The Art and Slog of Innovating

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 Berry: Welcome everybody to in
the interim, uh, Barry consultants

podcast about all things, science,
statistical and clinical trial and medical

sciences of a really cool guest today.

Uh, Mike Krams is our guest today and
for everybody's Uh, knowledge, he joined

Berry Consultants in January, but Mike
has a, a really interesting history in

clinical trials, drug development, he's
been 30 years at multiple pharmaceutical

companies, he's led quantitative sciences
departments at these, uh, he's done

innovation everywhere he's been, and now
we're thrilled to have him at Berry, but

we talk about innovation, how do we do
innovation at in, in drug development.

So Mike, welcome to In the Interim.

Mike Krams: Hi Scott, and hey, you
forgot to mention how it all works.

You write a letter to Professor Don Berry,
uh, 25 years ago, and think there'll

never be a response, but then there was
a response, and the rest is history.

Scott Berry: Yeah.

So, so, so this is a great source
of, we, we are celebrating Barry

Consultants 25th anniversary this year.

Uh, we're, we're marketing this
and you can see this everywhere,

but we've been working with Mike
Krams for well over 25 years.

Um, so I, I'd love to hear a little bit
about, and I know much of this, but tell

us about the Aston Stroke trial, which
was your first interaction with Don Barry.

Mike Krams: Yes, I was trained as a
neurologist and a stroke neurologist.

Um, you know, we, uh, worked in
one of the first stroke units in

Germany, uh, before I went to London
to do, um, functional brain imaging.

And in that functional brain imaging
environment, I learned a lot about being

clever on how to detect a weak signal
in functional MRI and PET studies, a

weak signal in a noisy environment.

I learned a lot of statistics.

And then I joined a pharmaceutical
R& D, uh, organization.

Um, they were looking for a stroke
neurologist with a background

in functional brain imaging,
interested in developing acute

ischemic neuroprotectins for stroke.

And so I was made for the job.

And then I looked at what was
already happening in that area.

And there were plenty of efforts
to create new treatments for

acute ischemic stroke patients.

And one failed after the other.

And, you know, when I, when I started
in my career in pharma R& D, I just

came from the Institute of Neurology
Functional Brain Imaging Unit.

Um, you know, the sophistication
of the, um, of the functional brain

imaging, experimental design and
analysis world contrasted with a

world in those days where you spend
millions of dollars into having two

groups and doing a t test at the end.

And it was just mind boggling.

And it also was pretty clear that
identifying the correct dose and

treatment regimen was a big, big issue
when people went from phase 2 to phase 3.

And so, I tried to figure out, you know,
if that's the problem, if we are not

very good in learning about the correct
dose and treatment regimen to take into

confirmative trials, what needs to change?

And that

Scott Berry: so let's, let's
set this up a little bit.

So you've got a, you've got a potential
neuro protectant for acute stroke.

You go to a traditional, uh, trial design
invariably your statistician is going

to say three doses, 80 patients a dose,

Mike Krams: Uh, maybe more

Scott Berry: two.

Mike Krams: yes, but, but yes, it's,
it's, it's like, uh, two or three, uh,

active, uh, treatment arms and a control
and pairwise comparisons and that's it.

And, uh, you know, just mind boggling,
uh, um, how, um, how, how one can do that

and not think about, uh, alternatives.

And so I, Uh, started educating
myself about what else one

could do when I came about.

There were a number of papers that were
really well written and the author was

this professor whom I imagined with a long
beard and probably would never talk to me.

And the name was Professor Don Barry.

Anyhow, I eventually found the
courage to write him a letter

and say, hey, I'm really I'm very
impressed with your, uh, papers.

There's this one paper about Bayesian
thinking in clinical research and,

you know, it totally resonates with me
because as a medic, that's how I function.

What you describe as the Bayesian
thinking, that's, that's how, how,

how you work, uh, in medicine.

And so, hey, can we meet and can you
perhaps, uh, uh, discuss with us, um,

what one could do to approach this,
uh, problem, namely, uh, identifying

the correct dose in a different manner.

