,127
And when you hear this talk,

105
00:03:47,127 --> 00:03:49,084
you'll learn more about those
categories

106
00:03:49,084 --> 00:03:50,797
and why they're the relevant
ones.

107
00:03:50,997 --> 00:03:53,133
But that's the assessment that
we did.

108
00:03:53,133 --> 00:03:55,361
And the data that everybody
focuses on

109
00:03:55,361 --> 00:03:57,237
from that paper are two things.

110
00:03:57,771 --> 00:04:01,074
One, that IRB chairs
tend to be really cautious.

111
00:04:01,274 --> 00:04:04,137
There is a big chunk of them
who said that even a single

112
00:04:04,137 --> 00:04:04,444
blood

113
00:04:04,444 --> 00:04:07,247
draw was more than minimum risk,
which it just isn't.

114
00:04:07,247 --> 00:04:08,582
That was the first thing.

115
00:04:08,582 --> 00:04:09,449
The second thing was that

116
00:04:09,449 --> 00:04:11,989
there was this dispersal,
just an amazing amount of

117
00:04:11,989 --> 00:04:12,686
disagreement.

118
00:04:12,886 --> 00:04:14,454
So for instance, allergy skin
testing,

119
00:04:14,454 --> 00:04:16,022
almost a third thought it was
minimal

120
00:04:16,022 --> 00:04:16,923
risk, almost a third

121
00:04:16,923 --> 00:04:19,647
thought it was a minor increase
over amount of risk, and almost

122
00:04:19,647 --> 00:04:19,993
a third

123
00:04:20,260 --> 00:04:22,444
thought it was more than a minor
increase

124
00:04:22,444 --> 00:04:23,563
over a minimal risk.

125
00:04:23,563 --> 00:04:26,666
So why are we getting that much
dispersion?

126
00:04:26,867 --> 00:04:29,042
Well, the other aspect of this
study,

127
00:04:29,042 --> 00:04:30,570
which we couldn't publish

128
00:04:30,570 --> 00:04:31,935
because we didn't have tests of
it,

129
00:04:31,935 --> 00:04:33,573
but I thought was especially
interesting.

130
00:04:34,040 --> 00:04:35,041
These were interviews.

131
00:04:35,041 --> 00:04:37,677
We called up 188 IRB
chairpersons.

132
00:04:37,677 --> 00:04:40,347
We gave them this list of
interventions.

133
00:04:40,347 --> 00:04:42,116
And we said, tell us, is this
minimal

134
00:04:42,116 --> 00:04:43,550
risk, minor increase our risk

135
00:04:43,550 --> 00:04:44,951
or more than my increase heart
intervals.

136
00:04:44,951 --> 00:04:47,531
So you multiply 188 times 8 or
9,

137
00:04:47,531 --> 00:04:50,657
whatever the heck that turns out
to be.

138
00:04:50,657 --> 00:04:50,957
Right?

139
00:04:50,957 --> 00:04:55,328
It's almost 50, 100, 2000
specific cases

140
00:04:55,328 --> 00:04:58,331
in which you're asking somebody
to categorize the risks

141
00:04:58,598 --> 00:05:02,417
of a procedure intervention,
and not, in one case, zero

142
00:05:02,417 --> 00:05:02,903
cases.

143
00:05:02,903 --> 00:05:05,405
Then any of the chairperson
say, I don't know.

144
00:05:07,674 --> 00:05:08,842
And that was really striking

145
00:05:08,842 --> 00:05:10,677
because a lot of these things
they didn't know about,

146
00:05:10,677 --> 00:05:13,680
they didn't have data on,
and they should have said, I

147
00:05:13,680 --> 00:05:14,347
don't know,

148
00:05:14,347 --> 00:05:18,151
because I can't do
what Belmont tells me to do.

149
00:05:18,151 --> 00:05:20,493
I can't do this systematic
assessment

150
00:05:20,493 --> 00:05:22,455
of a rail of the relevant data

151
00:05:22,455 --> 00:05:24,090
because I don't have it
and I don't know it.

152
00:05:24,090 --> 00:05:25,892
I'd have to look it up
and get back to you.

153
00:05:25,892 --> 00:05:27,928
Nobody did that.

154
00:05:27,928 --> 00:05:30,497
So what's going on?

155
00:05:30,497 --> 00:05:33,433
I think what people are doing is
and part of the problem is, and

156
00:05:33,433 --> 00:05:33,800
this is

157
00:05:34,367 --> 00:05:38,672
my previous career was spent
studying evolutionary biology

158
00:05:38,905 --> 00:05:40,540
and I think one of the things we
learned

159
00:05:40,540 --> 00:05:42,856
from evolution is that the
reason

160
00:05:42,856 --> 00:05:44,611
why we're here right now

161
00:05:44,845 --> 00:05:47,981
is because our predecessors were
very

162
00:05:48,348 --> 00:05:51,151
pretty good assessors of risks,

163
00:05:51,151 --> 00:05:54,254
and they made those assessments
really, really fast.

164
00:05:54,721 --> 00:05:56,189
So you imagine

165
00:05:57,524 --> 00:05:59,993
two populations

166
00:05:59,993 --> 00:06:02,162
saber tooth, their sounds
easily.

167
00:06:02,162 --> 00:06:03,997
It could be a saber tooth tiger.

168
00:06:03,997 --> 00:06:06,617
The one, the one group of our
ancestors

169
00:06:06,617 --> 00:06:08,902
were very thoughtful about
risks.

170
00:06:08,902 --> 00:06:11,753
And they started discussing it
and thinking, could this be a

171
00:06:11,753 --> 00:06:12,038
saber

172
00:06:12,038 --> 00:06:13,240
tooth tiger?

173
00:06:13,240 --> 00:06:15,493
I don't know, maybe
we should think about it for a

174
00:06:15,493 --> 00:06:15,809
while.

175
00:06:15,809 --> 00:06:17,210
Maybe we should investigate it.

176
00:06:17,210 --> 00:06:19,746
The other group ran away
when they heard the sound.

177
00:06:19,746 --> 00:06:23,637
Well, guess which one survived
and reproduced the second one,

178
00:06:23,637 --> 00:06:24,084
right.

179
00:06:24,084 --> 00:06:27,587
And so what we've got
is we've inherited this response

180
00:06:27,587 --> 00:06:29,941
where we just make risks
assessments

181
00:06:29,941 --> 00:06:30,790
really fast.

182
00:06:31,124 --> 00:06:33,298
That can be valuable in certain
cases,

183
00:06:33,298 --> 00:06:35,528
but in other cases it leads us
astray,

184
00:06:35,528 --> 00:06:38,123
particularly in research where
we're often

185
00:06:38,123 --> 00:06:40,100
dealing with unfamiliar things,

186
00:06:40,100 --> 00:06:42,359
where the heuristics that we
rely on

187
00:06:42,359 --> 00:06:44,304
aren't necessarily applicable.

188
00:06:44,304 --> 00:06:45,872
So that's IRB chairpersons.

189
00:06:45,872 --> 00:06:48,842
We're doing this
just give you one other example.

190
00:06:49,142 --> 00:06:51,378
I started doing this for a long
time.

191
00:06:51,378 --> 00:06:53,420
So people sometimes call me up
and say,

192
00:06:53,420 --> 00:06:55,148
could you help us decide whether

193
00:06:55,148 --> 00:06:56,866
this procedure, that procedure
is of

194
00:06:56,866 --> 00:06:58,918
minimal risk or greater than
minimal risk?

195
00:06:59,152 --> 00:07:00,720
These are some investigators

196
00:07:02,088 --> 00:07:02,555
who study

197
00:07:02,555 --> 00:07:05,693
metabolism, and
they often do IVF glucose

198
00:07:05,693 --> 00:07:06,993
tolerance tests.

199
00:07:07,427 --> 00:07:09,544
And the head of the national
consortium

200
00:07:09,544 --> 00:07:11,064
called me up once and said,

201
00:07:11,398 --> 00:07:12,465
Dave, you got to help us.

202
00:07:12,465 --> 00:07:14,991
All these IRBs think that IV
glucose

203
00:07:14,991 --> 00:07:17,937
tolerance test is more than
minimal risk.

204
00:07:18,171 --> 00:07:19,903
And you have to help us convince
them

205
00:07:19,903 --> 00:07:21,541
that they're actually minimal
risk

206
00:07:21,541 --> 00:07:24,411
because we're confident
that they are minimal risk.

207
00:07:24,411 --> 00:07:25,578
And I say, great.

208
00:07:25,578 --> 00:07:27,699
So when you submit these
protocols

209
00:07:27,699 --> 00:07:29,883
arguing that you glucose
tolerance

210
00:07:29,883 --> 00:07:33,460
tests are minimum risk,
what data do you present to the

211
00:07:33,460 --> 00:07:33,720
IRB

212
00:07:33,720 --> 00:07:35,755
to convince them
that these are minimal risk?

213
00:07:35,755 --> 00:07:37,390
You know what the guy said to
me?

214
00:07:37,390 --> 00:07:38,958
What are you talking about?

215
00:07:38,958 --> 00:07:40,794
Why would we present data?

216
00:07:40,794 --> 00:07:43,129
This is like a nationally
renowned

217
00:07:43,129 --> 00:07:46,749
clinical researcher
who spends his career collecting

218
00:07:46,749 --> 00:07:47,167
data.

219
00:07:47,701 --> 00:07:49,375
But then when it comes to
ethics, he's

220
00:07:49,375 --> 00:07:51,271
like, what's data got to do with
it, Dave?

221
00:07:51,271 --> 00:07:53,373
They're just minimal risk.

222
00:07:53,373 --> 00:07:54,708
So nobody's doing this.

223
00:07:54,708 --> 00:07:57,310
IRBs aren't doing it.
Investigators aren't doing this.

224
00:07:57,310 --> 00:08:00,013
And I should point out,
you can notice that paper

225
00:08:01,181 --> 00:08:02,148
30 years after

226
00:08:02,148 --> 00:08:05,819
Belmont, this example, almost 45
years

227
00:08:06,152 --> 00:08:07,885
after Belmont, it's been a long
time

228
00:08:07,885 --> 00:08:09,522
and people still aren't doing
it.

229
00:08:09,522 --> 00:08:13,760
So that's the set up for trying
to justify

230
00:08:13,760 --> 00:08:15,962
or give grounds
for why we need a framework.

231
00:08:15,962 --> 00:08:18,965
And then I'm going to propose
one and talk through it some.

232
00:08:18,965 --> 00:08:20,200
So just a really background.

233
00:08:20,200 --> 00:08:21,425
So this is just for the
philosophers

234
00:08:21,425 --> 00:08:21,901
in the crowd.

235
00:08:21,901 --> 00:08:23,603
Don't get too worked up about
this.

236
00:08:23,603 --> 00:08:26,439
So I'm gonna talk about things
like risks and benefits.

237
00:08:26,439 --> 00:08:28,041
And then people get confused.

238
00:08:28,041 --> 00:08:30,610
These are the terms
that the regulations use.

239
00:08:30,610 --> 00:08:33,498
But then people will say,
well, a risk is a chance of a

240
00:08:33,498 --> 00:08:33,813
harm.

241
00:08:34,114 --> 00:08:37,091
So if I'm going to stick a kid
with a needle, that's going to

242
00:08:37,091 --> 00:08:37,384
hurt,

243
00:08:37,951 --> 00:08:39,652
that must not count as a risk,
right?

244
00:08:39,652 --> 00:08:41,435
Because risks are some
probability

245
00:08:41,435 --> 00:08:43,690
of a harm where it's basically
guaranteed.

246
00:08:43,690 --> 00:08:44,773
If I stick a kid with a needle,
it's

247
00:08:44,773 --> 00:08:45,225
going to hurt.

248
00:08:45,225 --> 00:08:47,394
So that must not be a risk.
That must be something else.

249
00:08:47,394 --> 00:08:50,130
Now, these are actually
just terms of our risks.

250
00:08:51,331 --> 00:08:52,891
Apply to things where you're not
sure

251
00:08:52,891 --> 00:08:54,367
if a bad thing is going to
happen,

252
00:08:54,567 --> 00:08:56,834
and it's also where you're sure
some bad things are going to

253
00:08:56,834 --> 00:08:57,137
happen.

254
00:08:57,137 --> 00:09:00,190
And benefits aren't just things
that are good that definitely

255
00:09:00,190 --> 00:09:00,840
will happen.

256
00:09:00,840 --> 00:09:04,745
It's also potential benefits,
speculative benefits, chances of

257
00:09:04,745 --> 00:09:05,311
benefit.

258
00:09:05,512 --> 00:09:09,482
So think about these terms
very broadly okay.

259
00:09:09,482 --> 00:09:11,751
So here's the framework
that I'm going to go through.

260
00:09:11,751 --> 00:09:15,255
And we're talking about this
talk about our wonderful alums.

261
00:09:15,455 --> 00:09:16,256
The first author.

262
00:09:16,256 --> 00:09:18,015
And this was Annette a read who
was

263
00:09:18,015 --> 00:09:19,926
a fellow post-doc fellow at the
time.

264
00:09:20,193 --> 00:09:21,858
Right.
So she's now a member of the

265
00:09:21,858 --> 00:09:22,429
department.

266
00:09:22,429 --> 00:09:25,432
So what I'm going to do
is basically just go through

267
00:09:26,266 --> 00:09:28,334
each of the steps
in this framework, describe

268
00:09:28,334 --> 00:09:31,337
basically how I think it works
or should work.

