WEBVTT

NOTE
This file was generated by Descript 

00:00:00.000 --> 00:00:03.519
Judith: Welcome to Berry's In the
Interim podcast, where we explore the

00:00:03.519 --> 00:00:08.109
cutting edge of innovative clinical
trial design for the pharmaceutical and

00:00:08.109 --> 00:00:10.340
medical industries, and so much more.

00:00:10.909 --> 00:00:11.539
Let's dive in.

00:00:12.769 --> 00:00:15.049
Scott Berry: Well, welcome back everybody.

00:00:15.049 --> 00:00:18.529
To, in the interim, I am
your host, Scott Berry.

00:00:19.189 --> 00:00:20.959
Appreciate you joining me here.

00:00:21.079 --> 00:00:24.169
Uh, as, as on many of
these I have discussed.

00:00:25.159 --> 00:00:29.749
Uh, my wife Tammy, listens to all of
these and her comments to me about,

00:00:30.229 --> 00:00:32.509
uh, about the various podcasts.

00:00:33.019 --> 00:00:36.379
Uh, also asking, you know, where
am I looking in all of this?

00:00:36.379 --> 00:00:40.999
So I, you can see I've, I've this,
this sort of setup and I don't.

00:00:41.084 --> 00:00:44.324
I don't peer into the, the camera here.

00:00:44.654 --> 00:00:49.634
I'm looking at my large monitor
where I have my, my notes about the

00:00:49.634 --> 00:00:53.144
podcast and my, my other computer
down here and the camera's over here.

00:00:53.144 --> 00:00:55.874
So I, I know it's a little bit
odd that I'm always looking

00:00:55.874 --> 00:00:57.074
in the opposite direction.

00:00:57.554 --> 00:00:59.294
Maybe this is in part my.

00:00:59.749 --> 00:01:04.849
Uh, my introvertedness of being
a statistician, it's, it's

00:01:04.849 --> 00:01:06.709
hard to peer into the camera.

00:01:06.709 --> 00:01:08.899
There's nobody there, uh, in it.

00:01:09.199 --> 00:01:12.559
And that, that, that's the
old, that's the old joke.

00:01:12.559 --> 00:01:15.589
Uh, how do you tell an
introverted statistician from

00:01:15.589 --> 00:01:17.389
an extroverted statistician?

00:01:18.169 --> 00:01:20.989
The extrovert looks at
somebody else's shoes.

00:01:22.789 --> 00:01:27.649
Alright, so I, I think it's fair game
for, for me to criticize statisticians

00:01:27.649 --> 00:01:33.739
and I, I may do that again, uh, because
I am one, uh, in this years ago.

00:01:35.314 --> 00:01:36.664
We, I was involved.

00:01:36.664 --> 00:01:37.414
We were involved.

00:01:37.414 --> 00:01:38.854
Barry consultants were involved.

00:01:38.854 --> 00:01:43.804
A number of people, really wonderful
effort in a project called Adapt It

00:01:44.254 --> 00:01:51.484
Adapt it was funding from the FDA
and the NIH to investigate bringing

00:01:51.484 --> 00:01:57.604
innovative trial designs into
trials that were part of the net.

00:01:58.279 --> 00:02:01.129
Neurological emergency treatment trials.

00:02:01.519 --> 00:02:05.959
And we went in and we, we, we took what
were gonna be five different trials.

00:02:05.959 --> 00:02:08.329
It turned out to be six
trials we investigated.

00:02:08.419 --> 00:02:11.659
And what if we looked to
innovate the trial design?

00:02:13.129 --> 00:02:14.089
What was really.

00:02:14.569 --> 00:02:20.929
Great about the project is we
had, uh, behavioral scientists

00:02:20.929 --> 00:02:28.039
studying that process of bringing
innovative trial designs, uh, to

00:02:28.039 --> 00:02:30.499
these, the, these particular trials.

00:02:31.279 --> 00:02:36.019
One of 'em was, uh, which ended
up being ice cap cardiac arrest.

00:02:37.564 --> 00:02:39.634
Sub people subjects, uh, sorry.

00:02:39.934 --> 00:02:42.514
Patient suffers cardiac arrest.

00:02:42.514 --> 00:02:44.404
They're brought into the emergency room.

00:02:44.404 --> 00:02:48.874
All of these particular scenarios
centered on the emergency room.

00:02:49.234 --> 00:02:53.914
And the question is, does hypothermia
create better neurological outcomes?

00:02:53.914 --> 00:02:57.364
There's a period of time where the,
the blood is not flowing in the brain.

00:02:57.364 --> 00:02:59.584
Lack of oxygen and heat.

00:02:59.654 --> 00:03:03.974
Makes that worse by putting a
patient in a hypothermic state,

00:03:03.974 --> 00:03:05.324
does that create better outcomes?

00:03:05.324 --> 00:03:09.164
That was just one of six where
there was a stroke trial, there

00:03:09.164 --> 00:03:10.964
was a spinal cord injury trial.

00:03:10.964 --> 00:03:14.474
There were a number of particular
trials, so there were behavioral

00:03:14.474 --> 00:03:19.319
scientists that were studying bringing
innovative trial designs to that.

00:03:19.319 --> 00:03:22.004
We explored adaptive sample size.

00:03:22.004 --> 00:03:24.734
We explored response,
adaptive randomization.

00:03:24.734 --> 00:03:27.374
We explored futility, interim analyses.

00:03:27.729 --> 00:03:31.869
Multiple arms modeling, all
in the different scenarios.

00:03:32.409 --> 00:03:36.879
The behavioral scientists in
studying this part of their, their

00:03:36.879 --> 00:03:41.334
goal was what are the barriers to
bringing innovative trial design?

00:03:41.644 --> 00:03:42.064
Lines.

00:03:42.874 --> 00:03:44.434
Is it the funders?

00:03:44.614 --> 00:03:46.024
Is it clinicians?

00:03:46.144 --> 00:03:47.794
Is it regulators?

00:03:47.944 --> 00:03:49.504
Is it operations?

00:03:49.714 --> 00:03:51.544
Are it statisticians?

00:03:51.874 --> 00:03:54.784
Uh, it what, what are
the various barriers?

00:03:55.084 --> 00:04:00.814
Interestingly, the number one barrier
that they found worse, statistic.

00:04:02.719 --> 00:04:06.949
Which is kind of an interesting place
that, that interesting role, and I

00:04:06.949 --> 00:04:13.459
think it was a different time, this
was about 2005 to 2010 ish that we

00:04:13.459 --> 00:04:15.439
were investigating these trials.

00:04:15.679 --> 00:04:20.119
It was a different time, but they,
they were hurdles that you had

00:04:20.119 --> 00:04:24.889
to overcome that bringing these
innovations, the what ifs in all of that.

00:04:25.789 --> 00:04:27.529
Um, now.

00:04:29.374 --> 00:04:32.014
What are, what we, we do
this at Berry Consultants.

00:04:32.014 --> 00:04:34.264
We do a number of what
I'll call innovative.

00:04:34.264 --> 00:04:37.864
And, and now innovative is
a self glossing is better.

00:04:37.864 --> 00:04:42.214
And I I always get concerned saying,
oh, we design innovative trials By

00:04:42.364 --> 00:04:45.304
some sense you mean better trials,
smarter trials we're better.

