medicine is a wonderful art but it is in
trouble when we doctors go to work every
day
we know that one-third of the care that
we will provide will be completely non
value-added futile that over a third of
the care that we should provide will
actually not be given it will be missed
overlooked and that the third cause of
death in most Western countries is not
road accidents or diabetes it is the
honest mistakes we make as health
professionals going to work we all
physicians and patients deserve more
than that it’s time for us to inject
some precision some certainty into these
work processes imagine imagine that you
had a gift that you could see into the
future that you could look at someone
and know that something bad is going to
happen to him and that you could do
something about it you could give him a
different path that would take him out
of harm’s way
change his life course what would happen
to medicine and healthcare in the future
like this when people can actually see
into the future instead of going to see
your doctor with your ailments she will
call you one morning and say hey how are
you today
great you’ll say oh good please don’t be
alarmed but in 12 hours you will have a
heart attack so why don’t you come over
to the office and we’ll fix you before
the symptoms begin now I know you think
it’s science fiction but it is not it is
happening every day in this country as
well as in others in massive scale let
me give you one example of how such
pre-emptive medicine looks like some of
you have friends or family with kidney
failure going through dialysis you know
how painful and debilitating this
condition is and how wonderful it would
if you could have come earlier years
before before the symptoms began make
small difference and made them not go
through this painful path well in order
to do that you need to predict which
patients will actually go there in five
years they look fine now they feel fine
we should predict who it will be which
is exactly what we did about seven years
ago we used an massive amount of Health
Records that’s held by Israel’s biggest
health provider elite health services
and we used all of this abundance of
data to find those patients and we asked
our physicians to actually approach them
bring them in and give them care at such
early point and our data suggests that
the rate of dialysis have been
dramatically dropping in the last few
years after this intervention in this
subgroup now we’re trying to do the same
now for other diseases for diabetes for
pneumonia even for cancer we’re
fortunate to be able to do this this is
non-trivial you know in other countries
they do not have universal health care
they do not have such massive amount of
electronic healthcare records spanning
back for years and they surely don’t
have the technological hub that we have
in this country I don’t know if you know
but in this country there are more
startups medical patents scientists and
engineers than any other country around
the world so this is why people have
already termed the coin for this country
the digital health nation perhaps you
asked yourself how does this prediction
actually work now
not always can I see into the black box
but in some instances we can and I want
to give you a glimpse so at one time we
wanted to predict which patients would
eventually have osteoporotic fractures
that mean brittle bones that break and
we wanted to identify them early so we
did our prediction thing and the
computer insisted and one of the key
was i disease well obviously a mistake
because I know from med school do those
two have nothing in common
they’re not associated but the computer
insisted and then it hit us it’s not
that the eye disease make bones more
brittle it’s the visual impairments that
make pickup fall and when they fall they
break so computers are better than us in
identifying such subtle patterns and
accumulating them together and that’s
how it actually works so we want to do
more than that more than predicting
diseases we want to be able to select
the right treatment for every patient
but this is not a simple thing this is a
balancing act you see in deciding on the
intensity of treatment too much is bad
too little is dangerous and the optimal
point in the middle weight should stop
pushing is individual and this is
basically the art of medicine to make
the skull we try today to use evidence
but there’s a problem with the evidence
that we have because the evidence that
we get from the literature is average
but the patient in front of us is not
average none of us is average and this
is why medicine is an art every person
is different everyone is unique and they
need personal fitting it was one of the
greatest physicians of all times
that have said the following if it were
not for the great variability between
individuals that medicine might as well
be a science rather than art so it’s the
need to personalize care that what makes
medicine an art forget the average use
your clinical intuition to tailor fit I
know it’s sometimes there’s a risk of
human error and it’s really easy to
follow the average guidelines but this
is
what you should do look at the
individual clinical intuition is key or
is it can we automate even that process
can we do that can we turn the art into
science well think about the following
last time you purchase something at
Amazon at the bottom of the page they
said maybe you’d like to buy something
like this okay what they did is they
went for people like you that purchased