269
00:09:31,671 --> 00:09:33,897
And then as I mentioned at the
top,

270
00:09:33,897 --> 00:09:35,041
I'm also going to

271
00:09:35,041 --> 00:09:36,966
go into some of the more
complexities

272
00:09:36,966 --> 00:09:38,111
that get raised here.

273
00:09:38,111 --> 00:09:39,546
I think try to chemistry.

274
00:09:39,546 --> 00:09:41,654
I think they're
still really interesting

275
00:09:41,654 --> 00:09:42,182
questions

276
00:09:42,182 --> 00:09:44,099
to try to address here for
anybody

277
00:09:44,099 --> 00:09:46,186
interested in doing research
ethics.

278
00:09:46,619 --> 00:09:47,120
Okay.

279
00:09:47,120 --> 00:09:49,789
So to start, one thing to notice
is that

280
00:09:51,624 --> 00:09:53,793
basically every clinical trial

281
00:09:53,793 --> 00:09:56,007
is a composite or an aggregate
of lots

282
00:09:56,007 --> 00:09:57,230
of different things.

283
00:09:57,931 --> 00:09:59,909
You bring people in, you weigh
them, you

284
00:09:59,909 --> 00:10:02,035
measure them, you ask them some
questions.

285
00:10:02,035 --> 00:10:05,498
You take two blood draws,
you give them a drug, you do an

286
00:10:05,498 --> 00:10:06,105
MRI scan.

287
00:10:06,306 --> 00:10:09,242
So there's all these different
components

288
00:10:09,242 --> 00:10:11,611
that make up a given study.

289
00:10:11,611 --> 00:10:14,113
And so the point of this
then when it comes to risk

290
00:10:14,113 --> 00:10:16,182
benefit assessment is
well what are you doing.

291
00:10:16,182 --> 00:10:19,082
The risk benefit assessment on
if there's all these different

292
00:10:19,082 --> 00:10:19,652
components,

293
00:10:19,986 --> 00:10:24,157
the answer is what you should do
is you do first.

294
00:10:24,157 --> 00:10:27,202
So if we're talking
from the perspective of an IRB,

295
00:10:27,202 --> 00:10:27,560
first

296
00:10:27,560 --> 00:10:30,174
they should look at the risk
benefit of each individual

297
00:10:30,174 --> 00:10:30,697
component.

298
00:10:30,997 --> 00:10:32,732
Assess each one of those,

299
00:10:32,732 --> 00:10:35,835
make sure they're acceptable
from a risk benefit perspective.

300
00:10:35,835 --> 00:10:38,805
And then look at the aggregate
or the study as a whole.

301
00:10:38,805 --> 00:10:42,142
So do both a component
and the study as a whole.

302
00:10:44,310 --> 00:10:45,879
Now not only there are lots

303
00:10:45,879 --> 00:10:47,453
of interventions, but they're
often

304
00:10:47,453 --> 00:10:48,982
different kinds of
interventions.

305
00:10:48,982 --> 00:10:50,868
A lot of studies will include
clinical

306
00:10:50,868 --> 00:10:52,952
interventions and research
interventions.

307
00:10:52,952 --> 00:10:56,289
So this is what standard
they call an add on trial

308
00:10:56,523 --> 00:10:58,758
where you give everybody
a standard of care.

309
00:10:58,758 --> 00:11:01,661
And then for the one arm you add
something experimental on

310
00:11:01,661 --> 00:11:04,664
top to see if that improves
standard of care.

311
00:11:04,964 --> 00:11:06,966
And so for the most part,

312
00:11:06,966 --> 00:11:08,852
when I IRBs are doing this
component

313
00:11:08,852 --> 00:11:10,737
analysis, they should be
evaluating

314
00:11:10,737 --> 00:11:12,562
the components that are
research,

315
00:11:12,562 --> 00:11:14,941
not the components that are
clinical care.

316
00:11:15,441 --> 00:11:16,743
And this is a statement

317
00:11:16,743 --> 00:11:19,846
basically directing our base
to do that from the U.S.

318
00:11:19,846 --> 00:11:20,747
regulations.

319
00:11:20,747 --> 00:11:22,572
So what they should be looking
at

320
00:11:22,572 --> 00:11:24,951
is the risks and benefits of the
research.

321
00:11:25,318 --> 00:11:27,445
And this is distinguished from
things,

322
00:11:27,445 --> 00:11:28,788
for instance, therapies

323
00:11:28,788 --> 00:11:31,634
that subjects would get
even if they weren't in the

324
00:11:31,634 --> 00:11:32,191
research.

325
00:11:33,526 --> 00:11:35,195
So the thought here is why do
you do that?

326
00:11:35,195 --> 00:11:37,797
Well, the thought here is
if it's clinically indicated,

327
00:11:37,797 --> 00:11:39,657
then the risk benefit profile of
that

328
00:11:39,657 --> 00:11:41,668
intervention is going to be
acceptable.

329
00:11:41,668 --> 00:11:43,610
That's what it means essentially
for it

330
00:11:43,610 --> 00:11:45,004
to be clinically indicated.

331
00:11:45,004 --> 00:11:46,902
So for the most part, IRBs don't
have

332
00:11:46,902 --> 00:11:49,108
to worry about the clinical
interventions.

333
00:11:49,375 --> 00:11:50,176
The caveat here.

334
00:11:50,176 --> 00:11:52,216
So the arrow is going to be
different

335
00:11:52,216 --> 00:11:52,712
caveats.

336
00:11:52,712 --> 00:11:55,848
If we go through this talk here,
the worry is that

337
00:11:56,716 --> 00:11:58,728
what I at least need to be
worried about,

338
00:11:58,728 --> 00:12:00,053
though, is the possibility

339
00:12:00,053 --> 00:12:03,723
that that add on treatment
in the one arm of the study

340
00:12:03,957 --> 00:12:05,999
that that might influence the
risk

341
00:12:05,999 --> 00:12:08,461
benefit profile of the standard
of care,

342
00:12:08,761 --> 00:12:11,022
it might make it more risky,
it might make it less

343
00:12:11,022 --> 00:12:11,564
beneficial,

344
00:12:11,564 --> 00:12:13,316
it might make it more
beneficial.

345
00:12:13,316 --> 00:12:13,900
Who knows.

346
00:12:13,900 --> 00:12:16,216
But so you can't completely
break it off

347
00:12:16,216 --> 00:12:17,837
the clinical interventions.

348
00:12:17,837 --> 00:12:19,939
You have to think about
if there are any interactions

349
00:12:19,939 --> 00:12:22,048
between the research
interventions

350
00:12:22,048 --> 00:12:24,344
and the clinical interventions
okay.

351
00:12:24,344 --> 00:12:25,979
So now here's the framework.

352
00:12:25,979 --> 00:12:28,081
So first step is that.

353
00:12:30,450 --> 00:12:30,950
We have

354
00:12:30,950 --> 00:12:32,872
this is the stick for the
department

355
00:12:32,872 --> 00:12:33,886
clinical research.

356
00:12:33,886 --> 00:12:37,290
Is it ethically acceptable
unless it's socially valuable.

357
00:12:37,290 --> 00:12:39,552
So the way you apply that here
is

358
00:12:39,552 --> 00:12:42,362
you first make sure that for
each of the

359
00:12:42,362 --> 00:12:44,714
interventions, the research
interventions

360
00:12:44,714 --> 00:12:47,066
that there's a socially valuable
reason.

361
00:12:47,066 --> 00:12:49,802
Why is that intervention in the
study.

362
00:12:49,802 --> 00:12:52,620
So you ask the investigator
you have three MRI's in this

363
00:12:52,620 --> 00:12:52,972
study.

364
00:12:52,972 --> 00:12:54,841
Why are you doing three MRI?

365
00:12:54,841 --> 00:12:56,075
What are you going to learn
from?

366
00:12:56,075 --> 00:12:59,912
For instance, the third MRI that
you couldn't learn from the

367
00:12:59,912 --> 00:13:00,680
second MRI.

368
00:13:00,680 --> 00:13:03,783
And you want a good explanation
for that.

369
00:13:04,450 --> 00:13:06,973
What are the things I'm going to
emphasize

370
00:13:06,973 --> 00:13:07,453
is that

371
00:13:08,554 --> 00:13:10,623
I sit on IRBs

372
00:13:10,623 --> 00:13:13,259
when I talk to people in IRBs,
I talk about how important

373
00:13:13,259 --> 00:13:14,927
this stuff actually is in
practice.

374
00:13:14,927 --> 00:13:17,964
I rattle on about this in IRB
meetings

375
00:13:18,264 --> 00:13:20,108
and then what the IRB on
actually says,

376
00:13:20,108 --> 00:13:21,668
okay, here's the protocol, Dave.

377
00:13:21,668 --> 00:13:23,913
You tell us whether the risks
and

378
00:13:23,913 --> 00:13:26,839
the benefits of this study are
acceptable.

379
00:13:27,073 --> 00:13:31,277
And my answer is almost always,
I have absolutely no idea,

380
00:13:31,511 --> 00:13:33,314
because to make these
assessments

381
00:13:33,314 --> 00:13:35,281
like whether or not you really
need

382
00:13:35,581 --> 00:13:38,415
the third MRI,
whether testing this new kind of

383
00:13:38,415 --> 00:13:39,018
treatment

384
00:13:39,018 --> 00:13:40,909
for leukemia, whether those
things

385
00:13:40,909 --> 00:13:42,188
are socially valuable,

386
00:13:42,855 --> 00:13:44,439
being able to make those
judgments,

387
00:13:44,439 --> 00:13:45,525
you have to know a lot.

388
00:13:45,525 --> 00:13:46,659
You have to know a lot of stuff

389
00:13:46,659 --> 00:13:50,563
that I just don't know
about, for instance, leukemia.

390
00:13:50,763 --> 00:13:54,067
So whenever you do this,
you're really need people

391
00:13:54,067 --> 00:13:57,603
who have content expertise
to make these judgments.

392
00:13:57,603 --> 00:13:59,694
You can't you can rely on
philosophers

393
00:13:59,694 --> 00:14:00,740
for the framework,

394
00:14:00,740 --> 00:14:03,045
but you can't rely on
philosophers

395
00:14:03,045 --> 00:14:04,944
to implement the framework.

396
00:14:04,944 --> 00:14:07,317
So why don't you do
the individual interventions

397
00:14:07,317 --> 00:14:07,613
that,

398
00:14:07,613 --> 00:14:10,076
as I mentioned, you put them all
together

399
00:14:10,076 --> 00:14:11,818
as the package that they are

400
00:14:11,818 --> 00:14:14,628
and you evaluate the overall
study

401
00:14:14,628 --> 00:14:17,190
to see if it has social value.

402
00:14:18,825 --> 00:14:21,928
Typically, people will point out

403
00:14:21,928 --> 00:14:23,299
you hope to gather important
data

404
00:14:23,299 --> 00:14:24,297
from every given trial.

405
00:14:24,297 --> 00:14:25,098
But presumably,

406
00:14:25,098 --> 00:14:27,322
or at least it's rare that any
given trial

407
00:14:27,322 --> 00:14:29,335
is going to give you definitive
data.

408
00:14:29,335 --> 00:14:31,424
Typically, you have to combine
the data

409
00:14:31,424 --> 00:14:33,673
from one trial with data that
you collect

410
00:14:33,906 --> 00:14:35,341
from other trials.

411
00:14:35,341 --> 00:14:38,344
And again,
this requires a lot of expertise

412
00:14:39,479 --> 00:14:39,979
okay.

413
00:14:39,979 --> 00:14:42,777
So let's just talk about
some of the challenges at this

414
00:14:42,777 --> 00:14:43,082
step.

415
00:14:43,382 --> 00:14:45,304
So one of the things that I
think

416
00:14:45,304 --> 00:14:46,586
is always interesting

417
00:14:48,421 --> 00:14:50,315
and my experience of being on
IRB

418
00:14:50,315 --> 00:14:52,725
for 25 years is IRBs differ on
the answer

419
00:14:52,725 --> 00:14:56,210
to this question,
which is imagine that the

420
00:14:56,210 --> 00:14:57,263
investigator

421
00:14:58,397 --> 00:14:59,599
submits a protocol to you.

422
00:14:59,599 --> 00:15:00,800
So that's how I IRBs work.

423
00:15:00,800 --> 00:15:02,773
The investigator submits a
protocol,

424
00:15:02,773 --> 00:15:03,870
the IRB reviews it.

425
00:15:03,870 --> 00:15:04,670
So here's a question.

426
00:15:04,670 --> 00:15:06,839
What's the job of the IRB?

427
00:15:06,839 --> 00:15:08,074
It's the job of the IRB.

428
00:15:08,074 --> 00:15:11,978
Just to look at that study
that got submitted to them,

429
00:15:11,978 --> 00:15:14,117
and just see whether or not that
study

430
00:15:14,117 --> 00:15:15,581
is ethically appropriate,

431
00:15:15,882 --> 00:15:18,651
or should somebody do something
like,

432
00:15:18,651 --> 00:15:22,255
yeah, you want to do X
with this patient population.

433
00:15:22,255 --> 00:15:23,523
But here's this other thing

434
00:15:24,490 --> 00:15:25,747
you could do with the same
patient

435
00:15:25,747 --> 00:15:27,226
population that would be more
valuable,

436
00:15:27,226 --> 00:15:30,185
that would be more interesting,
that would give us more valuable

437
00:15:30,185 --> 00:15:30,463
data.