00:04:45.724 --> 00:04:50.464
Um, but what I mean by innovative
trials is, is things that are

00:04:50.464 --> 00:04:52.384
rarely or never been done before.

00:04:52.399 --> 00:04:52.459
Her.

00:04:53.179 --> 00:04:57.319
Now, you could argue whether they're
worse or better, so I wanna leave

00:04:57.319 --> 00:05:01.609
that aside, but I, but I'll bring
that we're doing things that are done

00:05:01.609 --> 00:05:06.199
relatively rarely or have never been
done before within trial designs, I'll

00:05:06.199 --> 00:05:08.059
call those innovative trial designs.

00:05:08.749 --> 00:05:08.779
Okay?

00:05:09.199 --> 00:05:12.259
What turns out to be barriers?

00:05:12.319 --> 00:05:17.149
Uh, I, I, you know, statisticians
in some cases can be barriers.

00:05:17.329 --> 00:05:19.489
I don't think there.

00:05:20.104 --> 00:05:21.184
They're barriers.

00:05:21.574 --> 00:05:25.624
Uh, I think you can present
and create scenarios.

00:05:25.924 --> 00:05:29.824
What are the way that we make it
more, more robust and handle that?

00:05:30.334 --> 00:05:38.614
What turns out to be a barrier to, to
me, I'll speak for myself, is the, the

00:05:38.614 --> 00:05:44.404
scientists who say, I can't define it.

00:05:44.464 --> 00:05:46.474
I'll know it when I see it.

00:05:48.454 --> 00:05:50.224
Now, what do I mean by that?

00:05:51.064 --> 00:05:56.524
The scenario is you're, you're
trying to look at adaptive designs

00:05:56.524 --> 00:06:00.754
or you're trying to measure is
design A better than design B?

00:06:01.834 --> 00:06:05.554
You have to be able to understand
what's the goal of the trial,

00:06:05.554 --> 00:06:07.054
what are we trying to do?

00:06:07.054 --> 00:06:11.014
Are we trying to find the
right dose, the right duration?

00:06:11.014 --> 00:06:12.664
Are we trying to find the right patients?

00:06:12.844 --> 00:06:14.224
And what is success?

00:06:14.884 --> 00:06:19.084
If we're trying to compare A to B, we've
gotta be able to say, A does a better

00:06:19.084 --> 00:06:24.934
job at fi, uh, uh, at being successful
than B does, or vice versa, or the amount

00:06:24.934 --> 00:06:26.974
of resources to get the right answer.

00:06:27.244 --> 00:06:32.224
But you have to be able to define
what is good, what's right what, what

00:06:32.224 --> 00:06:34.354
trials would you go to Phase three.

00:06:35.824 --> 00:06:39.694
And when you, when you want to simulate
that and you want to compare that and

00:06:39.694 --> 00:06:46.474
you want to say is a better than tr trial
B if, if the, if the scientists on the

00:06:46.474 --> 00:06:52.024
other side say, I can't define what's
better, show me the trial result and I'll

00:06:52.024 --> 00:06:53.584
tell you whether I'll go to phase three.

00:06:55.264 --> 00:06:58.684
That doesn't lend itself
to analytical comparison.

00:06:58.684 --> 00:07:02.824
It doesn't lend itself to
building a better design because

00:07:02.824 --> 00:07:04.564
I don't know what success is.

00:07:05.524 --> 00:07:06.694
When are we there?

00:07:06.694 --> 00:07:08.584
What, what, what is good and what is bad?

00:07:08.914 --> 00:07:15.154
That becomes really the enemy
to doing innovative designs.

00:07:15.889 --> 00:07:20.059
Now, uh, I'll talk as we go through
here about ways to combat that.

00:07:20.059 --> 00:07:26.524
I don't think that's, I, I think you can
get out of those individuals good answers.

00:07:28.234 --> 00:07:29.314
In that scenario.

00:07:29.464 --> 00:07:33.304
So I wanna bring up a common
scenario where this can happen.

00:07:33.964 --> 00:07:40.444
And this was, by the way, um, this was a
topic suggested to me as somebody said,

00:07:40.444 --> 00:07:42.844
oh, I'm sure you did a podcast on this.

00:07:42.844 --> 00:07:43.714
I'll go find it.

00:07:44.044 --> 00:07:44.944
And I hadn't.

00:07:44.974 --> 00:07:47.764
But it is, it's a ubiquitous thing for us.

00:07:47.764 --> 00:07:53.014
This comes up all the time, and that
is Here I'll, I'll, I'll, I'll do a

00:07:53.014 --> 00:07:56.044
setup and then I'll get to the topic is.

00:07:57.169 --> 00:08:02.629
We're a a, a sponsors looking
to design a phase two trial.

00:08:03.739 --> 00:08:08.479
The phase two trial has two doses
in a placebo, so it's a three arm

00:08:08.479 --> 00:08:14.089
trial, and they're thinking it's a
reasonably rare disease, and they're

00:08:14.089 --> 00:08:15.739
thinking it is a rare disease.

00:08:16.099 --> 00:08:21.709
They're thinking of running a
trial that is 30, 30 and 30, so 30

00:08:21.709 --> 00:08:25.999
patients on placebo, 30 on dose,
the low dose, 30 on the high dose.

00:08:27.154 --> 00:08:29.854
They want to run the phase two
trial and then they're gonna

00:08:29.854 --> 00:08:31.264
run the phase three trial.

00:08:31.864 --> 00:08:34.835
And this is a scenario
where we know the phase 3M.

00:08:35.074 --> 00:08:35.314
Point.

00:08:36.424 --> 00:08:38.884
Uh, there, there are some scenarios
where you're running phase two

00:08:38.884 --> 00:08:42.604
and it's really a wide open space
of what the endpoint might be.

00:08:42.814 --> 00:08:47.704
That the endpoint is, is somewhat dictated
by the area measuring clinical benefit.

00:08:48.034 --> 00:08:52.114
So the, the endpoint is known and
then they're gonna run a phase three,

00:08:52.114 --> 00:08:57.574
which is a one-to-one comparison of
the dose that they select to placebo.

00:08:58.234 --> 00:09:04.354
And they're thinking that trial might be
90 versus 90, that the effect size that

00:09:04.354 --> 00:09:09.364
they need to have, they, they have a
pretty good idea of that trial being 90

00:09:09.364 --> 00:09:12.634
versus 90 being powered at their effect.

00:09:13.564 --> 00:09:16.384
So they're looking at a
development pathway that is

00:09:16.384 --> 00:09:18.634
phase two, which is 30, 30, 30.

00:09:20.209 --> 00:09:24.019
They, uh, get full follow up on patients.

00:09:24.019 --> 00:09:26.389
They go through the analysis of that data.

00:09:26.389 --> 00:09:31.009
They look at that data, they make
a decision on going to phase three.

00:09:31.219 --> 00:09:33.499
They then submit the protocol.

00:09:33.499 --> 00:09:36.889
They go through all what I'll call
the white space between two and

00:09:36.889 --> 00:09:41.539
three, that they then run a trial
of 90 versus 90 in phase three.

00:09:42.949 --> 00:09:46.549
So we, we said, well, what
about a seamless two three trial

00:09:47.179 --> 00:09:50.839
where you enroll 30, 30, 30?