And so that's how it all started.

And then, uh, you know, we had, uh,
just an amazing time, uh, exploring

the opportunity space for what could
be done with fantastic colleagues in

the team and Don, of course, and others
Peter Muller and many others Andy Greve,

key person and, you know, over many
months we were able to Effectively,

in those days, not exactly invent, but
really apply for one of the first times

simulation guided clinical trial design.

And, uh, you know, that the, the
explanation that we gave, uh, to our

senior managers was look, the way we
currently work is we build an airport.

Um, we then build a plane.

We put peasant passengers on a plane
that's never been flown before.

As the plane is traveling, we
build another airport where it'll

land, and then it'll land, and by
that time, we'll have destroyed

the airport from where we started.

And then we do this over and over again.

And that's how we do clinical research.

How can we do things differently?

And, uh, you know, with Don's
help, and Andy Reeve's help,

and other people's help.

Um, we eventually came up with, uh, design
that's published and well presented.

Um, where we had, believe it or not,
16, uh, different treatment arms.

And what is called a response
adaptive allocation to treatments.

Uh, uh, uh, responding to
the data, uh, on, uh, stroke

scores coming in in real time.

So that's how it started.

Nope.

Scott Berry: uh, interestingly
that the time it took to do

those simulations and build this.

This is a time where
software is not available.

You're custom coding.

This is many months to
carry out these simulations.

But even today that trial 16 doses
response adaptive randomization the result

of that trial Which has been published
was a resounding success in that it

demonstrated clearly the drug didn't work

Mike Krams: it did.

And it did so, there was a futility
rule built in, and so it stopped

early for, um, for, for futility.

And so the design did what
it was supposed to do, yes.

And, but you mentioned, uh, the
time that it took in those days to

fine tune the, uh, Uh, the software
that was, uh, handmade, uh, custom

built really to, uh, support this.

And, you know, we were always
angry when we had to wait another

two or three weeks before the next
iteration of software came in.

Compare this to today, where you have,
you know, software packages like facts

or others, where on a push of a button,
you can plug in stuff and immediately

get a first sense of how things work.

So, hey, that, but that's,
that's how it started.

Yep.

Scott Berry: Yeah.

Yeah.

So, so we now have software
that can do those simulations,

can build trials like that.

But in doing that, the hard part
may not have been the simulations.

The hard part may not have
been the modeling, the response

adaptive randomization.

It helps when you have Peter
Mueller, uh, you know, brilliant

scientist, uh, doing all of that.

So I, I want to talk a little
bit about bringing change.

You go into a place where it's typically
fixed trial designs, enroll this, and come

back and show me the data in three years.

Now you're doing interim's monthly,
weekly, response adaptive randomization.

How do you bring change to this industry?

Is it, you know, so that's
what I'd like to talk about.

You know, this,

Mike Krams: sure.

And let's, let's just talk in
general terms rather than about one

particular trial in one particular
pharma R& D environment, because

the general aspects hold true,
uh, really, uh, wherever you look.

Overall, Uh, our industry is extremely
conservative and people haven't got

a lot of, uh, encouragement to think
out of the box and try things that

might perhaps require additional
interactions with health authorities,

regulatory scientists, et cetera.

So the going in position is everybody is.

Uh, doing what was done before
and, uh, is not that comfortable.

Uh, proposing quite different approaches
and the question, of course, is why do

something different if what we are already
doing is working, but it's not working.

It's so inefficient.

And so you are asking how
to bring about change.

Well, if you are in an environment that
is utterly against it and from a top down

perspective, you know, senior management
will not allow you to, uh, explore.

Uh, uh, innovative, um, uh, uh,
approaches, it's very tough.

So if you want to do this bottom up and
you haven't got a champion in a, um, in

a more senior position, it's, it's, you
know, uh, my experience tells me I've been

most successful in places where my boss
and the boss of my boss were champions

for the idea that we brought forward.

So it's important to have champions.

in senior positions to, to help
you achieve this, but then what's

equally important is to clearly
articulate the value proposition.