438
00:15:30,863 --> 00:15:33,564
Is it the job of the IRB or
somebody else

439
00:15:33,564 --> 00:15:35,935
to make these comparative
judgments

440
00:15:36,602 --> 00:15:38,679
across studies, right to this
study,

441
00:15:38,679 --> 00:15:40,006
rather than that study

442
00:15:41,474 --> 00:15:44,343
that at least I've sat on almost
never do?

443
00:15:44,343 --> 00:15:47,013
Maybe they should be doing it
or somebody should be doing it.

444
00:15:47,013 --> 00:15:48,953
And the other one is just
comparative

445
00:15:48,953 --> 00:15:50,316
judgments within a trial.

446
00:15:50,516 --> 00:15:53,886
Well, you designed it with just
two MRI's.

447
00:15:53,886 --> 00:15:56,790
But if you added this third MRI,
look at all this other

448
00:15:56,790 --> 00:15:57,423
information

449
00:15:57,857 --> 00:16:00,371
you could get that would make
your trial

450
00:16:00,371 --> 00:16:01,627
even more valuable.

451
00:16:02,061 --> 00:16:03,830
What should it be do in that
case?

452
00:16:03,830 --> 00:16:06,098
They think that the study

453
00:16:06,098 --> 00:16:09,101
that was proposed is okay,
but here's a much better study.

454
00:16:09,402 --> 00:16:11,370
Should they say you then do the
advisory?

455
00:16:11,370 --> 00:16:14,166
You can't do this one
because you could do this better

456
00:16:14,166 --> 00:16:14,373
one

457
00:16:14,640 --> 00:16:18,511
or does it just have to meet
some sufficiency standard?

458
00:16:18,511 --> 00:16:20,076
And then there's the question as
you go

459
00:16:20,076 --> 00:16:21,480
through this about who should
make

460
00:16:22,381 --> 00:16:23,115
these judgments.

461
00:16:23,115 --> 00:16:25,084
We could talk about
enhancements.

462
00:16:25,084 --> 00:16:27,416
So this Annette and I used to
have

463
00:16:27,416 --> 00:16:30,022
this debate when we wrote that
paper.

464
00:16:30,256 --> 00:16:32,191
So typically most clinical
research

465
00:16:32,191 --> 00:16:34,053
focuses on diseases and
illnesses

466
00:16:34,053 --> 00:16:36,028
like cancers and things like
that.

467
00:16:37,296 --> 00:16:38,965
But what about

468
00:16:38,965 --> 00:16:40,585
trials
to do what people call

469
00:16:40,585 --> 00:16:41,367
enhancements,

470
00:16:41,367 --> 00:16:44,303
which is not to treat an illness
or a disease,

471
00:16:44,303 --> 00:16:46,044
but to make people who are
healthy

472
00:16:46,044 --> 00:16:48,040
even better off, to make them
happier,

473
00:16:48,040 --> 00:16:50,368
to make them healthier, to make
them live

474
00:16:50,368 --> 00:16:52,411
longer, is that is equally
valuable

475
00:16:52,411 --> 00:16:55,448
as doing something like treating
cancer?

476
00:16:55,448 --> 00:16:57,941
And again,
should people be making those

477
00:16:57,941 --> 00:16:58,751
assessments,

478
00:16:59,185 --> 00:17:02,188
and if so, who should be making
them?

479
00:17:02,955 --> 00:17:04,991
So this is just some
suggestions.

480
00:17:04,991 --> 00:17:08,261
When I gave that previous slide,
I get a lot of pushback from IRB

481
00:17:08,261 --> 00:17:08,995
folks like that.

482
00:17:08,995 --> 00:17:10,257
You shouldn't
be making these comparative

483
00:17:10,257 --> 00:17:10,596
judgments.

484
00:17:10,596 --> 00:17:13,332
That's not what I IRB
are supposed to be doing.

485
00:17:13,332 --> 00:17:14,500
So what I started doing

486
00:17:14,500 --> 00:17:17,621
is trying to collect
at least some evidence to

487
00:17:17,621 --> 00:17:18,571
suggest this.

488
00:17:18,571 --> 00:17:21,265
In some cases,
I think IRBs should be doing

489
00:17:21,265 --> 00:17:21,641
that.

490
00:17:21,974 --> 00:17:24,611
And these are cases now,
interestingly,

491
00:17:24,611 --> 00:17:27,113
they're from pandemics or
epidemics.

492
00:17:27,113 --> 00:17:30,126
So the first one was in West
Africa

493
00:17:30,126 --> 00:17:31,417
during a Ebola

494
00:17:31,651 --> 00:17:35,794
where the problem was there's
only so many patients, and all

495
00:17:35,794 --> 00:17:36,622
of a sudden

496
00:17:36,622 --> 00:17:39,992
you've got all of this money and
resources

497
00:17:40,259 --> 00:17:43,429
going into the area to study
Ebola.

498
00:17:43,429 --> 00:17:46,232
And what these guys point was,
is that

499
00:17:46,232 --> 00:17:47,718
there's a limited number of
patients

500
00:17:47,718 --> 00:17:49,368
for all the studies that we want
to do.

501
00:17:49,602 --> 00:17:52,405
And so if we really want to get
the most social value

502
00:17:52,405 --> 00:17:54,971
out of this situation, we're
going to need

503
00:17:54,971 --> 00:17:57,109
to prioritize which of the
studies

504
00:17:57,109 --> 00:17:59,399
we allow act to access these
patients

505
00:17:59,399 --> 00:18:00,947
and which ones we don't.

506
00:18:01,147 --> 00:18:03,449
And then a similar thing
this is arguments.

507
00:18:03,449 --> 00:18:08,460
More recently during Covid where
there's arguments early in the

508
00:18:08,460 --> 00:18:09,255
pandemic,

509
00:18:09,255 --> 00:18:11,934
we wanted to learn as much as we
could

510
00:18:11,934 --> 00:18:14,260
about Covid, at least initially.

511
00:18:14,260 --> 00:18:16,702
Obviously, in the long run
there was lots and lots of

512
00:18:16,702 --> 00:18:17,163
patients,

513
00:18:17,163 --> 00:18:19,065
but initially there weren't that
many.

514
00:18:19,065 --> 00:18:20,967
And who should be?

515
00:18:20,967 --> 00:18:24,804
Should somebody be prioritizing
and deciding which studies to do

516
00:18:25,037 --> 00:18:26,973
and which studies not to do
these?

517
00:18:26,973 --> 00:18:29,175
This group argue
that at least within

518
00:18:29,175 --> 00:18:30,009
institutions.

519
00:18:32,078 --> 00:18:34,077
So, for instance, the NIH
Clinical Center

520
00:18:34,077 --> 00:18:35,247
should be prioritizing.

521
00:18:35,247 --> 00:18:37,090
Johns Hopkins should be
prioritizing,

522
00:18:37,090 --> 00:18:38,584
at least within institutions.

523
00:18:38,584 --> 00:18:40,753
There should be some
prioritization.

524
00:18:40,753 --> 00:18:42,722
Okay. So that's potential
benefits.

525
00:18:42,722 --> 00:18:44,757
Social value.

526
00:18:44,757 --> 00:18:47,393
So the next thing to look at is
the risks.

527
00:18:47,393 --> 00:18:49,495
So the idea is to identify the
risks

528
00:18:49,495 --> 00:18:51,831
of each of the individual
interventions

529
00:18:52,131 --> 00:18:55,166
and then try to reduce them to
the extent

530
00:18:55,166 --> 00:18:56,202
that you can.

531
00:18:56,602 --> 00:18:59,038
Now I'll come back to this in a
second.

532
00:18:59,038 --> 00:19:01,805
One of the things that everybody
agrees

533
00:19:01,805 --> 00:19:04,076
at this step is that you should

534
00:19:04,076 --> 00:19:06,481
look at all of the risks in a
very,

535
00:19:06,481 --> 00:19:07,580
very broad way.

536
00:19:07,580 --> 00:19:11,350
So obviously that means health
risks, like

537
00:19:11,584 --> 00:19:13,671
is this drug
potentially going to knock out

538
00:19:13,671 --> 00:19:14,253
my kidneys,

539
00:19:14,253 --> 00:19:15,697
and I'm going to have to be on
dialysis

540
00:19:15,697 --> 00:19:16,622
for the rest of my life.

541
00:19:16,622 --> 00:19:19,200
That's obviously
a risk of being in a clinical

542
00:19:19,200 --> 00:19:19,592
trial.

543
00:19:19,592 --> 00:19:22,918
But when people emphasize
there's other sorts of risks,

544
00:19:22,918 --> 00:19:23,462
there's,

545
00:19:23,462 --> 00:19:24,964
for instance, economic risks.

546
00:19:24,964 --> 00:19:29,101
What if knocking out my kidneys,
I have a copay of $1,000,

547
00:19:29,335 --> 00:19:31,155
and then I'm going to have to
start paying

548
00:19:31,155 --> 00:19:32,672
thousands of dollars out of
pocket

549
00:19:32,672 --> 00:19:35,098
to get that kidney problem
fixed,

550
00:19:35,098 --> 00:19:36,642
or at least treated.

551
00:19:36,942 --> 00:19:39,438
What about the psychological
harms

552
00:19:39,438 --> 00:19:40,980
of being on dialysis

553
00:19:40,980 --> 00:19:42,214
and how that makes me feel?

554
00:19:42,214 --> 00:19:44,380
What about stigma if I'm on
dialysis

555
00:19:44,380 --> 00:19:45,885
and how I'm stigmatized?

556
00:19:46,252 --> 00:19:48,854
The argument here with respect
to risks

557
00:19:48,854 --> 00:19:51,724
is that all of those should be
considered.

558
00:19:51,724 --> 00:19:53,192
I'm going to come back to that
then

559
00:19:53,192 --> 00:19:54,772
when we talk about potential
benefits,

560
00:19:54,772 --> 00:19:56,062
I mean, it seems right, right.

561
00:19:56,062 --> 00:19:57,660
People
care about more than just their

562
00:19:57,660 --> 00:19:57,997
bodies.

563
00:19:57,997 --> 00:20:00,142
They care about their bodies,
but they care about their bank

564
00:20:00,142 --> 00:20:00,499
accounts.

565
00:20:00,499 --> 00:20:01,575
Some people care about their
bank

566
00:20:01,575 --> 00:20:01,901
accounts.

567
00:20:01,901 --> 00:20:03,736
More than they do their bodies,

568
00:20:03,736 --> 00:20:07,118
so that we should be worried
about all the different types of

569
00:20:07,118 --> 00:20:07,506
risks.

570
00:20:10,509 --> 00:20:12,178
So there's a couple challenges
with this.

571
00:20:12,178 --> 00:20:15,660
And the basic idea
is that what this step says is

572
00:20:15,660 --> 00:20:16,015
that

573
00:20:16,515 --> 00:20:19,518
in order to decide
whether a study is acceptable

574
00:20:19,985 --> 00:20:21,753
before it can begin, the IRB has
to

575
00:20:21,753 --> 00:20:23,823
make sure that the risks are
acceptable.

576
00:20:24,623 --> 00:20:25,724
It's obviously a conundrum.

577
00:20:25,724 --> 00:20:27,816
There is that that's typically
the point

578
00:20:27,816 --> 00:20:29,228
of clinical trials, right?

579
00:20:29,228 --> 00:20:31,872
At least early phase
clinical trials are designed to

580
00:20:31,872 --> 00:20:32,431
figure out

581
00:20:32,731 --> 00:20:35,701
what the risks are
of a given intervention are.

582
00:20:35,701 --> 00:20:38,953
So if the goal of the trial
is to find out what the risks

583
00:20:38,953 --> 00:20:39,238
are,

584
00:20:39,238 --> 00:20:41,069
how can you prospectively
evaluate

585
00:20:41,069 --> 00:20:43,008
whether those risks are
acceptable?

586
00:20:43,008 --> 00:20:44,934
It seems like you got to do the
trial

587
00:20:44,934 --> 00:20:46,078
before you can decide

588
00:20:46,078 --> 00:20:48,814
whether the trial is ethical to
do.

589
00:20:48,814 --> 00:20:52,143
And so in this, basically
what IRBs need to do is they

590
00:20:52,143 --> 00:20:52,451
need

591
00:20:52,451 --> 00:20:54,592
to be creative, they need to
expand

592
00:20:54,592 --> 00:20:56,489
beyond the given intervention.

593
00:20:56,489 --> 00:20:58,724
If there's no data on it.
Sometimes there is.

594
00:20:58,724 --> 00:21:01,577
Sometimes this intervention
has been used before in other

595
00:21:01,577 --> 00:21:02,228
populations.

596
00:21:02,661 --> 00:21:05,131
See how you can extrapolate
to this population.

597
00:21:05,131 --> 00:21:07,900
Sometimes this specific drug
hasn't been given

598
00:21:07,900 --> 00:21:10,543
to this population, but drugs of
a similar

599
00:21:10,543 --> 00:21:12,304
same class have been given.

600
00:21:12,304 --> 00:21:14,673
And you can try to extrapolate
from those data.

601
00:21:14,673 --> 00:21:16,857
Or maybe it's just the mechanism
of action

602
00:21:16,857 --> 00:21:18,677
between this drug and other
drugs.

