The National Academy of Medicine’s 2016 Richard & Hinda Rosenthal Symposium



Ladies and gentlemen, good afternoon and welcome
to the University of Washington for the Richard and Hinda Rosenthal symposium. Please welcome the president of the National
Academy of Medicine, Dr. Victor Dzau. Applause. >> VICTOR DZAU: Thank you. I love this. Welcome everyone. Good afternoon. Welcome to 4:00 Seattle time. We're starting right on time. Thank you for your patience. First I want to welcome you to the Richard
and Hinda Rosenthal symposium. I'm Victor Dzau, president of the National
Academy of Medicine. I'd like to begin to thank UW and of course
Paul Ramsey for hosting this meeting. This endowed series which started in 1988 just in time. 1988 sponsored by the Rosenthal Family Foundation
has been an annual event at the National Academy of medicine. Used to be called the Institute of Medicine. This is an historic moment because for the
first time we decided that we should take it on the road and really get to have much
interaction with our members, academy, private sector, etcetera. And I'm so glad that we are able to bring
it to Seattle and again thank you very much Paul for hosting us. For those not familiar with the National Academy
of medicine we used to be called Institute of Medicine or IOM. Actually, it was the health arm of National
Academy of Sciences. But as of a year and a half ago, we undergo
underwent a reconstitution to form the National Academy of medicine so now there are three
national academies under one roof. Natural Academy of Sciences, engineering and
National Academy of medicine. So it's been really great in the sense that
we've been able to do a lot more interaction across the academies. And noncommercial many of the things that
we could now in this topic for example which I call convergence science, certainly we're
seeing disciplines come together across the national academies. Each year we take on a health policy topic
for this symposium and from the founding days where I think the first topic 1988 was universal
healthcare. And then I also look at we had antimicrobial
resistance and last year we celebrated 15th anniversary of the landmark study to err is
human. You can imagine we're revisiting issues but
life moves on the record an we're making progress. But now I think this year assist topic is
really exciting and I have to give Paul Ramsey the credit. He and I were saying I wanted to come to Seattle
to do this. He said, "I got the perfect topic." And Chris, I think had a lot to do with you. tell you about that later. Anyway, the whole idea of position population
health is so timely and before I do that, I do want to recognize the Rosenthal family,
Rick and Nancy just came in from Los Angeles. So welcome so much for your sponsorship and
your support. Thank you. (Applause.) You think about topics that's so important
to health and healthcare they've got to be that people are interested in and actually
been doing a lot of work for example if you look at population health certainly we've
been credited for pioneering the whole concept of population health because in 1988 we published
a report that says future of public health that presented a vision for public health
from the traditional model of surveillance, epidemiology and emergency response to a model
that talks about the public health and the need to be involved with community all together. Because the executive officer of the National
Academy of medicine wrote this very important paper back then that pointed out if you look
at health determinants healthcare is only 10% whereas, genetic behavioral social issues
are major determine ants. That set the stage of thinking of moving forward
for population health and came up with the idea that delivery has to be part of this. And with the change in the Affordable Care
Act, the whole idea of accountable care now came together. And we wrote a report in 2012 calling for
the it's good of care delivery of public health and the primary care. Recently we've been looking at how to introduce
into practice. We give lip service to it. We ask about it. But we actually sent a report that in response
to the request to integrate social difference of electronic health record in meaningful
use too. In terms of social risk factors so one can
look at how to care for the relevance to many of the social issues. Positional medicine in 1988 the National Research
Council of the national academies actually called for sequencing of a human genome. This was very controversial at the time. There were people who said that's way waste
of money, to expensive. Not a lot of information and of course subsequently
2013 as you know, the human genome was sequenced. In 2011 the National Academy rote a report
calling for precision medicine new taxonomy of diseases. So as you can see, these two fields hold very
close to us really close to our heart. And I think this symposium brings the two
together. It's what we think of convergence signs, converges
edge of different discipline to solve a problem in this case looking at position medicine
in population health. Let's look at public population health. The question we always ask is can public health
be conducted with more precision? If you think about the positions of public
health they are assessment, policy development and assurance. The question is can we have more precision
in measuring population health problems and risks if can we develop the right intervention
and the right population and can we more precise in delivering those inter interventions. As you think about the possibilities where
I'm sure this distinguished panel will talk about, we think about things like genomics,
biomarkers, which may be able to identify particularly the vulnerable population we
can think about targeting early preventive strategies and of course technology use in
the point of care diagnosis and particularly in emerging infections identification of pathogens
aged surveillance intervention. It's very exciting to imagine this coming
together and I K. think of a better person than our keynote speaker Susan DesmondHellman
to talk about this. Here are the challenges. Many people are saying certainly in the public
hath area things and maybe competition if not conflict between population and precision
medicine which talks about the individual. The question with increased complexity of
care? Would it drive up the cost and would increase
inequity for those who can afford to have access to the technology and those who do
not. And these are obviously very important questions
that need to be addressed. I thought for sure today this will be addressed
by the panel. So without further ado, I'd like to turn the
meeting over to Paul Ramsey who is going to introduce our keynote speaker and the panelists
and I look forward to learning a lot from all of you and engaging in a very lively discussion. Thank you very much. >> PAUL RAMSEY: Thank you very much, Victor. Good afternoon to everyone and welcome. Victor, it is a privilege for us at UW medicine
to hold the Rosenthal symposium here in Seattle on a bright sunny day. A little traffic outside the meeting here
otherwise a very nice day in Seattle. you also have given me an unexpected reunion
turns out that Rick Rosenthal who was been introduced is a good friend of mine I haven't
seen for 45 years. Was my college roommate more than 45 years
ago. So thank you for bringing the Rosenthal symposium
to Seattle and the Rosenthal sun, Rick. So welcome Rick and Nancy. It my privilege to introduce Susan DesmondHellman. She's a philanthropy and physician. She has devoted her career to improving the
human condition. Nice summary of your career, I think. Been a pioneer in healthcare for more than
30 years and has driven major developments toward the eradication of disease, poverty
and health inequity. Today as chief financial officer of the Bill
and Melinda Gates Foundation Dr. Desmond-Hellman leads the vision for a world where every person
has the incentive to lead a healthy and productive life. Drawing on her diverse experience in both
the public and private sectors, she has created an environment for talented and committed
individuals to help more children and young people survive and thrive. The work at the foundation is combating infectious
diseases that impact the lives of underserved people around the world. And Dr. Desmond-Hellman is committed to empowering
people, particularly women and young girls to transform their lives. Trained as an oncologist, Dr. Desmond-Hellman
spent 14 years at Genentech developing break through medicines including two of the first
gene targeted or precision therapies for cancer. Her time at Genentech put her at the forefront
of the precision medicine revolution and her current role she champions a similar approach
to global development. What we'll be discussing today precision public
health. We need to have the right interventions, deliver
those to the best population and in the right places to save lives, to enable individuals
to thrive. Immediately prior to joining the Gates Foundation
in 2014, Dr. Desmond-Hellman served as chancellor at San Francisco overseeing all aspects of
the university, medical center strategies and operations. It was her second stint at CSF having completed
her clinical training there. She is the recipient of numerous honors and
awards. She is listed among more tune magazine's most
50 top powerful women in business for seven years and was inducted into the American academy
of arts and sciences and elected to the Institute of Medicine now the National Academy of medicine. She served in many other capacities in addition
to her service as chief financial officer for the Gates Foundation including service
on the board of Facebook incorporated. Please join me in welcoming our keynote speaker
Dr. Susan DesmondHellman. Thank you for that nice introduction. Victor and Paul, I'm pleased to be with you. I particularly want to thank today's sponsors