00:09:52.219 --> 00:09:59.749
You select a dose from that, and the
trial seamlessly moves into phase

00:09:59.749 --> 00:10:04.939
three, which is randomization of
placebo to the dose that's selected.

00:10:07.099 --> 00:10:10.429
In phase three, and we can talk
about the size of that, but let's,

00:10:10.429 --> 00:10:17.089
for a minute, just think about the,
the separate trials was a 90 patient

00:10:17.089 --> 00:10:19.819
trial followed by 180 patients.

00:10:20.749 --> 00:10:24.589
The seamless II/III trial
could be 30, 30, 30.

00:10:25.369 --> 00:10:33.304
You pick a dose and you enroll 60 against
60 in the second part of the trial.

00:10:34.744 --> 00:10:38.854
And the patients in the first
part and the second part are

00:10:38.854 --> 00:10:40.744
part of the primary analysis.

00:10:41.134 --> 00:10:46.114
So the dose that's selected in the
seamless trial will have 90 patients

00:10:46.144 --> 00:10:51.244
against placebo, 90, and you'll
have 30 on the unselected dose.

00:10:52.474 --> 00:10:57.664
Now what, what does that mean
In the seamless trial, you

00:10:57.664 --> 00:10:59.854
do have to adjust the alpha.

00:11:00.574 --> 00:11:04.594
At the end, if you're gonna include
those patients because you, you quote

00:11:04.594 --> 00:11:09.574
unquote picked the better dose in phase
II what you cherry picked the better

00:11:09.574 --> 00:11:14.884
dose in that, and then you, you are
including that data at the end of phase

00:11:15.104 --> 00:11:21.514
III straightforward ways to adjust
alpha and do that, that, that, what's

00:11:21.514 --> 00:11:28.084
great about that and what is, is kind
of amazing is that you have more power.

00:11:29.269 --> 00:11:33.379
Even though it's still 90 patients,
than you will have, if you run a

00:11:33.379 --> 00:11:39.829
separate 90 vs 90 phase III the power
of the Seamless II/III trial, which is

00:11:39.829 --> 00:11:48.709
90 patients followed by 120 patients,
210 total compared to 270 total, it's

00:11:48.709 --> 00:11:56.854
higher power for demonstrating success
at the end of that phase III So huge

00:11:56.854 --> 00:11:59.824
benefits in that scenario for resources.

00:12:00.154 --> 00:12:04.324
But the other part of it that's
just incredible is the lack

00:12:04.324 --> 00:12:05.974
of white space between them.

00:12:06.724 --> 00:12:11.824
The time for that trial to read out,
the time to then select the dose.

00:12:11.824 --> 00:12:16.264
Start phase three, enroll
180, uh, another 180 patients.

00:12:16.834 --> 00:12:19.354
So we, we simulated that.

00:12:19.384 --> 00:12:23.164
And by the way, you run into this
case of how do we pick the dose?

00:12:24.379 --> 00:12:27.379
Between them, and I don't wanna
spend a ton of time on this.

00:12:27.379 --> 00:12:31.549
Maybe that's a whole separate
podcast of picking that dose.

00:12:31.579 --> 00:12:33.679
It can be purely on efficacy.

00:12:33.679 --> 00:12:36.349
It can be a combination
of efficacy and safety.

00:12:36.619 --> 00:12:41.779
It can be efficacy with A-D-S-M-B,
weighing in on whether there's a, a

00:12:41.779 --> 00:12:43.979
safety or tolerability issue on it.

00:12:44.884 --> 00:12:49.204
But for now, let's think about picking
the better dose based on efficacy.

00:12:49.204 --> 00:12:53.464
And in this particular case, the
primary endpoint at the end of phase

00:12:53.464 --> 00:12:55.534
three is really a combination of both.

00:12:55.534 --> 00:13:00.724
It's it's the clinical outcome of
the patient, so we're gonna pick one

00:13:00.724 --> 00:13:02.284
that has the better clinical outcome.

00:13:02.284 --> 00:13:07.144
That's a case by the way, that you could
run into this enemy that I referred to of,

00:13:07.594 --> 00:13:09.514
I can't tell you how I'll pick the dose.

00:13:09.514 --> 00:13:11.384
You need to show me the
data and then I'll pick it.

00:13:11.494 --> 00:13:14.344
It, but that's not the topic.

00:13:14.944 --> 00:13:21.394
When we presented this and we showed
that this trial of 210 patients has more

00:13:21.394 --> 00:13:29.254
power than 270 patients adjusting for
Alpha, the incredible part of this is

00:13:29.254 --> 00:13:35.644
they estimated the time between phase two
and phase three will be almost a year.

00:13:37.384 --> 00:13:40.924
That, that's just the natural
space that will happen if they run

00:13:40.924 --> 00:13:46.384
a separate two A and two B, that
space disappears in a seamless two.

00:13:46.384 --> 00:13:51.274
A uh, seamless two, three
trial, uh, in that setting.

00:13:52.789 --> 00:13:55.189
We, we asked them how long that will take.

00:13:55.189 --> 00:13:56.389
We looked at enrollment.

00:13:56.389 --> 00:14:02.479
We simulated, uh, uh, the trials
using the numbers presented.

00:14:02.899 --> 00:14:09.799
And again, one trial is 270 patients
and the average time to readout

00:14:09.859 --> 00:14:15.019
of the phase three portion of
that was approximately 50 months.

00:14:16.069 --> 00:14:21.559
It was a range to that, that we had a,
a, a confidence interval or a prediction

00:14:21.559 --> 00:14:23.419
interval for the range of that.

00:14:24.019 --> 00:14:28.579
And the, the median time
for the inferentially

00:14:28.579 --> 00:14:31.189
seamless trial was 25 months.

00:14:32.119 --> 00:14:38.059
Just call it 45 months to 25
months, 20 months different.

00:14:39.124 --> 00:14:45.484
Less resources, higher probability
of success, win, win win.

00:14:46.174 --> 00:14:47.284
In that scenario,

00:14:49.534 --> 00:14:54.034
the feedback was, no, we don't want to
do the inferentially seamless trial,

00:14:54.634 --> 00:14:58.534
and it's a case where it's sort of, I
have this, wait a minute, what, what?

00:14:58.534 --> 00:15:00.004
What do I not understand?

00:15:00.874 --> 00:15:05.734
This is a very common response
actually, in a scenario like this.

00:15:07.009 --> 00:15:12.919
Uh, another aspect of this that's so
important is the chain of communication.

00:15:13.969 --> 00:15:20.359
If the communication is somebody said no,
and that went through person A and person

00:15:20.359 --> 00:15:22.879
B, and then we hear the answer's, no.

00:15:23.509 --> 00:15:26.929
It's hard to understand why,
and it's hard to ask about it.

00:15:26.929 --> 00:15:29.059
It's hard to potentially understand.

00:15:29.329 --> 00:15:31.339
Is there a remedy to that?

00:15:31.339 --> 00:15:31.609
No.

00:15:32.449 --> 00:15:35.749
And many times over the years we
brought, I brought this up that

00:15:35.749 --> 00:15:40.549
we're trying to do innovative things,
things that are not commonly done.

00:15:41.269 --> 00:15:46.459
It's not infrequent that somebody
says, no operation says no.

00:15:46.639 --> 00:15:51.349
Regulators say no, somebody
says no, and understanding why.

