Tip:
Highlight text to annotate it
X
Eric Green: Well, thank you, Susan. We will now hear from
Dr. Robert Green. By the way no direct relation, of course we're all related, but --
[laughter]
-- we just both happen to have the same last name. Dr. Robert Green earned his M.D. from
the University of Virginia School of Medicine and completed his residency in Neurology at
Harvard Medical School's Longwood Neurology Program. And afterwards, he completed research
fellowships at the Beth Israel Hospital and Children's Hospital in Boston. Now, Dr. Green
is an associate professor of medicine in the Division of Genetics at Brigham and Women's
Hospital in Harvard Medical School, and also associate director for Research at the Partner's
Healthcare Center for Personalized Genetic Medicine. His research interests have evolved
from a focus on clinical trials in genetic epidemiology to a focus on translational genomics
and health outcomes. He has been continuously funded by the National Institutes of Health
for 21 years and has published over 300 articles. His key contributions have included the development
of risk estimates based on family history and genetic markers, leadership and analysis
of large multi-center treatment and prevention trials, including trials enriched through
family history, and design leadership and future planning of the first large-scale randomized
clinical trials in translational -- in genetics. He serves on a number of advisory, editorial,
and grant review boards. He's been invited to participate in NIH planning workshops in
the future of genomic medicine and has been a featured a plenary speaker on translational
genomics in personalized medicine at meetings, both national and international in scope.
So, if we will please ask Robert Green to please come up and we will hear from him.
[applause]
Robert Green: Well, thank you. Thank you to the organizers
of the Smithsonian. Thank you, Eric. Thank you, Professor Wolf, and thank you all for
coming out. It's an exciting night. I hope you were as impressed by Professor Wolf's
historical review as I was, and I think we can all agree that racism is wrong, the Holocaust
was evil, patients' rights should be respected, including privacy, including the right to
refuse treatments and tests, and, in your final and stirring ending, that patient privacy,
patient rights, and public trust are critically important. So, if that is where we're going
to go, there is no debate. We totally agree.
But, as we'll talk more about, I think, in the give-and-take portion, I believe Professor
Wolf has fundamentally misread the nature of the practice of medicine. And I'll go more
into that after I make my prepared remarks.
My prepared remarks, however, are that genetic information is fundamentally similar to other
types of medical information and that we need to beware the FUD. Now those of you with smart
phones, don't look it up right away, let me surprise you with it. I'm going to start my
talk by disclosing, as I talk -- disclose with every talk, that my research is primarily
funded by NIH, especially NHGRI, to which I am grateful that I have financially unentangled
collaborations with some of the direct consumers companies, that I occasionally speak and am
compensated for speaking, from Illumina, a maker of genome sequencing, and I occasionally
advise some other companies in this space.
But I'd also like to disclose something else. I grew up around here
[laughter]
I went to school in Winchester, Virginia, only a few miles away, and I'm so pleased
that some of my friends from Winchester have joined us here tonight. And we all got our
start with teachers and mentors, and Laura Robb [spelled phonetically], my fifth grade
teacher, and John and Marjorie Lewis [spelled phonetically], my sixth grade church group
mentors, are here in the audience, and before I continue, can you just give them a round
of applause?
[applause]
Keep your hand up a minute. Thanks. And I don't know if it was fifth grade or sixth
grade or seventh grade, but somewhere along the way, I, and most you, probably learned
about Gregor Mendel and his peas and plants. And this fundamental and early historical
exam pool reinforced in your mind that genetics was somehow deterministic, that if you put
the right traits together, you always got this trait, or if you have that funny recessive
thing going on, you know, you had that probabilistic situation with the traits, but fundamentally,
genetics was destiny. The genetics meant that you had this particular gene, you had the
particular pea color, or the particular disease.
But with the Human Genome Project, with the extraordinary knowledge that Eric outlined
and that we've learned, and with what has happened since 2003, we've actually learned
that genetics is very rarely deterministic. Sure, genetic variation is responsible for
everything, what we look like, what we, perhaps, to some degree, to our temperament, or at
least partly responsible for some of these things. And Eric mentioned to you that we've
got these -- most of our DNA is exactly the same, but we've got these variants throughout
the genome that make us slightly different. And most of the diseases that we have, especially
all the common diseases, are not the products of one variant or two variants, they're the
products of hundreds of variants. That deck is shuffled, and it's like throwing a deck
every time -- it's like throwing a new pair of dice every time you're born.
So, yes, there are a few mutations that are really deterministic, just like there are
a few situations in medicine that mean, yeah, you're really going to die. But they actually
turn out to be the exception. So, determinism in genetics and genomics is the exception,
not the rule. Now this principle, determinism, has somehow imbedded itself, whether it's
early, or it's being an elementary student and studying about Mendel, or it's a number
of other things that I'll tell you about, this principle has imbedded itself deeply
in our psyche and it is really hard to dislodge. We've all got it, and it -- certainly that
this principle has been abused, just like other principles have been abused.