the Rosenthal family, reunions and all. Gracious of you and congratulations to everyone
in the audience for finding a place to park. You get extra credit for being here. It is a real pleasure for me to come and talk
about this topic with all of you this afternoon. And you will hear from the panel members a
lot of the details aged technology thinking that go behind this topic of precision population
health. But I want to tell you a couple stories that
I hope will bring to life why I care so much about this topic. I'm not much of a cook. But on Sunday in my kitchen, I opened one
of the cupboards and looked at something that's actually never touched food in my kitchen. It's a cherry wood bowl. That was carved for me by a widower. When I look at that bowl that's now 25 years
old I see in it the life of a patient that widower's wife who I couldn't save. A young woman having just had a baby with
inflammatory breast cancer. We fought long and hard for six months against
a foe her inflammatory breast cancer that I couldn't fix. Five years after she lost that battle can
cancer and I completely failed her in making her feel better, only made her feel worse,
Herceptin was approved. Every time I see that beautiful bowl I think
if only five years earlier we had had that precise way of helping her and have avoided
all the ways I tried to help her and only made her feel worse? So when I got to move from being a real cancer
doctor on the front lines with patients too often having that experience of working with
patients to tolerate side effects without the precise tools we needed to help them feel
better, I felt like having something like Herceptin or Glivec or other such medicines
was a miracle. So the 2011 report that you just heard about
towards precision medicine, a new taxonomy of disease I got to cochair in 2011 with my
colleague Charles Sawyer was responsible for Glivec so much better treated with a precision
approach that Charles and I dreamt of a day that not just the patients who we had both
cared for who had had their hopes and dreams of treating their cancer improved through
this precision approach but patients with things like type one diabetes and congestive
heart failure and other diseases we struggled with their nonprecise treatment that their
diseases could be better understood better treated someday. When I moved to the Gates Foundation 2 1/2
years ago I thought those days were behind me but I knew they were in good hands. That precision medicine how much the world,
including President Obama and president Chi had accepted that precision medicine was something
to celebrate, to fund, and to drive. I thought I can go and work at the Gates Foundation
on philanthropy and on thinking now about public and global health. Those were good times working on precision
medicine. Little did I think that too often when I looked
around the halls of the Gates Foundation or started learning frantically from colleagues
and collaborators at the Bill and Melinda Gates Foundation I would come back full circle
to wishing the same wish I had when I was treating cancer patients that we had better
more precise tools to target the very things that are involved with population health because
that same sense of urgency I had and if only had something better five years earlier I
could have saved a patient's life. Now I think that at scale. If only we had something better for the populations
who need us to think about their health and scale we would save millions of lives or allow
millions of people the chance to live a healthy and productive life without suffering. So for me precision public health or precision
population health is a way of thinking that same way with pace that all the best that
we have in science and technology should be directed as efficiently as possible so that
all people can thrive. Bill and Melinda is Gates Foundation has not
thought of themselves as being involved in this area. It's not the tradition for us to think about
funding surveillance specifically. Or epidemiology specifically. But one of the things that's really transformed
how we think at the Bill and Melinda Gates Foundation is seeing and understanding both
the power of precision in surveillance and molecular epidemiology but also how much we're
wandering in the wilderness when we try and think about vaccine trials or new remedies
or new understandings of disease at the foundation. So in 2014, the foundation made a commitment
to a new way of thinking about how we think about innovation at the foundation. And this commitment in 2014 established a
team that was focused specifically on surveillance and epidemiology. That specific focus for the first time at
the Bill and Melinda Gates Foundation. We've always been numbers and data junkies. But this team focused on surveillance thought
about something that I think best exemplifies the concept of precision public health or
precision population health and what they did is they started to look at something that
is absolutely essential for us if we want to tackle the sustainable development goals
and further achieve reductions in childhood mortality. Mortality in children under 5. Particularly in children under one month of
age. That's where we're stuck. 2.6 million children still die every year
before they reach their first month birthday and we don't know why. So what the Gates Foundation funded and began
with many, many collaborators, most importantly the Emory University, the primary site is
a network called the CHAMPS network and that stands for Child Health And Mortality Prevention
Surveillance. The Child Health Mortality and Surveillance
Network is focused on asking a very specific question with a hypothesis and the hypothesis
was if we know why children die so early of preventable diseases, only then can we more
precisely prevent those deaths. It's based on the knowledge that today most
of those diagnoses are made through a conversation with the mother. A conversation with the family. Now those conversations people call verbal
autopsies. Verbal autopsy is a fancy way for us in science
to feel better about what in fact is an excruciating discussion with a mom who just lost their
baby. Was the baby vomiting, have a fever, look
sick? You can imagine how imprecise that kind of
conversation is. Particularly when it happens as much as three
months after the baby has died. So what CHAMPS allows us to do is to change
that into something that we're using minimally invasive tissue sampling, improving lab capabilities
and making real diagnoses so with that community, with that population we can begin to address
those causes of disease. We've got a pilot site in CHAMPS in Soweto. You can imagine we were pretty nervous about
this. Now the conversation with the family isn't
what were the baby's symptoms but it's a conversation about using a minimally invasive tissue sampling
which are biopsies so we can know why the baby died. To our delight families are not only interested
in this, one sign of how interested they are, is dad comes along for the conversation too. Because they know something they can use both
in their family and in their community can now be in their grasp and that is a more precise
understanding of the cause of death. An example is something that actually is already
shown up in Soweto group B streptococcus as a cause of death in the baby means that that
mom, when she has her next child has a seven times increased chance of that baby dying
of group B strep as well. Completely preventable with penicillin. We're not talking about rich world medicine
here. We're tying about precisely diagnosing when
in that community causes deaths in neonates. So CHAMPS is a good example of something that
allows us to be more precise working with communities in something that can improve
the health of their population. There are many other examples of work we do
in global health that fit into that same category. A couple examples that are very topical and
aren't just out sigh the U.S. is the Zika virus, when you look at Zika virus and you
look in Florida. The state of Florida didn't declare a Zika
problem. It was down to neighborhoods. So, if you need to use remedies or special
education or increased emphasis on family planning or vector control, you could be more
precise with the ways you can prevent Zika spread in mime Miami and the nearby areas. That's precision population or public health
in action. When we try and intervene in mother to child
transmission of H IV, another great example of precision public health, we're targeting
those areas in subSaharan Africa where HIV prevalence is the highest so that can more
efficiently tackle prevention of mother to child transmission and in doing so we decrease
my more than half the transition of mother to child HIV just in the last five years. So I've gotten kind of excited about this
concept initially I was skeptical as good scientists many of whom are in the audience,
you should be skeptical of jargon. Precision business. Everything is precision these days. I think you should have healthy skepticism
about terms like precision population health unless you believe that there's something
real there. So I want to try and convince you that there's
something real there in two ways. One is a bit of call to action. So I want to give you forward elements that
the world needs to make precision population health real. And this is specifically focused at low income
countries because that's where we operate. Register all births and deaths. That would be a start. We can't push ourselves to improve global
burden of disease unless we understand what that burden is. So all countries should know their births
and deaths. Track disease. Sounds simple, right? Probably preaching to the choir in this room
and track disease and make sure we have a detailed understanding of what we would call
epidemiology. Inflammatory incorporatiing laboratory analysis. We talk about capacity building. Capacity to do believable quality laboratory
work is deeply lacking globally and we need that laboratory work to enable precision population
health. Finally and most importantly train more people
and not just in things like statistics epidemiology, surveillance, laboratory medicine. But train more people on counseling. On community interA, community leadership
so that communities can drive their own population's health. These four attributes if brought to every
country in the world would be transformational. They would allow us to get smarter and more
efficiently tackle equity in diseases. So for me the dream of precision population
health is the dream of why I came to the Bill and Melinda Gates Foundation, the dream that
all lives have equal value and every child no matter what zip code, geography they're
born in has an equal chance of having a healthy and productive life. I want to tell you a story to close my comments
and I want to tell a story that really made me this about what precision population health
could mean and it's a United States story. Not a global health story it's a story we
could talk about in a place like Seattle but it comes from Cincinnati. I heard this story at a summit at UC San Francisco
on precision public health and I was so moved by this story that even though I'm telling
someone else's story, I know you would love the story as much as I did. This is a story about asthma and childhood
asthma in Cincinnati and it's a story I heard from Robert Conn who is professor of pediatrics
at Cincinnati Children's Hospital. He and his colleagues actually published this
in several journals but the story he told really brought to life for me what precision
is all about and why this isn't just sequencing. It can be, but it's not. Nearly 50% of children in Cincinnati live
below the poverty line. In Avondale 65% of the kids in Avondale live
below the poverty line and in 2014450 children in since initially were hospitalized for asthma. What Dr. Cahn and his colleagues did is
use geospatial mapping to understand the areas of the city where these kids are coming from. They started saying look we seem to see the
same kids over and over and they know each other. Maybe there's something in mapping where these
kids are who keep come coming back to the emergency room that teach us something about
asthma. They show two maps in their population map
number one is areas of recurrent emergency room vision its and admissions to the hospital
for asthma and map number 2 was Section 8 public housing. And map number 3 was Section 8 public housing
where there's been violations cockroaches, mold poor conditions in the public housing
and when you merge the map of poor conditions and frequent emergency room visits. They're the same. They're not similar they're the same. So in a world you might have changed their
beta agonist, given them more inhalers, sent them home, scolded mom and dad for compliance. That would be a tactic that that would be
a very disease oriented tactic or you can think about a precision population health
approach and think completely differently about what the data were telling you. So the remedy wasn't to go back and increase
their inhalers or put them on more steroids. The remedy was to go to Legal Aid Society
in Cincinnati. And so the great part of this story was working
with the Legal Aid Society, they started targeting landlords who had Section 8 housing that was
not allowing these kids to survive and thrive. And in doing so, had a mass masterful efficient
and effective remedy for recurring hospitalizations for recurrent asthma. I love this story so much I decided to go
to Cincinnati and see it for myself. I got to go to the Burton apartments in Avondale
and meet a guy called Michael Pinkston and his 4yearold daughter. So Michael is a father of 9, one of whom his
daughter lives with them at Burton, public housing and he's now a community organizer
working with physicians and public health people at Cincinnati's children hospital to
make sure burr tan housing is going to be a safe place for his daughter and other children
like his daughter who won't be going back to the hospital for allergic asthma. So I just thought that was a great story and
I hope you like that stories much as I do because that brings to life what population
health and communities are all about. When Michael told me what it meant for to
have a place for his daughter to grow up where she can survive and thrive and not worry about
mold and going to the emergency room because she can't breathe, it said to me we need to
redouble our unfortunates globally. On what it means for kids to have the right
kind of environment and what it means for us as physicians, as scientists, as people
who care about health equity. In using all the data at our disposal so that
no child no matter where they are in the world has to worry about things that interfere with
them leading a healthy and productive life. So I look forward to the panel discussion. Thank you again for sponsoring this talk and
thank you for having me. (Applause.) Thank you Sue. We will transition now to the panel and I'll
introduce the panel but please begin to take your seats. Dr. Desmond-Hellman will be one of our panel
participants along with three others. Dr. Peter Lee is a corporate vice president
at Microsoft Research. Dr. Lee leads an organization tasked with
creating new research powered technologies and products. His work focused in several areas including
artificial intelligence, machine learning and big data analytics. Recent examine of the work led by Dr. Lee
include artificial intelligence powered simulation language feature in Skype. The wildly popular social chat box, Microsoft's
global scale. Artificial intelligence super computer and
other areas. Dr. Lee is working on a new silicon technologies
and also on cybersecurity. He was recently assigned the task of mobilizing
Microsoft's research and data cape, to focus directly on CHAMPS. Dr. Lee joined Microsoft six years ago. Prior to that time he was a professor of computer
science at Carnegie melon, you know, university where he was head of the computer science
department. He's also been a director at the defense advanced
research projects agency, DARPA and this past year he was appointed by President Obama to
the commission on enhancing national cybersecurity. The commission's report was just released
this past week. Dr. Lee is a member of the national academies
of computer science and telecommunications board and former chair of the computing research
association. Next Dr. Lee is Dr. Christopher Murray. Dr. Murray is the founding director of the
institute for health metrics and evaluation known to us at UW medicine around the world
now is IHME. Dr. Murray is the professor of global health
at the University of Washington and adjunct professional of health services. He is a physician and health economist and
his work led to the development of a range of new methods and empirical studies to strengthen
the basis of population health measurement. Measurement of the performance of public health
and Medicare care systems and assessment of the cost effectiveness of health technologies. IHME is focused on the challenges of measurement
and evaluation and the areas of health outcomes. Health services, financial and human resources,
evaluations of policies, programs and systems and decision analytics. Dr. Murray's early work focused on tuberculosis
control and development of disease methods and application along with Dr. Allen Lopez. In this work they developed a new metric to
compare death and disability from various diseases and the contribution of risk factors
of disease in developing and developed countries. This pioneering effort has been held as a
major effort in public health and important foundation for policy formulation and priority
setting. Dr. Murray worked at the World Health Organization
from 1998 to 2003 where he served as executive director of the evidence and information for
policy cluster. From 2003 until 2007 he was the director of
the Harvard University for global health and Harvard center for population and development
studies as well as serving for director of public policy at the Harvard School of Public
Health. We were delighted in 2008 which he agreed
to join the faculty here at the University of Washington. Our fourth panelist is Dr. Jay Shendure. Science is experiencing a genomics revolution
and Jay Shendure is one of the innovators sustaining its momentum. Its work has helped DNA sequencing become
faster, less expensive and more informative. He is a methods developer at heart but his
deep medicine genetics focuses on new technologies that have a large impact on understanding
of biology and disease, Dr. Shendure is a professor of genome sciences at University
of Washington. In his Ph.D. work as recently as 2005, he
included one of the first successful demonstration of parallel or nextgeneration DNA sequencing. His research group in Seattle has made significant
contributions to technologies including X only sequencing and its application to identify
the basis of Mendelian disorders and autistic spectrum disorders. His work also included genome wide experimental
help typing to genome sequencing of a human fetus. He's worked in the area of parallel reporter
assays and saturation of genome editing and common editorial index are for single cell
analyses. He is a recipient of the 2012 Kurt stern award
from the American society of human genetics a 2013 NIH directors pioneer award and 2014
Hudson alpha life sciences prize. Moderating this distinguished panel is Dr. Robert Waterston. Dr. Waterston received his bachelors degree
from Princeton University and MD and Ph.D. in pathology from University of Chicago he
completed a post-doctoral fellowship at Cambridge where he worked on see elegance with Dr. Brener. Dr. Waterston then joined the faculty at
an institution sometimes confused with the University of Washington, that is Washington
university in St. Louis. In St. Louis he continued his work on elegance
eventually collaborating with Dr. John Solston and Dr. Colston on sequencing the worm genome. His lab went on to have a leading role in
the sequencing of the mouse and chimpanzee genomes and most importantly, and I still
have the original journal on my desk Bob, the human genome. From 2003 Dr. Waterston joined us at the
University of Washington as William gates the third endowed chair in biomedical science. As chair he's over seen the development of
Department of Genome sciences now a world leader in genomics. Thank you Bob for moderating the panel. Please join me am n welcoming all the panel
members. Okay. Make a few remarks to start and then I'll
start with a few questions for the panel and then toward the end, we're going to open this
up for questions for the audience. So we hope everybody will participate. Its given extra examples of environment and
health of individuals. Can provide new insights and new avenues for
intervention. And our ability today to collect and analyze
huge amounts of data. Both molecular and environmental data as well
as electronic health records creates an incredible opportunity in medicine. And so we're going to try to see what some
of those opportunities are today. The realization has led to lots of interest
in exploiting these huge amounts of data. To improve healthcare. The report from the NRC on entitled medicine
new taxonomy for disease has been mentioned. That was a key report. And then more recently the president's initiative
on precision medicine makes another strong case for precision medicine. As we move from using signs and symptoms of
disease and treating the symptoms, as we move from that to a deeper molecular understanding
of the root causes. We often find what was once considered a single
disease to be actually several different subtypes, each demanding their own intervention and
therapy. And coupled with that of course is the realization
that while we share most of our DNA sequence each of us carries our own variance which
can influence our susceptibility to disease and our response to intervention. And so today we want to try to talk about
the relationship of this individual assessment of healthcare needs and responses a lot to
the overall population health that Sue addressed. Okay. So in short, are we poised to take advantage
of these opportunities to improve the health of the public in ways that we couldn't even
have imagined a few years ago. So with that that's all I'll say for now and
I want to give each of the panel of the a chance to respond to Sue's talk as well as
I hope they'll address the question of are precision medicine population health management
complements or are they competitors? And does the healthcare system have to choose? So Jay, you want to start? We'll just move down
>> JAY SHENDURE: So I'll try to keep this relatively brief. I come at this as a geneticist and precision
medicine has its origins in some way in genetics, I think a lot of early examples including
ones that Sue was involved with in pharma come from this idea that we can convergence
the basis of molecular markers which have in turn their basis in genetic differences
try to try to improve the precision with which we treat particular diseases. So to try to add a few points in some of these
echo things that Sue talked about, one is that there's a risk of hype with the human
genome project I think with the intervention of medicine around this hype in the concept
I think it's important are to take a step back like you said and make sure in ask something
real. One I think challenge is that these are associated
with genetics in its origins whereas, in truth there's so many ways to think about patients
and how we subdivide them including personality, lifestyle, socioeconomic and so on and so
forth. And then of course ideally the intersection
of those things with I think Sue's talk was hit a number of great points along these
lines. Precision medicine just the terminology overly
emphasizes treatment when in fact many of the best opportunities may be in prevention. So a great example of just something that
we already do that sits nicely at the interface of genetics, precision medicine population
health is newborn screening which is widespread in this country screens for 30 or 40 very
rare disease. But university has enormous benefits for that
small group of patients where you're able to identify what metabolic condition they
have. I had third point I'll make is what we as
a field bring is not just genetics. There's this actually goes back to Bob
and the Human Genome Project before my time. But one of the most important things about
the Human Genome Project was this idea of very, very open data sharing not necessarily
at the time you published but at the time you produce data. That evening owes is moving forward in infecting
other parts of the clinical research community this came to a head in the New England Journal
apparently the term data parasites is being used to describe researchers who use but do
not produce data and I think the community appropriately pushed back. I think we are moving towards a world where
we can do research in a free and open way. That in itself is going to be transformative
but also essential for these kinds of concepts to move forward
Okay., I'll stop there. So couple of reflections on the two questions. The first is if you think about traditional
measurement and population health, you know, measuring something for the country of India
isn't terribly helpful. There's a billion+ people but that's the norm. That has been the norm in global health which
we think of population health measurement at the political entity unit. That's changing and the shift is being for
example in the work of the global burden of disease towards taking large countries and
doing analysis for very large groups of people. But the direction that that work is going,
the logical extension is to bring it down to the very local level and why not down to
the individual level. Populations are aggregations of individuals. The more we know about individual specific
health outcomes and the sort of vector of causes that relate to that, the better from
understanding individual health but also the better for noticing population health. There's no intrinsic conflict on the science
side. They're very synergistic. There's a lot of barriers to getting there
and we can in later questions talk about that and whether the vision of people not hoarding
data will get beyond certain subfields. We can all hope. There is this logical extension of getting
more precise about what are out comes and what are causes. There is this children that everybody talked
about and certainly strong in the health committee about too much emphasis on precision medicine
end of that story runs the risk of distracting from dealing with simpler easier things to
fix. And I think that that tension is not new. If you look historically and look at the data,
rich people in urban areas always get access to higher quality care. This is one of the most robust empirical findings
we find. We do a lot of work documenting this in global
burden of disease and it's a very clear finding. There's nothing new about precision medicine. Namely as you have more precise interventions
we should expect they're going to end up in the hands of urban outlets in different parts
of the world and so the concern isn't about a new tension there. It's about managing the ongoing always present
tension between how we do a better job of sharing resources whether within a country
or between countries to achieve better outcomes particularly for the disadvantaged. I'd box that concern as part of a much broader
concern for example Gates Foundation one of its central tenets is to try to be a force
against that general pattern. I'll stop there. >> PETER LEE: Let me say it's a pleasure to
be here. I'm feeling like an interloper with such accomplished
scholars in medicine. I should say in full disclosure I'm a graduate
of the University of Michigan but I'm fine with the Huskies taking the wolverines place
in the college football playoff You sound a little bitter. On the theme of too much focus on precision
medicine, there's also a lot of hype and focus on data and big data. Like prevention medicine genomics, there is
tremendous promise that has yet to be realized there. The hype is perfectly understandable. But as I was listening to Sue's talk, I think
another important element part of the talk at least my reflection is the importance of
engagement and social engagement and in fact, it relates to a paper that appeared in the
journal of by two colleagues Ryan white and Eric Horvitz where the minds anonymize binge
search logs, I hope you all know who binge is. It's Microsoft's search engine. Identifying clear evidence of users of binge
expressing the fact that they have clearly been diagnosed with pancreatic cancer and
then from that point looking backwards, through their search history and then in the context
of all of the data for the global population of binge users
A. Running data prior to diagnosis, that users had pancreatic cancer and also what the false
positive rate would be. And what was shocking and the reason for the
paper is up to 15% of the cases could be predicted ahead of time with a false positive rate of
less than one in 100,000. This is a surprising factor. But what was not reported in the paper was
the fact that if we had the same engagement of our competitor in the web service space,
the numbers would actually be better. Obviously one could extrapolate to global
engagement and expect even more to do. And so it sort of made me think of Sue's comments
in that light. That there is actually significant amount
of scientific rigor yet to be understood in the space. I think we will get a chance to talk more
about precision medicine big data as well. But just to open I thought I would imagine
about engagement and importance in the space. Sue, do you want to add anything. I wanted to comment on what Peter just said