You don't do innovation for innovation's
sake, but if there is a way of

convincing the, environment that
we work in, that there is a better

decision to be made at an earlier
time point in a more efficient manner.

Now we're talking.

And, being able to articulate, clearly
understanding the position of the person

at the other side of the table, and
bringing along the audience, that's key.

Let me give you some

Scott Berry: So, so, so can I, can I
talk about, I think this is so critical,

so we, I, I, you've got experience of
this, you know, in development programs,

uh, much more top level than I, so I
end up working a lot with teams and we

explore adaptive designs in all of this.

and in, in, even at that level,
you've gotta create a champion that's

gonna go up and fight, fight for it.

First of all, you've gotta
demonstrate to that person that

what you're doing is better.

They've gotta believe in it.

But I've also found that you need that,
uh, a little bit, that that person has

ownership in that what you've created

Mike Krams: Oh, absolutely.

Scott Berry: so that when, when
we're doing these sim and so.

The name, the consultant, we don't
go in and say you've got to do X

and we're kind of beating them down.

They're not a champion for you.

They don't necessarily believe in it.

They're not going to
go and fight for this.

But if you get them to believe what you're
doing better, they made the decisions.

You guided them and you
showed them the ramifications.

That's why clinical trial
simulation is beautiful, by the way.

You can see the ramifications of all that.

They become the champion.

They can go up and say here's why.

And, and, and.

If they're fighting for it, they're,
they're champion at that level.

And then what happens above that?

Mike Krams: Yeah, well, you made
it so clear that it can't be

that an external party comes and
says this is the way it's done.

Drug development is a team sport.

And the way I've experienced this with
Don and you and others has been that

you've empowered us to stand up with a
proposal that you helped develop, but

you really empower the team on point
within the farm R and D organization.

to then carry, uh, the
flag and make the points.

And, uh, what, what happens?

So, so you need to, uh, think about
who are the key decision makers.

Initially, in, uh, uh,
compound development teams,

you have a, uh, team leader.

Uh, you have, uh, a, uh, Uh, program or
project manager, you have a therapeutic

area head, you have, uh, functional heads,
lots of people who have a say in this.

And often what happens is that even
though you may have been successful in

bringing your own team, uh, around to,
uh, embracing, uh, innovative idea,

it's enough if just one person somewhere
higher up says, Hey, I don't like this.

This, this, this has
never been seen before.

I don't think that this will fly.

That's enough to derail it.

And then it's important, uh, not to
give up, but to try to understand

what is the motivation of that person
to say, um, hey, it won't work.

I'll give you an example.

When, um, we implemented, um,
response adaptive dose finding

studies in a broader way.

One of the arguments that we made
were, um, it's of value to have

more doses rather than less doses.

Um, and, uh, the person who
headed up the Pharmaceutical

Sciences Organization at the time.

They would have ultimately been
accountable for making more dose

strengths and enabling that whole
many doses rather than a few.

They hated this.

They absolutely hated it.

And then, you know, within learn
trials, we proposed something that

proved really, an amazing thing.

We invited the head of the Pharmaceutical
Sciences organization to be a silent

observer in the data monitoring committee.

So that person was now able to see
firsthand how the information evolved over

time and was sworn to secrecy, obviously,
but in the back of their minds, they

were clear on how the trajectory of a
particular program was more likely to go.

And of course, that might have
implications on, bigger strategic

thinking and some key decisions that
might have been in that person's mind.

So to understand where the resi the
resistance comes from, taking that very

seriously, but then also finding an
opportunity to bring that person into

a position where that person can also,
support and champion what we're doing.

And that.

That function ultimately became a big
supporter for us, because they understood

that we were more frequently stopping
things that shouldn't go forward, and

bringing forward things that should
go forward more, more, rapidly.

Scott Berry: Yeah.

Yeah.

Uh, so a little bit.

So the title of this, we've titled
this the art and slog of innovation.

And so a little bit of that seems
like the slog that invariably when

you're doing something different that
there's, there's ways this has been

done many times within a pharmaceutical
company within the industry.

