Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:02
A
0:02
few weeks ago, some new research made
0:05
a lot of noise. A new study
0:07
is raising some questions about colonoscopy
0:09
screenings and how much they actually reduced
0:11
deaths from colon cancer. In one
0:13
of the largest studies ever, European
0:16
researchers found colon ASKBE screenings
0:18
cut cancer risk by eighteen percent
0:21
and made no difference in death rates.
0:24
For
0:27
two decades, colonoscopy has
0:29
been a rite of passage for Americans over
0:31
fifty, though the starting age has
0:33
recently been lowered to forty five.
0:36
journalist Katie Kirk even televised
0:38
hers after her husband died
0:40
of colorectal cancer to emphasize
0:43
the importance of the procedure.
0:45
Physicians and public health experts have
0:47
long urged people to get screened
0:50
because the best available evidence suggested
0:53
not only could colonoscopy find
0:55
cancer, it could also prevent
0:57
it. Was everybody wrong?
1:01
Some doctors challenged the recent studies
1:03
findings, insisting that above all
1:05
else.
1:05
colonoscopy saves lives.
1:08
Does
1:08
it?
1:10
From
1:10
the Freakonomics Radio Network, this is
1:13
FreakonomicsMD. I'm Bob
1:15
Bugena. Today, I've got
1:17
a cold, so please forgive my voice.
1:19
It'll sound little different when I talk to
1:21
you than when you hear me in conversation with
1:24
our guests. First, Doctor
1:26
Michael Brett Howard, the lead author
1:28
of this new colonoscopy study, would
1:30
tell us what we should really take away
1:32
from his group's research on colorectal
1:35
cancer screening. So you could just say, well,
1:37
you have to close. There's no differences. It doesn't
1:39
work. I don't think it's that simple.
1:42
But first, doctor Osmotchcott will
1:44
explain why we rely so
1:46
much on colonoscopy in the US
1:48
compared to other countries and how
1:50
that can lead us astray.
1:52
Colonoscopy as effective
1:54
as it can be is heavily
1:56
heavily operator dependent.
2:11
Golent cancer is The third
2:13
most common cancer for both men and
2:16
women in the US accounts for
2:18
about a hundred and fifty thousand new cancer
2:20
cases in the US alone and
2:22
responsible for about ten percent
2:24
of all cancer related
2:25
deaths. So it truly is
2:27
a big It's common and it's
2:29
lethal.
2:30
Doctor Osmotchcott is a gastroenterologist
2:33
and a professor at New York University
2:35
School of Medicine. How I got into gastroenterology
2:38
was I wanted to do
2:40
something public health related and
2:43
I thought what are some of the large
2:45
public health questions that
2:47
we have some good evidence for, but
2:49
there's a lot of missing pieces where
2:51
we could truly make a big difference. has
2:54
tried to make a difference by focusing
2:56
on colorectal cancer screening. She's
2:59
published nearly two hundred articles on
3:01
this topic and other related ones.
3:03
And she was the lead author of the current
3:05
version of the College of
3:07
Gastroenterology's colorectal cancer
3:10
screening guidelines. In
3:11
writing guidelines, it's very
3:14
important to look at the evidence
3:16
We start with very specific
3:19
questions. For instance, one
3:21
of the questions might be at what
3:23
age should we start screening? When
3:25
does the benefit start to accrue?
3:27
A second question might be which modality
3:29
should we be using? A
3:31
third one might be how often we
3:33
should be screening and
3:35
what benefit can we expect. The
3:38
idea is then you take all the
3:40
evidence and kind of
3:42
weigh it to answer your questions and
3:45
come up with a recommendation. The
3:47
new guidelines, the ones Ospuka authored,
3:49
say that most people between forty
3:52
five and seventy five years old, should
3:54
get checked for colorectal cancer with
3:56
a colonoscopy every ten
3:58
years, or a stool sample
4:00
analysis once a year.
4:05
There's a good reason colorectal cancer
4:07
screening is so strongly recommended.
4:09
A
4:10
lot of colon cancers arise
4:12
in precursor lesions called polyps.
4:15
So the whole idea is
4:17
essentially to one, find cancers
4:19
at early stages before people are
4:21
symptomatic. Because when detected
4:23
at early stages, the prognosis
4:26
is excellent. In fact, it's one of the
4:28
cancers where we can actually use
4:30
the word cure. People can have a normal
4:32
life expectancy. if the cancer
4:34
is found and resected early.