603
00:21:18,677 --> 00:21:19,612
So basically they're trying to
look

604
00:21:19,612 --> 00:21:23,115
for relevantly similar cases
for which there are data

605
00:21:23,349 --> 00:21:25,254
and then seeing how they can
apply them

606
00:21:25,254 --> 00:21:26,719
to this one, with the thought

607
00:21:27,019 --> 00:21:29,612
that the further you're
extrapolating,

608
00:21:29,612 --> 00:21:31,590
the greater the uncertainty.

609
00:21:31,590 --> 00:21:34,583
And then you should factor
that uncertainty into the risk

610
00:21:34,583 --> 00:21:35,161
assessment

611
00:21:35,161 --> 00:21:37,229
with a thought
that the more uncertain you are,

612
00:21:37,229 --> 00:21:39,965
the riskier
the intervention or the study

613
00:21:39,965 --> 00:21:40,232
is.

614
00:21:40,866 --> 00:21:42,168
So here's another challenge.

615
00:21:45,171 --> 00:21:46,570
The first one was that you got
to know

616
00:21:46,570 --> 00:21:48,007
something about the intervention
here.

617
00:21:48,007 --> 00:21:49,842
The point is that you don't have
to know

618
00:21:49,842 --> 00:21:50,943
about the intervention.

619
00:21:50,943 --> 00:21:52,755
You have to know who the
intervention

620
00:21:52,755 --> 00:21:53,979
is going to be put into.

621
00:21:53,979 --> 00:21:54,446
Right.

622
00:21:54,446 --> 00:21:57,146
So the risks of a certain drug
can vary

623
00:21:57,146 --> 00:21:59,084
dramatically depending upon

624
00:21:59,084 --> 00:22:00,252
who's going to receive it.

625
00:22:00,252 --> 00:22:03,844
But you don't know who's going
to receive it until the trial

626
00:22:03,844 --> 00:22:04,323
starts.

627
00:22:04,323 --> 00:22:06,453
So the example I have here is
drugs

628
00:22:06,453 --> 00:22:08,460
that are cleared by the kidneys.

629
00:22:08,827 --> 00:22:10,780
They might be very safe for
people,

630
00:22:10,780 --> 00:22:12,231
have good kidney function

631
00:22:12,231 --> 00:22:14,354
who can excrete the drug
properly,

632
00:22:14,354 --> 00:22:16,602
but they might be really
dangerous.

633
00:22:16,602 --> 00:22:18,679
The drug might build up in the
systems of

634
00:22:18,679 --> 00:22:20,806
people who have impaired kidney
function.

635
00:22:20,806 --> 00:22:23,173
So another thing that I need to
be aware

636
00:22:23,173 --> 00:22:25,244
about is the extent to which
risks

637
00:22:25,244 --> 00:22:28,722
and potential benefits
depend upon who enrolls in the

638
00:22:28,722 --> 00:22:29,181
study.

639
00:22:29,481 --> 00:22:32,418
And so the way IRBs
try to handle this is with

640
00:22:33,986 --> 00:22:35,487
inclusion exclusion criteria.

641
00:22:35,487 --> 00:22:37,223
You try to figure out who's
going to face

642
00:22:37,223 --> 00:22:40,443
excessive risks
and then you try to exclude

643
00:22:40,443 --> 00:22:40,893
them.

644
00:22:41,126 --> 00:22:43,293
The other thing
I mention here is the

645
00:22:43,293 --> 00:22:43,996
importance.

646
00:22:43,996 --> 00:22:46,189
Sometimes you just don't know
how a given

647
00:22:46,189 --> 00:22:47,900
drug is going to affect people.

648
00:22:47,900 --> 00:22:52,171
And so there's this importance
of monitoring the trial as you

649
00:22:52,171 --> 00:22:52,871
go along.

650
00:22:52,871 --> 00:22:54,807
And this goes back to Holly's
talk.

651
00:22:54,807 --> 00:22:56,759
If you start monitoring it in
the third

652
00:22:56,759 --> 00:22:57,910
or fourth participant,

653
00:22:57,910 --> 00:23:00,682
you start seeing a trend of
significant

654
00:23:00,682 --> 00:23:01,180
harms.

655
00:23:01,180 --> 00:23:04,842
Then you need to stop the study
and reevaluate and see whether

656
00:23:04,842 --> 00:23:05,551
or not it's

657
00:23:05,551 --> 00:23:08,554
okay to keep going from.

658
00:23:09,555 --> 00:23:10,619
All right, so this is going to
be

659
00:23:10,619 --> 00:23:11,457
even more into the weeds.

660
00:23:11,457 --> 00:23:12,157
So this is the end.

661
00:23:12,157 --> 00:23:14,210
Another I did this one yesterday
just

662
00:23:14,210 --> 00:23:16,595
to make sure I really drive you
guys nuts.

663
00:23:16,895 --> 00:23:19,495
And so for the most part
when I've been thinking about

664
00:23:19,495 --> 00:23:19,832
risks,

665
00:23:19,832 --> 00:23:22,897
I've been talking about risks
to the individuals who are in

666
00:23:22,897 --> 00:23:23,469
the study.

667
00:23:24,003 --> 00:23:26,605
And that's the standard risk
benefit assessment.

668
00:23:26,605 --> 00:23:28,741
And clinical research is done.

669
00:23:28,741 --> 00:23:30,721
More recently, people have
started

670
00:23:30,721 --> 00:23:32,177
to worry about the risks

671
00:23:32,177 --> 00:23:34,242
to what are just generally
called

672
00:23:34,242 --> 00:23:35,180
third parties.

673
00:23:35,447 --> 00:23:37,149
So here's an obvious example.

674
00:23:39,251 --> 00:23:41,658
Challenge trials are
trials where you take an

675
00:23:41,658 --> 00:23:42,621
infectious agent,

676
00:23:42,621 --> 00:23:46,724
like a virus, and you inject it
into a couple of research

677
00:23:46,724 --> 00:23:47,660
participants

678
00:23:47,660 --> 00:23:48,627
and you see what happens.

679
00:23:48,627 --> 00:23:50,462
Maybe you're going
to test different treatments,

680
00:23:50,462 --> 00:23:53,258
or are you going to try to study
the natural history of the

681
00:23:53,258 --> 00:23:53,732
disorder?

682
00:23:54,033 --> 00:23:57,810
So obviously, injecting viruses
into the participants poses

683
00:23:57,810 --> 00:23:58,771
risks to them.

684
00:23:59,238 --> 00:24:01,040
That's what we've been talking
about.

685
00:24:01,040 --> 00:24:03,981
But the point here is that
that can also pose risks to

686
00:24:03,981 --> 00:24:04,743
other people.

687
00:24:04,743 --> 00:24:06,823
If you're infecting the
participants

688
00:24:06,823 --> 00:24:08,614
and then the participants work

689
00:24:08,614 --> 00:24:11,522
with the nurses, the
participants go on

690
00:24:11,522 --> 00:24:14,653
weekend passes, the participants
go home.

691
00:24:14,887 --> 00:24:17,756
They could transmit that
infection

692
00:24:17,756 --> 00:24:20,759
to lots of other people.

693
00:24:22,995 --> 00:24:23,662
So what do you think?

694
00:24:23,662 --> 00:24:26,479
So that seems,
that may I won't force people to

695
00:24:26,479 --> 00:24:26,899
think,

696
00:24:26,899 --> 00:24:30,002
but that seems like
I should take that into account.

697
00:24:30,002 --> 00:24:31,337
Right? Does everybody agree?

698
00:24:31,337 --> 00:24:32,705
Yeah, of course
you should take that into

699
00:24:32,705 --> 00:24:33,005
account.

700
00:24:33,005 --> 00:24:34,659
So it seems like you should take
risks

701
00:24:34,659 --> 00:24:36,008
to third parties into account.

702
00:24:36,408 --> 00:24:38,382
But now the question is, how
broad

703
00:24:38,382 --> 00:24:40,646
do you think that principle
should be.

704
00:24:40,646 --> 00:24:43,349
So these are actually some real
examples.

705
00:24:43,349 --> 00:24:46,352
And we did an ethics grand
rounds on this

706
00:24:46,952 --> 00:24:49,488
about ten years ago where

707
00:24:50,856 --> 00:24:52,725
there was a participant

708
00:24:52,725 --> 00:24:55,861
who was a perfect candidate for
the study,

709
00:24:56,428 --> 00:24:58,315
and the study offered the
participant

710
00:24:58,315 --> 00:24:59,999
the chance, a potential benefit.

711
00:25:00,499 --> 00:25:02,963
But it turned out that the
participant

712
00:25:02,963 --> 00:25:05,104
was the caretaker for two people

713
00:25:05,104 --> 00:25:07,396
who were really sick, and the
participant

714
00:25:07,396 --> 00:25:09,241
could leave those people behind.

715
00:25:10,309 --> 00:25:11,983
So what the potential
participant

716
00:25:11,983 --> 00:25:14,012
was going to do was, was going
to drive

717
00:25:14,012 --> 00:25:18,105
from the west to Bethesda
with those two people in the

718
00:25:18,105 --> 00:25:18,484
car,

719
00:25:19,918 --> 00:25:21,653
and the home physician was
really worried

720
00:25:21,653 --> 00:25:24,095
that that drive could be
dangerous

721
00:25:24,095 --> 00:25:25,891
to the other two people,

722
00:25:25,891 --> 00:25:28,894
not the participant,
but to the other two people.

723
00:25:30,295 --> 00:25:32,112
Another example we just heard
about

724
00:25:32,112 --> 00:25:34,032
in a recent ethics grant rounds
was,

725
00:25:35,200 --> 00:25:37,147
participant who was undergoing
bone

726
00:25:37,147 --> 00:25:39,204
marrow transplant that required
them

727
00:25:39,204 --> 00:25:41,223
to stay in the clinical center
here

728
00:25:41,223 --> 00:25:42,608
for about three months.

729
00:25:43,275 --> 00:25:44,925
It turned out they had little
kids

730
00:25:44,925 --> 00:25:46,478
who were very attached to them,

731
00:25:47,045 --> 00:25:49,848
and people were worried
that the participant

732
00:25:49,848 --> 00:25:52,431
being away
from these little kids for three

733
00:25:52,431 --> 00:25:52,851
months

734
00:25:53,118 --> 00:25:55,671
would be traumatic to their
little kids

735
00:25:55,671 --> 00:25:56,522
who couldn't

736
00:25:56,522 --> 00:25:58,453
come here and had to stay at
home

737
00:25:58,453 --> 00:26:00,559
and be watched by the
grandparents.

738
00:26:00,559 --> 00:26:03,462
It turns out, in that case,
as we heard the story, that it

739
00:26:03,462 --> 00:26:03,662
did

740
00:26:03,662 --> 00:26:05,130
turn out to be really they did
come.

741
00:26:05,130 --> 00:26:06,632
They participate in this study,

742
00:26:06,632 --> 00:26:09,635
and it was really traumatic
for their two little kids.

743
00:26:09,835 --> 00:26:11,363
And this is a couple years
later,

744
00:26:11,363 --> 00:26:12,938
and apparently the kids are
still

745
00:26:12,938 --> 00:26:14,506
experiencing the effects of it.

746
00:26:14,506 --> 00:26:17,321
So in the second bullet, we
said, yeah,

747
00:26:17,321 --> 00:26:18,043
of course

748
00:26:18,043 --> 00:26:19,570
you should take into account
risks

749
00:26:19,570 --> 00:26:20,379
to third parties.

750
00:26:20,379 --> 00:26:21,580
But now what do we do with
these?

751
00:26:21,580 --> 00:26:23,415
Do we say, well, we're gonna
have

752
00:26:23,415 --> 00:26:25,083
inclusion exclusion criteria.

753
00:26:25,083 --> 00:26:28,720
If you have lots of little kids,
you can't enroll in our trial.

754
00:26:29,121 --> 00:26:31,924
If you have to drive
more than 50 miles to get here.

755
00:26:31,924 --> 00:26:33,725
You can't be at our trial.

756
00:26:33,725 --> 00:26:35,861
Then we have to look at all
this information on people

757
00:26:35,861 --> 00:26:36,995
about how many kids they have

758
00:26:36,995 --> 00:26:39,681
and where they live,
and whether they fly or they

759
00:26:39,681 --> 00:26:40,065
drive.

760
00:26:40,065 --> 00:26:41,553
Not only does that seem
intrusive,

761
00:26:41,553 --> 00:26:42,734
but it starts to seem like

762
00:26:42,734 --> 00:26:43,535
that's just too much.

763
00:26:43,535 --> 00:26:45,737
RBC can't be doing that, right.

764
00:26:45,737 --> 00:26:47,950
So I think this is a really
interesting

765
00:26:47,950 --> 00:26:48,574
challenge.

766
00:26:48,574 --> 00:26:49,341
Both.

767
00:26:49,341 --> 00:26:52,608
It seems like something
we should be doing on the one

768
00:26:52,608 --> 00:26:52,978
hand,

769
00:26:52,978 --> 00:26:54,513
but then when you generalize

770
00:26:54,513 --> 00:26:55,790
the principle, it seems like
it's

771
00:26:55,790 --> 00:26:57,416
something that's almost
impossible to do.

772
00:26:57,416 --> 00:26:59,351
And I have absolutely no
idea what the answer to this

773
00:27:00,419 --> 00:27:01,019
puzzle is.