00:15:51.349 --> 00:15:58.519
A lot of times there's a reasonably
straightforward way to remedy that to

00:15:58.519 --> 00:16:00.979
then get cheaper, faster, stronger.

00:16:02.764 --> 00:16:07.234
The reason for this, which is a
very common reason, is funding

00:16:09.244 --> 00:16:13.624
the company is on the smaller side
of the pharmaceutical industry.

00:16:14.434 --> 00:16:21.454
A big pharma may be able to fund the phase
two and phase three, and they can set

00:16:21.454 --> 00:16:23.314
up those criteria and they can run it.

00:16:23.314 --> 00:16:25.264
What about a smaller biotech?

00:16:26.209 --> 00:16:31.339
They currently don't have the
money to finish the 210 patients.

00:16:31.339 --> 00:16:33.739
They can run the 90 patient trial.

00:16:34.129 --> 00:16:37.219
They have the funding for that,
but they need to get funding

00:16:37.219 --> 00:16:38.989
in order to enroll the rest.

00:16:39.979 --> 00:16:43.909
So that's a very common
response, a very real, uh, thing.

00:16:43.909 --> 00:16:47.509
I'm not gonna laugh at that in any
particular thing, but it, it, it's

00:16:47.509 --> 00:16:53.839
part of designing these trials,
innovative trials is can we combat that?

00:16:54.454 --> 00:16:57.994
So it's a very common
scenario, by the way.

00:16:57.994 --> 00:17:00.094
It's the importance of being there.

00:17:00.334 --> 00:17:03.784
When somebody says no, it says
maybe we have a remedy for that.

00:17:04.864 --> 00:17:10.174
So can we set up funding solutions?

00:17:11.104 --> 00:17:16.924
And triggers of the phase
two that triggers funding.

00:17:17.524 --> 00:17:22.594
And, and by the way, these funders could
be private equity, the funders could be

00:17:22.594 --> 00:17:27.994
a larger pharmaceutical company, could
be a range of individuals that might

00:17:27.994 --> 00:17:30.364
be interested in investing in the drug.

00:17:30.394 --> 00:17:35.764
They would like to invest in a drug that's
going to be successful in this scenario.

00:17:36.589 --> 00:17:42.979
So in the separate trials, what typically
happens is the phase two data reads

00:17:42.979 --> 00:17:48.769
out and they unblind themselves to that
data, and they can share that data with

00:17:48.769 --> 00:17:54.139
a potential funder, and the funder can
play with it, and they go back to this.

00:17:54.139 --> 00:17:54.769
We know it.

00:17:54.769 --> 00:18:00.619
When we see it, the funder then can
decide, will I invest X dollars in this?

00:18:01.129 --> 00:18:04.219
As we move forward for the
phase three trial, and then

00:18:04.219 --> 00:18:06.919
I'm a partner with the biotech.

00:18:07.099 --> 00:18:10.969
If this ends up being successful and
the drug makes money, we're invested

00:18:10.969 --> 00:18:16.609
in that and we bring in our experts
and we're we're able to decipher

00:18:16.609 --> 00:18:21.649
from the phase two data whether we're
gonna be successful in phase three,

00:18:21.649 --> 00:18:25.699
what the market share looks like,
and is this a good investment for us?

00:18:25.759 --> 00:18:28.789
And that's a very common
scenario, and that's what happens

00:18:28.789 --> 00:18:30.589
in a typical phase two truck.

00:18:31.594 --> 00:18:34.114
Now what are we looking to do?

00:18:34.114 --> 00:18:39.664
We're looking to take that phase
two trial and setting up objective

00:18:39.664 --> 00:18:42.994
criteria that would trigger funding.

00:18:44.314 --> 00:18:48.844
So you go to a funder and you say,
okay, you're interested in funding this.

00:18:49.204 --> 00:18:55.654
Can we define what you would need to
see in phase two to fund phase three?

00:18:57.514 --> 00:18:59.224
If we can do that.

00:19:00.049 --> 00:19:06.799
We can do that interim analysis
where we select the dose and we

00:19:06.799 --> 00:19:11.689
can also let the funder know, did
you hit your funding criteria?

00:19:13.699 --> 00:19:17.209
Sometimes this is done
separate from dose selection.

00:19:17.569 --> 00:19:21.229
By the way, this can be done in a
phase three trial, so this isn't

00:19:21.229 --> 00:19:23.269
unique to Seamless two three trial.

00:19:23.269 --> 00:19:26.309
That was just the particular recent
example that we're working on.

00:19:27.319 --> 00:19:32.659
Sometimes this happens where the company's
ready to run phase three, but they

00:19:32.659 --> 00:19:35.179
can't fund phase three in its entirety.

00:19:35.809 --> 00:19:39.289
And a funder says, you know,
I'm interested in this drug,

00:19:39.289 --> 00:19:40.639
but I need to see more.

00:19:40.639 --> 00:19:45.829
I need more proof of proof of concept
of this before I'm willing to invest

00:19:45.829 --> 00:19:47.929
the big pot of money for phase three.

00:19:49.279 --> 00:19:54.739
Sometimes you can start phase
three and do an interim analysis.

00:19:55.939 --> 00:20:04.369
Of whether we hit that proof of concept,
proof of funding that the funder says yes.

00:20:04.369 --> 00:20:09.169
If you could jump this hurdle, we
would fund the rest of phase three.

00:20:11.119 --> 00:20:13.909
If we can objectively set that up.

00:20:13.909 --> 00:20:18.919
Agreement can be made that an
analysis takes place during the trial.

00:20:19.219 --> 00:20:22.369
I'm not gonna call it an interim
analysis, but it, and I'll come back

00:20:22.369 --> 00:20:24.829
to that, an analysis takes place.

00:20:25.039 --> 00:20:30.079
It's conducted by an independent
statistical group, an independent

00:20:30.079 --> 00:20:34.909
CRO, and literally it's a checkbox.

00:20:36.349 --> 00:20:39.169
Did we hit the funding criteria?

00:20:39.169 --> 00:20:41.659
Maybe it's two or three check boxes.

00:20:42.169 --> 00:20:44.749
Did we jump the following hurdle?

00:20:44.749 --> 00:20:49.849
And that is signed by the independent
group and that is given to the

00:20:49.849 --> 00:20:54.079
sponsor and we'll come back to who,
who knows and who sees what when.

00:20:55.039 --> 00:20:59.269
But that checkbox could be presented
to the funder in agreement.

00:20:59.629 --> 00:21:03.679
That they've agreed to now fund
the rest of the trial based on the

00:21:03.679 --> 00:21:05.839
deal they have with the biotech.

00:21:06.649 --> 00:21:11.329
And they know they hit proof of
concept in phase two, even though

00:21:11.329 --> 00:21:13.339
they don't get to see the data.

00:21:14.749 --> 00:21:18.079
Now what that, that's the big
question of what you get to

00:21:18.079 --> 00:21:19.399
see during the course of it.

00:21:19.429 --> 00:21:19.969
So.

00:21:21.679 --> 00:21:27.709
The data in a phase three trial
with one of these analyses or

00:21:27.709 --> 00:21:29.299
a seamless two, three trial?

00:21:29.299 --> 00:21:34.429
It doesn't really, neither one is
different because as far as regulators

00:21:34.429 --> 00:21:39.019
in the scientific community are
concerned, those patients in the phase

00:21:39.019 --> 00:21:42.409
two part or the earlier part of phase
three, they're phase three patients.