4:35
And the second goal
4:37
is to detect these precursor lesions
4:40
so that by removing them,
4:42
we could derail that cancer. So we
4:44
can actually talk about cancer prevention in
4:47
that context. The
4:48
kind of screening asthma describes where
4:51
precancerous polyps are definitely
4:53
found and removed is colonoscopy.
4:55
She says colorectal cancer
4:58
stands alone across all of
5:00
oncology in terms of how precisely
5:02
we can look for it. Colon cancer
5:04
is unique compared
5:05
to other cancers where we can actually
5:08
detect it in these phenoplastic conditions.
5:11
that makes it very different from other
5:13
cancer where we actually look for the cancer
5:15
itself. So in that regard, it's
5:17
one of the more optimistic cancers
5:19
to screen for. Colonoscopy
5:22
has been recommended to screen
5:24
for colorectal cancer since
5:26
the mid nineteen nineties. Uptake
5:28
struggled initially and two thousand,
5:31
only around twenty percent of adults
5:33
over age fifty, then the
5:35
recommended starting age, said they'd
5:37
undergone the procedure. But by
5:39
twenty twenty, that number was
5:41
closer to seventy percent Despite
5:43
its popularity among gastroenterologists, and
5:46
its perceived value in terms of
5:48
finding and even preventing cancer.
5:50
It turns out there are still a lot
5:52
of remaining questions about
5:54
colonoscopy. Starting with
5:56
how beneficial is it compared to
5:58
other less invasive
5:59
screening methods like
6:01
say the fecal immunochemical test
6:04
or fit, which analyzes a
6:06
stool sample.
6:08
That is a million dollar question because
6:10
we don't know the answer. the two
6:12
modalities have never been compared
6:14
head to head in what we consider
6:17
a randomized clinical trial Having
6:19
said that, there are two trials
6:22
ongoing, and one of them
6:24
is actually in the US. I happen to
6:26
be a part of it. It's all across the
6:28
Veterans Affairs hospitals across
6:30
the country. We've enrolled fifty
6:32
thousand veterans and
6:35
randomized them to yearly
6:37
fit or the stool test versus
6:40
a colonoscopy every ten years.
6:42
And as you know, these studies take
6:44
a long time to get
6:45
done, And then the
6:47
outcome is going to be
6:49
risk of dying from colon cancer,
6:51
and we're looking to see if, say,
6:53
colon loss reduces
6:55
it more than fit screening. We
6:57
have to wait ten years before the results
6:59
are available. The last patient was
7:01
recruited in twenty seventeen. So
7:04
in about five years, we'll hopefully
7:06
have some results of which test
7:08
is best. So
7:09
you've been involved in devising guidelines
7:12
for colorectal cancer screening? What's
7:14
the evidence base for them?
7:16
Are we talking about randomized controlled trials?
7:18
Are we talking about observational studies? what
7:21
we look for is randomized controlled
7:23
clinical
7:23
trials because those are
7:25
considered gold standard of research
7:28
studies And in the realm of colon cancer
7:30
screening, fortunately, there have been
7:32
several since the nineteen
7:34
seventies, the largest being
7:37
one in the U. S. that I work with closely
7:39
called the Minnesota fecal alkyl blood trial,
7:41
which truly put screening on the map.
7:44
And then there have been trials
7:46
in Europe with stool testing
7:48
because it was the original
7:50
and the first modality. a
7:52
lot of evidence has gathered
7:54
around stool testing.
7:56
How do guidelines for colon cancer
7:58
screening differ between
7:59
the US and Europe?
8:01
The US tends to be an outlier for
8:04
pretty much the entire world
8:06
in that we use colonoscopy preferentially
8:09
as our colon cancer screening modality.
8:13
There are about fifteen million
8:15
colonoscopies done in the U. S. every
8:17
year everywhere else
8:19
Europe, Canada and other
8:22
places that have organized screening programs
8:24
such as Australia, parts of South
8:26
America, the predominant colon
8:29
cancer testing modality is
8:31
the stool test or the newer
8:33
version of that called fit. And
8:36
only because it's readily available,
8:38
it's affordable, it has great
8:40
evidence behind it, and it lends
8:42
itself nicely to programmatic
8:44
screening. So in Europe,
8:46
the predominant modality is the
8:49
fecal canal blood, and they actually do
8:51
it every other year. whereas in
8:53
the US,
8:54
our recommendations are every year.
8:56
And why is it that we rely more heavily on
8:58
colonoscopy in the US compared to
9:00
other countries? we're
9:01
a resource, wealth nation, and
9:04
we like to use the best and
9:06
the strongest resources we
9:08
have. And in two
9:10
thousand one, Medicare agreed to
9:12
pay for colonoscopy as a screening
9:14
tool, and truly that's
9:16
when the use soared.
9:18
Is
9:18
there variation in the
9:20
quality of colonoscopies or
9:23
in the people who perform them?
9:25
And useful for telling us something about whether
9:27
or not the identification and
9:30
removal of polyps may have
9:32
a causal effect on
9:33
outcomes. Yes. And that's a
9:36
very, very crucial point. So colonoscopy
9:38
as effective as it can
9:40
be is heavily, heavily
9:42
operator dependent. And as a
9:44
result, about twenty years
9:46
ago, we developed a set of quality
9:49
indicators for colonoscopy. One
9:51
of those indicators is, for
9:53
instance, completion of the colonoscopy
9:55
So we've set a bar that the completion rate
9:57
for screening colonoscopy needs to be
9:59
ninety five percent or higher for
10:02
an endoscopist. Another bar
10:04
is the number of precancerous
10:06
polyps detected by an
10:08
endoscopist needs to be
10:10
at the minimum. twenty five percent
10:12
or more. So those are all indicators
10:14
that tell us if the
10:17
exam was able to
10:19
detect the kinds of things that we want
10:21
to use the exam for.
10:23
Prior to the last month, what was the
10:25
evidence based for the
10:27
benefits of colonoscopy? there
10:29
was great observational data
10:31
from pretty large studies
10:33
that suggested the benefit
10:35
of screening
10:36
using colonoscopy
10:38
could reduce cancer
10:41
incidents by about eighty five
10:43
percent and the risk
10:45
of
10:45
dying from colon cancer by somewhere
10:47
between
10:47
sixty percent to eighty percent But
10:50
as, you know, observational studies
10:53
inherently have design issues,
10:56
and they have a lot of biases because
10:58
the comparison groups aren't
11:01
balanced. So therefore, some
11:03
of the results you're getting with the group
11:05
undergoing colonoscopy, maybe
11:07
not just the colonoscopy, but other
11:09
factors that make them healthier
11:12
or less like you develop cancer or
11:14
die from it. And the field
11:16
has been looking for randomized
11:18
controlled trial evidence on
11:20
colonoscopy.
11:23
And we've been waiting for a trial on
11:25
colonoscopy and asking for
11:27
one and our wish was answered.
11:29
and it's one of those be careful what you ask
11:31
or because you may or may
11:33
not wanna know what it says. What
11:35
does it say? After the break,
11:38
we'll talk with the lead author of this long
11:40
wish for randomized control
11:42
trial on colonoscopy that
11:44
has stirred strong emotions on
11:46
both sides of the debate. People
11:49
start doing things because they've are
11:51
convinced for some reason
11:53
with limited data that this is the right
11:55
thing to do. I'm about Pugena,
11:57
and this is freaking Abbott's MD.
12:03
Preeconomic
12:07
MD is sponsored by indeed
12:09
ambitious hiring goals for the last quarter
12:11
of twenty twenty two with
12:13
powerful hiring partner, big goals are
12:15
no big deal. You need indeed.
12:18
Indeed. Indeed is the hiring platform
12:20
where you can attract interview,
12:22
and hire all in one place.
12:25
Don't spend hours on
12:27
multiple job sites looking for
12:29
candidates with the right skills, when
12:31
you can do it all with indeed.
12:33
Find top talent fast with indeed
12:35
suite of powerful hiring tools,
12:37
like indeed instant match,
12:39
assessments, and virtual interviews.
12:41
Kate waiting,
12:42
Indeed's US data shows
12:45
over eighty percent of Indeed employers,
12:48
find quality candidates whose
12:50
resume on Indeed matches their job
12:52
description the moment they
12:54
sponsor a job. In
12:55
Star App can shine before the
12:57
interview with over one hundred and
12:59
thirty five, indeed, rated
13:01
assessment tests they can take.
13:03
From cooking, to coding,
13:06
joined over three million businesses
13:08
worldwide using indeed
13:10
to hire great talent
13:12
fast. Start hiring now with a
13:14
seventy five dollars sponsored job
13:16
credit to sponsor your job
13:18
post at indeed dot com slash
13:20
economics. offer good for a
13:22
limited time. Claim your
13:24
seventy five dollar credit now at
13:26
indeed dot com slash economics.
13:29
Indeed
13:29
dot com com slash beeconomics.
13:32
Terms and conditions apply. Need
13:34
to hire. You need indeed.
13:39
Take
13:40
charge of your purchases with a cash
13:42
back rewards credit card from Pathway's
13:44
Financial Credit Union. Make purchases
13:46
anywhere anytime. and
13:48
get one point five percent cash back
13:50
for each dollar you spend. That's
13:52
right. One
13:53
point five percent cash back for each
13:55
dollar you spend. Visit pathway
13:57
c
13:57
u dot com. That's pathway c u
13:59
dot
13:59
com to find out more about the
14:02
cashback
14:02
rewards credit card from Pathways. Pathway's
14:04
Financial Credit Union. Your true
14:06
financial
14:06
partner, member NCA, Equal
14:09
Housing Lender.
14:11
My
14:15
name is Michael Brett Hauraur.
14:17
I am a professor of medicine
14:19
here at University of Oslo and
14:21
the Oslo University Hospital where
14:23
I am also a gastroenterologist. Doctor
14:26
Michael Brett Howard, lives in Norway,
14:28
which is different from the US in
14:30
a lot of ways. But for the
14:33
purposes of this discussion, there's
14:35
one
14:35
difference that stands out. Most
14:37
countries here in Europe do not
14:40
actively recommend colonoscopy as
14:42
a primary screening test for the general
14:44
population because people think it's too
14:46
invasive It's too costly
14:49
and there was a lack of
14:51
randomized trials that can
14:53
quantify the
14:54
benefit of that screening test colonoscopy until
14:57
we published our study. That
14:59
study is titled, Effect of
15:01
colonoscopy screening, on
15:03
risks of colorectal cancer and
15:05
related death, and it was published in
15:07
October in the New England Journal
15:09
of Medicine. It's findings
15:11
rattle gastroenterology, especially
15:13
in the US because they suggest
15:15
colonoscopy may not be
15:17
as effective as previously
15:19
believed when it comes to detecting
15:21
colorectal cancers and
15:23
also reducing deaths from the disease.
15:26
On its face, This one's countered
15:28
to what previous research has shown.
15:31
Is this study a game changer?
15:34
As Osmo Shotcut told us earlier,
15:36
To do a proper randomized controlled
15:39
trial of colorectal cancer screening
15:41
methods, you'd need a lot of time.
15:43
Which Michael and his team had?
15:46
We
15:46
started to plan this trial fifteen
15:48
years ago. We had our first meeting in
15:50
two thousand five. And
15:52
already then, we
15:55
had a strong interest in
15:57
finding out what the benefits and
15:59
the harms are
15:59
of colonoscopy screening for
16:02
the general population to be used as
16:04
a collateral cancer screening tool. So we
16:06
wanted to find out because we thought,
16:08
well, we would like to know with the
16:10
best methodology available, which
16:12
are randomized trials. how large
16:14
is the benefit as compared to the
16:16
harms? And then we set
16:18
out to plan this trial, which happened to
16:20
be very, very large and needed
16:22
to run for a long, long time.
16:24
Tell me about the study design, where the
16:26
patients were, how they were
16:28
recruited, what was done,
16:30
all that. The
16:30
research question we ask is,
16:33
what are the benefits of
16:35
introducing a screening program for
16:37
a general population of people
16:39
around sixty years of age, which is usually
16:41
a good screening age for colorectal cancer.
16:44
What is the effect of such a
16:46
population screening program? we looked
16:48
for cities and
16:50
regions in Europe where
16:52
such screening was not yet introduced
16:54
because we wanted a control group
16:56
of people who did not get that
16:58
test. And we found four
17:00
regions in Europe, one region
17:02
in Norway and the south of the country and
17:04
one region in Sweden in the middle of Sweden.
17:07
One region in Poland and
17:09
two regions in the Netherlands. We
17:11
set up colonoscopy centers.
17:13
And we built an infrastructure
17:15
and then we randomized
17:18
all the people to either
17:20
get an offer for a colonoscopy
17:22
or no offer at all. And
17:24
there were about ninety five thousand
17:26
people living in these cities where we
17:28
were with the trial. and
17:30
one third got an invitation for a colonoscopy and
17:33
the other two thirds did not get an
17:35
invitation. And then, of
17:36
course, We did the examinations,
17:38
which took us about four years.
17:40
And then we
17:41
followed everybody in both groups, the people
17:43
who got the invitation and the people who
17:45
didn't get an patient followed them for
17:47
ten years with regard to the risk
17:49
of getting colorectal cancer and the risk
17:51
of dying from colorectal cancer.
17:53
And
17:53
what was the standard of care for patients in
17:56
those regions before the trial was
17:58
conducted? The
17:59
standard of
17:59
care was no screening which
18:02
was important for us. There was no
18:04
screening program set up by the
18:06
government or the local authorities, and
18:08
there were no private endoscopy
18:10
clinics. as you would have in the US. This
18:12
is different. This is Europe. So it's a public
18:14
healthcare system, and so there will be no
18:17
availability for people in the control group
18:19
to get
18:19
screening. Obviously, there were services if
18:21
they were referred by their JP
18:24
for some complaints that was available
18:26
but not colonoscopy is for screening in the
18:28
usual care group. So
18:29
in the usual care group, no access
18:31
to colonoscopy really unless they had
18:33
symptoms like bleeding from below or
18:35
something like Correct. Were they doing stool testing
18:37
or any other form of non invasive
18:40
testing? No. No.
18:41
No testing at all. No.
18:43
Nothing. many
18:44
people got colonoscopies in the
18:47
invitation group. The participation
18:48
rate of the people who got an invitation
18:50
was different in the four different countries.
18:52
So here in Norway, where I'm sitting right now,
18:54
sixty percent, six zero of
18:56
all the people who we invited
18:59
said yes, and underwent a colonoscopy, which
19:01
is a very high number. In the
19:03
other countries, it was lower. In Poland, it
19:05
was thirty three percent of
19:08
the people in Sweden, it was about forty two percent
19:10
and in the Netherlands, which is not part of
19:12
the current paper. It was down to
19:15
twenty two. So
19:15
between twenty two and sixty percent of the
19:17
people showed up for their colonoscopy. And
19:20
then what did you find?
19:21
We did all these
19:23
colonoscopies, about thirteen thousand
19:26
colonoscopies. We removed a lot of
19:28
polyps. We found some cancers, obviously,
19:30
some early cancers. over
19:32
here, we have all these registries
19:34
that follow people over time. So
19:36
it was easy for us to check
19:39
if people got cancer or if
19:41
they died of cancer. We found
19:43
that after ten years, the
19:45
people in the Sweden group the
19:47
folks who did not get an offer for screening
19:49
one point two two percent got colorectal
19:51
cancer as
19:52
compared to zero point nine eight percent
19:54
in the screening group, which is
19:56
a relative risk reduction of eighteen
19:58
percent, one eight
19:59
percent. That's the main outcome. So
20:03
that of course includes all the people
20:05
who were randomized to screening
20:07
but did not show up.
20:09
We also did so called pair
20:11
protocol analysis, where you only look at
20:13
the people basically got the
20:15
colonoscopy as compared to the people in the
20:17
control group. Now that is a
20:19
tricky analysis to do because
20:21
you always have election bias of the people
20:23
who show up versus the control
20:25
group. There is more uncertainty to these
20:27
estimates. However, the results were,
20:29
of course, more favorable for screening.
20:31
So for risk of colorectal cancer,
20:33
the effect was thirty one
20:35
percent reduction. Then we look at
20:37
death of colorectal cancer there, the risk
20:40
of all dying of colic cancer
20:42
was low. Even in the
20:44
control group, it was not higher
20:46
than zero point three percent,
20:48
which is
20:48
a very low risk. and the
20:50
reduction in the screening group was not
20:53
different and it
20:54
was zero point one five.
20:57
percent in
20:57
the per protocol analysis, which is
20:59
a fifty percent five-zero decrease,
21:02
but all on very low levels. The
21:04
risk of dying from molecules, it was
21:06
very low. in both groups.
21:08
And then what did you find for the
21:10
overall or all cause mortality?
21:12
The all cause mortality was not
21:14
different between the groups. About eleven percent
21:17
of people died of any cause
21:19
in both groups, so there
21:21
was no difference.
21:23
So how would you
21:24
then put together the findings? You
21:27
found an eighteen percent reduction
21:29
in colorectal cancer for
21:32
those people who invited, but you didn't
21:34
find a statistically significant reduction
21:36
in colon cancer mortality for
21:38
that group. What I would say for colon cancer
21:40
risk is that the real effect
21:42
is somewhere between eighteen percent and
21:44
thirty one percent risk reduction. And
21:47
then, I think it's very important to talk about the
21:49
absolute risks and the absolute risk
21:51
reductions. But I want to know, okay,
21:53
should I do this? I would like
21:55
to know, so what is my risk to start
21:57
with? What is my risk of getting this
22:00
disease? if I buy a car, for
22:02
example, when I walk into the shop
22:04
and the car dealer says, hey, you can buy this car
22:06
for fifty percent off, I would still like
22:08
to know the price of the car.