774
00:27:01,019 --> 00:27:03,121
And if anybody knows the answer,
I'd be very interested

775
00:27:03,121 --> 00:27:05,784
in hearing it afterwards
or during a question answer

776
00:27:05,784 --> 00:27:06,091
okay,

777
00:27:07,726 --> 00:27:10,028
now a little bit of philosophy

778
00:27:10,028 --> 00:27:11,835
or philosophy of science that
is,

779
00:27:11,835 --> 00:27:13,532
I think, practically relevant.

780
00:27:13,532 --> 00:27:16,219
So I think it's worth
at least thinking about a little

781
00:27:16,219 --> 00:27:16,468
bit.

782
00:27:16,468 --> 00:27:18,320
So we've been talking about
risks

783
00:27:18,320 --> 00:27:20,172
and benefits of an intervention,

784
00:27:20,172 --> 00:27:23,375
like giving somebody a drug
as if this a straightforward

785
00:27:23,375 --> 00:27:24,176
figuring out.

786
00:27:24,176 --> 00:27:24,910
Right.

787
00:27:24,910 --> 00:27:26,912
Inject somebody with a drug.

788
00:27:26,912 --> 00:27:27,679
What are the risks?

789
00:27:27,679 --> 00:27:31,116
Well, it's the potential
that bad things or harms

790
00:27:31,116 --> 00:27:33,552
are going to be experienced
as a result of that. Right.

791
00:27:33,552 --> 00:27:36,555
That seems pretty easy to
imagine.

792
00:27:36,889 --> 00:27:38,785
What's important to notice,
though

793
00:27:38,785 --> 00:27:40,626
in some cases it doesn't matter.

794
00:27:40,626 --> 00:27:41,460
It's implicit.

795
00:27:41,460 --> 00:27:43,452
But in some cases it's really
important,

796
00:27:43,452 --> 00:27:44,897
which is that whenever we do

797
00:27:44,897 --> 00:27:49,134
risk benefit assessment,
we're appealing to some baseline

798
00:27:49,134 --> 00:27:51,824
or some comparison against
which we're making those

799
00:27:51,824 --> 00:27:52,404
judgments.

800
00:27:52,404 --> 00:27:54,573
So here's an example
to try to show you that

801
00:27:54,573 --> 00:27:56,494
there's a phase two study of a
treatment

802
00:27:56,494 --> 00:27:57,743
that has been shown safe.

803
00:27:57,743 --> 00:27:59,545
So we know it's safe let's say,

804
00:27:59,545 --> 00:28:02,297
and offers a small chance
of helping participants

805
00:28:02,297 --> 00:28:02,915
medically.

806
00:28:02,915 --> 00:28:04,383
So they have a serious illness.

807
00:28:04,383 --> 00:28:07,119
And there's a small chance
this will help them.

808
00:28:07,119 --> 00:28:09,354
Is that a potential benefit
study.

809
00:28:09,354 --> 00:28:12,457
So we think about the risks
and benefits of that trial okay.

810
00:28:12,691 --> 00:28:14,426
So it's supposed
to be a trick question right?

811
00:28:16,595 --> 00:28:18,043
Anybody want to take a shot and
fall

812
00:28:18,043 --> 00:28:19,531
into the trick or address the
trick?

813
00:28:19,531 --> 00:28:22,301
Or should I just click to the
next slide.

814
00:28:22,301 --> 00:28:23,702
It depends. Right.

815
00:28:23,702 --> 00:28:25,737
So this is supposed to be
what the comparison is.

816
00:28:25,737 --> 00:28:26,905
It depends on what.

817
00:28:26,905 --> 00:28:28,006
Well it depends on

818
00:28:28,006 --> 00:28:29,697
what's going to happen to these
people

819
00:28:29,697 --> 00:28:31,209
if they don't enroll in my
trial.

820
00:28:31,209 --> 00:28:31,610
Right.

821
00:28:31,610 --> 00:28:33,312
If outside of my trial

822
00:28:33,312 --> 00:28:35,936
they wouldn't get anything,
they would just sit at home and

823
00:28:35,936 --> 00:28:36,381
watch TV.

824
00:28:36,882 --> 00:28:38,760
Then arguably, being in the
trial

825
00:28:38,760 --> 00:28:40,752
is potentially beneficial for
them

826
00:28:40,752 --> 00:28:42,257
because there's this small
chance

827
00:28:42,257 --> 00:28:43,989
of benefit, a small chance, a
benefit

828
00:28:43,989 --> 00:28:46,104
is better than no chance of
benefit

829
00:28:46,104 --> 00:28:47,192
from watching TV.

830
00:28:47,192 --> 00:28:49,658
So overall,
the comparison suggests

831
00:28:49,658 --> 00:28:50,996
potential benefit.

832
00:28:51,229 --> 00:28:53,105
But in contrast, imagine there's
some

833
00:28:53,105 --> 00:28:55,133
drug out there that cures this
disease.

834
00:28:55,133 --> 00:28:56,735
You take it once and it cures
you.

835
00:28:58,236 --> 00:28:58,604
If the

836
00:28:58,604 --> 00:29:00,275
choice is between taking that
drug

837
00:29:00,275 --> 00:29:02,341
and being in my trial, then it
looks like

838
00:29:02,341 --> 00:29:03,782
not only isn't the trial
beneficial,

839
00:29:03,782 --> 00:29:05,143
but it looks really risky,
right?

840
00:29:05,143 --> 00:29:07,523
Being in my trial is that nine
people

841
00:29:07,523 --> 00:29:09,581
the opportunity of being cured.

842
00:29:09,581 --> 00:29:11,483
That seems like a very bad
thing.

843
00:29:11,483 --> 00:29:13,719
So now it looks like a really
risky trial.

844
00:29:13,719 --> 00:29:17,758
So the point here is that it
depends on what you're comparing

845
00:29:17,758 --> 00:29:18,023
to.

846
00:29:18,423 --> 00:29:19,891
Okay. So that's important.

847
00:29:19,891 --> 00:29:21,193
That's clear

848
00:29:21,193 --> 00:29:22,959
now why this is practically
relevant,

849
00:29:22,959 --> 00:29:24,630
not just philosophically
relevant.

850
00:29:24,863 --> 00:29:27,866
Is this that

851
00:29:28,066 --> 00:29:31,436
what the comparison is
or what is available to people

852
00:29:31,436 --> 00:29:35,841
otherwise varies
dramatically in populations.

853
00:29:35,841 --> 00:29:38,136
It varies dramatically in the
U.S and

854
00:29:38,136 --> 00:29:40,679
it varies dramatically across
the world.

855
00:29:40,979 --> 00:29:43,188
And so what that means in the
end is that

856
00:29:43,188 --> 00:29:45,450
whether a study is potentially
beneficial

857
00:29:45,450 --> 00:29:48,720
or risky is going to depend upon
where you do it.

858
00:29:49,287 --> 00:29:50,868
And if you do it in certain
populations,

859
00:29:50,868 --> 00:29:52,290
it might be potentially
beneficial.

860
00:29:52,491 --> 00:29:55,894
If you do in other populations,
it might be really risky.

861
00:29:56,194 --> 00:29:57,462
That's complicated.

862
00:29:57,462 --> 00:30:00,932
And it also drives people
absolutely nuts.

863
00:30:00,932 --> 00:30:03,011
So to evaluate the risks of
research,

864
00:30:03,011 --> 00:30:04,302
it's important to know

865
00:30:05,170 --> 00:30:06,896
if these people were in our
trial,

866
00:30:06,896 --> 00:30:08,674
what would happen to them
instead,

867
00:30:09,007 --> 00:30:10,886
and how does what we're
proposing

868
00:30:10,886 --> 00:30:12,878
to do to them in the trial
compare

869
00:30:13,111 --> 00:30:15,280
to what would happen to them
instead of a trial

870
00:30:15,280 --> 00:30:18,330
may be risky in some places,
but potentially beneficial in

871
00:30:18,330 --> 00:30:18,750
others.

872
00:30:19,217 --> 00:30:20,185
So where does this get?

873
00:30:20,185 --> 00:30:22,621
Is there going to be a session
or an international?

874
00:30:22,621 --> 00:30:25,223
Okay, so somebody's probably
gonna talk about this becomes a

875
00:30:25,223 --> 00:30:25,657
big deal,

876
00:30:25,657 --> 00:30:28,082
especially when you talk about
research

877
00:30:28,082 --> 00:30:29,761
in international settings.

878
00:30:29,995 --> 00:30:33,548
And this was a big deal
in some HIV studies about 30

879
00:30:33,548 --> 00:30:34,299
years ago.

880
00:30:34,299 --> 00:30:37,035
They turn into huge fights where

881
00:30:38,904 --> 00:30:40,205
still is true.

882
00:30:40,205 --> 00:30:43,642
Not as much as it used to be,
unfortunately, but 20 years, 15,

883
00:30:43,642 --> 00:30:46,678
20 years ago,
if you lived in certain places,

884
00:30:46,678 --> 00:30:49,266
there were very, very good
treatments

885
00:30:49,266 --> 00:30:51,783
for HIV in other parts of the
world

886
00:30:51,783 --> 00:30:54,546
where there was lots of HIV,
those treatments weren't

887
00:30:54,546 --> 00:30:55,120
available.

888
00:30:56,555 --> 00:30:59,078
So now imagine you have this
idea of this

889
00:30:59,078 --> 00:31:01,293
super cheap drug that's not as
good

890
00:31:01,293 --> 00:31:03,788
as the fancy ones you can get in
Bethesda,

891
00:31:03,788 --> 00:31:06,164
Maryland, but it's better than
nothing.

892
00:31:07,532 --> 00:31:10,326
Now, notice
if you ran that trial in

893
00:31:10,326 --> 00:31:11,103
Bethesda,

894
00:31:11,103 --> 00:31:13,974
Maryland, it's
going to be like the example I

895
00:31:13,974 --> 00:31:14,740
gave before

896
00:31:14,740 --> 00:31:16,942
where there's a chance of a cure
outside,

897
00:31:16,942 --> 00:31:19,144
not a cure, but a really good
treatment.

898
00:31:19,377 --> 00:31:20,479
Whereas if you do it in
Bethesda,

899
00:31:20,479 --> 00:31:23,348
it looks like a risky
and maybe an unethical study.

900
00:31:23,348 --> 00:31:25,625
But if you do it in sub-Saharan
Africa,

901
00:31:25,625 --> 00:31:27,319
where they don't have access

902
00:31:27,319 --> 00:31:29,003
to those treatments, it looks
like

903
00:31:29,003 --> 00:31:30,589
this is potentially beneficial.

904
00:31:31,022 --> 00:31:33,992
So that's what risk benefit
suggests, right.

905
00:31:34,326 --> 00:31:36,366
But now what you're saying is
that a trial

906
00:31:36,366 --> 00:31:37,095
this unethical

907
00:31:37,095 --> 00:31:39,029
to do here because the risk is
okay

908
00:31:39,029 --> 00:31:40,632
to do in sub-Saharan Africa.

909
00:31:40,632 --> 00:31:42,846
And that made people really
nervous

910
00:31:42,846 --> 00:31:43,668
to say that.

911
00:31:43,869 --> 00:31:45,871
And some people thought that

912
00:31:45,871 --> 00:31:49,274
that made them so nervous
that it must be wrong somehow.

913
00:31:49,274 --> 00:31:51,468
And we have to figure out how.
And there were huge fights about

914
00:31:51,468 --> 00:31:51,676
this.

915
00:31:51,676 --> 00:31:53,197
So it sounds like maybe you guys
will talk

916
00:31:53,197 --> 00:31:53,812
more about this.

917
00:31:53,812 --> 00:31:55,981
So the important point now here
is just

918
00:31:55,981 --> 00:31:56,815
the comparison

919
00:31:56,815 --> 00:31:59,073
can have really practical
implications

920
00:31:59,073 --> 00:32:00,619
for some of these trials.

921
00:32:00,852 --> 00:32:01,319
Okay.

922
00:32:02,487 --> 00:32:03,948
There's
an end of the philosophy for a

923
00:32:03,948 --> 00:32:04,256
second.

924
00:32:04,256 --> 00:32:06,558
So now we've identified
the risks of the interventions.

925
00:32:06,558 --> 00:32:08,493
We want to minimize them

926
00:32:08,493 --> 00:32:11,806
I just I just do these
because this is personal trauma

927
00:32:11,806 --> 00:32:12,297
for me.

928
00:32:12,631 --> 00:32:15,567
I still get
very freaked out by blood draws.

929
00:32:15,567 --> 00:32:18,740
And they used to
I really had an incredibly hard

930
00:32:18,740 --> 00:32:19,070
time

931
00:32:19,271 --> 00:32:20,505
when I was a little kid.

932
00:32:20,505 --> 00:32:22,407
I still do, actually.

933
00:32:22,407 --> 00:32:24,342
And then

934
00:32:24,342 --> 00:32:26,645
I got tons of blood draws in my
life.

935
00:32:26,645 --> 00:32:28,213
And then at one point, some
nurse

936
00:32:28,213 --> 00:32:29,781
says to me, look out the window.

937
00:32:31,082 --> 00:32:33,718
I looked out the window, she did
it.

938
00:32:33,718 --> 00:32:35,287
And she and I said,
okay, you're going to do it.

939
00:32:35,287 --> 00:32:36,721
She says, no, we're done.

940
00:32:36,721 --> 00:32:38,290
And it turns out for at least
a lot of people

941
00:32:38,290 --> 00:32:40,625
like me, it's not the physical
sensation

942
00:32:40,625 --> 00:32:41,326
that's bad.