00:21:44.359 --> 00:21:49.699
It is, it is considered bad
form and we'll come back.

00:21:49.729 --> 00:21:50.239
I'll come back to that.

00:21:51.079 --> 00:21:55.099
To unblind the data during the
course of a phase three trial.

00:21:55.759 --> 00:22:01.489
Sponsors are remained blinded to what
is the effect size of the treatment.

00:22:01.489 --> 00:22:06.469
During the course of that, we don't
want Jane Doe to be unblinded.

00:22:06.469 --> 00:22:10.519
We don't want the pharmaceutical
company to know that Jane Doe or

00:22:10.519 --> 00:22:13.429
John Doe was treatment or placebo.

00:22:13.849 --> 00:22:14.774
It does the data.

00:22:15.614 --> 00:22:19.214
You know, is there altered, is
there some aspect of knowing that

00:22:19.214 --> 00:22:21.944
unblinded data that changes the trial?

00:22:22.034 --> 00:22:25.484
And, and this is this thing
called operational bias.

00:22:25.484 --> 00:22:28.994
So we'll come back to why, why,
typically you don't, but let's

00:22:28.994 --> 00:22:31.214
just assume that you can't.

00:22:31.549 --> 00:22:32.899
Unblind the data.

00:22:32.929 --> 00:22:35.449
Go to a conference present.

00:22:35.539 --> 00:22:38.989
Ooh, look how good our data are in
the middle of the phase three trial.

00:22:38.989 --> 00:22:40.489
Will you help fund the rest?

00:22:40.849 --> 00:22:44.389
Which is why a lot of times it's a
separate trial so that they can do

00:22:44.389 --> 00:22:46.789
that or, or, or present their data.

00:22:47.059 --> 00:22:50.989
Unblind the data in the middle
of a phase three with one of

00:22:50.989 --> 00:22:52.999
these analyses, or a two, three.

00:22:54.679 --> 00:22:58.249
You can't interrogate the data itself.

00:22:58.249 --> 00:23:03.739
You don't want to unblind yourself to
any individual patient, but you can come

00:23:03.739 --> 00:23:10.789
back with a limited set of information
of is the data progressing well at

00:23:10.789 --> 00:23:13.249
this point above a particular hurdle?

00:23:13.909 --> 00:23:17.599
Sometimes if futility analysis
are done in trials a lot, this is

00:23:17.599 --> 00:23:19.609
different than a futility analysis.

00:23:20.779 --> 00:23:22.549
What do we call this by the way?

00:23:23.149 --> 00:23:27.139
We tend to call it either
an administrative analysis

00:23:27.139 --> 00:23:28.939
or a financial analysis.

00:23:30.649 --> 00:23:34.369
We'll write into the protocol
that an administrative or

00:23:34.369 --> 00:23:36.439
financial analysis will be done.

00:23:36.709 --> 00:23:44.509
We don't want to call it futility because
writing in futility means if you don't hit

00:23:44.509 --> 00:23:47.959
that, we're gonna stop enrolling patients.

00:23:48.709 --> 00:23:50.509
We're gonna change the trial.

00:23:51.484 --> 00:23:53.944
But this analysis doesn't change.

00:23:53.944 --> 00:23:56.764
The trial at all doesn't change
the sample size doesn't change

00:23:56.764 --> 00:23:58.114
how a patient is treated.

00:23:58.804 --> 00:24:01.264
All it does is trigger funding.

00:24:01.294 --> 00:24:02.104
Potentially.

00:24:02.434 --> 00:24:06.394
It could be that that funder, it
doesn't trigger the funding, but

00:24:06.394 --> 00:24:09.394
the company gets funding another
way and the trial continues on.

00:24:10.984 --> 00:24:13.324
So this is not something that affects.

00:24:14.074 --> 00:24:17.134
Patients, the, the, the
scientific aspect of it.

00:24:17.524 --> 00:24:19.564
And so we don't want to call it futility.

00:24:19.564 --> 00:24:21.634
It's a misusing of futility.

00:24:21.634 --> 00:24:25.054
If you do that for funding,
call it administrative, call it

00:24:25.054 --> 00:24:26.554
financial, and have that done.

00:24:27.484 --> 00:24:32.644
So let's come back a little bit to, to
why we do that and who knows what, when.

00:24:33.259 --> 00:24:33.319
Ha.

00:24:34.039 --> 00:24:39.919
But if you read the FDA guidance
document on adaptive designs where

00:24:39.919 --> 00:24:43.489
interim analysis are done, they
have a great deal of concern of

00:24:43.489 --> 00:24:45.919
what they call operational bias.

00:24:46.969 --> 00:24:54.229
Operational bias is, is, it's a hard
thing to define largely, it's that by

00:24:54.229 --> 00:25:00.949
knowing the knowledge of the data, it
alters the future of the trial or the

00:25:00.949 --> 00:25:03.259
interpretation of the data in some way.

00:25:03.289 --> 00:25:03.319
Okay.

00:25:04.039 --> 00:25:10.549
Imagine in the middle of a trial if, if
patients find out that the investigational

00:25:10.549 --> 00:25:16.609
treatment isn't working very well and
many of them drop out of the trial, now

00:25:16.609 --> 00:25:19.129
you've got a huge missing data problem.

00:25:19.129 --> 00:25:22.699
The interpretation of the
trial, you've, you've.

00:25:23.479 --> 00:25:24.859
Uh, risk.

00:25:24.859 --> 00:25:29.539
The whole interpretation and the ability
to understand is a treatment beneficial.

00:25:29.989 --> 00:25:35.599
The opposite could happen if there's
a really positive signal that patients

00:25:35.599 --> 00:25:40.489
drop out and go elsewhere, that
future patients enrolled look very

00:25:40.489 --> 00:25:44.539
different, that the investigator will
no longer en enroll this patient or

00:25:44.539 --> 00:25:46.759
that patient, or something changes.

00:25:47.674 --> 00:25:51.934
Now this, these are loosely defined and
it's, it's, by the way, it's hard to

00:25:51.934 --> 00:25:54.064
know if you ever have operational bias.

00:25:54.214 --> 00:25:59.074
Uh, it's hard to defend that you didn't,
if somebody says, oh, you knew this, and

00:25:59.074 --> 00:26:01.174
the it, it, it, it changed your trial.

00:26:01.414 --> 00:26:03.454
It's hard to prove that it wasn't there.

00:26:03.814 --> 00:26:10.924
So we do a great deal to, to make sure
that there's no concerns about this

00:26:11.134 --> 00:26:13.714
mysterious thing called operational bias.

00:26:15.169 --> 00:26:19.669
So during the course of a phase three
trial, we don't want the pharmaceutical

00:26:19.669 --> 00:26:22.039
company, the sponsor to know the data.

00:26:22.039 --> 00:26:24.949
We don't want clinical sites
to know the data patients, to

00:26:24.949 --> 00:26:26.329
know the data and all that.

00:26:26.329 --> 00:26:29.869
It's important within the
scientific credibility of the trial.

00:26:30.829 --> 00:26:32.569
So in this circumstance.

00:26:34.279 --> 00:26:34.609
What?

00:26:34.609 --> 00:26:41.179
It's not uncommon that you get a piece of
paper that says, was this criteria met?