because I think one of the things that drives innovation in IT in technology is frustration. Not dissimilar to research council process
for putting together precision medicine report, there was a point in time when we just
as a committee started investigating and the venting started like this. And I go to buy a pair of socks, everything
I do for the next month is populated by sock ads. Am I the only one who had this experience. I'm trying to look for wool socks because
it's kind of colder here than California. So I have wool socks everywhere in my feed. Facebook wool socks, Twitter wool socks, I've
got wool socks everywhere and I bought the wool socks already. But when I go see my primary care doctor,
there's nothing resembling wool socks that she knows about me. That's just weird. So I think this pancreatic cancer example
is profoundly important. People who want to sell you stuff know everything
about you or many things about you depending on how privatized you've kept your devices
and it is astounding to me that we've walled off you and your caregiver from that information. I get the privacy and the creepy factor but
it's weird that you can do that with socks but you can't do that with pancreatic cancer. That's just weird. I heap coping my job at Microsoft doesn't
depend on that. Let me jump in here. Both Peter and Jay mentioned challenges in
large data and data sharing because it seems to me that to make this really effective
we have to have millions of data points. Millions of people involved the President's
initiative calls for a cohort of more than a million people, I think. So what are the big challenges that people see
in terms of just the acquisition of all this data and then sharing it. So question can maximize the benefit. I don't know who wants to start. Peter? No, Chris does. Okay. >> CHRISTOPHER MURRAY: We spent a lot of time
trying to obtain data. So we experience this on a daily basis. And I think the reality is that a really powerful
incentives against data sharing. Spectacular exceptions, human genome made
a commitment to that and that was spectacular. But it's rare. And why is it so hard? Well, there's enormous financial gains to
be made out of mining data or at least people believe that. So there's a scramble to obtain patient level
data. The tech companies, you know, some of the
consulting companies are out buying data from developing countries. We see that that's an impediment to data sharing. People think if I hold on to that data I can
either sell it or somebody else if I hold on and have proprietary rights I can sell
a product from that Let me comment on the Human Genome Project. When we started that was the attitude about
genes. That that in I could get this gene, I could
patent it and be rich I don't know how that relatively small group
you made that transformation but, if anything we're seeing the opposite which is less likely
than even a few years ago. Second there's politics. Some governments don't terribly want transparency
because it exposes especially at the finegrain local level exposes disparities. It exposes that they're not taking care of
certain minority groups or disadvantaged groups so there's a lot of political impediments
but there's spectacular examples of the opposite. Brazil government made an extraordinary commitment
to transparency scary because they give patient level data in the public doe plain because
they're more worried about history of dictatorship than risk to individuals. You can download individual data for Brazil
now. Lots of praises that isn't the case and the
other problem is privacy considerations. It's actually hard, even if you want to share,
there's all these restrictions it's a growing movement and not getting smaller. We see some places that used to share data
freely now having internal movements toward much more restriction on data. There are powerful incentives and we need
to think about those. Just to add one point to that the classic
model has been to assume that patients want to keep their data private without necessarily
asking them. So precision medicine for example they've
done, here at the U.S. one they've done a number of surveys including at various socioeconomic
backgrounds and so on and so forth and the clear message is that people are completely
excited to share their data with a relatively minimal set of protections. So part of it is there are a lot of incentives
not to share including in academia in this country, the system is not set up to incentivize
clinical researchers to share even if the patients want to share. And that's something that is culturally needs
to shift. Building from a commercial perspective one
thing promotes sharing, there are two elements. One is when individual consumers own their
data, have clear ownership and control and secondly then on top of that when there are
clear network effects. Network effect for example in the World Wide
Web created through an informal law called Metcalf's law and every new web page and new
user and every new node increases super linearly the total commercial value of the entire ecosystem. When you have that kind of effect, then you
get a commercial flywheel because consumers can benefit from that network effect and every
business in the world has an incentive to build on a network effect. Right now it is a puzzle from the a commercial
perspective how to create those two conditions and get them started. Susan: I think adding to that from the network
effect and moving from consumer to community, a community driven ambition to improve health
is a huge driver of sharing, overcoming privacy issues. So I'm really optimistic that the precision
medicine initiative can give us new insights on to how more communitydriven and community
ownership so there's a return on invest: If you look at a blood drive you know how much
there's a strong volunteer spirit in America and yet we don't tap into that because people
aren't seeing the community good or the social good. And the sense that somebody will capital on
it. The immortal life of Henrietta lax taught
people you have to be careful with that data. This social network effect of driving what
I want for my community and that's why I told the story about Cincinnati, you might say
sharing my asthma data or sharing my experience, if I think my community's health is going
to be improved and other kids like mine benefit, how that scales to Chris's point when you
go global and national, especially if you don't really want everybody knowing your data
is harder. But what I love about the population aspect
of this is don't think about doing it to a community. Think about a community doing what they need
to be healthy, driving it themselves. That really changes that dynamic. And if you can get the communities to network
there's a huge benefit in that potentially. So maybe we can push data collectors in the
right direction. We talked a little about precision medicine. Jay, I'm going to challenge you with defining
it a little more precisely. And in particular you mentioned genetics is
often strongly associated with it. Tell us more about what genetics really has
what role it has in precision medicine also in population health. >> JAY SHENDURE: So defining the term. As I understand it, the group that you cochaired. At the time the term that was used for this
concept was personalized medicine I've spoken I think to Earlson on the committee sometimes. He wasn't just on the committee. He was a force. we're so lucky he was on our committee. So my understanding is that part of the
it was a conscious choice to choose the word precision over personalized with the idea
that this really was about new kinds of taxonomies of individuals more than one person to define
a group to say something at the end of the day you're treating a patient but this isn't
personalized it's about being more precise and defining different groups of patients. So the definition I go back to is the historical
approach to medicine to treatment and to prevention has been designed for the quote average patient
and so some patients are going to do better than the average patient. Some worse. And it's really a simple concept which is
can we be more precise in terms of predicting which individuals will responsibility to a
treatment, whether it's a preventive strategy or whatever it is. So as to improve outcomes for the entire group. Right? Prettiest straightforward. In terms of genetics, I think precision medicine,
precision population health hopefully are much broader than genetics, we tend to associate
whole genome expensing and things like that are being contemplated in the context of the
national program but that's certainly not the central piece. It's ideally much bigger than that and the
examples of genetics interfacing and precision medicine are often places where we're not
doing whole genome sequencing, we're much more targeted in what we are looking at. Newborn screening is certainly one examine:
Another is this idea that we can come up with risks for common diseases and there are studies
showing this is added value beyond family history and other predictors of let's say
coronary heart disease. One example of a preventive strategy, there's
researchers have shown that you can predict response to tobacco cessation strategies on
the basis of nicotine metabolism there are manners for. Not sequencing someone's genome but relying
on a interface of picking the right patient doing the right task that may or may not involve
genetics and using that information to stratify and inform care. I used to have I applied early days of nextgeneration
sequencing, the cost has dropped from 10�million at the time that the completion of a Genome
Project to a little over a thousand now. If you take the fact we spend $5,000 a year
in healthcare in this country per person on average. If you have a thousand dollar genome and you
amortize that over your lifetime it's 10 or 20 bucks a year. But even at that price point you could argue
it still doesn't justify itself because we haven't learned enough. So I do think there's value in Jeanette irks,
I'm a geneticist. But it's not going to be a big single brush
approach. We have to be more judicious about how we
apply it. >> VICTOR DZAU: Anybody else want to amplify
or comment. >> Susan DesmondHellman: One comment for the
clinically oriented folks in the audience, in some ways they're trying to make it more
algorithmic, that's why taxonomy really applies here although we tried hard not to do a report
on taxonomy because we were sure no one would read it and precision medicine just captured
us more than taxonomy. That taxonomy implies that everybody knows
great clinicians. What great clinicians do when they go into
a room is they start binning where this patient should go. They're observing the patient, taking a history,
doing a physical exam and great diagnosticians, great clinicians are thinking Okay. Here's their age they're said teary, do they
smoke in what's their lifestyle. Great clinicians are narrowing the degrees
of freedom let's say a patient is short of breath or something, whether you learn medicine,
you're narrowing your degrees of freedom of what you think is wrong and then you do diagnostic
tests and precision medicine for me is really that taxonomy. Is saying now we have new great tools be it
genetic sequencing, or other data elements that can do a more precise job of what the
great clinicians of what's wrong with that patient from a diagnostic and therapeutic
point of view. The exciting thing about it is what I just
described is hard to do well as a patient who in front of you and ill. Think about trying to do that with a patient
in front of you who is well and you're trying to figure out how to keep them well. And that's what I this I is so fascinating
if you like public health and I do. Is taking all that talent that for too long
in this country has been Okay. Somebody's sick and I am now struggling to
try and help make them well and say what if we could use that precision and diagnostic
skills and use attributes of data and other things to narrow that to say for you, here
are things you could do that could keep you healthy. Now we know some of those like don't smoke. I'm telling you start there. But what if you could just do a little better
than that. That's exciting. One thing I should say here, and again, in
all humility end I think I'm roughly a first or second year graduate student in the field
at this point. But one observation. Precision medicine when I engage particularly
with practitioners, first of all, first one observation is that the cognitive load, what
we're asking for from practitioners, cognitively is extreme in many cases. There's one element of precision medicine
as Jay and Sue mentioned asking for more creativity based on more inputs. But the promise at least in my early observation
it still makes sense and is important is on the flip side it is also the promise of dramatically
reducing errors and in providing very concrete and maybe algorithmic support the idea for
human terms for a practitioner is not to go overboard on a cognitive load that is demanded
of a practitioner in the diagnostic process. But instead to eliminate or reduce sense of
error while giving more tools and more data on which to make diagnoses or decisions. Go ahead, Chris. >> CHRISTOPHER MURRAY: Two comments. One is your description of the great clinicians
put that in a Bayesian's framework. The great clinicians are the ones with great
priors and they are just much more efficient of coming up with the cite diagnosis. But to the logic of applying that to keep
patients well. We've had a project with the Dartmouth Hitchcock
health system which was interesting which is take all the insights from population health
about risk factors whether it's components of diet, whether it's environmental risks
or behavioral risks around you know obesity. Physical activity, smoking, etcetera. The 60 odd risks that we track. And see if you can inject that in clinical
practice. We wrote for them an algorithm that does a
pretty good job of predicting your risk of disease and death of they've been trying this
out on patients and doctor. When you tell patients their probability and
even how you can change it, don't like that. They've reverted from telling them to do smiley
faces. Smiley face, frowny face. More palatable version than you have 26% probability
of dying because you have CHF and all these risks. There's interesting things about you can actually
people in 23409 just us but others who are trying to deliver this already in clinical
practice and I think you can go much farther with that as you get more and more information
on the genetic testing. The interesting thing about that comment is
if you take into account location and health status, they're incredibly predictive. You don't need a lot more to be rather accurate
about what's going to happen. So the interesting thing about the newer tests
that comes along on prevention, not tailoring treatment. What's the value of information there. At the margin how much better are we able
to predict and that will tell us what are the extra things we should be adding into
that algorithm What you just said is one of the beautiful
things about newborn screening. So newborn screening is an amazing thing to read
about. It's like a miracle. People who haven't read it should read it. Part of the reason is it's done 50 different
ways. It's done 50 different ways in 50 states. But there's a group that has to have a discussion
that says Okay. We're doing 29 tests that's standard and then
say population whatever. To be number 30 to your point of saying it's
a tall order to be a predictive test that you have a remedy for and has a prevalent
tie up and quality and other things as attributes of the that's as it should be. Because nobody wants to hear something that's
scary if it's not actionable and if you can't rely on the data. So I think that's actually known in public
health from the newborn screening history. And a really great attribute of good public
health which is tall order to put this in your if you're in Ethiopia tall order to
put this in your basket of interventions because government has to pay for it and front line
health workers have to do it. If you're down the street, tall order to say
I'm going to add another test for newborn screening because there's costs and overhead
and you could be wrong. You need to have something actionable. Those are things that population precision
health needs to challenge ourselves with. What's the bar? What's the hurdle rate to say your intervention
or your test gets in the kit, you know, the must do
>> Great, discussion. From my perspective one of the potentials
is actually not only to aid in the diagnosis but to did a better job of stratifying diseases
so that I think of anemia 200 years ago. It was just that you were pale and we didn't
know what to do. And today we're in a situation with autism
for example where we're starting to get the genetic basis for some of the cases and depending
on what gene is affected, there might be different phenotypes and different therapies. And so precision medicine also has the possibility
(music) [Laughter]
I had the ringer off. Okay. Chris, then I'm going to ask you. >> CHRISTOPHER MURRAY: You mentioned the challenge
of precision medicine making healthcare less equitable. You were optimistic that it wasn't going to. But are there ways we can think about that
would keep it from getting worse? >> CHRISTOPHER MURRAY: Or I think that's pretty
straightforward. Which is it's the it's a good dose of transparency
about where resources are going relative to need now and then making that publicly known,
giving that so civil society, giving that to various groups so that becomes a part of
the discourse nationally about what are priorities for public health and healthcare. I don't think there's any other anecdote. Well, wishing in this domain doesn't really
work because the you know, the data is really compelling that this always happens. It always happens. >> CHRISTOPHER MURRAY: We State of Washington. Really interesting. If you look where the state dollars go for
public health and relate it to any metric you want and need, it's about inverse. So here we are. High income state, great out comes but there's
an international tendency for groups to capture resources in any political process. I think the anecdote there is just being much
clearer than we are now about where money goes. And to whom it goes. Think about the very local issue. But the last time the federal government reported
or helped states track spend by state was 2009. We're in 2016. So it's there's no data on state spending
on healthcare for the last seven years from the national accounts people. That's at a sort of rudimentary age sex level
and not at the detail we'd want to know about spending on autism. There's nothing around that. That's my simple answer to that one
>> VICTOR DZAU: But then we have to get people to collect and reveal the data. >> CHRISTOPHER MURRAY: Yeah, I think on the
spend side the federal government does a good job over 65 we have access to extraordinary
CMS type data. You have good survey programs. But there hasn't been a drive to pull the
built together it's a little about data sharing thing
People will sell you data. With some spend you can actually do this. So it's one of these funny things that there's
data out there. It's actionable and useful but for some reason
doesn't happen. Time and time again, you see publicly available
data sets that could inform the data precision population health idea that haven't been analyzed
and there's a little culture going of using data there and computational methods that
are available. This anybody else
I think two things to add from equity perspective I think you have to fund public health. Public health departments are so awfully underfunded
