We
expect
that
medical
therapy
will
change
and
evolve
with
time.
Good
treatments
will
replace
bad
ones,
and
then
better
ones
will
replace
those.
Antibiotics
have
replaced
arsenic,
and
anesthesia
has
replaced
a
bullet
held
bracingly
between
the
teeth.
Recently,
however,
change
has
occurred
in
surprising
ways.
If
you
have
followed
the
news
about
prostate
cancer
screening,
mammography
for
women
in
their
forties,
hormone
replacement,
cholesterol-‐lowering
medications,
and
stents
for
coronary-‐artery
disease,
you
might
think
doctors
cannot
get
anything
straight.
These
common
practices
were
not
replaced
by
better
therapies;
they
were
found
to
be
ineffective.
In
some
cases,
they
did
more
harm
than
good.
You
might
be
worried
that
some
medical
practices
are
nothing
more
than
fads.
In
some
cases,
you
would
be
right.
We
call
this
phenomenon
“medical
reversal.”
Instead
of
the
ideal,
which
is
replacement
of
good
medical
practices
by
better
ones,
medical
reversal
occurs
when
a
currently
accepted
therapy
is
overturned—found
to
be
no
better
than
the
therapy
it
replaced.
Now,
you
might
argue
that
this
is
how
science
is
supposed
to
proceed.
In
high
school,
we
learned
that
the
scientific
method
involves
proposing
a
hypothesis
and
testing
to
see
whether
it
is
right.
This
is
true.
But
what
has
happened
in
medicine
is
that
the
hypothesized
treatment
is
often
instituted
in
millions
of
people,
and
billions
of
dollars
are
spent,
before
adequate
research
is
done.
Not
surprisingly,
sometimes
the
research
demonstrates
that
the
hypothesis
was
incorrect
and
that
the
treatment,
which
is
already
being
used,
is
ineffective
or
harmful.
Cases
of
Reversal
A
few
people’s
stories
highlight
notorious
examples
of
reversal
in
medicine.
Consider
Samuel
Jones,
a
57-‐year-‐old
gentleman
who
had
a
heart
attack
in
1991.
He
was
admitted
to
the
hospital
and
treated
with
the
standard
medical
regimen,
which,
at
the
time,
included
flecainide,
a
drug
used
to
stabilize
his
heart
rhythm.
In
the
late
1980s
and
early
1990s,
flecainide
and
its
sibling
drugs
were
widely
used
with
the
intention
of
suppressing
extra
heartbeats
(called
premature
ventricular
contractions,
or
PVCs)
and
preventing
death.
The
logic
behind
the
use
of
these
drugs
was
ironclad.
PVCs
are
strongly
associated
with
sudden
death.
The
more
you
have,
the
more
likely
you
are
to
die.
And
there
was
no
better
medicine
for
suppressing
PVCs
than
flecainide.
How
could
it
not
save
lives?
Some
cardiologists
were
so
confident
in
the
drug
that
they
would
not
let
their
patients
enter
studies
of
this
medication.
Such
studies
required
that
some
patients
be
randomized
to
receive
a
placebo
instead
of
flecainide.
In
all
good
conscience,
how
could
they
risk
having
their
patients
deprived
of
this
lifesaving
drug?
Unfortunately,
flecainide
did
not
work.
In
1992,
a
large
study
called
the
Cardiac
Arrhythmia
Suppression
Trial
(or
CAST
trial)
showed
that
flecainide,
as
well
as
a
similar
drug,
decreased
PVCs
as
expected
but
also
increased
patients’
chance
of
dying.1
The
finding
was
devastating.
It
changed
not
only
how
we
treat
heart-‐attack
patients
but
how
we
evaluate
medical
therapies
altogether.
CAST
taught
us
that
even
the
most
careful
reasoning
and
the
best
scientific
models
do
not
guarantee
an
effective
clinical
treatment.
What
works
in
the
lab,
or
on
a
computer,
or
in
the
head
of
the
smartest
researcher
does
not
always
work
in
a
patient.