00:26:41.179 --> 00:26:43.939
And I'll come back to what
do those criteria look like?

00:26:43.999 --> 00:26:46.099
What kind of things do we do for that?

00:26:46.459 --> 00:26:47.569
But is it met?

00:26:47.569 --> 00:26:47.929
Yes.

00:26:47.929 --> 00:26:48.289
No.

00:26:49.024 --> 00:26:53.524
Is criteria one met or
criteria two met or no?

00:26:53.674 --> 00:26:54.964
So three check boxes.

00:26:54.964 --> 00:26:56.734
So what does that mean?

00:26:56.734 --> 00:27:00.964
Yes, you can reverse engineer
from the checkbox roughly.

00:27:01.144 --> 00:27:07.474
Are the data above or below this
largely that's not considered to

00:27:07.474 --> 00:27:10.114
create the risk of operational bias.

00:27:10.744 --> 00:27:17.104
You still want to keep that
information relatively confidential.

00:27:17.884 --> 00:27:23.074
You don't want to publicly disclose
if you can prevent it, that

00:27:23.314 --> 00:27:26.194
our effect size was above 20%.

00:27:26.194 --> 00:27:31.714
At the interim, does that cause
this potential operational bias?

00:27:32.614 --> 00:27:37.564
So setting up a relationship with a funder
that says you're gonna get this checkbox,

00:27:37.804 --> 00:27:43.474
the at the sponsor, maybe somebody in
the C-suite gets this as well, A CFO.

00:27:44.779 --> 00:27:50.959
Somebody who's not going to sites and
talking to clinicians and investigators

00:27:50.959 --> 00:27:54.829
that are enrolling patients, you'd
like to create that separation.

00:27:54.919 --> 00:27:59.929
You'd like the people involved in
running the trial to remain as blinded

00:27:59.929 --> 00:28:01.669
as possible during the course of it.

00:28:02.689 --> 00:28:06.919
So it might be that the, the CFO, the CEO.

00:28:07.169 --> 00:28:12.629
Seize that data, but otherwise it is
firewalled from the rest of the company.

00:28:12.779 --> 00:28:18.329
And then the funder can get this
information also agreeing not to

00:28:18.329 --> 00:28:22.619
publicly disclose it, but they agree
to write a check if that's there.

00:28:24.449 --> 00:28:28.829
Now, what kind of criteria do we
tend to use here in, in, in a phase

00:28:28.829 --> 00:28:34.319
three trial in that, uh, it's not
uncommon that we use the predictive

00:28:34.319 --> 00:28:36.119
probability of trial success.

00:28:37.864 --> 00:28:39.454
Now, what is that?

00:28:40.714 --> 00:28:46.564
That is a calculation that's integrating
over the uncertainty of the treatment.

00:28:46.774 --> 00:28:51.964
What's the probability that statistical
significance is gonna be met when

00:28:51.964 --> 00:28:54.154
we enroll out the full trial?

00:28:55.324 --> 00:29:00.934
Sometimes this is called assurance based
on a particular prior distribution.

00:29:01.759 --> 00:29:06.919
Uh, in that scenario, sometimes
conditional power is used given

00:29:06.919 --> 00:29:08.419
the effect size we've seen.

00:29:08.419 --> 00:29:10.909
Now, suppose that's the real effect size.

00:29:10.909 --> 00:29:12.829
What's the probability of success?

00:29:13.879 --> 00:29:18.259
The Bayesian predictive probability of
success doesn't assume a single effect.

00:29:18.259 --> 00:29:23.209
It assumes the uncertainty of that effect
given the data we have, it's a very

00:29:23.209 --> 00:29:28.699
natural Bayesian thing to integrate over
the uncertainty of the treatment effect.

00:29:29.854 --> 00:29:33.634
And calculate the probability
that we're gonna meet statistical

00:29:33.634 --> 00:29:35.344
significance at the end of the trial.

00:29:36.874 --> 00:29:41.524
What I've described that is a lot
of times, and, and you'll, you'll,

00:29:41.584 --> 00:29:44.494
you'll not be surprised that
I'll go to the sports world here.

00:29:44.914 --> 00:29:49.324
Uh, you can watch an NFL football
game, you can watch a basketball game.

00:29:49.564 --> 00:29:54.544
Uh, you can go to ESPN and it will tell
you the probability that one team's gonna

00:29:54.544 --> 00:29:57.454
win the game at any point in the, in the.

00:29:57.964 --> 00:29:58.894
The the game.

00:29:59.344 --> 00:30:03.424
So at halftime, one
team's ahead by 14 points.

00:30:03.874 --> 00:30:09.844
It might say the probability they win
this game is 92%, or if it's a tie

00:30:09.844 --> 00:30:13.084
game, might be 50/50 In that setting.

00:30:14.359 --> 00:30:17.269
That's the predictive
probability of success.

00:30:17.689 --> 00:30:21.709
That's exactly what we're doing in a
clinical trial, that we're calculating

00:30:21.709 --> 00:30:26.959
that given what we understand now and
the strength of the two teams, the

00:30:26.959 --> 00:30:30.919
strength of the two treatments, what's
the probability that we win the trial?

00:30:30.919 --> 00:30:34.999
Now, a clinical trial is not a
game where the high score wins.

00:30:35.209 --> 00:30:37.309
You've got to win by a point spread.

00:30:37.639 --> 00:30:40.459
You've gotta be statistically
significantly better.

00:30:41.179 --> 00:30:47.209
So if it's, if the two treatments are
tied at this point, the chance of winning

00:30:47.209 --> 00:30:49.369
this trial is gonna be less than 50%.

00:30:49.369 --> 00:30:51.739
It depends on what sample
size it is in that.

00:30:52.129 --> 00:30:56.029
If you are seeing exactly
the effect size you need.

00:30:56.374 --> 00:31:00.124
To show at the end of the trial,
that's where it's about 50 50,

00:31:00.454 --> 00:31:01.684
that you're gonna win the trial.

00:31:01.684 --> 00:31:03.604
If the data gets a little
bit better, you're gonna win.

00:31:03.604 --> 00:31:06.514
If it's a little bit worse, you're
gonna lose statistical significance.

00:31:06.844 --> 00:31:09.604
That would be a 50%
probability of success.

00:31:09.784 --> 00:31:14.374
That's pretty good data by the way,
depending on the, the, the disease.

00:31:14.584 --> 00:31:19.054
If this were an Alzheimer's
trial, for example, or a acute

00:31:19.054 --> 00:31:22.654
stroke trial, and you've got a
50 50 shot of winning the trial.

00:31:23.599 --> 00:31:25.459
You might want to invest in that trial.

00:31:25.459 --> 00:31:30.679
That's a really good particular
gamble in that scenario, uh,

00:31:30.679 --> 00:31:34.939
because the, the outcome is
such a positive outcome for you.

00:31:35.149 --> 00:31:40.909
The, the revenue to reap in that scenario
is, is positive enough to win the trial.

00:31:41.959 --> 00:31:47.209
Now that predictive probability of
success sometimes gets confused for power.

00:31:48.769 --> 00:31:54.229
Power is not something that changes
based on the data that you observed.

00:31:54.499 --> 00:31:58.609
Power is something that if the
treatment effect is Delta, here's

00:31:58.609 --> 00:32:00.049
the chance we win the trial.

00:32:00.499 --> 00:32:03.049
The trial's gonna be 80 or 90% power.