943
00:32:41,326 --> 00:32:43,508
It's the seeing this thing
getting stuck

944
00:32:43,508 --> 00:32:45,363
inside your body that's
horrible.

945
00:32:45,630 --> 00:32:48,300
And it just completely went
away.

946
00:32:48,300 --> 00:32:50,635
And so the point of this is just

947
00:32:50,635 --> 00:32:52,971
we need people
who know what they're doing,

948
00:32:52,971 --> 00:32:54,356
and we need people to be
creative

949
00:32:54,356 --> 00:32:55,574
when we're doing this stuff.

950
00:32:55,574 --> 00:32:56,942
This just shows a way in which,

951
00:32:58,777 --> 00:33:00,507
in which we talked about, I
don't people

952
00:33:00,507 --> 00:33:02,280
know the jiggling of the, the
cheek one.

953
00:33:02,280 --> 00:33:04,382
My dentist taught me this one.

954
00:33:04,382 --> 00:33:04,616
All right.

955
00:33:04,616 --> 00:33:07,052
So so I know these are my
tangent. Sorry.

956
00:33:07,052 --> 00:33:08,843
If I'm late, if I go over to my
phone,

957
00:33:08,843 --> 00:33:10,021
holly holic, yell at me.

958
00:33:10,622 --> 00:33:11,723
This one is really interesting.

959
00:33:11,723 --> 00:33:14,426
It turns out that
that nerve pathways. Right.

960
00:33:14,426 --> 00:33:16,491
So so if you stick somebody down
here,

961
00:33:16,491 --> 00:33:18,230
the reason why it hurts, right?

962
00:33:18,230 --> 00:33:20,799
They're all these neuro pathways
that goes up into your brain

963
00:33:20,799 --> 00:33:23,543
and your brain does some
complicated things and and it

964
00:33:23,543 --> 00:33:24,102
says ouch.

965
00:33:24,369 --> 00:33:27,172
Right. Well,

966
00:33:27,172 --> 00:33:29,908
these pathways
come together at various points,

967
00:33:29,908 --> 00:33:33,245
and they can't transmit
more than one signal at a time.

968
00:33:33,745 --> 00:33:37,182
And what's really interesting
and evolutionarily seems weird

969
00:33:37,515 --> 00:33:40,819
is that those are some of those
nerve pathways prioritize

970
00:33:40,819 --> 00:33:43,116
sort of dull, pounding
sensations

971
00:33:43,116 --> 00:33:44,856
over sharp, sticky ones.

972
00:33:45,891 --> 00:33:48,304
And so what that
does is, and what my dentist

973
00:33:48,304 --> 00:33:48,894
used to do

974
00:33:49,261 --> 00:33:52,330
when we did this was sticking
that big needle into my cheek.

975
00:33:52,597 --> 00:33:55,600
At the same time,
what you do is jiggle my cheek,

976
00:33:56,034 --> 00:34:00,099
and that sort of dull sensation
would get prioritized by the

977
00:34:00,099 --> 00:34:00,505
nerve

978
00:34:00,505 --> 00:34:02,084
pathways over there getting
stuck,

979
00:34:02,084 --> 00:34:03,942
and I wouldn't feel that getting
stuck.

980
00:34:03,942 --> 00:34:05,143
It was just fantastic.

981
00:34:05,143 --> 00:34:08,146
I think I was 40 before somebody
before somebody did that to me.

982
00:34:08,146 --> 00:34:10,248
So I was like, people should be
doing this

983
00:34:10,248 --> 00:34:11,149
more often, okay,

984
00:34:11,650 --> 00:34:13,818
second point is that

985
00:34:13,818 --> 00:34:15,754
sometimes the way you minimize
risk is do

986
00:34:15,754 --> 00:34:18,300
less tools,
fewer blood draws, do fewer

987
00:34:18,300 --> 00:34:18,757
MRI's.

988
00:34:19,190 --> 00:34:20,191
That can reduce risk,

989
00:34:20,191 --> 00:34:22,279
but it also can potentially
undermine

990
00:34:22,279 --> 00:34:24,029
the social value of the study.

991
00:34:24,029 --> 00:34:25,096
So then that's the point.

992
00:34:25,096 --> 00:34:28,099
We're going to have to judgment
and figure out how to balance

993
00:34:28,300 --> 00:34:29,935
these competing considerations.

994
00:34:29,935 --> 00:34:33,038
The fairness issue is also this
is one

995
00:34:33,038 --> 00:34:36,808
that we've had huge debates
in our department over years,

996
00:34:36,808 --> 00:34:39,612
which is here's another way to
minimize

997
00:34:39,612 --> 00:34:41,913
the aggregate risks of a study,

998
00:34:43,081 --> 00:34:46,069
exclude anybody
who faces really high risks from

999
00:34:46,069 --> 00:34:46,318
it.

1000
00:34:46,685 --> 00:34:49,965
If you do that, you're going
to really minimize the aggregate

1001
00:34:49,965 --> 00:34:50,288
risk,

1002
00:34:50,288 --> 00:34:52,629
but at the same time,
you're excluding people from

1003
00:34:52,629 --> 00:34:53,191
that study.

1004
00:34:53,191 --> 00:34:55,794
And if that study has a chance
for a potential benefit,

1005
00:34:55,794 --> 00:34:57,134
then there's a way in which it's
bad

1006
00:34:57,134 --> 00:34:58,363
for the people you're excluding.

1007
00:34:58,363 --> 00:35:01,366
And some people argue that
that's unfair.

1008
00:35:02,200 --> 00:35:04,803
Okay. So here's some examples.

1009
00:35:04,803 --> 00:35:07,234
Experimental drug for serious
conditions

1010
00:35:07,234 --> 00:35:09,240
with very few effective options.

1011
00:35:09,808 --> 00:35:12,544
This is this is a console
we had a couple of years ago.

1012
00:35:12,544 --> 00:35:14,813
The drug poses a risk of
significant

1013
00:35:14,813 --> 00:35:17,082
bleeding that requires
transfusion.

1014
00:35:17,382 --> 00:35:19,368
So the investigator comes to us
and says

1015
00:35:19,368 --> 00:35:21,453
I think this is a really
important trial.

1016
00:35:21,653 --> 00:35:23,353
This is a really important drug
to test,

1017
00:35:23,353 --> 00:35:24,756
but I'm terrified that I'm going

1018
00:35:24,756 --> 00:35:28,257
to kill some of my participants
because of this risk of

1019
00:35:28,257 --> 00:35:28,893
bleeding.

1020
00:35:29,127 --> 00:35:32,130
And what I want to do is I want
to exclude

1021
00:35:32,464 --> 00:35:35,533
anybody
who faces a really high risk for

1022
00:35:35,533 --> 00:35:36,301
bleeding.

1023
00:35:36,534 --> 00:35:37,802
And so who is that?

1024
00:35:37,802 --> 00:35:39,437
That's people are bleeding
disorders.

1025
00:35:39,437 --> 00:35:42,607
It's people have low platelets
for whatever reason.

1026
00:35:42,607 --> 00:35:45,543
And interestingly,
it's also people who declined

1027
00:35:45,543 --> 00:35:48,546
blood transfusions for religious
reasons.

1028
00:35:48,780 --> 00:35:50,889
And so the question was if you
want

1029
00:35:50,889 --> 00:35:53,118
to minimize the risks of this
study,

1030
00:35:53,118 --> 00:35:54,963
what you should do is absolutely
exclude

1031
00:35:54,963 --> 00:35:56,254
all of those people, right?

1032
00:35:56,821 --> 00:35:59,391
So is that the right thing to do
or is that unfair,

1033
00:35:59,391 --> 00:36:01,370
particularly given serious
condition,

1034
00:36:01,370 --> 00:36:02,494
no other treatments.

1035
00:36:02,494 --> 00:36:05,754
These are the kind of,
I think, really important

1036
00:36:05,754 --> 00:36:06,297
debates

1037
00:36:06,297 --> 00:36:07,961
and questions that arise when we
start

1038
00:36:07,961 --> 00:36:09,667
thinking about risk benefit
assessment

1039
00:36:09,667 --> 00:36:11,923
and balancing risk benefit
assessment

1040
00:36:11,923 --> 00:36:14,239
with other values like fairness
okay.

1041
00:36:14,572 --> 00:36:16,241
So potential benefits.

1042
00:36:16,241 --> 00:36:18,443
So we've done social value.
We've done risks.

1043
00:36:18,443 --> 00:36:20,661
Now we look at the potential
benefits

1044
00:36:20,661 --> 00:36:21,980
of the interventions.

1045
00:36:22,213 --> 00:36:25,490
And again as with risk
we look at the potential

1046
00:36:25,490 --> 00:36:26,117
benefits

1047
00:36:26,117 --> 00:36:28,254
above and beyond what people
would get

1048
00:36:28,254 --> 00:36:30,221
if they were outside of the
trial.

1049
00:36:32,190 --> 00:36:34,172
So now I want to come back to
the point

1050
00:36:34,172 --> 00:36:35,493
I was harping on earlier.

1051
00:36:35,493 --> 00:36:38,034
Remember when I talked
about financial risks,

1052
00:36:38,034 --> 00:36:39,164
psychological risk,

1053
00:36:39,164 --> 00:36:42,167
stigma, risks,

1054
00:36:43,168 --> 00:36:44,969
and we said you should count all
of them?

1055
00:36:44,969 --> 00:36:47,939
Seems obvious.
You should count all of them.

1056
00:36:47,939 --> 00:36:50,942
So here's the question is
does the analogous point apply

1057
00:36:50,942 --> 00:36:52,343
to potential benefits?

1058
00:36:52,343 --> 00:36:56,848
Should you include
not just the potential benefits

1059
00:36:56,848 --> 00:37:00,051
of clinically getting better,
but the potential benefits

1060
00:37:00,051 --> 00:37:02,887
of other sorts of things
that could happen to you?

1061
00:37:02,887 --> 00:37:04,809
So before we talked about
economic risk,

1062
00:37:04,809 --> 00:37:06,491
if there's a chance that it's
good

1063
00:37:06,758 --> 00:37:09,527
choice, had to knock out my
kidneys

1064
00:37:09,527 --> 00:37:11,690
and I'm going to need treatment,
I got to pay out of pocket for

1065
00:37:11,690 --> 00:37:11,896
that.

1066
00:37:11,896 --> 00:37:14,132
That's an economic
risk of being in this study.

1067
00:37:14,132 --> 00:37:16,012
Well, now I'll take a different
study that

1068
00:37:16,012 --> 00:37:17,802
offers everybody who
participates $500.

1069
00:37:18,870 --> 00:37:21,372
Is that a benefit of being in
the study?

1070
00:37:21,372 --> 00:37:22,674
Should that count as a benefit?

1071
00:37:22,674 --> 00:37:25,231
If there's economic risks over
here,

1072
00:37:25,231 --> 00:37:27,645
are economic benefits over
there.

1073
00:37:27,912 --> 00:37:31,416
And I'll just say that
the almost consensus view

1074
00:37:31,416 --> 00:37:34,283
in the literature
is, yeah, the economic risks

1075
00:37:34,283 --> 00:37:34,719
count.

1076
00:37:34,719 --> 00:37:36,737
If everybody's
got to take those into

1077
00:37:36,737 --> 00:37:37,555
consideration.

1078
00:37:37,555 --> 00:37:39,453
The economic benefits don't they
shouldn't

1079
00:37:39,453 --> 00:37:41,126
consider them or basically
shouldn't

1080
00:37:41,126 --> 00:37:43,061
consider them at all.

1081
00:37:43,061 --> 00:37:46,101
And so obviously there's this
question of

1082
00:37:46,101 --> 00:37:47,732
does that make sense?

1083
00:37:47,732 --> 00:37:49,459
Should we treat them
differently,

1084
00:37:49,459 --> 00:37:50,034
and if so,

1085
00:37:50,034 --> 00:37:52,036
why should we treat them
differently.

1086
00:37:52,036 --> 00:37:53,772
So I was just one of the other
questions.

1087
00:37:53,772 --> 00:37:56,774
I'm going to leave us with for
now.

1088
00:37:57,242 --> 00:37:58,276
So what's really identified?

1089
00:37:58,276 --> 00:38:01,466
The risks, identify the risks,
minimize them, potential

1090
00:38:01,466 --> 00:38:02,046
benefits,

1091
00:38:02,046 --> 00:38:04,505
identify them and try to make
them larger

1092
00:38:04,505 --> 00:38:05,884
to the extent you can.

1093
00:38:06,151 --> 00:38:10,488
And this puts us back into this
worry about fairness again.

1094
00:38:11,456 --> 00:38:14,626
When I was on the infectious
disease IRB,

1095
00:38:14,626 --> 00:38:18,263
they were just starting
to test protease inhibitors,

1096
00:38:18,263 --> 00:38:20,098
which people might know turned
out

1097
00:38:20,098 --> 00:38:22,407
to be like the initial
blockbuster

1098
00:38:22,407 --> 00:38:23,902
for treatment of HIV.

1099
00:38:23,902 --> 00:38:26,437
It really revolutionized the
treatment

1100
00:38:26,437 --> 00:38:28,973
of HIV, where HIV went from a
disease

1101
00:38:28,973 --> 00:38:32,229
that basically killed everybody
to a disease that people lived

1102
00:38:32,229 --> 00:38:32,544
with.

1103
00:38:32,911 --> 00:38:34,879
Now, we didn't know all that at
the time,

1104
00:38:34,879 --> 00:38:37,282
but we know
these drugs looked really good.