And there's a lot of copycat doing
what the last person did and all that.

And now you're saying, proposing
something different and invariably.

10 people are going to give a reason
not to as this goes from concept and

development and all of these will stop
it if you're not able to combat that

with with information and bring those
people along as well that you're not an

enemy that we're making better decisions.

That's a bit of the slog of this
innovation aspect of it, uh, from it.

That is just so hard to do it.

Mike Krams: Well, you know,
uh, it's actually hard work.

And you need to be at the table
where strategic decisions are made.

If you're not there, if you're
not present, uh, doing this Um,

from a distant with only punctual
interactions is much more difficult.

Um, so being embedded in discussions
that lead to better problem solving,

that lead to, uh, creating, uh, resource
efficiencies and ultimately that lead to

creating a new fun intellectual culture.

That's a key thing.

Uh, you know, is incredibly rewarding.

But you have to be part of
the inner discussion where the

strategic direction is being set.

That's why it's so important not just to
look at an individual trial in isolation,

but at the development strategy overall.

And then think.

Within that bigger picture, how do
we, um, uh, design the, uh, strategy?

Scott Berry: yeah, it's
it's so interesting.

So part of this is.

You hear this.

You need to be at the table.

The question is, how do
you get to the table?

You can't just blurt out,
I need to be at the table.

That generally doesn't work.

And in our circumstances, as a
statistician, we may have a number of

statisticians joining this, we have
scenarios where, you know, a client

may say, can you show us the power?

Can you show us this, this, and this?

How you present this is so critical.

And if you pass this to somebody else
who then brings it up there, and they're

bringing They present it differently.

They answer questions differently,
and it's so much the ability to

see how people react to this and
what is what do they need to see?

Is it about time?

Is it about cost?

Is it about another development
program that they want to

take these additional shots?

So being able to listen
and hear all of this?

And then be able to present exactly what's
going to help them make that decision

is the really hard part about this.

And it's earning your way at the table
that people think you help them and

you're presenting the right thing
and you understand the bigger part.

That's what's harder
for us as statisticians.

We're really good at calculating
something, but this is the hard

part to earn your way to the table.

Mike Krams: True.

Respect is earned, but I tell you, uh,
I have had the privilege to work with so

many, statistical experts and modeling
experts and mathematicians who are

absolutely brilliant in articulating
clearly in a way that the other, function

that might not be, mathematically
inclined still understands, and that

is a necessary condition for it.

Getting, the implementation
of innovation guaranteed.

So very good communication skills, but as
you pointed out, also very good listening

skills and psychological skills, where
does the other party come from, but,

there is also a need for putting your
foot down and, making sure that, others

understand what statistical experts do
is a contribution to strategic thinking.

It's not just a subservient
number crunching activity in

the background, absolutely not.

And insisting that there be that,
partnership, between clinical experts,

translational science experts,
regulators, and statistical and

others, experts, is, very important.

Yeah.

Scott Berry: my, my words I
associate with you and you say

this all the time is imagine.

Um, so, so let me, uh, you know,
imagine if we could do the following.

We started this off with, uh, 1990s,
a neuroprotectant for acute stroke.

We are now 2025.

We have no neuroprotectant
for acute stroke.

We have thrombolytics.

We have endovascular therapy.

That, by the way, is incredibly
effective in some patients.

We're still in search for
neuroprotectant, uh, in stroke.

Imagine if development
would have been different.

Are there neuroprotectants that we missed,
that we didn't take shots on goals?

We got the wrong dose.

Now, we're, we're, we're,
we as an industry, we're

developing really cool things.

They're out there, but this is a
really hard area where in some ways

we haven't made tremendous progress.

And does the development affect that?

Mike Krams: Yes.

So, hey, uh, this is so close
to my heart, but it's not

just in acute ischemic stroke.

It's in any, uh, important, uh,
disease where there aren't solutions.

Uh, you gave a talk once about statistics
and baseball, and it inspired me.

