I really used to laugh at those folks
are you still laugh at those folks you
know the type I mean to use gadgets like
this with ings and fitbit’s measuring
everything that went in and out of their
bodies in 2011 this field was called
quantified self it really took off and I
used to really laugh at these folks
until 2012 when I stepped on a scale and
I weighed 247 pounds I thought to myself
when I hit 250 they can measure me in
tons that’s an eighth of a ton this is
literally what I look like
during my 2012 TEDMED talk some people
are embarrassed to see their videos I
was really embarrassed to see this video
I thought to myself what happened to me
I’m driving everywhere in Silicon Valley
I used to walk a lot I bought everything
Fitbit bakes I started tracking my steps
like many of you do now but actually
more interesting I start to weigh myself
every day and I record everything I eat
I’ve recorded every meal every day for
the past five years I lost 50 pounds and
two and a half years
this is a day the driven weight-loss and
me as a data scientist this was an
incredible experience I learned a lot I
learned that it doesn’t really matter
about the quality of the food entries
you make I know the good days I know the
bad days but in a precision way I now
know which are the foods that add excess
calories to my waistline it doesn’t
matter if the data is inaccurate the
learning is precise now I’m not giving a
product testimonial for Fitbit not
everyone gets the same results I’m an
early adopter for a lot of these gadgets
I love the social aspects all of my
friends on Fitbit on Facebook know with
my Fitbit I lost 50 pounds but here’s
the interesting point I made the most
major health intervention of my life but
I when I went to go see my primary care
doctor at a major medical system near
here and they spent a billion dollars
installing an electronic health record
system my primary care doctor was the
last one to know I lost 50 pounds
because their billion-dollar system
doesn’t talk to my $40 gadget and that
has to change we have to make that
change and like that medicine is gonna
change from this bygone era of hands-on
practice into a new era of data-driven
medicine now data and computation have
changed many fields over the past few
decades it was only a matter of time
before this happened to medicine and
data by itself doesn’t do anything I as
a patient I want to share my data now
with those that care about me and those
that care for me but this is a
partnership now the physician of the
present or the physician of the future
needs to accept that data and send back
advice and recommendations that doesn’t
happen today so let’s not lament the
fact that data is driving medicine in
the future let’s celebrate it and let’s
use it
and let’s get more patients and
healthcare providers using the data
that’s out there we have so much data in
medicine today unbelievable amounts of
data we have data on individuals data on
patients data on physicians we have data
on the quantity quality and composition
of the care we’re delivering or the lack
of quality in the care that we’re
delivering we have data on performance
of physicians and nurses we have Dan on
the price of medical devices and
pharmaceuticals we’re entering an era of
big data in biomedicine some of the
biggest data actually comes from the
health systems and the physicians and
the data they generate with you
for example molecular data we’re getting
more and more measurements of molecules
where the proteins or DNA from you or
your tissues or your tumors someday on
the left is the Whitehead Institute a
picture of the Whitehead Institute one
of the centers that generated a good
chunk of the genome project that ended
in 2003 after 13 years and 3 billion
dollars of effort and on the right is my
hand holding a USB stick sequencer one
of these actually works on the space
station now it’s a no-brainer
we’re gonna be getting more molecular
data from patients in the future
there’s imaging data whether it’s x-rays
of chests or lungs or bones or CT scans
or MRIs these are enormous files ranging
from 3 megabytes to 3 gigabytes
depending on the type of study that your
physician might order and then there’s
claims data
what’s a claim a claim is how we and
medicine send a bill to get paid here’s
an example of a handwritten claim you
can find on the Internet
they’re electronic now of course that’s
a top you see some information about a
patient their address their age and then
you see codes at the bottom codes on
what do we do with these patients and
what disorders or conditions we think
the patient has these are how we and
medicine get paid when we see a patient
a claims can go only so far
to represent a patient even though these
are electronic and we’ve had them for
decades you can already see how limiting
they can be the ultimate data to have
are your medical records let me give you
an analogy of the difference between
claims and medical records let’s say I
told you that I really loved going to
this restaurant called bricks up in Napa
Valley maybe some of you have been there
and you’re curious why do I love this
restaurant so much does it really help
you to look at the bill you can see I
spend a decent amount of money there
maybe I had a nice party there but
wouldn’t you rather see what we ordered
at the restaurant you can see we ordered
the crab cakes you can see we ordered a
Balducci Pinot Noir which of these would
you rather have to actually understand
what are we doing what was our
experience at the restaurant and that’s
the difference between medical claims
and medical records medical records are
the ultimate record what actually is
happening to a patient what exactly is a
medical record we’re asked to think
about these all the time
when we sign forms what is in a medical
record
well medical records go back to 1907
they were invented at the Mayo Clinic so
it’s been a hundred and ten years since
we’ve had paper-based medical records
and when we think of medical records we
think of shelves like these you remember
seeing these maybe when you were growing
up
these are paper-based records in an
outpatient clinic nicely organized with
stickers and handwritten notes inside
but these are for outpatients the minute
you actually come into a hospital your
medical record grows they become
enormous and they’re stored in binders
like these and these binders were
strictly defined they’d start with the
medication orders than the admit note
and the progress notes flawed followed
by the flow sheets and the nursing notes
and then the lab test results and
radiology and pathology and then
somewhere in the back were the social
work notes it’s almost like we had a
pecking order in medicine
physically implemented through the
medical wreck
you can only put your notes in the right
place now of course today these are all
electronic but back then these were
enormous records in the newborn
intensive care units we used to
celebrate when our tiny baby patients
probably weighed more than their medical
records but now of course they’re
electronic and the only way electrons
the way electrons but they cost billions
of dollars you’re three examples Sutter
spent 1 billion dollars on their
electronic health record system partners
over in Boston spent 1.2 billion dollars
and Kaiser spent 4 billion dollars on
their electronic health records just a
tad bit more than the $0.99
you’re questioning whether to purchase
in the App Store for an app now it’s
kind of funny if you look at these
headlines the billion dollars Sutter
spent is written up as a boondoggle but
the four billion dollars Kaiser spent is
written up as winning a bet and they
bought the same software the pays to
have the right friends as reporters and
these electronic record systems are
costing us billions now instead of the
sections of the record we now have
tables up to 18,000 tables making up the
databases that feed the electronic
health record systems we have tables
that record every encounter in patient
now patient every drug ordered and
dispensed IV vital sign every lab test
result and of course the notes all of
those millions and millions of text
notes every admission note every
progress note every discharge note every
phone call gets documented in all of
these notes enormous amounts of data yet
despite all of this data we’re still
sometimes missing the most obviously
relevant pieces of data let me give you
an example if a big city hospital sends
you home after having major surgery to
get care by your regular primary care
doctor
the big city hospital might never
actually know what happened to you after
that surgery that’s called your outcome
your outcomes the only raw form of
outcomes the big city hospital might
have is called your discharge
disposition
you know what discharge disposition is
did you die or did you leave and if you
left did you go home did you go to a
hospice or did you go to another
hospital discharge disposition is like
that other Oak Ridge Boys song you don’t
have to go home but you can’t stay here
so imagine the very next thing that
happens to you whether you did better or
worse is stuck in someone else’s major
computer system let’s think about how
ridiculous that is for a moment let’s
imagine you’re learning how to play a
game called chess or teach a computer
how to play chess wouldn’t you think the
computer will want to know which moves
it made that led to a winning game or a
losing game now imagine trying to learn
how to play chess and you only get the
first half of the game or imagine trying
to learn to play chess and you want to
get the last half of the game isn’t it
obvious
how limiting that is if these systems
can’t talk to each other we got to do
better than this
we got to create new methods and more
poorly new business models to get health
systems to start to share this data with
each other we’re doing this in the
University of California but more
importantly we need to figure out how to
get this data from you because you know
if you’re doing better beyond your
surveys of whether you like to are care
or like the parking structure at our
facilities are you actually doing better
we need to do better than this
and we can do better than this we got to
actually build these systems and
actually start to tap into ways that you
can electronically communicate with us
to give us the data we need now we have
all this data and we spent billions to
acquire it but how are we going to get
this to the scientists and analysts and
software engineers who know what to do
with this data
in a safe respectful way it’s your
medical record data we have to
de-identify that data and to me it’s
really important for me to teach you
what does it mean to de-identify a
medical record data the identification
is not something we can make up it’s
coded in federal law appears in an
example of the federal law we have to
remove 18 different types of identifiers
from that medical record that make us
know it belongs to you for example at
the top it’s obvious we got to remove
your name from the record we got to
remove the dates that we saw you and so
much more
and we tell you we’re doing these things
and then after we’ve de-identified these
records and certified they’re
de-identified then we can get this to
the software engineers and analysts and
the scientists to figure out what’s
working in medicine and what might be
excessive in medicine and when we
combine that clinical data in molecular
data and wearable data and behavioral
data and your preferences when we
combine all that maybe then we can
customize the health care we deliver to
you and that is what we call precision
medicine the customization of healthcare
we can deliver to you based on all of
that data and not just from you but from
the entire population with the outcomes
to learn what is working and not working
in medicine and apply it to you now to
get there we’re going to need to build
an enormous database we’re gonna need a
lot of people sharing their data in a
safe de-identified way and a molecular
data and the clinical data and the
imaging data and then give that to
scientists and how are we gonna get
there well one way forward is through
the precision medicine initiative and
now called all of us research cohort
this is the goal that President Obama
set out years ago to recruit a million
or more volunteers maybe some of you to
contribute to medical record data and
your molecular data and blood tests and
D identify that so we can learn what
keeps healthy people healthy and what
can we do to customize to care for those
that are ill we have a long way to go to
build this but you’re going to start to
see health systems and physicians asking
you to start to participate in research
studies like this I do hope you start
to say yes what am I trying to do for
precision medicine in the University of
California I’m trying to build maps maps
of death and disease now compare that to
Google Maps which takes you to pleasant
destinations I’m trying to figure out
how Californians get diseases and die in
our state right here’s an example of a
map made by Hannah and J Hannah was a
researcher in my lab she runs her own
research group now in Korea and Jay
helped us with the visualizations here’s
an example of a map with real California
data on the top left circle that means
that we see patients showing up with
alcoholism and then if you got to the
bottom right a year later we see these
same patients showing up with liver
disease or liver cirrhosis and then if
you zip back to the left there they’re
showing up a liver abscesses and you can
do this zigzag or you could come
straight down the squares mean the
patients died of those two diseases so
it’s really rare to die of alcoholism
you typically die of those other two
disorders the maps can get very
complicated here on the Left patients
show up at acute myocardial infarction
that’s a heart attack you can die right
away of that but if you take the
northern route you end up next in a year
with heart failure that means the heart
muscle isn’t strong enough to pump the
blood it needs and you can die of that
or that leads to lung disease the orange
circle and the fluid starts to back up
because the heart can’t pump that blood
fast enough and the big blue circle on
the right is sepsis septicemia that’s a
blood infection now I’m a pediatrician I
never really saw many patients showing
up but a heart attack but I always
thought that if you had a heart attack
it’s the heart that does does you in in
the end but it’s actually an infection
that does he win at the end after three
years and if you don’t take the northern
route you can take the southern route
and you kill your kidneys first and then
end up with sepsis that way now what I’m
trying to do is to build these maps now
I love Google Maps I use Google Maps all
the time but they’re still one thing
Google Maps can’t do Google Maps can’t
show you where all the cars are on the
map I know it shows you the traffic
but it doesn’t show you each and every
little car on the map how could it ever
do that right but for me my vision for
the University of California isn’t just
to build the maps and show the maps I
actually want to see what all of our 15
million patients are doing on the map
where are they where are they going next
this is a prototype with real California
data you can see as patients are getting
older we’re making the colors brighter
the age is going up at the top right a
lot of them are trying to get diseases
and dye a whole bunch of that I’m gonna
hit that sepsis purple circle in a
moment then they’re gonna die there they
go and we can start to use this map
right we can start to predict what’s
gonna happen next what’s gonna happen in
the next 90 days what’s gonna happen in
the next year and what are we gonna do
about it
and that to me is going to be the new
definition for precision medicine how
can we actually use all that molecular
clinical behavior preference data to
accurately and accountably deliver
precision care to our 15 million
patients in University California and
that’s what we’re trying to build this
is the future so we’re surrounded by
data and medicine but data doesn’t do
anything by itself it’s stuck it’s
frozen in these ponds and puddles there
you’re puddles to go splashing in those
puddles learn what’s in your medical
record data and start to share your
medical record data with scientists and
researchers in these large research
cohorts we need everyone’s help to build
this new future a precision medicine
thank you
