Talkin' 'bout Sweet Time

In this episode of "In the Interim...", Dr. Scott Berry dissects prevailing concepts of “clinically meaningful difference” in clinical trials focusing on progressive diseases. Through detailed examples from pancreatic cancer (Ben Sasse, Revolution Medicines), emphysema (Elevair), Alzheimer’s disease (lecanemab), and IVF, Scott challenges the adequacy of the population-mean of a continuous outcome in reflecting true patient benefit. The episode discusses inconsistent usage and interpretation of acronyms such as MCID, CSD, and Target Product Profile (TPP). Dr. Berry advises trialists to resist interpreting the mean difference using patient-level minimal effects, and adopt responder analyses and cumulative probability approaches to enhance patient-level relevance. Guidance is offered for analyzing the effect of time-saved instead of a mean differences in a clinical endpoint at a single time for progressive diseases – measuring “sweet time.”

Key Highlights

• Focus on added time not the change from baseline as the most meaningful outcome for a progressive disease.
• In-depth evaluation of MCID, CSD, TPP, and risk of misinterpretation.
• Critique of trying to interpret mean-based endpoints for clinical meaningfulness such as six-minute walk distance and CDR sum of boxes.
• FDA advisory panel guidance on MCID for IVF live birth endpoints and dichotomous versus continuous endpoints.
• Advocacy for responder analyses and cumulative probability of achieving thresholds in reporting the clinical effect of a treatment.

For more, visit us at https://www.berryconsultants.com/

Creators and Guests

Host

Scott Berry

President and a Senior Statistical Scientist at Berry Consultants, LLC

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.

Well, welcome back
everybody to In The Interim.

I'm your host, Scott Berry.

Now, what, what is In The Interim?

In The Interim is a podcast.

Uh, I think you figured that part out.

We dive into clinical trial science.

We dive into all of the scientific
angles of them, the implications of

clinical trials, and the viewpoint is
typically from a statistical standpoint.

I am a statistician and, uh, been di-
designing trials for Berry Consult-

with Berry Consultants for 27 years now.

So I dive into some of
the experience of that.

Now, many of these podcasts, which we
do weekly podcasts, many of the topics

come up from things that happen la-
recent things that happen, things that

perhaps I, I, I, I wanna share from that.

This one, uh, was-- struck me.

I was reading a news article and it,
it brought back a, a, a really common

topic that I think is very interesting
from a statistical stand, from a

patient standpoint, and I'm gonna
introduce the topic today with a song.

And, um, by the, by the way, I, I do
want to say that the, um, the producers

of this show, Snapmarket, uh, Brandon
Giella at Stat Snapmarket, they do a

tremendous job producing this podcast,
and I, I wanna give a shout-out to them.

And I'm gonna do that because
I'm gonna mess them up here,

uh, right, right here in this.

So I'm gonna play this song, and I'm
sure there's a really easy way to do

it with, with the technology of this.

But, uh, I'm gonna play it anyway.

I'm gonna play this song
to introduce the topic

Said I was in my early 40s with a
lot of life before me when a moment

came that stopped me on a dime

I spent most of the next days looking
at the X-rays and talking about the

options and talking about sweet time.

I asked him when it sank in that
this might really be the real end.

How's it hit you when you
get that kind of news?

Man, what'd you do?

And he said, "I went skydiving,
I went Rocky Mountain-"

Point seven seconds on a bull named

He said, "Someday I hope you get the
chance to live like you were dyin'"

So I, I'll stop it there.

I'm sure most of you know this song.

It's of course Tim McGraw
singing Live Like You Were Dying.

And the part of that song that I
wanna go back to is when during this,

and of course it's this, uh, a song
about somebody in their early 40s and

presumably they're diagnosed with cancer.

And he says, "Talking about the
options and talking about sweet time."

So I wanna come back to that.

It's gonna be a bit of a theme of
the show, and that is sweet time.

There's, uh, a recent story, and
I was reading about this, that

resembles this song quite a bit.

And the, the, the topic is…

Well, uh, the, the story
is about Ben Sasse.