the same thing and they said people like
you tend to buy this also it works
really well why can’t we do it in
medicine we can actually look back at
libraries of millions of patients and
see what happened to them in the past
two patients just like them out of
millions and know what’s about to happen
to them imagine what will happen when we
add to all of this data genetic data as
well what we will be able to do but we
can do so much even today without the
genetic data on everyone we can do
precision medicine it is about a year
ago we created the tool as part of a
competition that was held by the New
England Journal of Medicine the two we
created which we called I predict was
actually trying to find whether the
average is fitting for most of the
people and in that specific trial what
we found is that it is not about a third
of the patients in that trial we’re
better off without the intensive
treatment despite recommendations and we
could know it through the models that we
created so the competition went pretty
well for us and now we are trying to
adapt this to many other diseases so it
seems to me that even if it will take
some time computers will account for the
biological variability between humans
better than doctors and so in such
circumstances is medicines to the heart
or is it just a branch of data science
what is the added value of a doctor his
data storage capacity and his processing
powers are pretty limited especially in
the time that he has for your visit so
imagine how much can be missed
imagine how much is missed every day the
path is quite clear slowly but steadily
we are looking at the end of the age of
clinical intuition we can resist the
thought all we want and it will take a
while but eventually like the Masters of
chess and go when it comes to driving
conclusions from the patient data the
greatest physicians of our time will
bend the knee to their electronic
sidekicks dr. house will yield to deep
mind Sherlock Holmes will be beaten by
Watson but I’m sure that you feel that
there is a point here that’s missing you
know the soul of this profession the art
is not just about precision
it’s about personalization and by that I
mean that the specific biological traits
do not make a person if you want to make
medicine personal you need to tailor it
to the person their goals their
preferences their needs it’s not just
their biology that’s what good doctors
do can algorithms do that actually the
answer is yes to some extent they can do
even that as the mathematician Cathy
O’Neill said in a TED talk just about a
year ago don’t mistake algorithms as
impartial judges they are not they are
just opinions
embedded in codes algorithms are
opinions embedded in codes and that’s
not a bad thing if we can get the
algorithms to look at the opinion of the
patient and put it into account and this
is of I think significant importance not
just the opinions of the scientists and
the doctors and what will happen if we
won’t do that well that’s the difference
between precise and personalized between
medicine that looks into your biology
and the one that looks at you as a whole
person it is a dramatic difference if
you fail to do that the most precise
advice that you will get will not be
followed do you know that less than 50%
of the drugs that we prescribe are
actually being taken by patients less
than 50% and that’s what’s happened when
you don’t tell your treatment to the
patient in front of you and what will
happen if you have a wonderful algorithm
extremely precise that will give you the
most precise treatment that would
actually be 3d printed in the physicians
room for you and you will not be part of
the decision-making process you will go
out of the physicians room you will take
the pill box and you will drop it in the
nearest trash can just like has
happening today in more than half of the
cases because it’s precise it’s not
personalized to them we’ve actually made
a little attempt to do this in our I
predict model and we allowed our
patients and their physician to adjust
their treatment according to their
preferences and the weights and we have
we found was a dramatic difference when
we accounted for their preferences the
algorithm recommendation changed so our
profession as doctors is about to be
truly disrupted the days of the knowing
all doctor relying on their clinic
intuition are all but over machines are
already doing much of this in an
extremely precise manner and will
override clinical intuition very soon
the art of medicine is not in precision
it is not in the deduction skills that
will be the key it is in the humane part
of health care in caring for our
patients and partnering with them to
figure out the right path for them among
many precise options we can and we are
training algorithms to the some of it
but only some and that part of the art
of medicine is going to last for quite a
while
so in an era where computers are already
decision aids and will become very soon
decision makers and when this is what
the virtual doctors office may look like
or feel like still the most important
technology in the physicians bag
although sometimes neglected a
technology that holds great healing
powers that may take decades to be
replaced by machines was and will be the
human touch I believe that in the future
even when will we lo launch the most
advanced version of I predict in order
to feel safe what we will truly want to
hear from our navigator in our time of
need
from our doctor are two words I hear thank you