you
are you still laugh at those folks you
know the type I mean to use gadgets like
this with ings and fitbit’s measuring
everything that went in and out of their
bodies in 2011 this field was called
quantified self it really took off and I
used to really laugh at these folks
until 2012 when I stepped on a scale and
I weighed 247 pounds I thought to myself
when I hit 250 they can measure me in
tons that’s an eighth of a ton this is
literally what I look like
during my 2012 TEDMED talk some people
are embarrassed to see their videos I
was really embarrassed to see this video
I thought to myself what happened to me
I’m driving everywhere in Silicon Valley
I used to walk a lot I bought everything
Fitbit bakes I started tracking my steps
like many of you do now but actually
more interesting I start to weigh myself
every day and I record everything I eat
I’ve recorded every meal every day for
the past five years I lost 50 pounds and
two and a half years
this is a day the driven weight-loss and
me as a data scientist this was an
incredible experience I learned a lot I
learned that it doesn’t really matter
about the quality of the food entries
you make I know the good days I know the
bad days but in a precision way I now
know which are the foods that add excess
calories to my waistline it doesn’t
matter if the data is inaccurate the
learning is precise now I’m not giving a
product testimonial for Fitbit not
everyone gets the same results I’m an
early adopter for a lot of these gadgets
I love the social aspects all of my
friends on Fitbit on Facebook know with
my Fitbit I lost 50 pounds but here’s
the interesting point I made the most
major health intervention of my life but
I when I went to go see my primary care
doctor at a major medical system near
here and they spent a billion dollars
installing an electronic health record
system my primary care doctor was the
last one to know I lost 50 pounds
because their billion-dollar system
doesn’t talk to my $40 gadget and that
has to change we have to make that
change and like that medicine is gonna
change from this bygone era of hands-on
practice into a new era of data-driven
medicine now data and computation have
changed many fields over the past few
decades it was only a matter of time
before this happened to medicine and
data by itself doesn’t do anything I as
a patient I want to share my data now
with those that care about me and those
that care for me but this is a
partnership now the physician of the
present or the physician of the future
needs to accept that data and send back
advice and recommendations that doesn’t
happen today so let’s not lament the
fact that data is driving medicine in
the future let’s celebrate it and let’s
use it
and let’s get more patients and
healthcare providers using the data
that’s out there we have so much data in
medicine today unbelievable amounts of
data we have data on individuals data on
patients data on physicians we have data
on the quantity quality and composition
of the care we’re delivering or the lack
of quality in the care that we’re
delivering we have data on performance
of physicians and nurses we have Dan on
the price of medical devices and
pharmaceuticals we’re entering an era of
big data in biomedicine some of the
biggest data actually comes from the
health systems and the physicians and
the data they generate with you
for example molecular data we’re getting
more and more measurements of molecules
where the proteins or DNA from you or
your tissues or your tumors someday on
the left is the Whitehead Institute a
picture of the Whitehead Institute one
of the centers that generated a good
chunk of the genome project that ended
in 2003 after 13 years and 3 billion
dollars of effort and on the right is my
hand holding a USB stick sequencer one
of these actually works on the space
station now it’s a no-brainer
we’re gonna be getting more molecular
data from patients in the future
there’s imaging data whether it’s x-rays
of chests or lungs or bones or CT scans
or MRIs these are enormous files ranging
from 3 megabytes to 3 gigabytes
depending on the type of study that your
physician might order and then there’s
claims data
what’s a claim a claim is how we and
medicine send a bill to get paid here’s
an example of a handwritten claim you
can find on the Internet
they’re electronic now of course that’s
a top you see some information about a
patient their address their age and then
you see codes at the bottom codes on
what do we do with these patients and
what disorders or conditions we think
the patient has these are how we and
medicine get paid when we see a patient
a claims can go only so far
to represent a patient even though these
are electronic and we’ve had them for
decades you can already see how limiting
they can be the ultimate data to have
are your medical records let me give you
an analogy of the difference between
claims and medical records let’s say I
told you that I really loved going to
this restaurant called bricks up in Napa
Valley maybe some of you have been there
and you’re curious why do I love this
restaurant so much does it really help
you to look at the bill you can see I
spend a decent amount of money there
maybe I had a nice party there but
wouldn’t you rather see what we ordered
at the restaurant you can see we ordered
the crab cakes you can see we ordered a
Balducci Pinot Noir which of these would
you rather have to actually understand
what are we doing what was our
experience at the restaurant and that’s