But, in fact, there's a bunch of historical antecedents to genomic medicine that have
stressed determinism and stressed caution. Some of these are justified, some are less
justified, but it has been an insistent drum beat of determinism and caution, caution and
determinism, as we have moved forward with genetics. Some for very good reasons. We learned
about Mendel and his peas early. Preconception and prenatal testing around the birth of a
baby are highly charged. If a woman is going to terminate a pregnancy, she has to do so
in a limited amount of time with tremendous emotional stress. Caution is warranted there.
And those are some of the first examples of applying genetics and genomics. Newborn screening
is one of the greatest public health successes of all time, and newborn screening involves
testing for some very, very rare deterministic mutations.
Misattributed paternity. It turns out there's a small percentage of babies whose father
is not who they think they are --
[laughter]
-- and that's throughout most of societies around the world. And guess what? Genetics
reveals that pretty easily. That's a sensitive subject. Talk about patient partnership, the
current caution attributed to this in genetics, I was shocked to find, actually has geneticists
and genetic counselors being advocating to misrepresent the truth to families on whom
they stumble across this. I'm troubled by that. That's -- it's an unhappy truth, perhaps,
but it's still the truth.
Huntington's disease was the first of these for which we had a reputable genetic test,
and for many years, it was the only disease for which we had a test, a predictive test,
and it is, by far, the exception. It is a horrible disease that comes on in adult life,
is always fatal, and there's absolutely nothing you can do about it. That example colored
our perception until today. The complexity that Eric referred to is pretty daunting,
and it colors our view toward the entire field of genetics and genomics, even in situations
where things are pretty straightforward.
We've had a certain amount of optimism about genetics. I've shared in it. Susan says she
shares in it. We've even probably used it to sell genetics to the Congress and to the
public. Maybe we've oversold it. Part of that overselling, at times, or at least oversold
the timeframe, some of that is a reflection of determinism: "Hey, we can tell you everything
about yourself because genetics is so deterministic." And, as Susan said, the budget of the Human
Genome Project had a set-aside, earmarked, millions of dollars earmarked for ethicists
to study the ethics of genetics. Well, listen, that's a good thing. Don't get me wrong. I've
benefited from this, Susan has benefited from this, the world has benefited from this. But
if you're an ethicist coming into a field, searching for funds, are you going to say
"Everything's fine, go ahead"? No. You're going to say, "This is bad. This is dangerous.
Be careful about that, watch out for that."
So there has been, I would propose, all of the historical antecedents have created a
bias toward FUD. What is FUD? Fear, uncertainty, and doubt. And so I lay this groundwork because
I really believe that, in the long run, I believe that some of these early examples
have created certain sensitivities that we absolutely need to respect. But in the long
run, as genomics becomes part of medicine and part of our everyday life, fear, uncertainty,
and doubt will do more harm than it will do good.
FUD Number 1: Genetic information is different because it's going to frighten people. Well,
14 years ago, every Alzheimer specialist in the world said, "You should not disclose the
genetic risk for Alzheimer's Disease because it will terrify people, there will be catastrophic
psychological reactions. There's a gene called ApoE, and if you -- one in four of you actually
carry that -- the allele that puts you at increased risk for Alzheimer's Disease. That's
maybe, I don't know, 25 of you, 30 of you. So, 30 of the people sitting in this audience
are at increased risk for Alzheimer's Disease, about three times the risk of other people.
How many of you would like to know if you're going to be at that increased risk? How many
of you would like to know the answer? Okay. Well, it's usually higher, but then usually
Susan isn't speaking in front of me.
[laughter]
I don't think you should have to know either, actually. But I think you should have the
right to know. And pretty much every ethicist of the day 14 years ago said, "Under no circumstance
should this even be offered." Well, we put that to the test in a series of randomized
controlled trials over 10 years in which we enrolled over 1,000 people in something called
the "Reveal Study" and people who wanted this information did very well. I'll refer you
to all the papers if you'd like, but that's the bottom line.
FUD Number 2: Genetic information is different because it's too complicated to be understandable.
You heard themes of this. We didn't get to see each other's slides before this, but I
kind of thought she'd touch on some of these. It's just too complicated. People are going
to misunderstand. There's going to be errors. Well, it's not like there's no errors in the
rest of medical testing, right?
[laughter]
I mean, really. And it's not like your doctor understands all the complications of every
technology that he or she is using. It's not like every patient understands everything
that the doctor is saying. In fact, we have hands-on evidence from the Coriell Institute's
Personalized Medicine Initiative where they're giving thousands of common complex SNPs back
to doctors and patients together; from the radically progressive Personal Genome Project
where George Church is sequencing hundreds of individuals and putting their entire sequence
on the Internet for anyone in the world to use; from the Multiplex -- am I up there now?