trouble when we doctors go to work every
day
we know that one-third of the care that
we will provide will be completely non
value-added futile that over a third of
the care that we should provide will
actually not be given it will be missed
overlooked and that the third cause of
death in most Western countries is not
road accidents or diabetes it is the
honest mistakes we make as health
professionals going to work we all
physicians and patients deserve more
than that it’s time for us to inject
some precision some certainty into these
work processes imagine imagine that you
had a gift that you could see into the
future that you could look at someone
and know that something bad is going to
happen to him and that you could do
something about it you could give him a
different path that would take him out
of harm’s way
change his life course what would happen
to medicine and healthcare in the future
like this when people can actually see
into the future instead of going to see
your doctor with your ailments she will
call you one morning and say hey how are
you today
great you’ll say oh good please don’t be
alarmed but in 12 hours you will have a
heart attack so why don’t you come over
to the office and we’ll fix you before
the symptoms begin now I know you think
it’s science fiction but it is not it is
happening every day in this country as
well as in others in massive scale let
me give you one example of how such
pre-emptive medicine looks like some of
you have friends or family with kidney
failure going through dialysis you know
how painful and debilitating this
condition is and how wonderful it would
if you could have come earlier years
before before the symptoms began make
small difference and made them not go
through this painful path well in order
to do that you need to predict which
patients will actually go there in five
years they look fine now they feel fine
we should predict who it will be which
is exactly what we did about seven years
ago we used an massive amount of Health
Records that’s held by Israel’s biggest
health provider elite health services
and we used all of this abundance of
data to find those patients and we asked
our physicians to actually approach them
bring them in and give them care at such
early point and our data suggests that
the rate of dialysis have been
dramatically dropping in the last few
years after this intervention in this
subgroup now we’re trying to do the same
now for other diseases for diabetes for
pneumonia even for cancer we’re
fortunate to be able to do this this is
non-trivial you know in other countries
they do not have universal health care
they do not have such massive amount of
electronic healthcare records spanning
back for years and they surely don’t
have the technological hub that we have
in this country I don’t know if you know
but in this country there are more
startups medical patents scientists and
engineers than any other country around
the world so this is why people have
already termed the coin for this country
the digital health nation perhaps you
asked yourself how does this prediction
actually work now
not always can I see into the black box
but in some instances we can and I want
to give you a glimpse so at one time we
wanted to predict which patients would
eventually have osteoporotic fractures
that mean brittle bones that break and
we wanted to identify them early so we
did our prediction thing and the
computer insisted and one of the key
was i disease well obviously a mistake
because I know from med school do those
two have nothing in common
they’re not associated but the computer
insisted and then it hit us it’s not
that the eye disease make bones more
brittle it’s the visual impairments that
make pickup fall and when they fall they
break so computers are better than us in
identifying such subtle patterns and
accumulating them together and that’s
how it actually works so we want to do
more than that more than predicting
diseases we want to be able to select
the right treatment for every patient
but this is not a simple thing this is a
balancing act you see in deciding on the
intensity of treatment too much is bad
too little is dangerous and the optimal
point in the middle weight should stop
pushing is individual and this is
basically the art of medicine to make
the skull we try today to use evidence
but there’s a problem with the evidence
that we have because the evidence that
we get from the literature is average
but the patient in front of us is not
average none of us is average and this
is why medicine is an art every person
is different everyone is unique and they
need personal fitting it was one of the
greatest physicians of all times
that have said the following if it were
not for the great variability between
individuals that medicine might as well
be a science rather than art so it’s the
need to personalize care that what makes
medicine an art forget the average use
your clinical intuition to tailor fit I
know it’s sometimes there’s a risk of
human error and it’s really easy to
follow the average guidelines but this
is
what you should do look at the
individual clinical intuition is key or
is it can we automate even that process
can we do that can we turn the art into
science well think about the following
last time you purchase something at
Amazon at the bottom of the page they
said maybe you’d like to buy something
like this okay what they did is they
went for people like you that purchased
the same thing and they said people like