But
now,
more
than
20
years
later,
this
is
a
lesson
that
many
physicians
and
leading
researchers
still
have
not
really
learned.
Let’s
consider
another
medication,
this
one
for
high
blood
pressure
(hypertension).
Hypertension
remains
one
of
the
most
common
ailments
in
America.
Although
most
people
with
hypertension
are
entirely
without
symptoms,
there
is
no
doubt
that
hypertension
strongly
predicts
an
increased
risk
of
stroke,
heart
disease,
and
death.
It
is
truly
a
silent
killer.
People
with
naturally
lower
blood
pressure
live
longer,
and
people
who
use
certain
medications
to
lower
their
blood
pressure
can
decrease
their
risk
of
death.
One
of
the
first
drugs
used
to
treat
hypertension
was
atenolol.
Atenolol,
a
member
of
the
beta-‐blocker
class
of
antihypertensives,
dramatically
lowered
blood
pressure,
and
for
decades
it
achieved
the
rarefied
status
of
“trial
standard,”
meaning
that
you
had
to
show
your
new
drug
was
as
good
as
atenolol
in
order
to
get
it
on
the
market.
In
2002
something
really
unsettling
(or
maybe
even
terrifying)
happened.
After
atenolol
had
been
used
to
treat
hypertension
for
nearly
20
years,
the
results
of
the
LIFE
trial
were
published.
This
trial
compared
atenolol
to
a
newer
drug,
losartan.
People
who
took
losartan
had
fewer
strokes
and
lived
longer
than
those
that
took
atenolol.
At
first
glance
it
seemed
that
we
had
come
up
with
a
better
antihypertensive;
this
seemed
like
an
example
of
replacement.
Somewhat
surprising
was
that
both
drugs
lowered
blood
pressure
the
same
amount.
The
question
was,
did
losartan
beat
atenolol
because
losartan
is
better
or
because
atenolol
is
actually
ineffective?
The
tantalizing
finding
in
the
LIFE
study
was
extended
in
2004,
when
a
pooled
analysis
of
all
people
who
took
atenolol
versus
sugar
pills
(placebos)
in
trials
showed
that
atenolol
was
no
better
than
the
placebo.
Atenolol
did
lower
people’s
blood
pressure,
but
it
did
not
decrease
people’s
risk
of
dying
or
of
having
a
heart
attack.
Let
that
sink
in.
A
drug
that
was
widely
used,
accepted
as
the
standard
of
care,
had
made
millions
and
millions
of
dollars
for
its
manufacturers,
and
had
made
high
blood
pressure
the
subject
of
dinnertime
conversation,
did
not
make
you
live
a
single
day
longer.
A
recent
study
showed
that
metoprolol
(another
beta-‐blocker)
is
no
better
than
atenolol.
If
you
have
taken
a
beta-‐blocker
for
high
blood
pressure,
you
may
have
shaved
a
few
percentage
points
off
your
risk
of
stroke,
but
you
did
not
extend
your
life.
In
the
world
of
blood-‐pressure
medicines,
you
took
a
pill
that
does
not
work.2
If
the
only
place
modern
medicine
erred
was
with
medications,
we
would
be
fortunate.
But
there
are
also
medical
procedures,
some
used
for
decades
by
practitioners
who
have
reaped
huge
financial
rewards,
that
have
been
shown
not
to
work.
Let’s
examine
another
person.
Anthony
Baker
is
a
55-‐year-‐old
mechanic
who
thinks
of
himself
as
active.
His
job
in
a
garage
requires
him
to
be
on
his
feet
most
of
the
day.
In
the
summer
he
mows
his
lawn
every
week—with
a
push
mower,
no
less—and
in
the
winter
he
shovels
snow
from
his
walk
after
storms.
Anthony
does
not
have
any
health
problems—well,
at
least
that
he
knows
of—and
he
sees
his
doctor
yearly.
At
his
last
checkup,
he
was
told
that
he
was
in
fine
health,
aside
from
the
yearly
advice
to
quit
smoking.
Over
the
past
few
months,
however,
Anthony
has
noticed
that
when
he
exerts
himself,
he
feels
an
ache
in
his
chest.
This
“pressure”
resolves
after
about
five
minutes
of
rest.
Anthony
calls
his
doctor
to
tell
him
about
the
chest
tightness.
His
doc
is
concerned
that
Anthony
might
have
developed
coronary-‐artery
disease
and
that
his
symptoms
represent
angina.
Angina
is
cardiac
pain
that
occurs
when
the
heart’s
demand
for
oxygen
outstrips
its
supply.
It
often
happens
during
exertion
because
that
is
when
the
demand
for
oxygen
is
greatest.
Anthony’s
doctor
orders
a
stress
test,
a
test
that
examines
the
heart’s
function
during
exercise.
The
test
shows
that
Anthony
has
a
narrowing
in
his
circumflex
coronary
artery
(one
of
the
three
coronary
arteries
that
supply
oxygenated
blood
to
the
heart).
Based
on
the
results
of
the
test,
the
doctor
diagnoses
Anthony
with
coronary-‐artery
disease,
tells
him
to
stop
smoking,
starts
him
on
a
few
medications
(metoprolol,
a
beta-‐blocker,
to
lessen
the
heart’s
oxygen
demand;
atorvastatin,
a
cholesterol-‐lowering
medication;
and
aspirin).
He
also
schedules
him
to
see
a
cardiologist
the
following
week.
When
Anthony
sees
the
cardiologist,
he
has
not
yet
started
his
medications
and
reports
one
episode
of
chest
pain
that
occurred
while
he
was
walking
his
dog
over
the
weekend.
The
cardiologist
schedules
Anthony
for
a
coronary
angiogram.
During
an
angiogram
dye
is
injected
into
the
coronary
arteries
to
assess
their
patency—to
determine
whether
or
not
there
are
blockages.
The
angiogram
shows
a
90
percent
blockage
of
his
circumflex
artery.
During
the
procedure
the
cardiologist
places
a
stent,
a
metal
tube,
in
the
artery
and
the
90
percent
blockage
disappears.
He
gives
Anthony
another
prescription,
this
one
for
clopidogrel,
a
blood
thinner,
to
add
to
the
rest
of
his
medications.
One
week
later
Anthony
sees
his
regular
doctor
and
is
pleased
to
say
that,
despite
having
shoveled
six
inches
of
snow
the
previous
day,
he
is
feeling
the
best
he
has
felt
in
years.
He
has
no
pain
and,
in
fact,
he
recognizes
that
he
was
probably
having
more
chest
symptoms
than
he
was
aware
of
before
the
procedure.
So,
was
Anthony
cured
by
the
cardiologist?
It
is
undeniable
that
Anthony
feels
better.3
You
may
also
be
thinking
that
Anthony
narrowly
made
it—that
if
he
had
not
had
the
artery
opened,
he
would
have
had
a
heart
attack.
That
because
of
the
stent,
Anthony
will
live
longer.
And
though
the
procedure
cost
more
than
$10,000,
on
the
whole,
it
must
be
worth
it.
But
what
if
we
told
you
that
Anthony
is
no
less
likely
to
have
a
heart
attack.
What
if
we
add
that
Anthony
will
not
live
a
single
day
longer.
And,
perhaps,
most
concerning,
what
if
we
tell
you
that
in
12
months,
Anthony’s
chest
pain
will
be
the
same,
even
though
he
had
the
stent.
Did
Anthony
really
benefit?
There
is
a
long
list
of
medical
procedures
whose
use
is
based
on
scientific
evidence
that
is,
at
best,
suspect.
Vertebroplasty
is
another
great
example.
For
years,
doctors
struggled
to
treat
people
(mostly
women)
who
suffered
osteoporotic
fractures
of
the
spine.
These
fractures
can
lead
to
chronic
back
pain.
In
the
late
1990s,
a
couple
of
radiologists
had
a
brilliant
idea.
Why
not
insert
a
needle
into
the
fractured
bone
and
inject
medical-‐grade
cement.