the difference between medical claims
and medical records medical records are
the ultimate record what actually is
happening to a patient what exactly is a
medical record we’re asked to think
about these all the time
when we sign forms what is in a medical
record
well medical records go back to 1907
they were invented at the Mayo Clinic so
it’s been a hundred and ten years since
we’ve had paper-based medical records
and when we think of medical records we
think of shelves like these you remember
seeing these maybe when you were growing
up
these are paper-based records in an
outpatient clinic nicely organized with
stickers and handwritten notes inside
but these are for outpatients the minute
you actually come into a hospital your
medical record grows they become
enormous and they’re stored in binders
like these and these binders were
strictly defined they’d start with the
medication orders than the admit note
and the progress notes flawed followed
by the flow sheets and the nursing notes
and then the lab test results and
radiology and pathology and then
somewhere in the back were the social
work notes it’s almost like we had a
pecking order in medicine
physically implemented through the
medical wreck
you can only put your notes in the right
place now of course today these are all
electronic but back then these were
enormous records in the newborn
intensive care units we used to
celebrate when our tiny baby patients
probably weighed more than their medical
records but now of course they’re
electronic and the only way electrons
the way electrons but they cost billions
of dollars you’re three examples Sutter
spent 1 billion dollars on their
electronic health record system partners
over in Boston spent 1.2 billion dollars
and Kaiser spent 4 billion dollars on
their electronic health records just a
tad bit more than the $0.99
you’re questioning whether to purchase
in the App Store for an app now it’s
kind of funny if you look at these
headlines the billion dollars Sutter
spent is written up as a boondoggle but
the four billion dollars Kaiser spent is
written up as winning a bet and they
bought the same software the pays to
have the right friends as reporters and
these electronic record systems are
costing us billions now instead of the
sections of the record we now have
tables up to 18,000 tables making up the
databases that feed the electronic
health record systems we have tables
that record every encounter in patient
now patient every drug ordered and
dispensed IV vital sign every lab test
result and of course the notes all of
those millions and millions of text
notes every admission note every
progress note every discharge note every
phone call gets documented in all of
these notes enormous amounts of data yet
despite all of this data we’re still
sometimes missing the most obviously
relevant pieces of data let me give you
an example if a big city hospital sends
you home after having major surgery to
get care by your regular primary care
doctor
the big city hospital might never
actually know what happened to you after
that surgery that’s called your outcome
your outcomes the only raw form of
outcomes the big city hospital might
have is called your discharge
disposition
you know what discharge disposition is
did you die or did you leave and if you
left did you go home did you go to a
hospice or did you go to another
hospital discharge disposition is like
that other Oak Ridge Boys song you don’t
have to go home but you can’t stay here
so imagine the very next thing that
happens to you whether you did better or
worse is stuck in someone else’s major
computer system let’s think about how
ridiculous that is for a moment let’s
imagine you’re learning how to play a
game called chess or teach a computer
how to play chess wouldn’t you think the
computer will want to know which moves
it made that led to a winning game or a
losing game now imagine trying to learn
how to play chess and you only get the
first half of the game or imagine trying
to learn to play chess and you want to
get the last half of the game isn’t it
obvious
how limiting that is if these systems
can’t talk to each other we got to do
better than this
we got to create new methods and more
poorly new business models to get health
systems to start to share this data with
each other we’re doing this in the
University of California but more
importantly we need to figure out how to
get this data from you because you know
if you’re doing better beyond your
surveys of whether you like to are care
or like the parking structure at our
facilities are you actually doing better
we need to do better than this
and we can do better than this we got to
actually build these systems and
actually start to tap into ways that you
can electronically communicate with us
to give us the data we need now we have
all this data and we spent billions to
acquire it but how are we going to get
this to the scientists and analysts and
software engineers who know what to do
with this data
in a safe respectful way it’s your
medical record data we have to
de-identify that data and to me it’s
really important for me to teach you
what does it mean to de-identify a
medical record data the identification
is not something we can make up it’s
coded in federal law appears in an
example of the federal law we have to
remove 18 different types of identifiers
from that medical record that make us
know it belongs to you for example at
the top it’s obvious we got to remove
your name from the record we got to
remove the dates that we saw you and so
much more
and we tell you we’re doing these things
and then after we’ve de-identified these
records and certified they’re
de-identified then we can get this to
the software engineers and analysts and
the scientists to figure out what’s
working in medicine and what might be
excessive in medicine and when we
combine that clinical data in molecular
data and wearable data and behavioral
data and your preferences when we
combine all that maybe then we can
customize the health care we deliver to
you and that is what we call precision
medicine the customization of healthcare
we can deliver to you based on all of
that data and not just from you but from
the entire population with the outcomes
to learn what is working and not working
in medicine and apply it to you now to
get there we’re going to need to build
an enormous database we’re gonna need a
lot of people sharing their data in a
safe de-identified way and a molecular
data and the clinical data and the
imaging data and then give that to
scientists and how are we gonna get
there well one way forward is through
the precision medicine initiative and
now called all of us research cohort
this is the goal that President Obama
set out years ago to recruit a million
or more volunteers maybe some of you to
contribute to medical record data and
your molecular data and blood tests and
D identify that so we can learn what
keeps healthy people healthy and what
can we do to customize to care for those
that are ill we have a long way to go to
build this but you’re going to start to
see health systems and physicians asking
you to start to participate in research
studies like this I do hope you start
to say yes what am I trying to do for
precision medicine in the University of
California I’m trying to build maps maps
of death and disease now compare that to
Google Maps which takes you to pleasant
destinations I’m trying to figure out
how Californians get diseases and die in
our state right here’s an example of a
map made by Hannah and J Hannah was a
researcher in my lab she runs her own
research group now in Korea and Jay
helped us with the visualizations here’s
an example of a map with real California
data on the top left circle that means
that we see patients showing up with
alcoholism and then if you got to the
bottom right a year later we see these
same patients showing up with liver
disease or liver cirrhosis and then if
you zip back to the left there they’re
showing up a liver abscesses and you can
do this zigzag or you could come
straight down the squares mean the
patients died of those two diseases so
it’s really rare to die of alcoholism
you typically die of those other two
disorders the maps can get very
complicated here on the Left patients
show up at acute myocardial infarction
that’s a heart attack you can die right
away of that but if you take the
northern route you end up next in a year
with heart failure that means the heart
muscle isn’t strong enough to pump the
blood it needs and you can die of that
or that leads to lung disease the orange
circle and the fluid starts to back up
because the heart can’t pump that blood
fast enough and the big blue circle on
the right is sepsis septicemia that’s a
blood infection now I’m a pediatrician I
never really saw many patients showing
up but a heart attack but I always
thought that if you had a heart attack
it’s the heart that does does you in in
the end but it’s actually an infection
that does he win at the end after three
years and if you don’t take the northern
route you can take the southern route
and you kill your kidneys first and then
end up with sepsis that way now what I’m
trying to do is to build these maps now
I love Google Maps I use Google Maps all
the time but they’re still one thing
Google Maps can’t do Google Maps can’t
show you where all the cars are on the
map I know it shows you the traffic
but it doesn’t show you each and every
little car on the map how could it ever
do that right but for me my vision for
the University of California isn’t just
to build the maps and show the maps I
actually want to see what all of our 15
million patients are doing on the map
where are they where are they going next
this is a prototype with real California
data you can see as patients are getting
older we’re making the colors brighter
the age is going up at the top right a
lot of them are trying to get diseases
and dye a whole bunch of that I’m gonna
hit that sepsis purple circle in a
moment then they’re gonna die there they
go and we can start to use this map
right we can start to predict what’s
gonna happen next what’s gonna happen in
the next 90 days what’s gonna happen in
the next year and what are we gonna do
about it
and that to me is going to be the new
definition for precision medicine how
can we actually use all that molecular
clinical behavior preference data to
accurately and accountably deliver
precision care to our 15 million
patients in University California and
that’s what we’re trying to build this
is the future so we’re surrounded by
data and medicine but data doesn’t do
anything by itself it’s stuck it’s
frozen in these ponds and puddles there
you’re puddles to go splashing in those
puddles learn what’s in your medical
record data and start to share your
medical record data with scientists and
researchers in these large research
cohorts we need everyone’s help to build
this new future a precision medicine
thank you
you
Please follow and like us: