Subscribe
Share
Share
Embed
In this episode of "In the Interim…", Dr. Scott Berry explores the origins of REMAP-CAP with Prof. Steve Webb, former chair of the REMAP-CAP International Trial Steering Committee. This episode examines how pandemic preparedness efforts after 2009 H1N1 shaped the design of an international, adaptive platform trial to be able to respond rapidly to new infectious threats. Steve and Scott explain the sequence of strategy meetings, the role of the PREPARE consortium in securing EU funding and subsequent federation across Australia and Canada. The discussion details REMAP-CAP’s technical foundations: a modular master protocol, domain architecture, Bayesian adaptive methods, and frequent interim analyses. When COVID-19 emerged, these core elements permitted immediate platform activation to combat the pandemic infection with assessment of treatments across multiple domains—including steroids, immune modulation, and anticoagulation—generating actionable evidence in weeks. The episode also addresses international data harmonization, multi-platform trial collaboration, and the capacity to adapt trial structure as infectious disease threats evolve.
Key Highlights
Key Highlights
- Response to H1N1 and feckless pandemic trials
- International strategy meetings—origins of platform concept
- PREPARE consortium and cross-continental funding
- Modular master protocol, factorial allocation, and domain-specific appendices
- Bayesian triggers and response adaptive randomization
- Pivot to COVID-19 and rapid data generation
- Multi-platform international collaboration
For more, visit us at https://www.berryconsultants.com/
Creators and Guests
What is In the Interim...?
A podcast on statistical science and clinical trials.
Explore the intricacies of Bayesian statistics and adaptive clinical trials. Uncover methods that push beyond conventional paradigms, ushering in data-driven insights that enhance trial outcomes while ensuring safety and efficacy. Join us as we dive into complex medical challenges and regulatory landscapes, offering innovative solutions tailored for pharma pioneers. Featuring expertise from industry leaders, each episode is crafted to provide clarity, foster debate, and challenge mainstream perspectives, ensuring you remain at the forefront of clinical trial excellence.
Judith: Welcome to Berry's In the
Interim podcast, where we explore the
cutting edge of innovative clinical
trial design for the pharmaceutical and
medical industries, and so much more.
Let's dive in.
Scott: Welcome everybody
back to In The Interim.
I'm your host, Scott Berry, and I've got
a, a, a wonderful topic today and I was, I
was chatting with Steve just before we hit
go here on, th- we're episode, I believe
episode 65 of In The Interim, and I
think, uh, at, you know, spending t- I've
been doing clinical trials for 27 years.
If I were to stop tomorrow and
somebody said, "Scott, what was
your biggest accomplishment?"
It would be Remap-Cap.
Remap-Cap to me will be, you know,
one of the incredible things in
my life, the experience of it.
It in of itself as a scientific
beast, the impact it's had, the,
the so many aspects of it, the
friendships that have grown out of it.
So it's, it's the most impactful and
here we are, episode 65, and this is the
first episode on the Remap-Cap trial.
Part of it is it's such a
beast to get your arms around.
So I, I'll introduce and
we'll, we'll talk about this.
This is going to be the
origins of the Remap-Cap trial.
In and of itself is such a cool
story of it, and my guest today
was there, creation of this.
Also a very good friend,
uh, uh, of mine, Steve Webb.
Steve is a professor, uh, is
intensive care specialist and
a professor of critical care
research at Monash University.
He's also the executive
director of Empiric, which
is a whole separate podcast.
Uh, interestingly, you know I do a lot
of sports things and, and in sports they
talk about a first ballot Hall of Famer.
Steve, in 2014, the Australian
government estabuli- established the
Australian Academy of Health and Medical
Sciences, and he was part of the 14
fellows that initiated the academy.
So very well-respected researcher.
He's, his clinical trials he's
designed have enrolled more than 70,000
patients, 200 million in funding,
more than 30 papers in The New England
Journal and JAMA, more than 75,000
references, huge impact on patients.
He's also one of the global leaders of
the Remap-Cap platform trial and was
the head of the International Trial
Steering Committee when COVID hit, and
that'll be a big part of our story.
So Steve, welcome to In The Interim
Steve Webb: Thank you very much, Scott.
It's a, pleasure to be
chatting about The Beast
Scott: Yes, the beast.
All right.
So I was curious going back and, and
remembering this, and I got in a bit of
a, of, of a rabbit hole, uh, with this
and looking back at all the emails.
I wanna know, when was the first emailâ¦
I have all my emails ever from
Berry Consultants, which is, is
fun, fun things going back to.
The first email that ever
showed up with Steve Webb on it.
So I look back, and this was, uh,
March 13th, 2012, is the first email
where you and I are both on the
email, and it's from Derek Angus.
It's a follow-up to a Pittsburgh
meeting that is talking
about potential p- pandemic.
Sorry, I wasn't there, but, uh,
the aftermath of this came about.
Do you wanna describe this meeting and
the subsequent meetings and pandemics?
Steve Webb: So the story of Remap Cap
really starts in the southern hemisphere
winter of 2009, when the swine flu
pandemic, uh, spread around the world.
to Australia, New Zealand, other parts
of the southern hemisphere, uh, first,
and it, it was a decent pandemic.
Um, uh, in the Royal Perth Intensive
Care Unit, where I was a, a specialist,
uh, at the peak of the pandemic,
half of the intensive care unit
was occupied by patients with swine
flu, all of whom were ventilated,
all of whom, uh, were very sick.
Um, but, uh, swine flu, um, came and went,
um, uh, uh, over the space, uh, of two
winters, one southern, one, one northern.
And, uh, um, the world sort of, um, took
a breath and said, "Well, that was bad,
but it could have been a lot worse."
And so a bunch of people got together
from the intensive care world to
think about how could we be better
prepared for future pandemics.
Um, in Australia, New Zealand, uh, we did
some observational, um, uh, work, which
was, uh, quite valuable from a public
health perspective, we had no chance
of putting together a clinical trial
in the time span that was available.
A wave would hit, it would last about
eight weeks, um, and you just can't launch
trials, um, uh, at least, uh, at that
time, over that sort of, uh, timeframe.
was a, a meeting in 2010 in Toronto,
uh, where Derek Angus, uh, who was
also, uh, um, one of the, um, together
with you, me, and Roger, probably
the midwives, uh, of this thing.
Um, and your dad spoke at it, Scott.
Um,
Scott: Yep.
Yep
Steve Webb: talked about adaptive platform
trial methods, and it made a lot of sense.
Although I've got to say, I had to
listen to talks from Don, from you,
from Roger, at least six times before
I, um, finally thought, "I-- Now
I understand this, uh, this thing
called an adaptive platform trial."
Um, and then the 2012 meeting that
you referred to was in Pittsburgh,
and Roger, uh, Roger Lewis was there.
Um, and, um, uh, that was really, a
series of meetings about, you know, nine
months apart, we decided that adaptive
platform trials were the only thing
that made sense for future pandemics.
Because they create, um, um, warm
infrastructure can pivot and adapt rapidly
to whatever is needed for a disease
that can't be predicted in advance.
Scott: Uh, that, that, that
I think comes to fruition.
So there's these d- lots of
researchers meeting, ISARIC.
Uh, John Marshall was at a number
of these meetings, who ends up
being a, a, a, a very important
player in the REMAP-CAP story.
And I, I think myâ¦
I enter this in, uh, on July 2nd, 2012.
There's an ISARIC meeting
in Annecy, France.
I believe you were there.
Yeah.
So I believe it's probably the
first time we met, uh, in that,
and I presented on adaptive trials,
and it was the first sort of what
could this look like in a pandemic.
There was a two-day working group from
that, and I think from that largely
grew the pr- grew the PREPARE network
Steve Webb: Co-correct.
Herman Goossens, uh, was a, um, very, um,
inspirational and influential infectious
diseases physician based in Belgium.
uh, he understood the damage that
pandemics could, uh, do to the world if
one turned up, and the critical need of
being able to generate evidence rapidly
about, uh, um, optimal treatment in
a situation where, um, the clinical
trial, uh, design and infrastructure
needs to be constructed, um, in the
presence of enormous uncertainty.
And so Herman, uh, um, uh, um, uh,
proposed, uh, to the European Union
that there would be a funding call for a
consortium around pandemic preparedness.
And it was a fantastic program of work.
There was lots of laboratory science.
There was stuff about, uh, diagnostics.
There was stuff about consumer and
community engagement, about how you
prepare the community for a future
pandemic in terms of the research that
needs to be, uh, conducted at the time.
And one of the work packages, uh,
work package five, as I recall,
was an adaptive platform trial, uh,
for, um, any future, uh, pandemic
Scott: Uh, yep.
Uh, and, and that's sort of where, where
I, I sort of really remember starting
on this, and it was called Prepare.
And we were talking about at the
time, I think we were talking
about a platform for antibiotics.
It, it was relatively generic treatment,
but we, we hadn't jumped into the,
the, the multifactorial, the huge
aspect, which is really interesting
part of the pandemic was the abil-
was the factor- multifactorial story
of that, but we'll get to that.
But it
Steve Webb: Yeah, no, I,
think,
Scott: it wasâ¦
Steve Webb: I, I
Scott: Yeah
Steve Webb: I think it, I think it
was multifactorial from the get-go.
But part of, um, uh, the discussion and
the concept at the time that the only way
to be ready for a, a, a pandemic was to
have that were already recruiting to other
questions that were-- that are relevant,
uh, to improving patient outcomes.
and so part of the concept was to
construct this, uh, platform that
would, uh, do good things, uh, during
the inter-pandemic, uh, period.
And, uh, um, patients with pandemic
influenza, or as it turned out, pandemic
COVID, have, um, essentially a diagnosis
of severe acute respiratory illness
or community-acquired pneumonia.
so the plan was to construct a platform
that would recruit patients, uh, with
community-acquired pneumonia, ask
questions that were relevant to improving,
uh, patient outcomes, because CAP is
a, um, a major driver of, um, um, uh,
bad outcomes, um, and be able to pivot,
uh, when a pandemic, uh, turned up.
So having succeeded in getting that grant
call, uh, included within, um, what was
then the, uh, FP7 scheme, subsequently
has become the Horizon, uh, scheme.
Um, the, um, uh, um, put together a
really impressive, uh, uh, consortia.
Um, uh, Derek, myself, you, all
contributed to, uh, the design
of what was included within work
package fee-- uh, work package five.
it
Scott: And
Steve Webb: uh, I don't know,
it was maybe a page and a half
in this, uh, massive grant
Scott: Yeah.
Steve Webb: And, um, uh, after
the wheels turned inside the
European Union, e- I can't even
remember exactly what year is it.
Might have been 2015 or 2016 now.
grant is, uh, successful, and we're away
Scott: Yeah, and interestingly, as part
of this story, the federation of this,
this trial grows into something that has
multiple sponsors globally and a single,
uh, platform that, that has the federated,
and we, we talk a little bit about that.
But I found an email in December
of 2013 where you send out that you
now have funding from the NHMRC,
uh, to extend this to Australia
and, and what this would look like.
And this starts a little bit the
federation of what becomes Remap-Cap
Steve Webb: Yeah.
So that money was largely used to,
um, uh, pay for the preliminary work
that was necessary to write a grant
application to the NHMRC, uh, for what
was called various things back then.
Uh, it was-- Once upon a time, it
was GACSARI, then it was ADCAP,
and, uh, then it was OPTIMIZE-CAP.
but that NHMRC grant application,
uh, uh, was, was successful.
And the preceding FP7 grant, I think, made
a massive difference to the likelihood of
success of the Australian grant because
it had already been through peer review
in another location, though it was a
relatively small part of this big, massive
pandemic, uh, preparedness consortia.
But that gave it credibility,
and I think that likely helped,
uh, get it over the line.
So that had us funded in two
countries, in, in two regions.
And
next to come on board was John Marshall,
who somewhere through that time period
was successful at getting, uh, it
funded, uh, through the CIHR in Canada.
And, um, really with Derek's
intercession, um, we settled
on the name of, uh, REMAP-CAP.
Uh, uh, De-Derek had had a, um, uh,
a viewpoint article in JAMA through
that area-- through that era, he
had, um, proposed the term REMAP
to describe, uh, you know, the, the
generic description of the beast.
A randomized embedded
adaptive, uh, platform trial.
And of course, it was also remapping
the way that clinical trial evidence,
uh, should be, uh, should be generated.
Uh, uh, Derek's, uh, uh,
well, he's, he's got-- Uh, uh,
Scott: Yeah.
Steve Webb: Derek's, Derek's
not short of marketing ability,
and it matches his substance.
Scott: Yeah, yeah, and
a, and a fabulous name.
So a little bit, you touched
on this a little bit.
Let's, let's talk about the sort of
scientific structure of this thing
that was created, which started
enrolling patients, I believe, in 2015.
The first patient was enrolled in this.
Uh, uh, so it was '15 or
'16, it was relatively early.
Uh, it-- we, we weren't, we didn't have a
super mature protocol, and we'll get into
the, the development of what I think is a
Tolkien-esque document in and of itself.
But the Prepare solution, you talked
about this, was to start a standing
master protocol where in the back,
in the backdrop to this, we called
it a sleeping strata, we would
talk about a potential pandemic.
So if there was a swine flu,
if there was Pandemic X, we
could immediately pivot to that.
But meanwhile, there's a t- a trial
going on, and I believe the first,
the first therapies being w- used
were five different antibiotics
as an antibiotic strategy, so a
patient could be randomized to that.
A question of extending macrolide to
four days or 10 days, so that was a
second question that a patient could
be randomized at in a factorial way, as
long as the antibiotic had macrolide.
Uh, and then was this strange beast,
hydrocortisone, um, uh, which I, I'm so
fascinated by corticosteroids and do they
work, do they not work, uh, question.
But, uh, patient-- the question was
should you give patients steroids,
uh, uh, severe patients, uh, in these?
Originally, I believe this
was somewhat severe patients.
This was ICU-based severe patients,
and these were the three questions that
were enrolling at the launch of this
Prepare trial that migrated in name
to Remap-Cap in a factorial experiment
Steve Webb: There was an initial
protocol, um, which, um, to the
best of my recollection, uh, did
get us, uh, recruiting in a small
number of locations that didn't
include Australia at that stage.
But I think it was apparent that, um,
the protocol that had been written
initially was, um, um, w-was a standalone
master protocol containing everything.
uh, if one looked as to how this is going
to evolve, adapt, new questions coming
in, um, there was a feeling that this sort
of, um, single protocol document wasn't
going to be, uh, effective over time.
And that led us to the development of the
modular protocol, which, uh, we've used,
I think, quite successfully ever since.
And the principle here is to have a
core or master protocol that contains
the eligibility criteria at platform
level and the primary endpoint that's
going to be used at platform level
across, uh, all of, uh, the domains.
Little bit about the disease of
interest in terms of background.
Quite a bit about the statistical
methods that are going to be used
in the platform, there's a separate
statistical analysis appendix.
But nothing about the, uh,
interventions that are going to be
evaluated, uh, within the platform.
So we had to invent a word.
We invented the word domain to
describe sets of mutually exclusive,
uh, interventions that would
be, um, uh, represent the, um,
comparisons that, uh, existed, around
specific, uh, um, modes of therapy.
And the protocol has this structure
of a master protocol onto which are
bolted, multiple domain-specific
appendices, and they contain the
information about, uh, the background
information about why that's a relevant
question and the, uh, appropriateness
of the interventions and what's known.
domain-specific eligibility, uh, criteria,
appropriate, domain-specific secondary
endpoints, my favorite section of every
REMAP-CAP protocol, the statistical, uh,
approach to this particular domain, which
is like selecting from a menu that is
set out in the master protocol, in the
core protocol, about what statistical
triggers, what combination of states and
strata apply to, uh, the, to this domain.
Um, so, um, uh, that modular protocol
structure I think has been a critically
important part of the success of
REMAP-CAP, it was a substantial advance,
and it proved its worth in the pandemic,
where with this modular structure,
could write a new domain-specific
appendix, do the associated data
management changes, and have a new
domain running in three or four weeks
Scott: Yeah, I'm, yeah, so, uh, it,
we- we'll get to the pandemic, but
I agree with you that the previous
document wasn't ready for that.
And I remember we had this sequence,
there were a number of players involved.
Colin McArthur was involved,
Alastair w- was involved.
But it was a sequence of about
a, a three weeks to a month
where you, Derek, and I wereâ¦
It was morning in the
US, it was your evening.
Derek and I would have coffee, and
you would have a glass of wine next
to you, and we'd be working on it.
The- then all of a sudden, you'd wake up
the next morning, and it would be night,
and Derek has a nice big, uh, glass
of wine, and you've got your coffee.
And it was morning and night, and
morning and night writing this.
And I refer to it almost as
Tolkien-esque because Tolkien, when
he wrote The Lord of the Rings and
all this, he invented languages for
the Elvish, he invented songs, and we
had to invent all this terminology.
What, what works for this
description of a common set of
interventions being compared?
And you can't reuse a word that means
something else in clinical trials because
then everybody confuses what you're doing.
And we spent a huge amount of this
inventing terminology that allowed this
modular thing, that you had to build this
entire protocol infrastructure so that
you could plug in a DSA and remove it,
and the protocol doesn't change at all.
I, I mean, it's built that you plug it
and pull it out, and it functions, and
as you say, incredibly quickly we're ans-
we're, we're addressing that question.
Uh, it's part of the database.
And so the, the master protocol
itself is this amazing thing that has
come out of Remap-Cap that, by the
way, many people have used, copied,
which is fantastic, modeled after.
It in and of itself is an incredible
sort of scientific achieve- achievement
Steve Webb: Um, and, um, uh, I've spent
a lot of time in the last couple of
years, talking to clinicians or even
clinical trialists who've got experience
with conventional clinical trial designs
about adaptive platform trials, and
it's the concept of the domain, which
is one of the most difficult concepts
to try and articulate and explain
in a way that, uh, people get it.
But I've got a new technique which is
working really well, which is that I
tell them, if the goal of the platform
trial is to try and design the best
possible Big Mac, then what you have is
the components of the Big Mac as domains.
So there's a pickle domain,
yes or no for pickles.
There's a beef patty domain.
Do- is it flame grilled or fried?
Um, and, uh, people suddenly get
the idea of what a domain is.
But we had to go to a
Big Mac to make progress.
Scott: And we then had to name the Big Mac
is when you get all of the pieces of the
domain that work, what do we call that?
And we called that a regimen.
So each piece together, the Big
Mac itself is another regimen.
We, we have patient subgroups
where there may be differential
effect, and so we named this and
named all these pieces to it.
S- so we went through this amazing thing.
We set up the statistics.
We, we talked about it.
We set up response adaptive randomization
can be done within the trial.
We set up the entire thing as Bayesian.
Uh, and interestingly, a foreshadowing
of, of medical papers that come
out of this have zero P values.
All Bayesian triggers in all of this.
So this is kind of all set up, and as
you d-- as you, as you've said, when
we had meetings maybe two thousand
seventeen, everybody would come together.
We've got this new protocol, um,
uh, in this Genevieve O'Neill, who's
not working on Remap CAP anymore.
She was on every one of these
calls, I think, in, in making
this go forward from Australia.
But that we had more investigators
than patients at the time.
Uh, we're enrolling CAP, and then
Steve Webb: there, there, there, there
was celebrations, there were fireworks,
uh, there was champagne popping
when we finally had more patients
enrolled than we had investigators
Scott: Yeah.
And by the way, we still, uh,
after the pandemic's over, we're
back to this question of CAP.
These patients are absolutely
informing conclusions, and we've
got exciting conclusions within CAP.
But meanwhile, um, a, a-- this Wuhan
disease comes up, and it was interesting
looking back at the, the genesis
of this from early 2020 in January.
There's this weird thing going on until
March when it became part of REMAP-CAP
Steve Webb: So one of my favorite
emails from, uh, Remap, uh, Scott,
is the one I sent to, um, Green,
who is the extraordinary, amazing
project manager that, um, uh, carried
Remap CAP through the pandemic.
Uh, Cameron sends me, um, news links
to this new disease in Wuhan, and he
copies, uh, it to Colin MacArthur, and,
uh, Cameron says, uh, "Do you think,
uh, we need to be, uh, worried about
what's happening in, uh, in, in Wuhan?"
Um, and, um, uh, my response to Cam
is, "Oh, you know, Cam, in the last
couple of years there have been
these sort of little mini outbreaks.
They, um, uh, cause a little bit
of excitement and then they, uh,
disappear, uh, reasonably quickly.
That's what I expect will
happen, uh, to this one.
besides, um, no pandemic in 2020 please.
We're close but we're not ready."
Scott: Yes, yes.
I, I, amazingly, we've said many
times, "Oh, we would've been so
much better off if it was COVID,
uh, 20, uh, or one year later."
Now, we were never gonna be ready for what
it was, uh, and it was going to be around
the clock, but, but it was, uh, it wasâ¦
I- Iâ¦
And then it hit.
And one of my other favorite emails was,
uh, right about that time, we were now
starting to get ready to run this trial.
We had enrolled a set of patients.
We were talking about doing analyses.
Um, and I think we had done several
analyses up until this point.
So I brought on two statisticians
to help me, uh, Lindsey Berry and
Elizabeth Lorenzi, Liz Lorenzi.
And, um, they have this email
of saying, "Yeah, we've got
this little thing in Wuhan, but,
you know, don't worry about it.
Um, uh, it's not a big deal."
And they look back at this, and
now they've, they've been heroic in
the, the analyses that have happened
within Remap Cap as well, and this
has now become, uh, uh, an amazing
story of their involvement as well.
Okay, there's an i-
Steve Webb: don't know what-- I, I don't
know how much you want to talk about the
pandemic, but, um, I think it's worth
just w- at least one anecdote, uh, Scott.
The, the speed with which the platform
was able to do things was extraordinary.
We talked about being able to launch
a new domain in three or four weeks.
although there were times during the
pandemic when even that felt too slow.
But
Scott: Yeah.
Yeah
Steve Webb: able to launch a new trial
for, for, um, a condition that's never
existed for, um, interventions, for new
set of interventions in three or four
weeks was, was, um, uh, was unprecedented.
But, um, we had this
result in November of 2020
Scott: no, let's hang on.
Let's not, l- let, let's sort
of set this up a little bit.
I do wanna shift completely
to the pandemic now, so let,
let's talk about the pandemic.
Let's talk about the, the start of it,
and I've got an email from you, you're the
chair of the ITSC, where you met with, uh,
uh, groups, uh, monitoring the pandemic.
There are patients starting to
trickle in at sites where Remap-Cap
is a trial there, and I've got it on
March 2nd, 2020, that you declare,
"We are open for the pandemic,"
Steve Webb: We had
Scott: at this time
Steve Webb: we had a sleeping strata
was in the first version of the
modular protocol, and the question
to the ITSC was should we activate,
uh, this, uh, sleeping, uh, strata?
there was a, um, a unanimity
within the ITSC that that was
exactly what we needed to do
Scott: So now this trial that had been
built, been, as you described, all the way
back in 2009, uh, I'm, I'm sure pandemic
preparedness has gone back further than
that, but this whole idea of a platform
trial being generated from that, getting
funded by the EU, moving on to the
different Australia thing, this idea,
all of a sudden, March 2nd, 2020, the
trial is now enrolling in the pandemic.
So we've avoided this whole
issue of, okay, now we have to
design a trial for this question.
We're already enrolling.
By the way, we have
patients coming into this.
And so the things that were
already in the trial, some of
these shift over into COVID.
Uh, all three of them shift over.
I think antibiotics became a not very
interesting question within them, but
macrolides became very interesting.
Steroids became s- s- very interesting.
Now, this is the part that you
sort of touched on, is now the
ability within the modular approach
to start to add questions to it.
So now we're open for business.
We already have therapies we're, we're,
we're interested in, we're shifted.
Now, what does this process look like?
Steve Webb: So the ITSC establishes, um,
processes to consider what interventions
we evaluate in this, uh, new disease.
We're learning about this news-- new
disease in a very, uh, ad hoc way.
um, um, uh, uh, my recollection,
um, quite a bit in the p- in the,
in the general media about all of
these patients ending up in ICUs on
ventilators, of, um, arterial and venous
thromboembolic, uh, complications.
There's actually relatively little
published, even in terms of, um, uh, case
reports and, uh, cohort, uh, studies.
Um, uh, a-and so we have to
start making decisions about
what's going to be evaluated.
Um, we have an existing
corticosteroid do-domain, which
includes, uh, hydrocortisone.
We make some modifications to
that domain, including, um, a d- a
different dose of, uh, hydrocortisone.
And we actually managed to recruit the
first COVID patient into a COVID-specific
domain for corticosteroids, I think
on March the 6th, which is before the
WHO officially declared a pandemic.
So that I think reflects the, uh, the
flexibility that the adaptive platform
trial provided in this, uh, scenario.
But we think about what else
should be evaluated, relatively
quickly, we put together domains to
evaluate, um, uh, drugs which are
thought to have antiviral activity
against the, the SARS-CoV-2 virus.
Um, there's a domain proposed,
uh, that, uh, comprises, uh,
different immune modulators.
interleukin-6 receptor antagonists like
tocilizumab, uh, anakinra, uh, interferon.
there's also a question of whether
or not patients should be routinely
therapeutically anticoagulated, uh,
with heparin, and that had, um, um,
entered practice just seen as being, um,
appropriate to prevent arterial and venous
thromboembolism, given the high incidence
that was, uh, was being observed.
Sorry.
Excuse me.
Um, um, uh, there's
also
Scott: so let's
Steve Webb: of
Scott: let's set time to this.
I believe, so March 2nd, uh, uh, REMAP-CAP
says we're now enrolling patients
COVID, suspected or proven COVID.
March 6th, four days later, you're
randomizing to something that
was already there, the steroids.
Added the, uh, immune modulation
domain before the end of March.
The convalescent plasma domain,
I think, was added in April.
Anticoagulation, as you described,
is early May, uh, within that.
So I believe this, this idea that you
said every couple weeks we're adding
them, and to some extent that felt
slow, but we're doing this and adding
those, uh, before just as May starts,
we've added all these particular
questions in a multifactorial thing.
So some patients are being
randomized to four and five
questions at this point, which is
Steve Webb: Y-y-yes, I think that,
Scott: just, amazing
Steve Webb: I think at the height
of the pandemic, um, um, the,
the, the largest number of, um,
interventions that were assigned to
a patient in Remap-Cap was seven.
Scott: Mm-hmm.
Mm-hmm.
Steve Webb: that's m-m-that's
multi-multi-factorial
Scott: Yeah.
Yes, yes.
Okay, so now it's we're, we're,
we're, we're getting set up.
And meanwhile, uh, as you
describe, it would've been nice
had this been another year.
We're getting the
operational part of this.
We have multiple databases
that are contributing to this.
We want to be doing interim analyses
very frequently in this trial.
So it is a great deal of work by a,
a, a huge number of people to get
this ready to start doing analyses.
And then mid-June happens.
Yep
Steve Webb: many, many, many people
literally working 14, 16-hour days,
seven days a week across many countries.
Um,
Scott: ãã
Steve Webb: Green, um, Colin, uh,
the, the, uh, Lisa Higgins, who was
putting the data together at Monash
at that stage, the teams at Berry
Scott: Yep.
Yep.
Uh, the UK enrolling patients,
Tony Gordon, Lenny Deirdre,
Steve Webb: Yes
Scott: all of this, uh, amazing.
So mid-June happens where, uh,
we've done interim analyses, but
nothing is triggered at this point.
We've got triggers set up.
Mid-June, the RECOVERY trial, which
by the way, in and of itself is an
amazing story of the RECOVERY trial.
We're-- REMAP-CAP's
largely enrolling severe.
Meanwhile, we, during I think July,
we open what's called a moderate
state, uh, a step different.
So we're adding to the patient
population in a modular way as well,
but we're enrolling only severe
at this point, defined as being on
intensive care-based organ support.
Uh, and we went through a whole bunch
of things at this point is what is
the definition of an ICU, uh, because
people are using non-ICU things, and
we, we had to lose the notion of what,
whether you're located in an ICU or
do you have I- uh, ICU-level care.
So there were things we had to redefine
because of the surges and all of that.
But the RECOVERY trial comes
out with an announcement based
largely on moderate patients that
steroids are, uh, saving lives.
And so REMAP-CAP is enrolling steroids
largely in a severe population.
So REMAP-CAP does what at this point?
Steve Webb: Um, so I just want
to, uh, uh, comment about, uh,
Scott: Yep
Steve Webb: amazing trial Recovery was.
It was put together
incredibly, uh, quickly.
It, it was in some ways also an
offshoot of the same group of people
who'd been working on, um, um, uh,
Prepare, uh, with, with, with Herman,
uh, Herman Goosens, and as you said,
was predominantly focusing on, um,
uh, ward patients, whereas Remap-Cap
was exclusively, at least initially,
uh, uh, critically ill patients.
So, as soon as the Recovery result,
uh, um, uh, comes out, the, um,
corticosteroid component of Remap-Cap,
uh, in patients with COVID, um, uh,
ceases recruitment, and we start
this huge amount of work to try and
get, uh, our results, um, analyzed.
Um, only really-- O-one thing that all
clinical trialists, um, um, intuitively,
um, know deep in their bones um, there is
nothing better when your trial has, um,
uh, reached a certain conclusion, another
trial reach the same, uh, conclusion.
Your anxiety levels drop,
uh, um, uh, enormously.
Um, so, um, the WHO put together a
process to try and do an individual
patient meta-analysis across five
or six trials that had been, uh,
recruiting patients, uh, to steroids.
And, um, that all happened,
uh, at light speed.
Um, the Recovery result, um, was, um,
uh, released, um, uh, changed global
practice instantly, um, in June.
there was a, a manuscript, I think,
um, I'm not sure precisely when it was
through, uh, July or August, and but
it was September when, um, the WHO,
um, IPDMA was published together with
the primary manuscript of these other
steroid trials, including Remap-Cap.
And, um, uh, the Remap-Cap results
were strongly supportive of the result
that had come from, uh, Recovery, uh,
with five or six hundred patients.
Uh, my recollection, we were
sitting on a, uh, uh, posterior
probability of ninety-three percent,
uh, superiority for hydrocortisone
over, uh, no hydrocortisone.
we wouldn't have been too far away
from, uh, triggering with some,
uh, additional recruitment, but it
provided strong supportive evidence of
the benefit of this cheap, available
treatment for this dreadful disease.
Scott: Um, uh, it was another
really neat thing of the Bayesian
approach that at that time we
stopped because care had changed.
We couldn't give a control anymore.
And so we got to read this out, and it
was this 93% probability of superiority.
It wasn't that we're non-significant,
it was the, the-- in the paper, we
didn't need adjectives, moderate.
It was there's a 93% probability this
is beneficial, uh, which was such a
cool part and a simple thing and such
a straightforward scientific thing
about a clinical trial that was, was
a positive of the Bayesian thing.
Steve Webb: We, we, uh, Scott,
Scott: Now,
Steve Webb: two
Scott: yeah
Steve Webb: we, we need,
we need two T-shirts.
One is, "This is a p-value-free
zone," and the other is,
Scott: Yes
Steve Webb: treatment, we
estimate treatment effect."
Scott: Yes.
Oh, I thought you were gonna go to the
T-shirt that you wore much of 2020,
um, that I, I will always remember.
Uh, it had a picture,
it was emoticon-based.
It had a picture of a magnet and
then a pile of, uh, shit next to
it, uh, which was, which was quite
apropos for what was going on, yes.
Um, okay.
So at that time then, Remap
Cap adds the moderate state.
It, um, it's now triggering a number
of conclusions, uh, and we'll get to
some really important ones of those.
But it's also, um, cooperating
with other platform trials in a
multi-platform randomized trial.
So it's working with Operation Warp
Speed, um, and the trials, it's working
with a Canadian one where we all pool
our data together to have a single
conclusion and, and a result, and
we create a new thing for a trial.
Uh, and then we open up a new
domain, simvastatin is a- open,
antiplatelet domain is opened.
Um, and then we start coming in
with really interesting conclusions
at the end of 2020 into early 2021
Steve Webb: Yeah.
So, uh, um, one of those triggered
on Christmas Eve, Scott, and
Scott: Yep
Steve Webb: of us worked all the
way through Christmas Day, doing
the things that were necessary.
I thi- I, I, I, Iâ¦
That was one of the anticoagulation
results, but I can't remember,
uh, which one it was that, uh,
triggered on, uh, December 24
Scott: one was the, um,
uh, likely harm of severe.
Steve Webb: Yep.
Yep.
Scott: yep, yep
Steve Webb: But before, but before
that we'd had, um, again, a f- a,
a, a c- a, a flurry of activity in
about mid-November we'd triggered
for the, um, superiority of, uh, uh,
tocilizumab, an interleukin-6 receptor
antagonist, compared to, uh, control.
And I think the timeline is
something like as follows.
A week after, or it might have been,
uh, 10 days after the statistical
trigger, the then British Prime
Minister, Boris Johnson, announces
this result, uh, in a press release,
uh, or in, in a press conference.
Um, the following day, the
treatment becomes standard care
throughout the National Health
Service, uh, in the United Kingdom.
Uh, it's, it's, it's mid-November.
Well, no, uh, uh, sorry,
it's late, late November.
We got thatâ¦
I don't know what date we got
that manuscript to the New
England Journal of Medicine.
Uh, it, it-- I don't know if it
was in December or January, but it
was published in early February.
We're talking about eight or nine weeks
from statistical trigger to manuscript
that's been through peer review, it's
had revisions done, and it's published.
Uh, uh, it, it, the, it is, it is light
speed in the world of clinical trials
Scott: Yeah, uh, a- and likely
saved tens of thousands of lives.
Uh, uh, incredible therapy, uh, in it.
It, it, it's amazing now the spreadâ¦
Oh, so t- touching on, so tocilizumab
comes out, um, within that.
The anticoagulation being
harmful in severe, beneficial
in moderate, uh, comes out.
Common therapies given,
very impactful within this.
The-- what's interesting, something you
said, and I remember the, uh, uh, not a
panic, but we triggered the tocilizumab
result of superiority, and we had had
a surge in the UK of patients, and
so we had a good, solid result, but
there were like half, uh, half the
patients weren't even to 28 days yet.
And so we triggered this, and
the whole f- "What's gonna happen
when all of these came in?"
And of course, it came in almost the same.
Uh, incredibly beneficial is one of
this, you know, what's gonna happen.
You touched on this.
I think every result that REMAP-CAP
came up with, with its adaptive
design, its Bayesian, its RAR, its,
uh, its multifactorial, was backed
up by a result in a different trial.
Typically after that, I think all of these
results bore out w- with it, whether it
was antivirals being harmful for severe
patients, uh, h- hydroxychloroquine
within this, the benefit of steroids
within it, the benefits of the IL-6
receptor antagonist, the anticoagulation
result was backed up, the ACEs and ARBs.
Largely all of these things, vitamin C
largely doesn't seem to be a benefit.
I mean, it was amazing
how it kinda got it right.
You know, and I don't, I don't mean
to say in a doubting way, but oh my
God, it kinda got everything right
Steve Webb: Well, um, in, in two and
a half to three years, the platform
evaluated, fully tested repurposed
medicines for a disease that, um, um,
had never existed, uh, uh, before.
There, there are a couple of results.
I don't think the anticoagulation in
critically ill patients was ever, um,
uh, reproduced, uh, by another trial.
I don't think there were other trials,
or if there, if there were, they
were, um, uh, um, had insufficient
sample size to a-address the question.
Um, one
Scott: the multi-platform.
It was, it was the Canadian attack trial
and, and, yeah, the combining together.
Right, right.
Yep.
Yep.