1105
00:38:37,282 --> 00:38:39,386
And the question was, who do we
do

1106
00:38:39,386 --> 00:38:41,553
the early phase trials of
protease

1107
00:38:41,553 --> 00:38:42,821
inhibitors in

1108
00:38:42,821 --> 00:38:45,623
do we do it in people
who are relatively healthy,

1109
00:38:45,623 --> 00:38:47,826
with the assumption
it will be less risk to them

1110
00:38:47,826 --> 00:38:50,828
because they're healthier,
they're in better shape?

1111
00:38:51,229 --> 00:38:54,232
Or do we do it in people
who are really, really sick,

1112
00:38:54,566 --> 00:38:56,675
who probably face
greater risks from an

1113
00:38:56,675 --> 00:38:57,702
experimental drug?

1114
00:38:57,702 --> 00:38:59,837
But if it helps them,
if it actually works,

1115
00:38:59,837 --> 00:39:01,462
they're going to gain a lot more
than

1116
00:39:01,462 --> 00:39:03,174
the people who are relatively
healthy.

1117
00:39:03,174 --> 00:39:04,642
So we actually debated that.

1118
00:39:04,642 --> 00:39:06,211
And it's interesting.

1119
00:39:06,211 --> 00:39:08,304
So I think about it now, I
always say that

1120
00:39:08,304 --> 00:39:10,148
IRBs tend to be on the cautious
side

1121
00:39:10,148 --> 00:39:11,714
where they want us to reduce
risk,

1122
00:39:11,714 --> 00:39:13,051
but in this case, we didn't.

1123
00:39:13,051 --> 00:39:13,551
In this case,

1124
00:39:13,551 --> 00:39:15,756
we actually mandated that the
studies

1125
00:39:15,756 --> 00:39:17,722
be done in people who were sick.

1126
00:39:17,722 --> 00:39:20,191
So we we did people who had low
CD4

1127
00:39:20,191 --> 00:39:24,128
counts, I think under 50 to be
to potentially do that trial.

1128
00:39:24,128 --> 00:39:26,136
So we focused on the potential
benefit

1129
00:39:26,136 --> 00:39:27,932
rather than minimizing the
risks.

1130
00:39:28,333 --> 00:39:30,068
Okay. So now we know what the
risks are.

1131
00:39:30,068 --> 00:39:32,570
We know what the benefits are.
We know what the social value

1132
00:39:32,570 --> 00:39:32,737
is.

1133
00:39:32,737 --> 00:39:34,772
So we're supposed to put it all
together.

1134
00:39:34,772 --> 00:39:37,427
And here this step is just
basically

1135
00:39:37,427 --> 00:39:39,344
to the potential benefits

1136
00:39:40,345 --> 00:39:42,780
to the participants plus the
social value.

1137
00:39:42,780 --> 00:39:44,449
Put those together to those.

1138
00:39:44,449 --> 00:39:46,750
Justify the risks and the
burdens

1139
00:39:46,750 --> 00:39:49,120
and the costs of doing the
study.

1140
00:39:49,354 --> 00:39:51,151
If the answer is yes, then at
least

1141
00:39:51,151 --> 00:39:53,258
from a risk benefit perspective,
you got

1142
00:39:53,691 --> 00:39:54,826
you have an acceptable study.

1143
00:39:54,826 --> 00:39:56,425
You could basically stop doing a
risk

1144
00:39:56,425 --> 00:39:57,895
benefit assessment at this
point.

1145
00:39:57,895 --> 00:40:00,365
If no, then you need a couple
more steps.

1146
00:40:00,365 --> 00:40:03,368
So I'll talk about those
additional steps

1147
00:40:04,369 --> 00:40:04,802
in a minute.

1148
00:40:04,802 --> 00:40:07,805
The first thing is so people
agonize over this all the time.

1149
00:40:08,006 --> 00:40:09,440
And this is a case.

1150
00:40:09,440 --> 00:40:10,825
So there's this there's this big
debate

1151
00:40:10,825 --> 00:40:12,210
that always goes on in research
ethics

1152
00:40:12,210 --> 00:40:14,205
about the way in
which you should treat clinical

1153
00:40:14,205 --> 00:40:14,579
research

1154
00:40:14,579 --> 00:40:16,613
like clinical care, and the ways
in which

1155
00:40:16,613 --> 00:40:18,349
you should treat them
differently.

1156
00:40:19,550 --> 00:40:20,839
I think
it depends, obviously, on the

1157
00:40:20,839 --> 00:40:21,152
context.

1158
00:40:21,152 --> 00:40:21,919
This is a case

1159
00:40:21,919 --> 00:40:24,023
where I think people get too
confused

1160
00:40:24,023 --> 00:40:25,957
when it comes to research
ethics.

1161
00:40:26,190 --> 00:40:28,893
So when I say this, the IAB
should look at whether the

1162
00:40:28,893 --> 00:40:29,394
potential

1163
00:40:29,394 --> 00:40:31,611
benefits of the drug to the
participants

1164
00:40:31,611 --> 00:40:32,664
justify the risks.

1165
00:40:33,031 --> 00:40:35,293
So like,
how do you do that outweigh

1166
00:40:35,293 --> 00:40:35,733
risks?

1167
00:40:35,733 --> 00:40:36,701
What does outweigh mean?

1168
00:40:36,701 --> 00:40:39,771
That's a matter for confidence
say for justified.

1169
00:40:39,771 --> 00:40:42,206
What do any of these things
mean.

1170
00:40:42,206 --> 00:40:45,898
And my response to that is,
yeah, I know this is hard, but

1171
00:40:45,898 --> 00:40:46,344
notice

1172
00:40:46,544 --> 00:40:48,648
that's what doctors and
clinicians

1173
00:40:48,648 --> 00:40:49,947
do every day, right?

1174
00:40:49,947 --> 00:40:53,885
That's the job of clinicians,
is to ask themselves for this

1175
00:40:53,885 --> 00:40:54,419
patient

1176
00:40:55,019 --> 00:40:57,212
to the potential benefits of
this drug

1177
00:40:57,212 --> 00:40:59,290
outweigh the risks to this
patient.

1178
00:40:59,490 --> 00:41:00,994
Now, it's
harder often to do it in

1179
00:41:00,994 --> 00:41:01,392
research

1180
00:41:01,392 --> 00:41:03,461
because we don't have as much
data.

1181
00:41:03,461 --> 00:41:06,208
But at least I think that the
approach,

1182
00:41:06,208 --> 00:41:07,899
the sort of basic idea,

1183
00:41:07,899 --> 00:41:09,434
is exactly the same.

1184
00:41:09,434 --> 00:41:12,503
And so I've termed this
the informed clinician test.

1185
00:41:12,737 --> 00:41:15,287
And basically what the should do
is either

1186
00:41:15,287 --> 00:41:15,773
pretend

1187
00:41:15,973 --> 00:41:19,649
they're an informed clinician
or call one up and actually ask

1188
00:41:19,649 --> 00:41:20,011
them.

1189
00:41:20,478 --> 00:41:21,546
Here's the situation.

1190
00:41:21,546 --> 00:41:22,962
Here are the patients. Here's
the drug

1191
00:41:22,962 --> 00:41:23,781
we want to give them.

1192
00:41:23,781 --> 00:41:25,149
Here's what we know about it.

1193
00:41:25,149 --> 00:41:27,511
Do you think that the potential
benefits

1194
00:41:27,511 --> 00:41:29,754
for these patients justify the
risks?

1195
00:41:29,754 --> 00:41:32,590
If they say yes, it's
a prospect of benefit study.

1196
00:41:32,590 --> 00:41:36,060
If they say no, it's a net,
what I call a net risk study.

1197
00:41:39,330 --> 00:41:43,368
A little bit more philosophy
here as so.

1198
00:41:44,569 --> 00:41:45,856
Well, I've been saying as you
look

1199
00:41:45,856 --> 00:41:47,105
at the individual interventions.

1200
00:41:47,105 --> 00:41:47,505
Right.

1201
00:41:47,505 --> 00:41:50,508
And then you look at the whole
package.

1202
00:41:52,777 --> 00:41:54,212
What are you supposed to do?

1203
00:41:54,212 --> 00:41:55,213
And I sort of highlighted this.

1204
00:41:55,213 --> 00:41:56,550
What do you suppose are we
looking

1205
00:41:56,550 --> 00:41:57,849
at the individual interventions.

1206
00:41:57,849 --> 00:41:59,984
And here's the question

1207
00:41:59,984 --> 00:42:02,987
to the potential benefits
of each intervention

1208
00:42:03,321 --> 00:42:05,923
have to justify the risks.

1209
00:42:05,923 --> 00:42:08,426
So imagine a study like this.

1210
00:42:08,426 --> 00:42:11,496
Imagine a study
where I'm going to give people

1211
00:42:11,496 --> 00:42:13,734
a potential treatment for kidney
disease

1212
00:42:13,734 --> 00:42:14,966
or say, liver disease

1213
00:42:16,300 --> 00:42:17,435
really could help them.

1214
00:42:17,435 --> 00:42:18,536
Looks really promising.

1215
00:42:18,536 --> 00:42:20,505
Lots of potential benefits.

1216
00:42:20,505 --> 00:42:23,845
But in order to do the study
scientifically, I have to do a

1217
00:42:23,845 --> 00:42:24,242
biopsy

1218
00:42:24,442 --> 00:42:26,921
like a liver biopsy or a kidney
biopsy

1219
00:42:26,921 --> 00:42:28,813
to pose a significant risks.

1220
00:42:29,013 --> 00:42:32,216
And I'm only doing
it for research purposes.

1221
00:42:33,184 --> 00:42:33,885
If you look at the

1222
00:42:33,885 --> 00:42:37,155
liver biopsy
or the kidney biopsy on its own,

1223
00:42:37,889 --> 00:42:41,360
it looks like that's very risky,
not potential benefit

1224
00:42:41,360 --> 00:42:42,260
intervention.

1225
00:42:42,794 --> 00:42:45,797
And at least saying children,
you wouldn't be allowed to do it

1226
00:42:45,797 --> 00:42:48,720
and people would probably
be really worried about doing

1227
00:42:48,720 --> 00:42:48,933
it.

1228
00:42:48,933 --> 00:42:50,846
Adults. So is that the way
you're supposed

1229
00:42:50,846 --> 00:42:51,302
to do it?

1230
00:42:51,302 --> 00:42:54,300
Each individual intervention
potential benefits have to

1231
00:42:54,300 --> 00:42:55,173
justify for us?

1232
00:42:55,406 --> 00:42:57,199
Or do you put the whole thing
together

1233
00:42:57,199 --> 00:42:58,709
and say, yeah, no, I know that.

1234
00:42:58,709 --> 00:43:01,712
But look, the potential benefits
of getting the drug,

1235
00:43:02,213 --> 00:43:06,150
they justify the risks of
not only just the drug,

1236
00:43:06,150 --> 00:43:08,786
but also the risks of the liver
or kidney biopsy.

1237
00:43:08,786 --> 00:43:12,390
So that being in the study
overall is prospect

1238
00:43:12,390 --> 00:43:15,560
a benefit for the people
who are going to get into it.

1239
00:43:15,560 --> 00:43:17,566
So that's a question of whether
or not

1240
00:43:17,566 --> 00:43:19,096
you have to look at each one

1241
00:43:19,096 --> 00:43:19,964
just individually.

1242
00:43:19,964 --> 00:43:22,150
Or you can have the potential
benefits

1243
00:43:22,150 --> 00:43:23,301
of one intervention

1244
00:43:23,301 --> 00:43:26,304
in a study
justifying the risks of another.

1245
00:43:26,304 --> 00:43:27,572
So hopefully that's clear.

1246
00:43:27,572 --> 00:43:29,273
If it's not, we can definitely
come.

1247
00:43:29,273 --> 00:43:30,641
We can definitely come back to
it.

1248
00:43:30,641 --> 00:43:32,343
I think that's a
really interesting challenge.

1249
00:43:34,679 --> 00:43:36,614
Here's a practical example of
this.

1250
00:43:36,614 --> 00:43:38,149
We wrote a paper on this.

1251
00:43:38,149 --> 00:43:42,320
Imagine a trial in kids
placebo controlled trial.

1252
00:43:42,320 --> 00:43:44,489
Some people get this

1253
00:43:44,489 --> 00:43:46,346
potentially beneficial
intervention

1254
00:43:46,346 --> 00:43:47,992
and some people get a placebo.

1255
00:43:48,993 --> 00:43:51,050
What's the risk benefit
assessment

1256
00:43:51,050 --> 00:43:53,531
of or the profile of that
overall study?

1257
00:43:54,131 --> 00:43:57,068
What some people will say as
well

1258
00:43:57,068 --> 00:43:59,337
at the outset before
randomization.

1259
00:43:59,337 --> 00:44:01,305
There's a chance you'll get the
drug.

1260
00:44:01,305 --> 00:44:03,895
And if you get the drug
and it's beneficial, then that

1261
00:44:03,895 --> 00:44:04,375
justifies

1262
00:44:05,443 --> 00:44:07,879
even being randomized to the
placebo arm.

1263
00:44:07,879 --> 00:44:09,757
So the overall study is
prospective

1264
00:44:09,757 --> 00:44:10,348
a benefit.

1265
00:44:10,348 --> 00:44:12,783
What we argued
is actually that's a mistake.