00:32:03.079 --> 00:32:05.419
Every trial is 80 or 90% power.

00:32:05.539 --> 00:32:07.969
You just figure out what
Delta makes it powered.

00:32:08.779 --> 00:32:13.669
So there are times where they say,
oh, to invest, we want it to be 90%.

00:32:14.674 --> 00:32:17.554
That's a really, really
high predictive probability.

00:32:17.554 --> 00:32:19.054
Think of that sports game.

00:32:19.264 --> 00:32:22.714
So you, this is, this
is exactly the scenario.

00:32:22.714 --> 00:32:23.374
You're sitting there.

00:32:23.374 --> 00:32:29.014
You could make a gamble on team
A or team B, and at halftime,

00:32:29.014 --> 00:32:31.414
a team has a 90% chance to win.

00:32:31.834 --> 00:32:36.304
If, if you're investing a and
you get a double A at the end,

00:32:36.304 --> 00:32:41.194
of course you're gonna invest in
that, uh, uh, in that scenario Now.

00:32:41.974 --> 00:32:46.504
The, the people that the group that
you're investing in understands

00:32:46.504 --> 00:32:50.344
the odds and the the payout of
that is going to be adjusted by it.

00:32:50.344 --> 00:32:51.514
At the time you make that.

00:32:52.804 --> 00:32:53.974
Now this is different.

00:32:53.974 --> 00:32:55.954
You get to make that investment on it.

00:32:55.984 --> 00:33:00.514
Now if you say, sure, I'll invest
it 90%, it's pretty unlikely.

00:33:00.754 --> 00:33:04.864
Well, I, it may be unlikely that
the company can hit that early on.

00:33:04.894 --> 00:33:10.474
It, but maybe they can hit 50% or 30%,
maybe that's where I would jump in.

00:33:10.804 --> 00:33:15.934
You can go back to thinking about,
suppose I had a separate phase two trial.

00:33:16.144 --> 00:33:17.824
When would I invest?

00:33:18.244 --> 00:33:21.814
What effect size would I need
to see before I'd be willing

00:33:21.814 --> 00:33:23.344
to invest in phase three?

00:33:24.544 --> 00:33:29.224
That can all be incorporated within
this administrative financial analysis.

00:33:29.224 --> 00:33:34.894
Is, and again, the enemy of that is,
I can't tell you when I would invest.

00:33:34.894 --> 00:33:37.714
You have to show me the data now.

00:33:38.899 --> 00:33:40.669
Many scenarios.

00:33:40.819 --> 00:33:45.199
This is kind of a weird concept, this
predictive probability of success

00:33:45.199 --> 00:33:48.949
in the trial to the people who are
so used to building a trial where

00:33:48.949 --> 00:33:53.209
they're used to talking about power
and type one error and all of that.

00:33:53.209 --> 00:33:55.009
This is all of a sudden
a little bit different.

00:33:55.159 --> 00:33:58.489
They may be familiar with conditional
power and they may be familiar

00:33:58.489 --> 00:34:02.269
with futility analyses, uh, but
this predictive probability of

00:34:02.269 --> 00:34:03.619
success is a little bit of a.

00:34:03.839 --> 00:34:06.839
Different sort of beast in the trial.

00:34:06.989 --> 00:34:11.069
So sometimes they're concerned that
funders might not understand that.

00:34:11.849 --> 00:34:14.519
The interesting thing is funders
understand that immediately.

00:34:14.519 --> 00:34:16.229
That's the world of funding.

00:34:17.399 --> 00:34:18.869
What do I invest?

00:34:19.079 --> 00:34:22.884
What do I make, and what's
the likelihood that I make it?

00:34:23.764 --> 00:34:26.434
And that can evaluate that investment.

00:34:26.464 --> 00:34:28.174
Incredibly straightforward.

00:34:28.174 --> 00:34:31.144
So this is a powerful thing to funders.

00:34:31.924 --> 00:34:38.944
Now, my why might a, a, a funder, uh,
agree to this objective criteria in it?

00:34:39.964 --> 00:34:46.264
Well, let's go back to the beginning
of this, is if they say No, run

00:34:46.264 --> 00:34:50.254
a separate phase two and show me
the data, and then I might invest.

00:34:50.869 --> 00:34:56.929
They're investing something that's
45 months to the readout of potential

00:34:56.929 --> 00:35:00.949
success and it's a bigger investment.

00:35:01.399 --> 00:35:07.669
270 patients will, will is
the investment and it's 45

00:35:07.669 --> 00:35:09.799
months as opposed to 25 months.

00:35:09.799 --> 00:35:12.589
In a smaller investment, we get it sooner.

00:35:13.279 --> 00:35:14.869
It's a win, win, win.

00:35:14.869 --> 00:35:18.794
If we can agree to this potential
funding, it's a win for.

00:35:19.594 --> 00:35:21.214
The pharmaceutical company.

00:35:21.514 --> 00:35:23.224
It's a win for the drug.

00:35:23.224 --> 00:35:24.634
It's a win for investors.

00:35:24.634 --> 00:35:25.714
It's a win for patients.

00:35:25.714 --> 00:35:30.304
It's a win for everybody in this
agreement if that can be agreed upon.

00:35:30.874 --> 00:35:32.944
Now it comes back to this.

00:35:33.214 --> 00:35:34.564
I know it when I see it.

00:35:34.564 --> 00:35:37.654
Can we define when we
would invest in that?

00:35:37.654 --> 00:35:41.554
It might even be a function of, if
we hit this higher hurdle, we'll

00:35:41.554 --> 00:35:43.294
invest, and here's the outcome.

00:35:43.294 --> 00:35:46.804
If we hit the lower hurdle, maybe
they get a higher fraction of the

00:35:46.804 --> 00:35:48.544
outcome because it's a bigger risk.

00:35:49.594 --> 00:35:54.424
I, I don't live in that world, but you
can imagine fairly straightforwardly

00:35:54.424 --> 00:35:57.124
this being a part of this agreement.

00:35:58.084 --> 00:36:00.364
This is a very common situation.

00:36:00.364 --> 00:36:05.974
Many groups are in this, even big
pharma needs to get big pharma.

00:36:06.859 --> 00:36:12.139
To agree to it and maybe large pharma
doesn't want to say we're gonna

00:36:12.139 --> 00:36:14.239
invest in a 1500 patient trial.

00:36:14.239 --> 00:36:18.769
We haven't seen enough individually
to, uh, at, at the corporate level

00:36:18.769 --> 00:36:20.659
to say, we're gonna invest in 1500.

00:36:21.049 --> 00:36:25.819
Run a phase two trial, well turn it into
a phase three if you're ready for that.

00:36:25.819 --> 00:36:29.449
Or a seamless two, three where
you have a funding go, no,

00:36:29.449 --> 00:36:30.919
go internally in the company.

00:36:30.919 --> 00:36:32.929
That's gotta be much easier to do.

00:36:33.529 --> 00:36:35.539
And again, it's a win-win, win.

00:36:35.974 --> 00:36:40.534
For the pharmaceutical company in
this scenario, if you can define

00:36:40.534 --> 00:36:44.434
it and create it, it's a very
straightforward thing to do in the trial.

00:36:45.454 --> 00:36:47.974
Important part in the
innovation part of it.