The
theory
was
that
the
cement
would
plump
up
the
bone,
the
nerves
would
get
some
extra
breathing
room
as
the
fractured
bone
was
lifted
away,
and
the
pain
would
dissipate.
When
these
pioneers
performed
the
procedure,
vertebroplasty,
on
a
few
dozen
patients,
they
were
amazed
with
the
results.
Patients
immediately
felt
better.
Patients
and
doctors
were
convinced.
In
the
very
early
2000s,
interested
parties
lobbied
Medicare
to
pay
for
vertebroplasty,
and
their
request
was
granted.
In
a
few
years,
tens
of
thousands
of
people
were
having
the
procedure
each
year.
By
the
end
of
the
decade,
vertebroplasty
was
a
billion-‐dollar-‐a-‐year
industry.
There
were
complications—rarely,
the
cement
would
go
somewhere
it
was
not
supposed
to—but
on
the
whole,
this
procedure
seemed
like
a
real
advance.
In
2009
two
groups
of
brave
investigators
put
vertebroplasty
to
the
true
test.
They
enrolled
200
patients.
Half
of
the
patients
got
vertebroplasty;
the
other
half
were
taken
to
the
procedure
room
and
prepped
for
the
procedure;
the
cement
was
opened,
so
patients
could
smell
it;
and
salt
water
was
injected.
Both
groups
of
patients,
those
that
got
vertebroplasty
and
those
that
had
a
sham
procedure,
had
identical
improvements.
Vertebroplasty,
as
it
turned
out,
is
no
better
than
a
placebo.
This
outcome
alone
may
not
convince
everyone
that
vertebroplasty
is
a
bad
thing.
Who
cares
if
the
effect
is
“in
your
head,”
as
long
as
it
works?
Some
with
expertise
in
the
placebo
effect
would
hold
this
position.
Although
vertebroplasty
was
no
better
than
a
sham
procedure,
either
may
very
well
be
better
than
doing
nothing.
In
this
book
we
examine
other
interventions
that
claim
to
decrease
pain
but
perform
no
better
than
a
sham
procedure.
In
many
cases,
both
are
better
than
doing
nothing.
What
does
this
mean?
There
is
something
about
the
acts
of
a
medical
intervention—the
thoughtful
counsel
of
the
doctor,
the
skilled
support
of
the
nurses
and
staff,
the
psychological
comfort
of
acting—that
collectively
make
us
feel
better.
The
downside
is
only
that
the
procedure
is
costly,
may
involve
risk
(either
because
of
the
procedure
itself
or
by
delaying
an
intervention
that
has
intrinsic
benefit),
and
involves
deception.
The
question
then
becomes,
Can
we
reap
the
benefits
of
placebo
treatments
without
the
deception,
the
risk,
and
the
cost?
In
chapter
2
we
discuss
the
placebo
effect
in
more
detail.
For
now,
let
us
at
least
agree
that
the
money
spent
on
the
cement
used
in
vertebroplasty—often
thousands
of
dollars
per
procedure—was
wasted.
Salt
water
would
have
sufficed.
When
medical
reversal
involves
pills
or
procedures,
it
affects
patients
unlucky
enough
to
have
an
illness
and
be
prescribed
a
faulty
therapy.
When
reversal
involves
prevention
efforts
and
screening
campaigns,
millions
of
healthy
people
can
be
affected.
In
the
past
five
years,
two
major
public
health
efforts,
breast
and
prostate
cancer
screening,
have,
to
a
large
extent,
been
overturned.
Let’s
consider
mammography.
In
2009
the
U.S.
Preventive
Services
Task
Force,
generally
considered
the
most
impartial
of
the
hundreds
of
groups
that
produce
guidelines
for
physicians
to
follow
in
their
clinical
practice,
changed
its
recommendation
on
whether
women
in
their
forties
should
undergo
screening
mammography.
The
old
recommendation
was
to
do
a
mammogram
in
this
age
group
every
other
year.
The
new
recommendation
is
not
to
do
it
at
all.
This
reversal
made
a
lot
of
people
very
angry.
An
article
from
November
of
that
year
in
the
Seattle
Times
had
the
provocative
title
“Mammogram
Mania:
Risking
Lives
or
Dollars?
Physicians’
Community
Speaks
Up,
against
New
Recommendations.”
Radiologists,
who
had
the
most
to
lose
(at
least
financially),
protested
the
loudest.
They
accused
the
task
force
of
trying
to
save
money
instead
of
lives.
Was
any
of
this
a
fair
characterization?
Were
the
changes
to
the
guidelines
a
shock?
Not
to
those
who
had
followed
the
medical
literature.
The
truth
is
that
the
guideline
change
was
three
years
in
the
making.
In
2006
a
huge
study
on
exactly
the
question
of
mammographic
screening
for
40-‐somethings
was
published
in
Lancet.
More
than
160,000
women
in
their
forties
were
randomized
to
mammography
screening,
or
not,
and
they
were
followed
for,
on
average,
10.7
years.
It
is
worth
pausing
to
consider
just
how
impressive
a
feat
this
study
was.
The
study
authors
got
more
than
100,000
women
to
participate
in
a
study
of
a
medical
treatment
that
required
yearly
visits
for
over
a
decade.
For
reasons
we
discuss
later,
as
far
as
evidence
in
medicine
goes,
this
is
about
as
good
as
it
gets.
The
authors
of
the
trial
found
that
there
were
fewer
deaths
from
breast
cancer
among
the
women
who
received
mammograms,
compared
to
those
who
did
not,
but
this
difference
did
not
reach
statistical
significance.
In
other
words,
the
difference
was
so
small
that
we
cannot
be
sure
the
difference
was
not
due
to
chance
alone.
What
this
means
is
that
a
few
women
who
got
mammograms
died
of
breast
cancer
and
a
few
women
who
did
not
get
mammograms
also
died.
Neither
group
had
more
women
who
died
of
breast
cancer.
If
you
look
at
the
study
for
what
is
truly
important,
dying,
from
any
cause,
not
just
breast
cancer,
you
find
that
the
death
rate
was
identical
in
the
two
groups.
Therefore,
we
should
not
screen
these
women
with
mammography,
because
in
the
history
of
medicine
no
one
has
ever
shown
that
doing
so
saves
lives.
In
aggregate.
On
the
whole.
Regardless
of
the
financial
impact.
Put
another
way,
what
if
all
the
women
in
their
forties
of
Washington,
D.C.,
got
mammography
and
all
the
women
in
their
forties
of
Chicago
did
not.
Well,
nearly
all
the
women
in
this
age
group
in
both
cities
would
be
alive
at
50,
and
a
few
unlucky
ones
in
each
city
would
have
died
of
breast
cancer.
And
all
the
statisticians
at
the
NIH
in
Bethesda,
Maryland,
and
at
the
University
of
Chicago
would
be
hard-‐pressed
to
say
for
certain
which
city
got
mammograms
if
you
did
not
tell
them.
If
the
statisticians
looked
at
death
rates
for
40-‐something
women
in
both
cities,
they
would
be
identical.
What
led
to
all
the
hand-‐wringing
about
the
guideline
changes
was
not
the
reversal
of
screening
mammography;
it
was
that
doctors
had
recommended
mammography
for
women
in
their
forties
for
all
those
years.
A
costly,
ambitious
medical
program
was
conducted
on
a
national
level
for
years
before
we
had
data
that
said
that
it
worked.
What
we
know
for
sure
is
that
because
of
false-‐positive
mammograms
and
a
phenomenon
called
“overdiagnosis”
(more
in
chapter
4)
the
program
gave
many
women
a
cancer
scare,
a
needle
biopsy,
radiation
exposure,
surgery,
or
worse.
We
believe
that
reversal
is
the
most
important
problem
in
medicine
today.
When
we
doctors
flip-‐flop
on
our
advice
to
patients,
it
usually
is
not
because
the
treatments
stopped
working.
It
usually
is
not
because
someone
discovered
a
harm
no
one
had
previously
noticed.
It
is
usually
because
the
practice
never
worked—we
were
wrong
all
along.