1266
00:44:12,783 --> 00:44:15,678
And you need to individually
look at the

1267
00:44:15,678 --> 00:44:18,789
risk benefit of the two arms of
the trial.

1268
00:44:18,789 --> 00:44:22,838
And that means that the risks
of the placebo arm have to be

1269
00:44:22,838 --> 00:44:23,661
acceptable.

1270
00:44:23,861 --> 00:44:27,324
And you can't justify them
by the fact that there was a

1271
00:44:27,324 --> 00:44:27,765
chance

1272
00:44:27,765 --> 00:44:29,639
the person could have gotten
into,

1273
00:44:29,639 --> 00:44:30,301
but didn't.

1274
00:44:30,301 --> 00:44:31,536
The active treatment are.

1275
00:44:31,536 --> 00:44:32,670
So that's just what exist.

1276
00:44:32,670 --> 00:44:34,906
Specific example of that.

1277
00:44:34,906 --> 00:44:35,406
Okay.

1278
00:44:35,406 --> 00:44:36,507
So we're just talking about

1279
00:44:36,507 --> 00:44:38,535
whether or not the potential
benefits

1280
00:44:38,535 --> 00:44:39,577
justify the risks.

1281
00:44:39,577 --> 00:44:42,036
If they do,
you have a prospect of benefit

1282
00:44:42,036 --> 00:44:42,446
study.

1283
00:44:42,446 --> 00:44:45,556
If they don't, you have
what I just call a net risk

1284
00:44:45,556 --> 00:44:45,983
trial.

1285
00:44:45,983 --> 00:44:48,626
So then the next last challenge
here

1286
00:44:48,626 --> 00:44:50,755
is thinking about net risks.

1287
00:44:51,055 --> 00:44:52,590
When are net risks acceptable?

1288
00:44:52,590 --> 00:44:55,593
How much net risk is acceptable.

1289
00:44:55,593 --> 00:44:59,230
So when it comes to individuals
who can't consent like kids,

1290
00:44:59,697 --> 00:45:01,881
there are some people who think
it's just

1291
00:45:01,881 --> 00:45:02,466
unethical.

1292
00:45:02,466 --> 00:45:05,469
Sister, it's been going on for
6100 years.

1293
00:45:05,770 --> 00:45:08,739
Some people say it's just
unacceptable.

1294
00:45:08,739 --> 00:45:11,336
It's unethical to pose any risks
to kids

1295
00:45:11,336 --> 00:45:13,477
for the benefit of other people.

1296
00:45:13,477 --> 00:45:15,253
It's always exploitation if they
can't

1297
00:45:15,253 --> 00:45:17,214
give information, we can talk
about that.

1298
00:45:17,214 --> 00:45:19,050
If people are interested.

1299
00:45:19,050 --> 00:45:21,085
Well, lots of people have that
worry.

1300
00:45:21,085 --> 00:45:22,392
In fact,
if you look at almost all

1301
00:45:22,392 --> 00:45:22,853
guidelines,

1302
00:45:22,853 --> 00:45:26,023
I know around the world
they allow some risks.

1303
00:45:26,023 --> 00:45:28,306
And kids particularly, and this
is the

1304
00:45:28,306 --> 00:45:30,828
buzzword you'll hear more about
from CMO,

1305
00:45:30,828 --> 00:45:33,014
whether or not the risks are
minimal

1306
00:45:33,014 --> 00:45:34,532
or greater than minimal.

1307
00:45:34,799 --> 00:45:35,833
And I'll just go through this.

1308
00:45:35,833 --> 00:45:37,386
I think you guys are going to
hear

1309
00:45:37,386 --> 00:45:38,436
a lot more about this.

1310
00:45:38,436 --> 00:45:40,871
So what's the minimal risk?

1311
00:45:40,871 --> 00:45:42,406
The standard based on the U.S.

1312
00:45:42,406 --> 00:45:43,007
regulations

1313
00:45:43,007 --> 00:45:44,884
that a lot of people have
adopted, is

1314
00:45:44,884 --> 00:45:46,711
what our bodies are supposed to
do.

1315
00:45:47,111 --> 00:45:49,814
It's supposed to cover the risks
of the intervention of the study

1316
00:45:49,814 --> 00:45:54,085
to the risks that are ordinarily
encountered in daily life,

1317
00:45:54,452 --> 00:45:56,254
and if the risks of the
intervention

1318
00:45:56,254 --> 00:45:57,655
are equal to or lower, then

1319
00:45:58,356 --> 00:45:59,792
the risks
or the other encountered in

1320
00:45:59,792 --> 00:46:00,257
daily life,

1321
00:46:00,257 --> 00:46:02,878
then it counts
as a minimal risk intervention

1322
00:46:02,878 --> 00:46:03,461
or study.

1323
00:46:03,461 --> 00:46:04,562
And if they're greater,

1324
00:46:04,562 --> 00:46:05,853
it counts as a more than a
medium

1325
00:46:05,853 --> 00:46:06,831
risk intervention study.

1326
00:46:06,831 --> 00:46:09,532
That's the standard
that when I was the study I

1327
00:46:09,532 --> 00:46:09,934
talked

1328
00:46:09,934 --> 00:46:11,525
about before that we did with
CMO,

1329
00:46:11,525 --> 00:46:13,070
we're talking to the IRB chairs.

1330
00:46:13,070 --> 00:46:15,172
That's the question we were
asking them.

1331
00:46:15,172 --> 00:46:17,535
And you notice how
if I just say to you, allergy

1332
00:46:17,535 --> 00:46:18,175
skin testing

1333
00:46:19,043 --> 00:46:20,865
or the risk of, I was your skin
testing,

1334
00:46:20,865 --> 00:46:22,413
11 year olds less than or
greater

1335
00:46:22,413 --> 00:46:24,030
than the risk of daily life, I
hope

1336
00:46:24,030 --> 00:46:26,017
all of you guys would say, I
have no idea.

1337
00:46:26,517 --> 00:46:28,211
You're going to tell me what
the risk of hours of skin

1338
00:46:28,211 --> 00:46:28,619
testing are.

1339
00:46:28,619 --> 00:46:30,290
You got to tell me what the
risks of daily

1340
00:46:30,290 --> 00:46:31,722
life are before I can compare
them.

1341
00:46:31,722 --> 00:46:32,456
And that's the thing.

1342
00:46:32,456 --> 00:46:35,413
I was pointing out that
none of those IRB chairs did,

1343
00:46:35,413 --> 00:46:36,027
even ones.

1344
00:46:37,094 --> 00:46:38,396
Okay, so

1345
00:46:38,396 --> 00:46:40,152
this is a this is a study we
finished

1346
00:46:40,152 --> 00:46:42,099
a couple of years ago with Will
Shukman,

1347
00:46:42,466 --> 00:46:46,370
who was now in sociology program
at UCLA,

1348
00:46:46,370 --> 00:46:50,207
who was a post back here,
and we wanted to get data.

1349
00:46:50,207 --> 00:46:51,881
There's been this debate for a
long time,

1350
00:46:51,881 --> 00:46:53,310
but there haven't been any data
on

1351
00:46:53,310 --> 00:46:55,607
what the public thinks about
exposing kids

1352
00:46:55,607 --> 00:46:57,248
to risk or benefit of others.

1353
00:46:57,248 --> 00:46:59,957
So we did a survey of this,
a nationally representative

1354
00:46:59,957 --> 00:47:00,351
sample.

1355
00:47:00,551 --> 00:47:03,492
And what we found is really
strong support

1356
00:47:03,492 --> 00:47:06,223
amongst the public, including
parents,

1357
00:47:06,424 --> 00:47:08,251
for exposing kids to research
risks

1358
00:47:08,251 --> 00:47:09,660
for the benefit of others,

1359
00:47:09,660 --> 00:47:12,708
including things like a bone
marrow biopsy, which poses more

1360
00:47:12,708 --> 00:47:13,064
risks,

1361
00:47:13,064 --> 00:47:16,233
I think, than most IRBs
would be willing to countenance.

1362
00:47:18,469 --> 00:47:19,904
Okay, that's kids.

1363
00:47:19,904 --> 00:47:21,172
You're going to get more of
that.

1364
00:47:21,172 --> 00:47:23,845
So that leaves the question
about what about competent

1365
00:47:23,845 --> 00:47:24,241
adults?

1366
00:47:24,241 --> 00:47:27,174
Should there be limits on the
extent

1367
00:47:27,174 --> 00:47:29,780
to which researchers can expose

1368
00:47:29,780 --> 00:47:32,600
competent adults to net risks,
to risks

1369
00:47:32,600 --> 00:47:34,552
for the benefit of others?

1370
00:47:34,885 --> 00:47:37,321
And how do we decide that?

1371
00:47:37,321 --> 00:47:38,689
And does it depend on things

1372
00:47:38,689 --> 00:47:40,944
like whether the participants
are healthy

1373
00:47:40,944 --> 00:47:42,593
or the participants are sick?

1374
00:47:42,593 --> 00:47:44,261
People debate that a lot.

1375
00:47:45,362 --> 00:47:47,098
Here are some common answers.

1376
00:47:47,098 --> 00:47:48,733
You'll see that it runs the
whole range.

1377
00:47:48,733 --> 00:47:50,658
So if you're sort of libertarian
minded,

1378
00:47:50,658 --> 00:47:51,669
you might say, look,

1379
00:47:51,669 --> 00:47:53,282
you said competent adults,
competent

1380
00:47:53,282 --> 00:47:54,805
adults get to decide what they
do

1381
00:47:54,805 --> 00:47:56,868
with their lives, including
facing risks

1382
00:47:56,868 --> 00:47:58,209
for the benefit of other.

1383
00:47:58,209 --> 00:48:00,644
So if they're competent, it's
just up to them.

1384
00:48:00,644 --> 00:48:02,747
Other people have really strict
standards.

1385
00:48:02,747 --> 00:48:06,117
They think, no, actually,
I think IRBs sometimes do this.

1386
00:48:06,450 --> 00:48:08,422
The net risks to which we expose
cannot

1387
00:48:08,422 --> 00:48:09,787
or shouldn't be much more,

1388
00:48:10,321 --> 00:48:12,912
if any, greater than the risks
of which we're willing to expose

1389
00:48:12,912 --> 00:48:13,324
children.

1390
00:48:13,324 --> 00:48:15,178
Some people are really strict
about this,

1391
00:48:15,178 --> 00:48:16,761
and then other people have
argued.

1392
00:48:16,761 --> 00:48:19,764
So Frank Miller
used to be in this department,

1393
00:48:19,997 --> 00:48:23,033
and he argues that,
well, let's look at research.

1394
00:48:23,033 --> 00:48:25,305
Participation
is kind of an altruistic

1395
00:48:25,305 --> 00:48:25,903
endeavor.

1396
00:48:25,903 --> 00:48:28,172
So let's look at what risks
we think are acceptable

1397
00:48:28,172 --> 00:48:31,342
in other altruistic endeavors
like organ donation.

1398
00:48:31,342 --> 00:48:33,257
Now notice that looks like it
would allow

1399
00:48:33,257 --> 00:48:34,145
pretty high risks.

1400
00:48:34,145 --> 00:48:36,179
It's interesting at the end of
this paper,

1401
00:48:36,179 --> 00:48:37,148
although these guys

1402
00:48:37,148 --> 00:48:39,011
suggest that standard, they're
kind of

1403
00:48:39,011 --> 00:48:40,384
like maybe that's too high.

1404
00:48:40,384 --> 00:48:41,051
We're not sure.

1405
00:48:43,554 --> 00:48:45,556
Alex Lundin, who's at Carnegie
Mellon,

1406
00:48:45,556 --> 00:48:47,757
suggests similar thing, but
usually like

1407
00:48:47,757 --> 00:48:49,794
a public service, like
firefighters.

1408
00:48:49,794 --> 00:48:51,595
And what do we think are risks
acceptable

1409
00:48:51,595 --> 00:48:53,801
and that those are going to set
the limits

1410
00:48:53,801 --> 00:48:54,799
for research risks

1411
00:48:54,799 --> 00:48:57,101
with competent adults. Okay.

1412
00:48:57,101 --> 00:49:00,104
So just a very brief summary.

1413
00:49:01,906 --> 00:49:03,240
Doing a systematic risk

1414
00:49:03,240 --> 00:49:05,342
benefit assessment is right by
the Belmont

1415
00:49:05,342 --> 00:49:06,243
Report was right.

1416
00:49:06,444 --> 00:49:08,412
It's important to do this right.

1417
00:49:08,412 --> 00:49:10,881
I don't think people are doing
it right.

1418
00:49:10,881 --> 00:49:12,923
And I think that
if we did it more

1419
00:49:12,923 --> 00:49:13,884
systematically,

1420
00:49:13,884 --> 00:49:15,548
it would make the research
better,

1421
00:49:15,548 --> 00:49:17,555
it would make the research more
ethical.

1422
00:49:17,788 --> 00:49:20,327
And then the second part of it,
which I hope didn't get too

1423
00:49:20,327 --> 00:49:20,758
confusing

1424
00:49:20,758 --> 00:49:21,258
and detract

1425
00:49:21,258 --> 00:49:23,248
from understanding the
framework, is that

1426
00:49:23,248 --> 00:49:24,995
I think there's really
interesting,

1427
00:49:25,262 --> 00:49:28,071
challenging questions here
for us, for anybody who's

1428
00:49:28,071 --> 00:49:28,666
interested

1429
00:49:28,999 --> 00:49:30,501
in research ethics, to figure
out.