namely that you need enough money to collect data and do that for equity. The other thing being done is not to have
data sets particularly for sequencing projects that are European males. To the precision medicine credit they've sought
communities that are not traditional communities that participate in research projects but
we have lots of sequencing on one kind of patient or human being
That was a strong theme in the PMI report. And broadly representative cohort. Another question. I will start with Sue on this one. Do you think precision medicine is likely
to improve population health or is it sort of beyond outside the realm of importance? How can we bridge the two. You talked a little about this. Susan DesmondHellman: I think it must improve
population health. If precision medicine disease oriented medicine,
sort of what we called when I started my training subspecialties, I think if subspecialties
will be relevant in the future, they may only be relevant if they feedback into population
health. I don't think it will be possible to operate
otherwise one of the things I think is often underused having been in the early days of
precision medicine is the ability to learn when you're wrong. I like whoever made the comment about humility,
we should have so much humility and so little hype about precision medicine we can do. I've been guilty. When we had Herceptin in 1998, I'm sure I
gave a speech saying everything will change. We will solve one cancer at a time and I'm
retiring. Here's the thing that's possible. Every time we have a new insight be it how
your immune system protects you from cancer, those new insights we're hopeful they will
help patients in precision medicine. They often have unintended consequences either
a safety thing or we don't help as many people as readily as we thought. There's so much rich biology there. That biology helped us understand population
health. It helps us understand the limitations of
what we think in a reductionist way we can do with precision medicine. And so I think there is there are vast
areas of he are verse translation that can be tapped medicine part of what I'm excited
about precision population health is getting the communities of practitioners in global
public and population health in the same setting like we're doing today with people who spend
their time in a lab or with a computer. That's a massive opportunity. >> VICTOR DZAU: Anybody else? How are we doing on time? Maybe I'll ask a final question of the whole
panel. How would you what will define success? For precision medicine or population health
management? What would be a good what would be the definition
of success. Chris, you want to start? >> CHRISTOPHER MURRAY: Sure, that's easy. I would think before this in the standard
factual way. There's trajectory we're on with where we
have with precision medicine I think success is better health outcomes compared to trajectory. You have to tease apart both aspects of that. What do you think would happen without further
advances in precision medicine. Then you've got to sort of either track it
and compare to what the scenario would have been. So grounding it in health outcomes. But no particular endpoint. >> CHRISTOPHER MURRAY: We've got views to
measure public health but not the right setting. I came across this definition of measurable
outcome when I was trying to fill out the section for a grant. But the definition is you'll know you succeeded
in doing something transformational when it's completely obvious that you succeeded manage
its selfevident you could say. I would like to think we're scratching our
heads trying to come up with little examples that it would not have succeeded but rather
if there are 100 things that we can point to using Google searches. It's not going to be a thousand different
things. I'll just add one example because I feel like
it's another great example to the Cincinnati one and the big search engine is I learned
about this while being involved in PMIs, I had no idea this was going on using phones
and individuals at risk for major depressive disorder and simply on the basis of number
of social contacts they're having at any given moment using that to predict when they're
going to have a major depressive episode. It's the kind of thing that I think you're
right, Chris, that there are many maces where the data that we do have in front of us is
enough. And having all this additional data won't
necessarily help very much. But I think there's also a lot of other place
where's we have nothing right now. You can use data in innovative ways to do
things you can't do in the framework of the current way the medical system operates. Hopefully obvious an this will be
I can't begin to define what success looks like. One thing Jay triggers a thought is I think
what the National Academy of Medicine is trying to do here in defining concept of precision
mental health is to ensure that a field like that gets established is really on a sound
scientific foundation and maybe this is part of what Chris was saying also. There's always a danger of a lot of hype around
any kind of advances in better pattern detection, pattern recognition. If I talk about the search industry as an
example. There are many hard lessons learned. Ultimately you find out you have to be scientific. You can't just detect pattern and from that
make suppositions that direct the direction of how you rank search results. You have to establish controls. You have to run many, many trials and develop
a scientifically sound foundation because time and again if you don't do that, you get
punished with results. Worse results. Not just for people depending on the search
service but worse results also in terms of the business value. Kind of ensuring there is a clear scientific
foundation that might have susceptibility to hype seems to be an incredibly important
thing here. >> Susan Desmond-Hellman: I have two increasing
the global burden of disease is an ultimate Uber metric. I would endorse that. But my two favorite ones are that the sustainable
development goal for under 5 mortality that we get there fatter and with a greater magnitude
of improvement on preventable denials in kids no matter where they are in the world. That is an outcome I'm most passionate about. But I would put a second thing in there and
this is one where I think folks like Peter and Microsoft or Facebook or am will or Google
or IBM, whatever the technology is that says something like Peter and transform health
IT quick. There is the hype prater of it. I think in this hype and deflation cycle we've
been NFF health IT, there is something special in there. There are insights in there. There is help for busy overwhelmed clinicians
in there. There is an opportunity on our healthcare
costs in there. I would love to see that happen. If it's good for society and good business,
that's sustainable and something special can happen. Great. So we're going to open this up for questions
from the audience, I think. We'll start with vic
Excuse me. >> VICTOR DZAU: Thank you for a spectacular
discussion. Panelist discussion. And many of the issues that you raise are
so spot on. One thing that I thought that the panel did
not actually tackle as much is the issue of cost. You kind of touched on it. So the question that's raised all the time
is that more and more innovative tests are coming in. Some extend three months of life. More precisely extend three moves of life
and others argue hepatitis C that you can actually cure the disease and downstream you
can see a lot of cost/benefit. Although not fully proven. So the question for all of you is really where
do you set that kind of conversation in terms of when you say what's the really outcome
you want to see. Great outcome and the contents of cost effectiveness
or what is it? How do we put that into the equation? Who wants to take that on? >> So you're asking where cost comes into
the story above and beyond the health dimension. That's one of the root concerns people have
about any technology. Historically technology hasn't reduced costs. It's increased costs. There's exceptions but that's generally the
rule and likely this is the difference so it does seem like we will there will be an
army of people doing this analysis as hep C is a great example of trying to model out
the different cost streams and outcome streams and it's going to be a societal debate about
those tradeoffs and I don't but it's unavoidable that's going to be part of the discussion,
I think. >> VICTOR DZAU: My question is with the potential
collecting so much information data, should there be in fact a post marketing assessment
whether in fact all those cost assumptions and economic assumption that you put in and
different drugs to be looked at again in other words a form of provisional approval? This is one of my bug bears here. So there's almost no ex post evaluation of
cause and effect. I keep asking for and I'm slightly skeptical
of the cost effectiveness universe because of this problem. Not because of the idea but because there
is very little respective assessment of bias if you take the ex Auntie assessment and look
at the ex post assessment. Take H IV, if you go back to 1999 the science
paper, it said all we have to do is spend $4 billion and we will stop the HIV epidemic,
we have prevention strategies, they've been costed up. $70 billion later of spending on prevention,
incidences barely budged. It's come from 3 million to 2 million
infections. Not a lot of success except prevention of
child transition, lot of success there but not in the rest. Has anybody held threes people to task for
their ex ante evaluation? No. I think we should be doing this and if the
National Academy would sponsor that, I think that would be a great contribution, Victor. [Laughter]
I would like to challenge the panel what would success look like by referring to the 2013
National Academy of Medicine report. U.S. health and international perspective,
shorter lives, poorer health. The title says it entirely. We, according to Dr. Murray's studies a
few years ago, we're about 17, 18, 19 years behind the healthiest country in the world. That is, if present trends continue, we'll
achieve the health status of Japan or one of the Scandinavian countries in a couple
decades. This is a major problem in this country. We are dead first according to the data. What would precision health do to change that? A lot. Potentially, right? Like we don't if you look within the
I mean, you know and actually we happen to have a bunch of papers coming in JAMA this
month and next about this topic of looking at specific outcomes at the county level. Again taking signal to noise processing methods,
data that's been in the public domain for decades, processing it and figuring out age,
sex, cause by location, it turns out that it's not a single story. It's not that if you live in the southeast,
everything's bad and if you live on New England, everything's good. It depends on the cause. Suicide is higher out west. Uterine cancer is in the rust belt. It depends on the cancer, on the disease and
by having that precision population it's not that precise, it's more precise than anything
we had before. You open up a different discussion about what
are the root causes whether they're social cause or healthcare access and the other thing
you can do is you can take causes say testicular cancer where we know the treatment is almost
perfectly effective. 99% 5year survival. You can look at a map and see it's Hispanic
immigrants because it's along the border of Mexico where you have higher testicular cancer
rates and that opens up obvious strategies for getting people access. But that needs to be done in a precise way
by cause. And I think now we're entering a realm where
the data and public domain, Bayesian methods are there and the computational power is cheap
that becomes quite real and if you add in electronic medical records, claims data you
can go a long distance in trying to target a solution that's not this sort of generic
gosh the U.S. is really bad. We need to go beyond that to find a solution
Yeah. What I'm about to say may be too philosophical. But I guess we're in the auspices of the national
academies and national campus I feel allowed. It is related. In the tech industry when you're trying to
create something that creates a disruption, somebody or some sector or something is being
disrupted. There are winners and losers. And I think that the you know, the whole
movement towards focus on outcomes is trying to create the business condition where the
costs of disease of poor outcomes will be the loser. Trying to measure and understand that a price
tag on it and creates therefore, the business opportunity so that can be disrupted. That's a very philosophical statement but,
if that is taken seriously and if we remain as optimistic and as progressive minded about
the possibilities of new technology as we have been, then we operate under the assumption
that we ought to be able to disrupt those and understand the financial opportunity that
comes out of there I think part of the challenge is clearly getting
people to use this data to redirect resources. We should be able to do it more precisely
and effectively by the means Chris is talking about. But we have to use the data to convince people
that we know what the answer is and that these resources will address the problem. You need both the insight that Chris is talking
about and you need the incentive structure that Peter is talking about. In the absence of those two, it won't happen. One of the experiences that most of us share
in the room is publication in peer reviewed journals and high end are asking for access
to your data. Posting it on a Web site in an accessible
format. Have they gone far enough? If not, how much further should the journals
go relative to the issue of accessibility to data and transparency. And so forth. I think it's a relevant question for this
issue of precision population health. I think there's a gap between stated journals
policies and what happens in practice. I've seen this over and over again. It is a state the journal policy that data
won't be available at the time of funnel investigation but you go and try to access the data and
it's not there, there is it's there but id's behind so many firewalls and approvals
that you don't get access to but usually it's just not there. And there's no you talk to they say Okay. Make sure it's there before it's published
and it just doesn't happen. It's getting better probably. But you do think there's some amount of public
shaming that will hopefully help move the ball
The journals don't have any followon clout. Right? You publish and you get your paper in and
Jay can complain to the journal but the paper is out. We have to think of other avenues or get the
journals to communicate with one another. So other questions? George: If I remember correctly in your introduction,
Victor, you mentioned that the first symposium sponsored by the Rosenthal family was on universal
healthcare and that reminded me that during the run up to the presidential elections people
like Bernie Sanders and Jill Stein mentioned all the obvious advantages, we all can agree
if you really want to do something to improve healthcare in this country you will have a
publicly funded universal healthcare system. But what they never mentioned is the potential
of universal healthcare to produce electronic records that are standardized that lead to
real population based epidemiology and precision population health studies. You can find something is going on that's
peculiar there in New Mexico somewhere and you're on it right away. So I wonder if people like Victor could
and Chris Murray would be the expert on it. How important do you think that would be? It goes to some extent to other countries. Do you think that would be a major point in
trying to fight for universal healthcare, publicly funded. This is like that tantalizing thing that's
just out you can just get. Poster child is New Zealand or Finland or
Sweden where they have extraordinary potential to link. But you can privacy concerns are so high in
those countries that lead time on getting permission to do a particular analysis is
about two to three years. We work a lot with Norway. It's extraordinary. You have incredible potential but there's
very few places that have actually delivered on that potential. U.K. is a very complicated structure. We think of it as universal system but GPs
are private. And so government doesn't have access to GP
records. They have to buy it from the private sector
and the vendor who sold the software to the GP is the one who owns the data and sells
it back to the government. So you have bizarre circumstance even in settings. You have to have a lot of incentive about
how you design that but the potential is very great but there's not a lot of great success
stories out there because the people who have the capacity also have these very stringent
privacy controls The values that drive, you know, investigators
healthcare also drive heightened awareness of privacy issues ooh. They travel together which actually is really
challenging. One thing I've argued time from time in different
venues and different venues there's a bulcanization of privilege to data sets. If brought together for research purposes
to greater social good. It ought to be possible to create a situation
where there's particular data commons where stakeholders, where owners of privileged data
sets to could in a low risk way pool and share data for specific research purposes, academic
research purposes. And there would be a shared risk pool for
privacy incidences that come out of that shared commons. It would be a very difficult thing to create
the first one and perhaps health data might be the challenging place to start. But this has been a discussion for example
in the National Research Council on telecommunications board where there's been recognition of tremendous
scientific value in data sets. But if you look at companies like Microsoft
or Facebook or Google. The tremendous risks to their businesses if
there are privacy incidents are extreme so how do you pool those risks in some way and
still provide access for legitimate scientific purposes. We have time for one last question. Al
I wonder if members of the panel can speculate a little bit on where support for research
in this area is likely to come in the future particularly in the evaluation of the outcomes
and the costs of these interventions I'm aware of one organization within the CDC, the national
Office of Public health genomics that had its budget cut by 95% in the last few years,
so I wonder with the importance of this and the enormous potential where is the research
support going to come from? So we live in interesting times. I think there is there is a baseline uncertainty
about the funding or the overall research funding situation in the country. I think that uncertainty is only increased. You know, the precision medicine initiative
is one thing that did have broad bipartisan support. But that doesn't address what you're asking
which is funding for public health and outcomes to research and things like that which the
precision medicine initiative isn't exactly that. That is uncertainty
minimal practice. I think we're at a loss. Greatly appreciate the National Academy of
Medicine decision to host the Rosenthal symposium. Thank you very much Victor and academy and
please join me in recognizing them and also I hope you are convinced that this is a really
expert panel. To discuss this and finally a thank you
to the audience. We have Victor members of the National Academy
of Medicine. And a few other academies. Round of applause to thank everyone here. (Applause.) There's time for conversation just outside
of our room. We have a reception so please join the panelists
for some discussion of precision population health. Thank you.

Leave a Reply

(*) Required, Your email will not be published