And in it was, uh, The motion of the
time machine to be able to compare

the thing that happened in the past
and move seamlessly to the presence

and, and build models around that.

And then I heard your dad, uh, Don
Barry, uh, talk about platform trials

and, you know, I've worked in different
companies on individual projects.

observing what the competition did.

And it really, at times, drove me nuts
how mistakes were reinvented without

comparing notes and working together.

Now Imagine that at the center
of the universe was not, an

individual, investigational compound,
but the need of the patient.

And imagine that everybody was
very keen to get to the right

solution at the earliest time point.

Of course, one would think about
how to bring things together.

And what I love to imagine are these
integrated research platforms, where

on one hand, on an ongoing basis, You
captivate all learnings on how to observe

patients in methodology like studies.

But then you build on top of
that the exploration of new

pharmacological entities.

A little bit as is done in platform
trials such as iSpy2 or the many others

that have been developed by you guys.

And so I feel So, that is the future.

And, uh, there are some, uh,
questions on how to achieve that.

Um, but, hey, uh, that's
where we need to go.

Yeah.

Scott Berry: to put in a plug for the
STEP platform trial, NIH funded, NINDS

funded that's trying to do exactly this.

Not rebuilding the airplane, airport every
time, uh, you know, so this, this is a

fantastic effort and we're seeing much
more of this, uh, in platform trials.

So, uh, but I'll come back to the
sports things and, you know, I almost.

I interpret things in sports.

I go back to this and a lot
of this is very similar.

So, so baseball for years, uh,
nobody would do anything different.

The manager, if you did what
everybody else did and the team

lost, it was the player's fault.

If you do something innovative
and you lose, it's your fault.

If you did something against
the grain and all that.

Uh, and it's the same in drug development.

If you do what everybody else
did, it's the drug's fault.

But if you do something
innovative, maybe it was my fault.

Maybe I did it wrong and all that.

The fascinating thing, of course,
in baseball is analytics has

completely changed the game.

And in 15 years, the way the
game is played, the way it's

set up, is entirely different.

So, innovation completely changed it, but
it's also done that in drug development.

Platform trials, adaptive
designs in 15 years.

Now, if you don't do these things, maybe
they'll say, Why weren't you doing that?

Why weren't you doing that?

And maybe it's a similar thing.

Mike Krams: you're absolutely right.

We've come a very long way.

And what is so amazing, and that is
something that statistical experts

can take a lead in, is the power
of simulations to be used as a

tool to bringing people around the
table and then playing ping pong

with arguments and saying, Hey, if.

What if, and then you try that out
and, uh, and I think the, there's a lot

of openness to, uh, to, uh, innovate.

I want to say this also with
experts in regulatory science.

You know, a lot of people say,
Oh, regulators don't like it.

It's absolutely not my experience.

My experience is that there have
been so incredible interactions with

health authorities that have helped
shape on how to go about innovating.

Scott Berry: Yeah, no, I agree completely.

And people don't see 10
trials done like this.

It's not because agencies
said you can't do that, but

they're not being brought that.

So I think regulators have have played
a huge part, actually, in a lot of the

innovations that's happened in this.

Yep.

But that's that's part of it again.

The slog part of it that when you
present this, well, regulators

aren't going to like this.

Well, drug develop, uh, drug
supply is not going to like this.

Well, CRO is not going to like
this, you know, and that's a bit

of the slog of, uh, and the hard
work in doing something different.

Mike Krams: yep, yep.

You know, there's the saying,
culture eats strategy for lunch.

But if you, uh, make it a culture to
have fun because you are intellectually

challenging each other, and it's not
innovation for innovation's sake, but

it's simply the question, what's the best
thing in the name of future patients?

Then everything else will follow.

Scott Berry: Yeah, yeah.

Fantastic.

Fantastic.

Well, boy, this has been been wonderful.

I'm incredibly excited as as we move
forward and part of Berry Consultants,

but it's an incredible look at
innovation, the art, and the slog.

Appreciate it very much, Mike.

Mike Krams: Hey, thank you.

Scott Berry: Yeah.

Awesome.

Thanks.