We
promoted
it
before
we
had
properly
studied
it.
We
knew
it
had
some
harms,
sure,
but
we
never
thought
it
lacked
benefits.
This
problem
underlies
people’s
distrust
of
the
medical
establishment
and
is
a
very
important
reason
that
health-‐care
costs
are
soaring
without
any
improvement
in
people’s
health.
Now
reversal
is
not
the
only
problem
in
medicine,
but
it
is
a
common
channel
through
which
many
of
the
problems
run.
Yes,
pharmaceutical
companies
manipulate
data
to
get
drugs
and
devices
approved
and
used.
Yes,
physicians
overtest
and
overtreat
for
fear
of
malpractice
suits
and
because
of
the
financial
benefits
that
this
behavior
brings.
Yes,
hospitals
and
clinics
market
tests
and
treatments
that
offer
little
benefit
because
they
believe
they
will
attract
well-‐insured
patients.
However,
if
we
find
a
solution
to
the
problem
of
medical
reversal,
a
way
of
preventing
the
use
of
therapies
that
do
not
work,
many
of
these
other
problems
will
cease
to
exist.
What
Is
to
Come
In
this
first
part
of
the
book,
we
describe
many
reversals
and
estimate
the
frequency
of
medical
reversal.
We
describe
the
harms
of
our
broken
system.
Many
are
obvious,
but
there
are
some
surprising
harms
of
medical
reversal.
The
repercussions
may
go
further
than
you
imagine.
In
part
2,
we
provide
you
with
an
understanding
of
evidence-‐based
medicine,
an
understanding
that
will
prepare
you
for
the
remainder
of
the
book
and
allow
you
to
question
the
next
report
about
a
“revolutionary
new
treatment.”
Part
3
focuses
on
the
causes
of
medical
reversal.
Money,
of
course,
is
involved,
but
in
this
case
it
is
not
the
entire
problem.
The
seeds
of
reversal
are
sown
at
the
start
of
medical
school
and
are
nourished
throughout
the
medical
industry.
In
part
4
we
propose
solutions
for
the
problem
of
medical
reversal.
Our
solutions
are
not
a
checklist
and
a
vague
call
for
you
to
be
prepared
when
you
see
the
doctor.
We
show
just
where
changes
can
be
made
to
medical
education,
medical
research,
and
the
process
by
which
treatments
are
approved,
so
as
to
eliminate
ineffective
therapies.
We
also
offer
suggestions
about
how
individuals
can
become
better
consumers
of
health
care,
how
to
be
people
who
are
unlikely
to
want,
or
be
given,
treatments
that
do
not
work.
Because
fundamentalism,
in
all
its
forms,
is
dangerous,
we
end
by
discussing
those
times
when
it
is
not
necessary
to
demand
that
practices
be
based
on
ironclad
data.
We
hope
that
through
reading
this
book
you
will
understand
the
scope
of
the
problem
of
medical
reversal—how
reversal
affects
every
part
of
the
doctor-‐patient
relationship.
This
understanding
may
very
well
change
your
opinion
of
“feeling
better.”
It
will
certainly
affect
what
you
do
when
you
feel
well,
in
order
to
stay
healthy,
and
what
you
do
when
you
feel
unwell.
NOTES
1 As
you
will
see
throughout
this
book,
one
of
the
great
sports
in
academic
medicine
is
to
choose
a
name
for
your
trial
that
can
be
reduced
to
a
catchy,
memorable
acronym.
In
our
careers
we
have
seen
not
only
CAST
and
LIFE,
discussed
in
this
chapter,
but
also
NICE-‐SUGAR,
CLEOPATRA,
and
our
favorite,
CABG
Patch.
2
We
are
only
talking
about
hypertension.
If
you
have
heart
failure,
you
really
need
your
carvedilol
or
metoprolol
succinate!
3
In
later
chapters
we’ll
learn
about
other
procedures
that
make
people
feel
better
but
provide
no
real
benefit.
These
examples
will
make
you
reconsider
the
whole
“feeling
better”
issue.