Ben Sasse is a former United States
senator from Nebraska, served in

the US Senate from 2015 to 2023.

At that point, he became the president
of the University of Florida.

He's actually a, a PhD
from Yale in history.

Um, and in October 2025, he was
diagnosed with stage four pancreatic

cancer, very similar to the song.

And so I was watching a story,
an interview of him, um, and,

and he's, uh, uh, recently in
the news, and this is, um…

We, we are August of 2026, and he's, he's
got a number of interviews out there.

And he was on "The Whole Story"
with Anderson Cooper on CNN.

And on there he, he talks about…

First of all, really interesting, he's
in a clinical trial for Revolution

Medicines, highly publicized, exciting
treatment for pancreatic cancer.

Uh, pancreatic cancer has been an
incredibly hard disease to treat.

We've made amazing progress in other
cancers, multiple myeloma, for example,

amazing success stories in breast
cancer, multiple my- multiple myeloma,

lung cancer, um, not pancreatic cancer.

Uh, glioblastoma's a number…

another one that we've not
been very successful at all.

And now there's this drug,
deruxin-rasib, sorry, uh, deruxin-sab.

Um, maybe the people from Snap Market
can cut that part out, uh, in their

incredible, uh, uh, producing of the show.

And they're…

In the, in…

They've have…

Press release came out April 13th of
2026, where the top-line results of

the randomized controlled phase III
trial that in the intent to treat

population, the treatment had shown
a median overall survival of 13.2

months, again highlighting the The,
the deadly nature of this disease.

This was, uh, a recurrent, uh,
second-line plus pancreatic cancer, 13.2

months overall survival on
the treatment compared to 6.7

months overall survival, uh, for
standard of care chemotherapy.

Hazard ratio of 0.4,

highly statistically significant.

That's a press release that came out.

And Ben Sasse is taking this treatment,
and in many ways, he, he attributes him

being alive today to this treatment.

And he, he was reading about this
in his discussions about this, and

a particular quote, and I think this
quote is actually from 60 Minutes,

where he did a, um, an interview.

He says that his additional time
is the result of providence,

prayer, and a miracle drug.

And he's really talking about this,
the, the, this providence aspect is

time, and the meaning to him is the time
he now has because of this treatment,

where perhaps without this treatment, he
doesn't have that time Now, what does that

have to do with clinical trial science?

Obviously, he's in a clinical trial.

He's a patient in a clinical trial.

I, I wanna, uh, uh, pivot a little
bit, and I'll come back to the song.

I, I, I hope this becomes clear why,
why the song during the course of this.

I wanna talk about MCID,

CSD, TPP.

Now, we have a big problem in clinical
trials that we have so many acronyms, and

it actually gets in the way of the science
at times, where when somebody says an

acronym, I say CSD, somebody else hears
that, and it ha- it has a meaning to them.

That may not be the meaning
that I gave to that, and I'll,

I'll come back to that as well.

So what are these, uh, uh, acronyms?

CSD is clinic- clinically
significant difference.

Uh, cl- s- uh, s- uh, c- uh, clinically
meaningful, uh, minimally clinically

meaningful difference is another one.

TPP is target product profile, and
we'll talk about what that means.

The, the goal of this CSD or MCID,
minimally clinically significant

difference, is really about thinking
about an endpoint and what value of that

endpoint clinically makes a difference.

And this is a concept that shows up
over and over and over again, and

it's used again in many different
ways within clinical trials.

So I wanna talk a little bit
about this concept of a CSD.

So I'll go back to 2018.

Uh, I was an advisor to a company,
BTG, uh, who had acquired, uh,

PneumoRx, was a device company, and
they had a device called, uh, Elevair.

It's a treatment for emphysema
And the, the treatment, emphysema

is where part of the lungs, uh,
starts to not function well.

You're not…

Your, your ability to breathe gets
worse, uh, and part of the lung really

becomes non-functional, and they have
a device where they put in-- they

insert these coils inside of the lung.

And I'm sorry if I get the, the medical
part exactly right, exact- if I get it

wrong here, but this is the general idea.

That allows the air to flow across
the lung and make it function better.

That's the idea of these, and there
are multiple, uh, devices like this

that their function is to move the
air, uh, to allow this residual air

in certain parts of the lung to move.

And so it's a device,
uh, to treat emphysema.

And they had run a pivotal trial.

Many times in devices, phase three
is called piv- are pivotal trials.

A pivotal trial of their device
against a sham procedure.

And, uh, it-- in that, they
had an endpoint at 12 months.

And by the way, it's my least favorite
endpoint of all clinical trials.

Six-minute walk distance
was the primary endpoint.

What is six-minute walk distance?

You get six minutes to
see how far you can walk.

Um, and the idea behind it
is it's about ambulation.

It's your ability for ambulation.

Uh, it, it's about, um, a more…

It, it's used in some trials, even
like Duchenne muscular dystrophy for

kids, uh, small boys with Duchenne.

They, their ambulation isn't
very good, so they, they…

It tests their ability to ambulate.

But it's also a test about endurance.

For six minutes to see how far you
can walk, so somebody with emphysema,

this is particularly challenging.

Now, feedback from patienc-
patients, they hate this test.

Uh, as you can imagine, this is, you
know, uh, in, in, in elementary school

when your teacher makes you go out
and run, you know, the 440 or, or

400 meters or go out and run, uh, the
endurance, it, it, you know, it's a

punishment most of the time, um, to that.

And my daughter who ran cross country,
it was interesting that her, her, um,

uh, sport was everybody else's somewhat
punishment or training part to it.

But so the six-minute walk, and the
trial showed a statistically significant

benefit in six-minute walk distance.

The average distance increase
for the device was 14.6

meters, 15 meters Statistically
significant, the primary endpoint.

Now there are a whole other host of
endpoints here: exacerbation rates,

hospitalization for exacerbation
rates, forced expiratory volume,

the ability to, to breathe more.

So lots of endpoints, but this is
the primary endpoint in the trial.

And the big question was, is that,
while it's statistically significant,

is it clinically meaningful?

Now, this endpoint in a number of
settings, there was a number attached

to this endpoint that twenty-five meters
is considered clinically meaningful.

Now, we w- uh, uh, b- you know,
how do they come up with that?

What, what does that number mean?

Uh, but it's a number thrown around a
lot, and it was presented at the public

advisory meeting that their effect
of fifteen meters was less than the

clinically meaningful difference of
twenty-five meters, questioning should

the device be approved, uh, within that.

Okay, let me give you a- another example.

Lecanemab, uh, I think Leqembi is
the, the brand name of this, but

lecanemab was the, was the drug.

It was actually banned two four
oh one, uh, uh, when we started

phase two trials of this.

But the phase three trial of lecanemab,
which is I believe one of two treatments,

disease-modifying treatments for
Alzheimer's that have been approved, um,

donanemab from Eli Lilly is the other one.

And the press release that came out from
the phase three trial, seventeen hundred

patients, uh, uh, randomized control,
blinded phase three trial, eighteen-month

exposure during the double-blind period,
says lecanemab treatment met the primary

endpoint and reduced clinical decline on
the global cognitive and functional scale,

CDR sum of boxes, compared with placebo at
eighteen months by twenty-seven percent.

I'm gonna come back to that.

So it, it…

by twenty-seven percent, which
represents a treatment difference

in the score change of .45

points.

That lecanemab was better than placebo at
eighteen months by a, an average of .45

points on the sum of boxes.

Just briefly, what is a sum of boxes?

It's a, this-- It's a test where
you get tests on various things.

It might be memory, cognitive, uh,
different functional things you carry

out, and usually each one of them
has a three-point scale from, uh,

zero might be, uh, uh, no deficit
to three would be a severe deficit.

And each one of these is
considered a worse deficit.

That's a box, and you
sum the number of boxes.

So this is a, a half a box difference
on average between the groups.

Highly statistically
significant, a P value of .00005

Um, and it, it got approved.

Now, a lot of people talk about
what is a clinically meaningful

difference on the change from placebo
for the treatment at 18 months.

And in the same press release,
they present a 27% slowing.

Now, how do they calculate that?

Um, they calculate that by looking
at what is the average amount of

decline for placebo over 18 months.

Then they look at the average amount of
decline on treatment, and they take the

ratio of that, and the ratio is 0.73.

So the treatment over that 18-month
period declined, and my hand's going down,

but this is actually going up as bad.

So it declined, um, at a rate
that was 73% of what placebo did.

So a 27% slowing in decline
is what's presented.

So I ask you to think about
the, the clinically meaningful

aspect of this treatment.

Is it a half a box different on
average at 18 months, or is it

that it slows decline by 27%?

Okay.

Another recent one is we've been involved
in several clinical trials where the

clinical trial's goal is to improve
the, uh, IVF implantation success rate.

So an embryo is implanted in the,
in the woman, and the question is

whether that becomes a live birth.

That's the goal of that implantation,
that the woman becomes pregnant

and, and a baby is born.

Typically, when they're implanted,
it's somewhere in the 30 to 60%,

uh, roughly success rates in that.

Now, it's, it's expensive,
it's challenging.

It, it can be while you're
harvesting these eggs, they,

they can be hard to come by.

Uh, viable, uh, embryos, um, uh, in
that, and they can be very challenging

to go through this procedure to get them.

So huge impact that a, a live
birth is, is made from that.

So there are, there are clinical
trials trying to improve that rate,

and let's call it 50%, uh, from that.

So there's, there's a very strange, um,
thing that the FDA did, and they came

out with an advisory panel to talk about
what is an MCID for this exact treatment.

And they came out with a measure that
says an absolute difference of 8%

in improving the
implantation to a live birth.

Uh, some of these trials
use pregnancy rates.

Uh, the pregnancy to live
birth is, is 95% plus.

They're, they're essentially
one and the same.

Um, uh, so live births in this
case, that improving it 8% is

minimally clinically important.

Now, that's very, very different
than the Alzheimer's case where

that's a continuous scale.

It's actually an ordinal scale,
but we don't want to get there.

A continuous scale that looks
at a score and talking about is

one point clinically meaningful?

Is two points, is a half a
point, uh, on a continuous scale?

This is a dichotomous outcome,
and they're saying 8%.

So here's ano- There are a number of
examples referring to this MCID concept.

Now- What does this mean?

Let's go back to the Alzheimer's
circumstance and, and almost to Ben

Sasse's circumstance or, or, or the,
the, the live like you were dying

song is in a progressive disease like
Alzheimer's or frontal temporal dementia,

Duchenne muscular dystrophy, ALS.

Many diseases, syndromes are progressive.

Cardiovascular disease, MASH, um,
these are all progressive diseases.

What does progressive mean?

That you continually decline.

Now, there may be perturbations that
from, uh, over a three-month period

you actually jump up, that y- you have
a good three months, but the, the, the

pattern is clearly a decline in that.

Now, humans decline.

Aging is a progressive syndrome.

Life is a progressive syndrome.

So many of the things we're trying to
treat in trials are progressive And

a huge amount of effort is put into
talking about a disease like Alzheimer's,

which is progressive, it is fatal,
but it's the progressive cognitive and

functional decline that is the s- you
know, the scourge of Alzheimer's disease.

And on this CDR sum of boxes, a huge
amount of effort goes in to think

about what's clinically meaningful.

Now, part of it is what,
why, why do we do this?

Part of it is you could run a very,
very large trial, 10,000 patients,

and get statistical significance.

And then the question is, well,
but should somebody take this?

Uh, if you have a 0.5%

slowing of Alzheimer's disease,
nobody would notice of that,

but you might be able to show a
statistical difference on that.

Now, I said a change of 0.5%.

Why, why did I say that to perhaps present
to you something that is a small effect?

It was about time.

Somebody with Alzheimer's, if
you're trying to say what is

a MCID for CDR sum of boxes?