From the Multiplex Study, an NHGRI study that has given SNP information, that's common risk
information, to hundreds of individuals about common diseases. And from the incredible ClinSeq
Project, another internal NHGRI project lead by Les Biesecker, where nearly a thousand
people have been sequenced, and make no bones about it, he is maybe sequencing them for
some cardiac risk issues, but he's telling them whatever he finds, and he's found some
fascinating things, including making diagnoses on people who had problems and never knew
they had the problems.
Sequencing had allowed us to peer inside previously completely unsuspected diagnoses. And our
own NHGRI-funded Impact of Personal Genomics Study, where we've questioned thousands of
people who took these online direct-to-consumer genetic tests and have all done pretty well.
Do you know how many people have taken these tests? Most people don't know. Couple hundred?
Couple thousand? Ten thousand? Hundred thousand? More? Seven hundred thousand. Do you not think
we would have heard of terrifically tragic anecdotes or there wouldn't have been lawsuits
if there had been?
FUD 2A: Genetic information is different because people won't act upon it, or, if you want
to turn that around, people will act on it, but they'll do so inappropriately. They won't
take their doctor's input. Well, in that very same PGEN [spelled phonetically] study, we
found that people are acting all right. Most of the things they learn are encouraging the
two to do basic public health improvements, and 30 percent of the people who get direct-to-consumer
genetic testing, something that's been pretty roundly criticized by many ethicists, make
improvements in their diet or exercise. And you know what else? Only about 1 percent of
these direct-to-consumer customers actually changed their prescription medicine without
consulting their doctor. That was another concern.
So, fear, uncertainty, and doubt raises these questions persistently, over and over and
over again. But like the myths, evidence dispels them.
FUD Number 3: Genetic information is dangerous because it's going to confuse regular doctors.
And sure enough, there is a lot of information in the genome, and it's a pretty big challenge
to figure out how to condense it, collapse it, distill it, and you've got to think about
a lot of things where there are errors, what's been proven, what hasn't. We've been funded
by NHGRI to do something called the MedSeq Project, which is a study in which we're giving
genome information back to doctors. Before we could even get started in that, what, are
we going to hand them a hard drive with 3 billion bits of information? No. Before we
do that, we created a one-page summary report that a primary care doctor can understand
and work with that has the segments on the more deterministic mutations, the more risk
mutations, the pharmacogenomic variants. And we're testing that right now. It's not perfect,
but you know what? Our primary care doctors, after only six hours of orientation, are getting
it, and working with it, and talking to it about their patients. You don't have to be
a nuclear physicist to get an MRI report. You don't have to be a radiologist to get
an x-ray report. All you have to do is put the report in a format that a doctor can understand
and work with.
FUD Number 4: Genetic information is particularly different for children. There's all sorts
of hypotheses about bonding problems. If we tell the parent of a child that the child
is at risk for an adult onset genetic disease, we're somehow changing the nature of the child
in the eyes of the parent. Well, that's a reasonable hypothesis. There's only one problem.
The evidence doesn't support it. What the evidence does support is that a parent dying
is really bad for a child. So, if you sequence a child and you find out something about an
adult onset risk, say of cancer, one of those parents passed on that mutation. You've just
given that parent a chance to save their life and be around for that child. Do parents want
this? You bet they do.
I went onto the newborn unit where parents had just given birth and I said, "Would you
be interested in genome screening for your baby?" and over 80 percent were somewhat,
very, or extremely interested. Now my friends in ethics would say to me, "Yeah, but they're
at a vulnerable moment. They've just given birth. They don't know what they're saying,"
[laughter]
Right? I repeated it three months later and nothing changed. My friends in ethics would
say, "Ah, yeah, but you haven't really educated them about what this is." I went through a
whole process of telling them the worst things they could possibly find out, and not just
the worst in terms of medical outcome, but the most uncertain. Exactly the same number
still wanted the information. And so, we've been funded to do something called the BabySeq
Project --
[laughter]
-- where we will be sequencing newborn children because, if you believe, as I do, that some
predictive information is good as an adult, then perhaps it's even better as a child.
I think that's a testable question. I'm not sure that's right, but we're gathering evidence.
And as Susan pointed out, there's a lot of information in your genome, and one of the
questions -- key questions of the day is about incidental findings. And Susan and others
would have you believe that a standardized process for seeking and reporting incidental
findings should be different in genomic medicine than for the rest of medicine. And I would
argue that's simply not true. That doesn't do you any good, that doesn't do me any good,
and, by the way, it has nothing to do with the Holocaust. It's about the practice of
medicine and doing well for the patient, and possibly, just possibly, saving that patient's
life. What is the right analogy? I continue to maintain that if I fall off my mountain
bike and get a chest x-ray for a possible broken rib, and you see that little nodule,
the radiologist sees that little nodule up in the corner there, they need to tell my
doctor about it. They, in fact, they'd get sued for malpractice if they didn't.