22:10
Right? I
22:10
don't buy the car just with that piece
22:12
of information that it's fifty percent reduced.
22:15
So I think we should start at saying, well,
22:17
look, the risk
22:18
of getting
22:19
this disease over ten years
22:21
is one point two percent. If you find
22:23
this risk interesting enough to do something about
22:25
it, if you go to colonoscopy, you can probably
22:27
reduce it to let's say, zero point
22:30
eight or zero point nine, somewhere around
22:32
there. And that
22:32
I think is the information we need to give
22:34
to patients or to people. conscious decision
22:37
about if they want to do this
22:39
or not. It's a hallmark of a
22:41
clinical study when I see
22:44
economists tweeting about a
22:46
study. And one of
22:48
the points that I saw a lot of my
22:50
colleagues making that,
22:52
so forty two percent of those people
22:54
ended up getting a colonoscopy. If
22:56
you look at that group on average and you compare
22:58
it to the usual care group on average,
23:01
and you don't find
23:03
a statistically significant reduction
23:05
in colon cancer mortality
23:08
is it because the intervention, in this
23:10
case, the colonoscopy wasn't effective,
23:12
or because only forty two
23:14
percent of people in the group got it.
23:16
And in economics, the way we deal with
23:18
that is we view that randomization as
23:21
an instrument or an instrumental
23:23
variable. It's not perfect. But you
23:26
basically are randomizing people to a higher
23:28
probability of getting the
23:30
treatment. And if you do that,
23:32
the quote unquote effect of
23:34
colonoscopy screening is
23:36
actually quite large. I
23:37
understand that economists think
23:39
mostly about death mortality, and we
23:41
have no difference there. So you could just say,
23:43
well, except to close, there's no differences, doesn't
23:46
work. I don't think it's that
23:48
simple. Here in this study,
23:50
The most interesting endpoint I think
23:52
is incidence. So the risk
23:54
of disease. And why do I think that? Where?
23:56
Two reasons. Number one, I
23:59
think it may be a little early to
24:01
see the full effect of
24:03
colonoscopy on death because it takes time from
24:05
people getting the disease until
24:07
dying from it. everybody thought when we designed
24:09
the trial that ten years would be enough,
24:11
we were probably wrong, not just
24:13
we making the study, but everybody in the
24:15
field, it probably takes longer. we
24:17
will see that because we will follow these
24:19
people longer. The other argument, however, is
24:21
that colonoscopy screening
24:23
is intended to
24:26
prevent colorectal cancer.
24:28
So
24:28
it's intended to reduce the risk
24:30
of getting the disease.
24:33
Therefore,
24:33
the death endpoint is only a logical
24:36
consequence of reducing the
24:38
incidents. So
24:38
incidence is for me the
24:40
more interesting endpoint for this
24:43
study at this time with this
24:45
screening instrument. I would guess that if you follow these patients over
24:47
five years or ten years, you might
24:49
actually find in that case that
24:51
there is a statistically significant
24:54
reduction in death from colorectal
24:56
cancer. And I'm curious, what's your
24:58
prediction? My
24:58
prediction also would be for colorectal cancer
25:01
death that the effect will
25:03
get larger. Until a certain point in time,
25:05
I don't know where that time point is. If it's twelve
25:07
years or fifteen years or twenty years, nobody
25:09
knows. But my guess would
25:11
be that We
25:12
may see one in two years time,
25:14
one in five years time. That will be
25:16
my prediction, although I'm far
25:19
from certain. And how large that will be? I have
25:21
no idea. What has been the
25:23
feedback
25:23
that you've gotten about the work?
25:25
Like, what are people saying? And how would you
25:27
respond? There
25:28
were some heated discussions the first days after
25:30
the publication, especially in the
25:32
United States, but things have calmed down and we
25:34
have had some nice conversations. What
25:37
everybody says and I certainly appreciate is
25:40
that this is a good
25:42
study and
25:43
we did it well and it was well designed and
25:46
executed which we like
25:48
obviously because we put a lot of
25:50
thinking and emphasis into
25:52
that and that it moves the
25:54
field forward because this is the first randomized trial
25:56
with this screening tool. I think
25:58
in the GI
25:59
community, there was disappointment. was disappointment
26:02
Because
26:03
especially in the US, everybody
26:06
thought at least in the field that
26:08
colonoscopy would have a
26:10
far higher benefit.
26:12
People
26:12
start doing things because they are
26:14
convinced for some reason with
26:17
limited data that this is the right thing
26:19
to do. The
26:21
study
26:22
was what we have been
26:24
waiting for and asking for
26:26
the biggest criticism about colonoscopy
26:29
screening is that we did not
26:31
have randomized clinical trials
26:34
showing the benefit or how
26:36
effective it was. Dr. Asma Shahcott
26:38
again. Observation studies are
26:40
always optimistic and more
26:42
rosy than real life. And we've
26:44
seen this time and time again. So
26:46
that's why we do randomized trials. We
26:48
do expect the differences from
26:51
observational studies to shrink, but nobody
26:53
quite expected it to be as low
26:55
as twenty percent
26:57
The lack of reduction
26:59
in colon cancer mortality, I think,
27:01
is
27:01
somewhat premature
27:03
and
27:03
perhaps shouldn't
27:06
have been included in the main study
27:08
because for colon
27:10
cancer mortality, the study is growsly
27:13
underpowered. So just
27:15
because we don't see a difference doesn't mean
27:17
there isn't one, but
27:19
that's lost
27:19
when you read the results.
27:21
Do you think people are going
27:24
to shun colonoscopy now or
27:26
could it even have, like, the opposite effect because
27:28
of all the attention they got? So
27:29
there was a large
27:32
public health outcry and also from the
27:34
medical community. What I'm seeing
27:36
is exactly what you alluded to
27:38
is such a strong
27:40
reaction that it might put screening
27:42
on the forefront of people's minds
27:44
and invoke a discussion
27:46
with their providers, give them that extra
27:49
nudge to think about it and then perhaps
27:51
even schedule it. So we might
27:53
see some unanticipated benefits
27:55
of this. And how
27:56
would doctor Michael Brett Hauer
27:59
advise his
27:59
patients, I would start
28:02
with explaining
28:03
them their risk. And
28:06
they're very good calculators online where
28:08
you can actually calculate your risk
28:10
of getting colorectal cancer when they would
28:12
come up with the risk of, let's say, one
28:14
percent over the next ten years. If
28:17
the same, tell me what I can do and how much
28:19
I can reduce, then I would say, okay, you can go
28:21
from one percent to let's say,
28:23
a zero point eight percent which is a twenty percent
28:25
risk reduction if you come to me next week in
28:27
the office and we do a colonoscopy. And
28:29
then I will explain to them,
28:31
the harms they need to take into
28:34
consideration, so perforation and
28:36
bleeding. And then finally, I would explain
28:38
to them what a colonoscopy entails.
28:40
all the things that are involved with it.
28:42
And then I think at the end, I will
28:44
tell them now you make your decision.
28:47
And some people would say yes, and some
28:49
people would say no. And
28:50
that is really what is called
28:53
shared decision making between a
28:55
doctor and patient. They need to understand the harms and the
28:57
burdens, and then they need to make a decision. And
28:59
that's not my decision. If that's
29:01
theirs.
29:03
So what happened when it was
29:05
his decision?
29:06
I had a colonoscopy. I
29:09
thought for me personally
29:11
it was a
29:11
deal that I was comfortable with with the
29:13
numbers that we just talked about, it was
29:16
negative. So if if I'm gonna do another one in
29:18
ten years, I'm not sure I will look
29:20
up the numbers
29:21
then and decide. As
29:24
doctor Michael Brett Howard said,
29:26
once you start doing something like
29:28
colonoscopy based on limited
29:30
data, it can be hard to slow
29:32
that momentum. Colonoscopy is
29:34
pretty safe overall, but it
29:36
does come with risks. both
29:38
from the test itself and from the anesthesia.
29:41
There's also the preparation which can
29:43
be unpleasant. We should hone
29:45
invasive tests like colonoscopies
29:48
to the same standard we hold drugs.
29:51
Randomized controlled trials. Studies
29:53
like Michaels and the research
29:56
Ospuma is doing, bring us
29:58
closer. For that, we owe them
29:59
and their colleagues a dead of
30:02
gratitude. That's it for
30:04
today's show. I'd like to thank my guests, Ospo
30:07
Shawcutt and Michael Brett Howard. And
30:09
here's an idea for you based on my conversations
30:11
with them. If you wanna study
30:13
the effectiveness of this chronoscopy, you
30:16
need randomization, either in a
30:18
trial like Michael did or
30:20
natural randomization, like we talk
30:22
about all the time on this
30:24
show. So what about this? We
30:26
know doctors are affected when their patients
30:28
have a bad outcome. In
30:30
reaction, they might change how they
30:32
practice. When a patient is diagnosed with
30:35
colon cancer, their physician
30:37
might start encouraging colon
30:39
cancer screening more often
30:41
to other patients. of the saliance
30:43
of that recent diagnosis. You
30:46
might then see higher
30:48
colonoscopy rates in these
30:50
compared to eligible patients the doctors
30:52
saw before the colon cancer
30:55
case. If that happens, you'd have a
30:57
natural experiment to
30:59
study if that greater screening led
31:01
to lower cancer related mortality.
31:03
You'd need lots of data over lots
31:05
of years, but it's doable.
31:08
what ideas do you have? Did you hear about the colonoscopy
31:11
study in the news? Did it change
31:13
the way you thought about screening?
31:15
Email me your thoughts because I'd love to hear them.
31:17
I'm at boppu at freakonomix
31:19
dot com. That's BAPU
31:22
at freakonomix dot
31:24
com. In speaking of research,
31:26
we know that people typically take
31:28
advice on which podcasts to
31:30
listen to based on recommendations from
31:33
people they know and trust.
31:35
So go find someone who knows
31:37
you and trust you and tell them to
31:39
subscribe to FreakonomicsMD. coming
31:42
up next week. We had been caught flat footed.
31:45
We had missed a window of opportunity to
31:47
really get this virus under control. If
31:49
that voice
31:49
sounds familiar, it's probably because you've
31:52
heard doctor Ashishja before.
31:54
First, as a public health
31:56
expert on TV and Twitter, and
31:58
now as the White House's COVID
32:00
nineteen response
32:02
coordinator. He's been
32:04
sending pandemic for a long time.
32:06
And next week, he'll tell us what
32:08
it's really like to try to manage
32:10
one from the front lines.
32:12
One of the things that you realize is you can have
32:14
app absolutely brilliant ideas
32:16
that just cannot be implemented. That's
32:18
all coming up next week on FreakonomicsMD.
32:20
Thanks again for listening.
32:22
Fri
32:24
reopenomics MD is part of
32:26
the Freakonomics Radio Network, which
32:28
also includes freak Economics Radio, no
32:31
stupid questions, and
32:33
people I mostly admire. All
32:35
our shows are produced by Stitcher
32:37
and run by radio. You can find us on Twitter
32:39
at doctor BabuPod. This
32:42
episode was produced by Julie
32:44
offer and mixed by Jasmine Klinger, with help
32:46
from Jeremy Johnston. Our staff
32:48
also includes
32:49
Neil Carruth, Gabriel Rob,
32:51
Greg Ripon, Leark
32:52
Foutage, Rebecca Lee Douglas, Morgan
32:55
Levy, Zach LaPinski, Ryan
32:57
Kelly, Catherine Mancurr, Eleanor
32:59
Osbourne, Daria Klineert, Emma
33:01
Terrell, Alina Coleman and
33:03
Steven Dubner, original
33:05
music composed by Louise Gara. If
33:07
you like this show or any
33:09
other show in the Freakonomics Radio
33:11
Network, please recommend it to your family
33:13
and friends. That's the best way
33:15
to support the podcast you love.
33:18
As always, thanks for
33:20
listening. The
33:24
actor Ryan
33:24
Reynolds got a colonoscopy. And I
33:26
was thinking, I'm gonna look at what happens
33:28
the colonoscopy rates after Ryan
33:31
Reynolds got his broadcast
33:33
at colonoscopy. His experience
33:35
went
33:35
really well and there were no
33:37
complications that we know of
33:38
then for what it's worth. He's protected
33:41
for a large number of years
33:43
to come. The
33:47
Freakonomics Radio Network,
33:48
the hidden
33:49
side of everything.
33:52
stitcher.
33:56
What
33:59
will it take
33:59
to
33:59
conserve ten billion acres
34:02
of ocean one point six billion
34:04
acres of land and over six
34:06
hundred thousand miles of river. What will it
34:08
take to protect and restore
34:10
natural habitats in over
34:12
seventy countries around the world and in
34:14
all fifty states here at home.
34:16
What will it take?
34:17
You. Together,
34:18
the dinner We will
34:19
make it happen. It's in our nature.
34:22
See how your gift can help. At
34:24
nature dot org, the
34:25
nature conservancy. protecting nature,
34:30
preserving life.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More