I think it's the wrong question.

Two-point change.

No patient says, "Oh, in 18 months I wanna
have one point better on that scale,"

or, "two points better on that scale,"
or, "five points better on that scale."

They're talking about sweet time.

They want more time to be functional.

They want more time of cognition
to enjoy life, whether it's, um, go

skydiving or rocky mountain climbing,
uh, riding a bull named, named Fu

Manchu in that, but whatever it is in
their time, grandchildren, children,

enjoy life, travel, those things.

They want time.

What did Ben Sasse say in that?

That the providence was the time
he was given And so it, it-- I

think it's the wrong question that
we do in these trials to ever talk

about an MCID for CDR sum of boxes.

I think it's a dumb
question, quite frankly.

Um, I'm sorry if I'm insulting that.

I think it's the wrong one.

I think it's all about sweet time.

Duchenne muscular dystrophy.

We do quite a bit of research and work
in clinical trials, disease modeling

in Duchenne muscular dystrophy.

This is a, uh, disease that affects boys.

Uh, exon skipping which leads to, uh, uh,
lack of dystrophin, which is essentially

a, a fuel for building muscles.

And so they get atrophy of
muscles throughout the body.

They lose ambulation, they lose their
ability in their arms, eventually

lose the ability to breathe.

And we're, we're doing better,
uh, with treatments on this.

We have treatments that change dystrophin.

It's still huge questions about
the, the clinical effects of

these various treatments, but
we're treating them better.

But s- you know, many times ambulation
is lost at about 10, 12 years old,

and these boys don't live beyond
sort of 18, 20 years old, and

maybe if they're lucky in that.

Now, there are other forms of, um,
uh, other dystrophies, uh, within it.

There are less severe forms,
Becker's disease and that, but

largely Duchenne, Duchenne muscular
dystrophy is this progressive

disease We use endpoints in that.

Six-minute walk is used in that.

I think that's gone out of
favor, thankfully, um, in it.

By the way, I, I dislike it
as a statistician, um, because

it's hugely sc- uh, skewed.

There, there may be a time where
somebody struggles with it and they

only can walk 50 meters, and then two
months later they can walk 400 meters.

Hugely skewed.

Now, we have techniques for handling
all that, but it's, it, it's a

brutal endpoint statistically.

But in, in Duchenne muscular
dystrophy, we do things like the North

Star Ambulation, uh, Assessment--
Ambulatory Assessment, NSAA.

It, it measures a number of, uh,
uh, functional things for the boys.

Um, uh, by the way, it's an X-linked
disease, which is why it's incredibly rare

that a female, um, uh, has this disease.

Both of their Xs would have to be
affected, uh, within it, otherwise they

have appropriate levels of dystrophin.

Um, time to rise is a recent thing
that a lot of, uh, is that, where how

long does it take the boy to stand up?

And we, we talk about a velocity.

So it's this velocity measure and their
ability to stand, um, uh, when it…

There's a number of, of

endpoints used.

Again, ask a 10-year-old boy what's
a clinically meaningful difference

on their velocity to stand.

You know, uh, what a silly question.

And, and if, if the goal at the end of
the day is about measuring clinically

meaningful, it's the wrong question.

Now, I don't have Duchenne, so I, I
don't wanna speak for the patients.

There's incredible patient groups
in this, but they want time.

They want sweet time.

More time with ambulation, more time with
upper, uh, uh, uh, body strength to carry

out, to work on their iPad, to do things.

They wanna live for as long as they can.

They want time.

So do we ever talk about time as that?

I mean, it's really, if you wanna measure
clinically meaningful, it's about time.

Now-

Does a test of means, another part
to this that is, is interesting, and

going back to that advisory committee
meeting for emphysema on six-minute walk

I think the whole idea of talking about
a clinically meaningful difference

on a continuous measure that we talk
about the mean of a population is

a completely different question So
I did not like that the FDA said

that this 15 meters is less than 25.

What is-- It's not clinically meaningful.

15 meters is the average
across a population.

A- And, and to talk about what an average
effect across a population is in terms

of clinically meaningful, clinically
meaningful is a patient level thing.

A population level thing, there
are many ways to get to 15 meters.

Some of which I think people would
say is not clinically meaningful,

and others would say they're
incredibly clinically meaningful.

Uh, a, a neat analogy, I do lots of
sports, and you know that, is late

in LeBron James' career, I read
a- an amazing stat that his, his

average points was 27 points a game.

Average assists and rebounds,
I think, was eight and seven.

Don't remember.

It was 27, seven, eight.

Was-- This was somewhat
late in his career.

I know he's still playing, um, in that.

But that was the per game average.

Not one game in his career, many hundreds
of games, if not thousands at this time,

did he ever go 27, eight, and seven.

That was his mean, but
he never once did that.

Now, it was his energized mean in that.

So again, the mean doesn't
really tell us much.

Here's an example.

15 meters.

Suppose you have a, a, a trial
or a truth about a treatment is

that every patient on the placebo
declines 25 meters in emphysema.

It's a, it's a progressive disease.

They progressively get worse.

Every patient on placebo exactly declines
25 meters, and 50% of those patients

on treatment have a-- have no decline.

That's a 25-meter difference
for those 50% of patients.

No decline, so that's a 25-meter
difference over 12 months.

And the other 50% have a 25-meter decline.

Presumably no benefit.

50% of the patients have no decline and
a 25-meter difference, and they meet

whatever and however somebody came up with
25 meters being clinically meaningful.

50% on treatment, 0% on placebo.

Few treatments we do have,
have that kind of benefit.

Stunning benefit.

That's, that's approvable without a doubt
What if everybody on placebo declines 200

meters over 12 months and every single
patient on the treatment declines-- By

the way, the average of that is 12.5

meters.

That 50%, 25, if you took the average
acro- across the population, it's 12.5

meters.

Now, nobody does 12.5.

Half do zero and half do 25.

I think that's an incredible treatment

Another one could be that the
placebo declines 200 meters.

Every single one of them, every single
patient on treatment declines 187.5

meters.

That's a 12.5-meter

difference on that.

The, the difference in
those 200 meters to 187.5,

somebody could very much say that
that's not clinically meaningful.

By the way, you haven't
bought them any time.

In the other case, 50% of
patients don't decline for 12

months, you've bought them time.

In this other case, 200 to 187.5,

the exact same mean difference,
you haven't really bought

them any time at all.

Probably an innoticeable
am-amount of time.

Non-noticeable.

Sorry, innoticeable is a word.

So there's a number of aspects to that.

So that means if you're doing a test
of means, a T-test, an MMRM in a

disease like this, I don't think you
can talk about a clinically meaningful

difference on a population level.

Now, okay, but, but what, what does
that mean, uh, in this setting?

Now, I wa-wanna contrast that to the
in vitro fertilization, the IVF That

endpoint is not a continuous scale.

Uh, you'll hear me on all these
podcasts rail against binary outcomes.

This is a binary outcome.

It's live birth or not.

It, it, it, in, yeah, I-- In some
circumstances, you could talk about

the health of the baby and all that,
but th-this is about a live birth.

It's a binary outcome.

There's not a continuous measure here.

Now, every single one of those, by
definition, is clinically meaningful.

And if you show a statistically
significant increase in the clinically

meaningful difference, I, I, I
mean, by definition, you've shown

a clinically meaningful effect.

To say an eight-point change in
clinically meaningful difference,

I mean, that's what it is.

Every one of these is
clinically meaningful.

Now, you could argue that it's not
clear on a continuous scale whether

one meter is all that different, and
I think that's sort of the issue.

So somewhere, somehow, somebody thought
there should be an MCID for a case where

every single binary outcome, it's, it
by definition is clinically meaningful.

The whole idea of trying to define a
difference in the percent of clinically

meaningful outcomes that's clinically
meaningful, it makes no sense.

I, I, I don't know how this ever was
thought to be a meaningful thing.

Now, I understand if you ran
an enormous trial and you saw

a 50% effect on one and a 50.1%

effect on the other one, and it's
statistically significant, very

reasonable question whether somebody
should pay for it or not and all that.

Um, it, it's somewhat
of a different question.

But the idea of assigning this
8%, uh, just seems like the…

By the way, I think it's a strange
answer to somebody who has an 18-month

grandchild and a 12-month grandchild.

Um, I don't like the eight if you're
gonna ask the question, but I think

it's, so I think it's the wrong,
wrong answer to the wrong question.

But it more important thing is
I think it's the wrong question.

It's a very, very different thing
to a continuous outcome, which

at least I get the goal of that.

Now, in clinical trial designs, these
acronyms show up all the time, and your

job as a trial designer, whether you're
statistician, whether you're clinician,

uh, whatever your role, a regulator in
this, is what, what do these things mean?

And I, I, I, in my time interacting with
people, they mean very different things.

So Remember, when we run a clinical
trial, so an Alzheimer's trial or a

trial in time to rise for Duchenne
muscular dystrophy, uh, overall survival

in a, a pancreatic cancer trial, we
power the trial for a particular delta.

So in time to rise, it might be
we're looking for a particular

effect on that velocity measure
or a percent slowing on that.

Or in Alzheimer's, we're looking
for a particular effect or a

hazard ratio in a clinical trial.

That delta that we power the
trial for, lots of people call

that thing something different.

Now, pharmaceutical companies love
this name, target product profile, TPP.

Sometimes they call that the TPP.

That's the p- product
profile we're targeting.

Doesn't necessarily mean they
believe the effect is delta,

but they're targeting that.

Others think that that's the--
defines the MCID, uh, in the setting,

and they might call it the MCID
or the CSD within that setting.

Now, think about that because when
you run a trial and you power it

for delta by-- for 90%, for example,
the-- you win the trial if you see

about two-thirds of that effect

So that powering a trial for delta does
not mean you're posting that delta is

the clinically meaningful difference.

Many, many trials are successful where
the effect is between delta and two-thirds

delta, which is statistical significance.

If you believe delta is the effect that
has to be observed for this treatment

to be on the market, you should
power it at, uh, four-thirds delta.

Uh, I don't know if I did the math right
there, but, uh, three, three-halves

delta, so that when you win the trial,
it corresponds to the right number,

uh, uh, that's observed at the end
that you say, "Aha, we, we hit that."

So when a, a lot of times I'll be sitting
down for a trial design and the, the,

the, the people we're working with on
the design say, "Oh, our TPP is 25%

Now, don't accept that.

Don't go along and think you
know what that means, 'cause your

interpretation of that is different
than the person who said it.

They might mean it to be the
effect needed to win the trial.

They might be, "That's
what we wanna power it on."

They might not mean any of that at all,
but that's what a competitor did, and we

wanna know are, are we better than that?

We need it-- Uh, so it could be
a number of different things.

The agency might talk about
it, and a regulator might talk

about it as the effect we need
to see to approve the treatment.

And so those things, acronyms,
somebody throws out an acronym, you

interpret it a particular way, but it
was intended in a very different way.

So try to get out of that.

Try to make sure you understand when
somebody says that, what-- in what

part-- what, what do they mean by that?

And sometimes it's just a simple
reaction that, "Well, if I tell my

statistician it's 25%, they can go off
and design the trial and we're done."

Careful, you, you may build something
that, you know, at the end of the day,

doesn't meet the goals of the trial.

Um, the other thing is, uh, uh, doing
a number of these public advisory panel

meetings where you present your data
to a, uh, a group, uh, to get approval,

and you've done a test of means,
continuous measure, it's going to be

really hard for people to interpret.

That 15 meters out of 25 meters
is not a very meaningful thing.

Again, it's an average
across a population, a lot

of ways to get to 15 meters.

You should put in responder analyses
about how many people get to that CSD.

What proportion of
patients achieve 25 meters?

And in that trial, for example, the
device showed 20% improvement in the

number of patients that got to 25 meters.

I think that measures at a patient level
what are we doing, and even more so

prospectively des- defined, you'll present
the cumulative probability of achieving

thresholds across a much wider scale.

Because the person you're presenting
to might think 50 meters is important,

um, in that, and now they can see the
whole scale of that, and it, it, it's

a way to understand the clinically
meaningful effect at a patient level.

Clinically meaningful is
a patient-level thing.

Really hard at a population level
The other part of how I started this

is this idea of in a progressive
disease using a change from baseline.

It just feels like the wrong S demand
that I don't think is patient-centered.

I don't think it's even
regulator-centered.

I don't think it's anything centered.

Uh, we do that.

And again, I do think it's about time.

Every single progressive disease, and many
of them use an MMRM model, a T-test, um,

where they-- Over this, this Alzheimer's
trial through 18 months, they didn't

tell us what the effect was at 12 months,
at, uh, 15 months, at nine months.

They said at 18 months, and then they
took the ratio of the percent slowing.

The S demand of interest in
almost every progressive disease

for disease-modifying effects.

There are symptomatic effects, so there
are donepezil as a treatment that, that

aids, um, in, in memory and function.

It doesn't change rate of decline
at all, so you wouldn't use that

for a treatment that's symptomatic.

But for something intending to be
disease-modifying, it's about sweet time.

The S demand should be the percent
slowing Then it brings about the

really interesting thing is we
analyze progressive diseases using a

model that doesn't take advantage of
understanding the disease is progressive.

We use a mixed model of repeated
measures, which makes no assumptions

over time within that setting.

We lose power by not doing that.

In an oncology trial of PFS of
OS, we model it as progressive.

You can only go one way.

You can only go to progression to death,
and it's modeled as progressive, and

we talk about hazard ratio is time.

Hazard ratio is how do I
change my event rate over time?

For mortality is about extending life in
that they presented the median survival of

thirteen months to six months in, in that
phase three trial, very patient-centered.

Now, I still wanna see the curves
of how they get there, but this is

talking about the thing that I think
patients care about, which is time.

Same thing in progressive diseases.

So make sure you understand when
somebody talks about an MCID,

what are they talking about?

Make sure it's on a scale that actually
you can measure clinical benefit.

And you've heard my opinion in, in
various, uh, progressive diseases where

we use continuous measures of that.

Now, don't accept the acronyms.

Show us example trial outcome.

Okay, we run this trial.

Here's an example outcome.

It's statistically significant
and shows fifteen-meter difference

on the population level.

Is this gonna meet your goals?

Doing bigger sample sizes, smaller
sample sizes, different ways to

analyze the endpoint, different
ways to present them, this set up

in a prospective manner to that.

Nothing is better for making sure
you're talking about the right

things and not acronyms than
showing example trials, real data.

You kind of lose the acronym part, and
you talk about the real effects of that,

how to measure the real effects of that.

So if you're a designer, if you're
involved in the design or analyses

of this, envision presenting to
an advisory committee meeting.

Even if, even if this is a trial
that's not gonna go to an advisory

committee meeting, you know,
think about in a journal, how are

we going to explain this effect?

Do I have the right design for that?

Is the d- the design
appropriate within that?

Okay

So we presented this, we talked
about Ben Sasse, 54 years old.

He's, he's-- finds himself all
of a sudden a patient in this.

He, uh, it was interesting in, in him
talking about his particular case,

incredibly analogous to Tim McGraw's song.

He was training for a sprint
triathlon when he started to have

symptoms, went in and was diagnosed
with stage four pancreatic cancer.

He found himself as a patient.

We are all patients.

Humans are patients.

You eventually, um, uh,
will become a patient.

We're, uh, in this setting,
we're all on this same journey,

and it's all about time.

So i-incredibly impactful work you're
doing, uh, reflected by Ben Sasse, uh, in

this setting, and thank you for that work.

It's incredibly impactful,
and I appreciate you joining

me today for this discussion.

Until next time, sweet time
will be here in the interim