Susan and I actually agree on a lot of things, but one thing she said is, I believe, fundamentally
wrong. It is not stumbling across this nodule. It is not noticing this nodule. If you look
in radiology textbooks, radiologists are required to systematically review the information in
front of them. They are required to go down the ribs one by one. They're required to look
at the shape and size of the heart, and they are required to look at the lung field and
tell your doctor in their professional report that there's something there, whether it's
related to the reason that the x-ray was put up in front of them or not, period. Fast and
simple, no questions asked. I admit there's a real question here. Is the proper analogy
a chest x-ray or is the proper analogy a total body scan? Critics have argued that the point
we're making is more like a total body scan when you only needed a chest x-ray, but I
reject that. Total body scans aren't bad because of incidental findings. Total body scans are
bad because they're overly expensive and they expose you to unnecessary radiation. It's
apples and oranges.
Yes, we came out after a 14-month process with 16 experts, plus 20 outside experts,
plus the total board of the American College of Medical Genetics and Genomics in an exercise
of imaging a future in which not 5,000 people, not 10,000 people, but 20 million people are
getting sequenced week in and week out. How should you think about managing incidental
findings? Do you really want a system where someone who walks in the door out of fear
gets sequenced and says, "Oh, I don't want any of that cancer stuff, no, no, no." Do
you make that easy? It's not a question of whether they have a right to refuse; as we'll
see in a minute, they have a right to refuse. But do you make it easy for the frightened
patient or the lazy doctor to avoid all that? "Don't look for those incidental findings."
No, you don't make it easy. You make it hard. They should have a right to refuse, but they
should have to be thoughtful about it.
This is a minimum list. It's basically standardized as the search and reporting of a very few
variants. It's consistent with the practice of medicine, and this is what it really is
about. We've got about 45,000 exome variants in this bar on the left, and if you take just
the ones that are associated with diseases, you've got about, oh 15,000. And then if you
take just the variants that are in the 56 genes, you've got that little purple line
at the top, and if you take only the variants that we're pretty darn sure are really bad,
the very pathogenic variants, you've got a tiny, tiny, tiny fraction of the genome. That's
all. It's the tiniest fraction of the genome that we think actually means something. You
know what the criteria we used in this working group? I can tell you, we didn't sit around
saying, "Let's violate patients' privacy and their rights." We sat around thinking, "If
we did a genome on our patient, and they had a mutation in one of these variants of this
kind and we didn't tell them, what would keep us up at night? What would make it so we could
not sleep?" That was the criteria. And we rejected lots of nominations.
So, again, if there's, let's say, 100 people in this audience, two of you are carrying
one of these 56 variants, a pathogenic variant. Two of you are carrying that. Now, I'm not
going to take a poll, but the situation you have to ask yourself is this: You go in for
a genome sequence, for, let's say, a heart issue. Or your child goes in for one. And
the genome is done, and a cancer pre-disposition that could affect you and your entire family
is uncovered. How easy do you want to make that for us to bury that information? And
how problematic is it? Well, Baylor, which is the most advanced laboratory in the world
for doing clinical sequencing today -- not research, but this is a real patient, they've
got a real problem, let's send it to Baylor to get sequenced. They have all this information.
They have, without hesitation, implemented the ACMG recommendation, and they put it up
black and white on their requisition: If you get your genome here, you get these back,
or your doctor gets these back in their report.
And by the way -- we'll get into this in the talking -- but your doctor doesn't actually
have to follow the recommendations, and your doctor doesn't actually have to tell you if
you make it that important for him or her not to tell you. We'll get into that. It's
hard; we wanted to make it hard. But it's possible.
So to conclude, I urge you to reject the FUD, reject the fear, the uncertainty, and doubt.
Certainly there have been abuses of genetics in the past, certainly there are sensitivities.
I don't think those abuses are specific to genetics and genomics. I don't think those
sensitivities are specific to genomic medicine. *** testing, life and death decisions of all
types. But I'll tell you what -- Susan has been polite enough not to use this term, but
other critics of these recommendations use the term "mandatory," have used the term "coercive"
-- I'll tell you that when you look up this term, FUD -- fear, uncertainty, and doubt
-- its origin was in companies trying to get a competitive advantage over each other in
-- and people reviewing -- business ethicists considered fear, uncertainty, and doubt to
be coercive. Think about that. It's kind of a weird way to think of it, but think about
that.
I want to thank all of you for being here. I want to thank NHGRI and all my colleagues
for supporting the research that we've done and that we're doing. Thank you.