Steve Webb: the, um, uh,
contributions from ACTA4 and,
Scott: Yep
Steve Webb: a-and, and uh, uh, CAP.
Attack, sorry.
Um, but, um, um, um, uh, Recovery
reported their antiplatelet, uh,
question, um, my recollection is
before we, uh, reported, uh, ours.
And they stopped having recruited about
15,000, uh, patients, uh, which is a,
you know, extraordinary, uh, sample size.
And, um, their results were, um, um,
no difference in outcome with 28-day
mortality as, uh, the primary endpoint.
we had this interesting result in,
uh, Remap-Cap, which is a rabbit
hole that we shouldn't go down, of
po-possibly there being non-proportional
treatment effect across the ordinal
scale, uh, that we were using.
But my recollection is that there was
something like a 93% or 95% probability
of s- of superiority for aspirin with
90-day mortality, uh, as the endpoint.
if we looked at the recovery
survival curves, the Kaplan-Meier
curves, were really only starting
to separate around, uh, uh, day 30.
So it's interesting how choice of
endpoints and the meaningfulness
of endpoints, um, can impact
on the appropriate, uh, or the
generation of evidence to, um, be
applied to improve patient care
Scott: Yeah.
Uh, another, uh, sort of amazing thing in
the trial and, um, with these 17 results
in COVID that you talked about is they
largely, most of them had the same design.
That we're going to enroll till
there's a 99% probability of
superiority or a 95% probability
that the effect was moderate at best.
Fu- we, futility, that it, it didn't
achieve above a certain threshold.
And well, many times we don't
have a maximum sample size.
We just go till one of
those results are hit.
And we had sample sizes of these
largely with these 17 different
things, different designs.
We had sample sizes for the
arms from about 100, and a good
number of them at 100, 200, 300,
all the way up to over 1,000.
And the time at which it triggered and
said, "Here's the answer," varied from,
uh, 100 patients to 1,000 patients
all spread out, and it just depending
on the effect of when it happened.
This incredible demonstration of
the power of adaptive learning and
adaple- adaptive sample size when
time was so incredibly critical
Steve Webb: Yep.
Yep.
Um, um, uh, um, the, the number of
patients dying from COVID in any one
week meant that, um, a week's difference
in when a result became available
was important, uh, to public health.
Um, one of the things which, um, uh,
I, I love about the Bayesian approach
is, is, is just this opportunity set
a statistical, uh, trigger of the
amount of confidence, uh, the amount
of credibility that is necessary
to, um, conclude a question and then
keep recruiting, frequently so that
as soon as there's sufficient, uh,
credibility to the result, um, the
experiment, uh, uh, can be concluded.
and I think if, um, I think about all of
the things I learnt, uh, when I was doing,
uh, biostats and epi as a master's student
and thought all of these things
about, um, um, uh, not answering the
question until you've, uh, recruited
your, uh, your proposed sample size
based off a, um, a power calculation.
And I thought, "Oh, these
are really, really important,
uh, features of, uh, design."
Um, I, I, I now understandâ¦
I, I still understand why
they're important, but I do
think there's a better way
Scott: Mm-hmm.
Steve Webb: Um,
Scott: Fantastic.
And so
Steve Webb: uh, uh, when it's appreciated
so frequently the, um, uh, estimates
about the size of treatment or effect
necessary to run clinical trials using
that approach are, um, guesses that
are capable of being just wrong in a
quantifiable way, wrong in the direction
of the treatment effect as well.
And one of the really good things that
came out of REMAP-CAP was, we tested
a variety of treatments because they
were being used, uh, off-label where
the treatment turned out to be harmful.
And the way
REMAP-CAP statistical triggers are set
up, particularly futility, if there's
an intervention that's very harmful,
it triggers very quickly for harm.
It's not necessary to statistically
prove if that's, uh, um, a valid
concept that something is harmful.
You just need to know it's not effective.
Some of the
biggest public health impact in
REMAP-CAP came from early identification
of harmful interventions that
were in standard care at the time.
Scott: Yeah.
Yeah.
Incredible.
So the, the story continues.
The REMAP-CAP story continues.
Um, w- it is still, it is
still enrolling patients.
I think I looked today, there
are 240 sites enrolling.
It's enrolling
community-acquired pneumonia.
It's evolving.
It is ready for a pandemic, uh, onto virus
or, or, or whatever happens to be next.
We are ready, uh, within this.
We also are getting results in
community-acquired pneumonia, influenza.
Uh, we're hoping to have on this show,
uh, to, to talk about some of the results
that are coming out, I think June 10th in,
uh, Belfast, there are results coming out,
so it is still doing incredible things.
Steve Webb: Um, uh,
Scott: s- yeah
Steve Webb: it just,
it, it, it just rolls on
Scott: Yeah.
So in-incredible.
I, I appreciate you helping me start
to get our hands around this beast.
The, the beginning of this, the, the--
There's a lot to talk about in this,
and I think we're gonna come back with
lessons learned, different results,
specific statistical things, response
adaptive randomization results.
So lots more to talk about Remap CAP.
Steve, thank you very much for
joining me on the origins of Remap CAP
Steve Webb: Thank you,
Scott: And
Steve Webb: uh, great to chat
Scott: And everybody else, thank you
all for listening, and until next
time, we'll be here in the interim