00:36:49.219 --> 00:36:54.049
A little bit of a warning here is
some of the terminology in this

00:36:54.049 --> 00:37:00.799
that you'll hear, uh, companies talk
about a TPP target product profile.

00:37:01.309 --> 00:37:06.349
They're going into it where they know
to be a competitive in the market.

00:37:06.589 --> 00:37:11.149
They need to see a 20%
improvement, a 10% improvement.

00:37:11.149 --> 00:37:15.139
What, whatever that 50% improvement,
whatever that is, that's what

00:37:15.139 --> 00:37:17.089
we're targeting for that product.

00:37:18.649 --> 00:37:22.309
I have to tell you, lots of people
talk about that and they all

00:37:22.309 --> 00:37:23.959
interpret it very differently.

00:37:25.609 --> 00:37:28.309
Sometimes it's the delta.

00:37:28.309 --> 00:37:37.669
I want to power my trial to win on that,
and so if the effect is 30% difference,

00:37:38.569 --> 00:37:43.579
I want the trial to be 90% powered for
that 30% difference, but you could win.

00:37:43.579 --> 00:37:47.089
If you see 20, about two thirds of
that effect, you can win the trial.

00:37:48.094 --> 00:37:53.434
Some people call the TPP, that
20, that when the trial's over,

00:37:53.434 --> 00:37:57.694
we need to be that value or above
to be competitive in the market.

00:37:58.504 --> 00:38:02.284
Those are two very, very different
things, and then they're two very,

00:38:02.284 --> 00:38:05.884
very different things for funding.

00:38:06.694 --> 00:38:10.534
So if 30% is what you're
powering the trial for.

00:38:11.209 --> 00:38:14.899
And at the interim you say,
we need to see 30% or above.

00:38:14.929 --> 00:38:16.759
It's disconnected with the trial.

00:38:16.789 --> 00:38:19.999
'cause you can win at 20,
halfway through the trial.

00:38:19.999 --> 00:38:25.249
If you see 30 and it's power,
you see 29, that's awesome.

00:38:25.909 --> 00:38:28.279
You are likely to win the trial again.

00:38:28.459 --> 00:38:32.149
Your probability of success
is probably hovering over 90%

00:38:32.149 --> 00:38:33.469
at that particular time point.

00:38:34.819 --> 00:38:38.824
Now, if somebody says, oh, but you're
not meeting TPP, that's the part where.

00:38:39.544 --> 00:38:43.354
The terminology gets in the way
of good communication for that.

00:38:43.954 --> 00:38:49.384
So one of the ways to combat this,
I'll know it when I see it, is

00:38:49.384 --> 00:38:56.194
to show it to them show them a
dataset and say, if we have this

00:38:56.194 --> 00:38:58.684
data in phase II would you fund us?

00:39:00.094 --> 00:39:07.984
Because it's hard for, for, uh, lots
of people, non-statisticians To think

00:39:08.019 --> 00:39:11.884
about the quantitative aspects of what
is a predictive probability of success?

00:39:11.884 --> 00:39:12.754
What does this mean?

00:39:12.754 --> 00:39:13.474
What does that mean?

00:39:13.834 --> 00:39:18.634
But if I can see the data, if I can
see the effect size that's meaningful

00:39:18.634 --> 00:39:23.884
to me, clinically meaningful, so we
show them a dataset and say, would you

00:39:23.884 --> 00:39:25.984
fund it if we showed you this data set?

00:39:26.014 --> 00:39:28.594
Sometimes they say, you know,
I also need to see this.

00:39:28.624 --> 00:39:30.004
Okay, we'll show you that as well.

00:39:30.004 --> 00:39:30.724
Would you fund it?

00:39:31.084 --> 00:39:33.364
Say, you know, if we see
this data, we would fund it.

00:39:34.564 --> 00:39:34.924
Okay.

00:39:34.924 --> 00:39:36.004
How about this dataset?

00:39:36.004 --> 00:39:37.414
Would you fund this dataset?

00:39:38.404 --> 00:39:43.144
And we show them example
datasets Kaplan-Meier plots

00:39:43.144 --> 00:39:44.974
of what data might look like.

00:39:45.334 --> 00:39:50.704
Uh, estimated hazard ratio, difference
in proportions, mean changes.

00:39:51.784 --> 00:39:54.874
We show them these data and
then you show them thousands.

00:39:54.874 --> 00:39:58.109
You show them scatter plots
of the various effect sizes.

00:39:58.789 --> 00:40:02.059
And then you say, by the way, this
point right here that you would fund

00:40:02.059 --> 00:40:08.239
the predictive probability of success
is 82%, or it's 63%, or it's this,

00:40:08.329 --> 00:40:10.219
and now here's this other data set.

00:40:10.219 --> 00:40:12.169
And you say, here's the predictive prob.

00:40:12.379 --> 00:40:15.679
And they get to understand
it it's meaningful to them.

00:40:16.789 --> 00:40:22.039
I know it when I see it through that
quantity because it's now familiar to me.

00:40:23.794 --> 00:40:28.414
So while that initially can be thought
of as this is the enemy, that person

00:40:28.894 --> 00:40:32.854
is somebody that you can work with
in simulating this, showing this

00:40:32.854 --> 00:40:37.744
example, showing example data sets,
tying it to a particular criteria,

00:40:37.894 --> 00:40:39.754
sometimes it changes the criteria.

00:40:40.264 --> 00:40:43.684
Let's just do the observed
effect at that time in that.

00:40:45.559 --> 00:40:49.339
Many times it can iterate and a lot
of times, by the way, you're making

00:40:49.339 --> 00:40:53.149
this decision in a seamless two
three trial where you don't have

00:40:53.149 --> 00:40:55.189
complete data on the phase two part.

00:40:56.629 --> 00:40:57.409
That's okay.

00:40:58.039 --> 00:41:01.609
You might be building
longitudinal modeling of this.

00:41:01.789 --> 00:41:08.119
You might over enroll phase two a little
bit to get enough completers because

00:41:08.119 --> 00:41:10.069
those patients count in phase three.

00:41:10.159 --> 00:41:11.449
They're not wasted.

00:41:12.409 --> 00:41:17.059
Maybe if you are enrolling one to
one to one, and you over enroll the

00:41:17.449 --> 00:41:23.419
three, do the three arms, two of those
arms would continue in phase three.

00:41:24.244 --> 00:41:29.134
So if you don't continue the trial and
you stop, they've been over-enrolled.

00:41:29.494 --> 00:41:33.304
But if you continue, only a third of
those patients have been over-enrolled.

00:41:33.304 --> 00:41:36.124
The other ones contribute to the
end of the phase three analysis.

00:41:36.664 --> 00:41:40.564
These are all the things that
go into these types of trials

00:41:40.564 --> 00:41:42.244
and these types of comparisons.

00:41:42.604 --> 00:41:46.084
In order to tackle what is a
ubiquitous problem, you've, you

00:41:46.084 --> 00:41:47.524
have to be able to fund the trial.

00:41:51.094 --> 00:41:58.114
All right, so I, I, I hope you've enjoyed
the discussion of funding Jumping Hurdles,

00:41:58.444 --> 00:42:01.594
combating the I'll know it when I see it.

00:42:02.614 --> 00:42:04.654
And thank you for joining me.

00:42:05.494 --> 00:42:09.214
Until next time, we'll
be here in the interim.