you tend to buy this also it works
really well why can’t we do it in
medicine we can actually look back at
libraries of millions of patients and
see what happened to them in the past
two patients just like them out of
millions and know what’s about to happen
to them imagine what will happen when we
add to all of this data genetic data as
well what we will be able to do but we
can do so much even today without the
genetic data on everyone we can do
precision medicine it is about a year
ago we created the tool as part of a
competition that was held by the New
England Journal of Medicine the two we
created which we called I predict was
actually trying to find whether the
average is fitting for most of the
people and in that specific trial what
we found is that it is not about a third
of the patients in that trial we’re
better off without the intensive
treatment despite recommendations and we
could know it through the models that we
created so the competition went pretty
well for us and now we are trying to
adapt this to many other diseases so it
seems to me that even if it will take
some time computers will account for the
biological variability between humans
better than doctors and so in such
circumstances is medicines to the heart
or is it just a branch of data science
what is the added value of a doctor his
data storage capacity and his processing
powers are pretty limited especially in
the time that he has for your visit so
imagine how much can be missed
imagine how much is missed every day the
path is quite clear slowly but steadily
we are looking at the end of the age of
clinical intuition we can resist the
thought all we want and it will take a
while but eventually like the Masters of
chess and go when it comes to driving
conclusions from the patient data the
greatest physicians of our time will
bend the knee to their electronic
sidekicks dr. house will yield to deep
mind Sherlock Holmes will be beaten by
Watson but I’m sure that you feel that
there is a point here that’s missing you
know the soul of this profession the art
is not just about precision
it’s about personalization and by that I
mean that the specific biological traits
do not make a person if you want to make
medicine personal you need to tailor it
to the person their goals their
preferences their needs it’s not just
their biology that’s what good doctors
do can algorithms do that actually the
answer is yes to some extent they can do
even that as the mathematician Cathy
O’Neill said in a TED talk just about a
year ago don’t mistake algorithms as
impartial judges they are not they are
just opinions
embedded in codes algorithms are
opinions embedded in codes and that’s
not a bad thing if we can get the
algorithms to look at the opinion of the
patient and put it into account and this
is of I think significant importance not
just the opinions of the scientists and
the doctors and what will happen if we
won’t do that well that’s the difference
between precise and personalized between
medicine that looks into your biology
and the one that looks at you as a whole
person it is a dramatic difference if
you fail to do that the most precise
advice that you will get will not be
followed do you know that less than 50%
of the drugs that we prescribe are
actually being taken by patients less
than 50% and that’s what’s happened when
you don’t tell your treatment to the
patient in front of you and what will
happen if you have a wonderful algorithm
extremely precise that will give you the
most precise treatment that would
actually be 3d printed in the physicians
room for you and you will not be part of
the decision-making process you will go
out of the physicians room you will take
the pill box and you will drop it in the
nearest trash can just like has
happening today in more than half of the
cases because it’s precise it’s not
personalized to them we’ve actually made
a little attempt to do this in our I
predict model and we allowed our
patients and their physician to adjust
their treatment according to their
preferences and the weights and we have
we found was a dramatic difference when
we accounted for their preferences the
algorithm recommendation changed so our
profession as doctors is about to be
truly disrupted the days of the knowing
all doctor relying on their clinic
intuition are all but over machines are
already doing much of this in an
extremely precise manner and will
override clinical intuition very soon
the art of medicine is not in precision
it is not in the deduction skills that
will be the key it is in the humane part
of health care in caring for our
patients and partnering with them to
figure out the right path for them among
many precise options we can and we are
training algorithms to the some of it
but only some and that part of the art
of medicine is going to last for quite a
while
so in an era where computers are already
decision aids and will become very soon
decision makers and when this is what
the virtual doctors office may look like
or feel like still the most important
technology in the physicians bag
although sometimes neglected a
technology that holds great healing
powers that may take decades to be
replaced by machines was and will be the
human touch I believe that in the future
even when will we lo launch the most
advanced version of I predict in order
to feel safe what we will truly want to
hear from our navigator in our time of
need
from our doctor are two words I hear thank you
Please follow and like us: