Panel Discussion: Risks and Benefits to Individuals and Communities

Panel Discussion: Risks and Benefits to Individuals and Communities


So we have one question for Bob. I think we can take that. While we’re doing that, I’d like
to invite all the other people from session
three up to the stage. We also have Misha Angrist
graciously agreed to pinch it since Michelle McGowan
was not able to make it. Welcome to him
and Deven McGraw as well, one of our keynote speakers. Also just a preface since we’ve
received a comment from online. If you could please
introduce yourself and speak clearly
into the mic. I will try to take that
into advisement as well so that people online
can hear you. They’re very interested to know
what the questions are. This is Manuel Corpas. I am chief scientist for
Cambridge Precision Medicine. My question to Bob was you are talking
about an aggregation tool that brings data from a very
heterogeneous set of sources. So how do you manage to harmonize
the different data fields in order to present some kind of
coherent, persistent interface in the way
you present those data? It’s complicated, right? Again, we’ve got two
really different problems here. One problem is we’ve got
these data structures that are in these databases. And the database formats
are not completely compatible. So we’ve talked
to all the labs that are contributing data
to ClinVar and all that stuff. And look,
when we talk to labs where they’re generating
the data that are being used
for clinical inference, two universally resources.
ClinVar and NOMAD. Everybody’s using them.
They’re completely open. They’re incredibly, incredibly
important global resources. They are global public goods. Everybody’s using those, but everybody’s also
complaining about, well, you know what,
we have to dumb down our data from our database at our lab
to send it off to ClinVar, and then they put it
in an Excel spreadsheet and then it may show up
five months later in ClinVar. So there’s lots of complaining
about the data formats, the technical issues. You know what?
We can solve those problems. It’s going to take a while, but ultimately
the information we need is what variant do you have
and what happened to you. Right? We can get rid of all
the private stuff in between. It doesn’t need to be
in these public databases. That can stay where it is
in the very private spaces in individual labs, in individual hospitals
in some cases. But we need to build
the infrastructure. That means standards.
That means sharing practices. And it means incentives
for people to do it. It’s mainly the incentives
for actually sharing the data that has become
the biggest barrier. That’s another
thing incidentally about the BRCA situation.
Think about it. Myriad has a database
based on two million tests that they’ve got in
a proprietary database. But you have Ambry, GeneDx,
Color, Veritas, LabCorp, Quest. They’re all entering
the market. Guess what?
They have to catch up, and they have to be able
to interpret those data. So they have a strong incentive
to share that we wouldn’t always
have for other genes. So the commons has some of
the business dynamics that would say, hey, yeah,
we can solve this problem. We’ll solve it collectively. At the same time,
you have an outlier who has a continuing financial
incentive not to share its data. And I don’t think we’re going
to solve that problem except by doing the same thing we did with the human
reference sequence. A few years after the human
reference sequence was posted, Celera dumped its data
into the public database. Why? Because it cost money
to maintain a database. And if people are not using it
and it costs money for you, why would you keep maintaining
it in your own data set when you could just dump it
on the public domain and make NLM pay
for maintaining it. So I think the same thing
will happen to these variants as we discover them
in gene after gene. And it will probably start
with BRCA1 and 2. So that’s problem one. Problem two is the rich,
rich data that are in these online
networks are free texts and, gosh, which mutation
do you have, how did you make your decision,
oh my gosh, that’s a mess. But again, it’s a mess that
we’re beginning to deal with. And the thing that we need
there is the sociology. Jill Holdren who is
part of this network that Amy has talked about and that KJ talked about
in his talk, we went out to the people
that are on these networks and said where do you
get your information. 70 percent of their
medical information is coming from
the online networks. The moderators of those groups
are like, hey, guys, we really
want good information. Please give us a pathway
to the best clinicians and the best scientists who know
how to interpret those data. Let’s build some bridges
among these constituencies. That’s a sociological process
and it’s a money process. And that’s a bunch of volunteers who don’t get paid a dime
to do that work. So that’s a sociologic problem
that we have to fix. And we need the pathway
for the information to flow from the databases
to the users and from the users into the
databases and solve problems. But we will solve it. So there were a couple
of related questions that came in online,
most for you, Bob. The first one was,
can a patient themselves submit their own data to ClinVar? And the second question is, how many women have shared
their Myriad data via the free your data
initiative? Yeah. A little wrinkle
in the story. There are a couple of pathways
for getting data into ClinVar. They are working hard
to make it an easy interface. It isn’t quite yet, frankly,
but you can get it in. The easiest way is actually
to do it through a lab that actually contributes its
data to ClinVar automatically. So that’s Ambry, GeneDx,
Invitae, Color. Most of the labs… all of
the labs that we talked to… Myriad did not respond
to our survey. But all the other labs
in the U.S. that we talked to are sharing. A lot of the
international labs are not, and that’s not because
they don’t want to. It’s because it’s a pain
in the butt for them to do it and they don’t have the money
and the staffing to do it. But in the United States, if you get tested
at another place — and there are all these
other portals that are being
built up around it. Amy, if you want to pipe in,
this is your wheelhouse. So tell us if there are
any other pathways for people to get data directly in. Thank you. Yeah. My name’s Amy Coffin.
I work for the BRCA Exchange. I think ClinVar is currently
the best pathway. In many ways, in my opinion, what’s really needed
is some sort of protocol to educate the
clinical community on the difference between
these testing companies. That’s kind of outside the scope
of my work at BRCA Exchange, but there’s really not
a clear delineation of the data
sharing practices. There’s not a lot
of transparency around the data
sharing practices. So some sort of call
to transparency would be a first step. Yeah. ClinVar is the main
repository that we use, and it depends on where.
Yeah. It just all depends on where
you’re located in the world and what kind of data
you have because there’s also
scientific information that’s emerging in literature
that can also help. So and the second question
was about how much data flowed in through the free
the data movement. And there was a program
that was called the Sharing Clinical
Data Project that Bob Nussbaum started
when he was at UC San Francisco and has continued. That’s kind of
a trickle frankly. The idea was there you get
your report back from your lab, usually Myriad because up
until 2013, it was the only lab
that was doing it. And your provider or you would
send it to a bunch of people that would take the form
and get what was your variant and what was your diagnosis
and get that into the system. And that system operated
for a while, but really it was
a pretty small trickle. And it’s a very mechanical,
labor intensive process. And so I think the wind
came out of the sails to some extent
on that project, and I don’t think that’s
the pathway we can rely on in the future. I was wondering if there was
the ability to go on and share queries or distribute
queries across these things. Let the people keep their data. Simply go on and have
a query processing agreement and come up with a distributed
query infrastructure. Has that been looked at? So one of the cool things
about BRCA Exchange is they’re trying
to set it up in such a way that there’s actually
a phone app where you can… and incidentally,
it’s a safe phone app because they don’t keep track
of your IP address or anything. They dump the data
onto your phone so you can do
your own interpretation. And if there’s been a change
in your variant that you’re interested in
and link back to say, hey, we just found something
about that that’s changed. That’s a limited version
of that. But in fact, to get that
kind of interpretation and stuff like that, you’re going to have to take
the data about that variant and take it to another place where they can do
the interpretation. [inaudible] Yeah. You can, but you’re not going to do that
through this interface, which is really basically what
do you know about my variant. [inaudible] That’s exactly what we’re doing. Can I ask the panelists
[inaudible]. We had one more online. Yeah. There was a
follow-up question about couldn’t Myriad develop
a secure access. So it’s a somewhat
related question. So the question is,
have people tried to engage Myriad around a model to provide
access to the data online? Sort of a Yale
open data access model. Yeah. So Deven answered
that question to some extent. Yeah. You can get your own data.
Absolutely. And you can actually say
to any lab, not just Myriad. You can say to any lab give me
my data and give it to them. So that pathway does exist.
That started in October of 2014. And so that’s
an open pathway now. It’s laborious.
One individual at a time is probably not the right way
to solve that problem, but individuals
who do want to solve that problem
that way can do it. Stacy Gray, City of Hope. This is a question
for all of the panelists, maybe to start with Greg. So you spoke a little bit
about the differences between opt-in
and opt-out consent and thinking about
how to structure consent in a way where people
are truly informed about what they’re doing
with many of these issues which are quite complex. So one of the questions
just to everyone is particularly
in the context of people who are sick and patients
who are dealing with a lot, how can we think
about really reviving and transforming
informed consent so that people understand what they’re consenting
to at the same time not being so burdensome
that we restrict research? In the little bit of research
that we’re doing on privacy where I’m actually having
an undergraduate student call various hospitals
across the country and ask them what is
your informed consent policy and how would I withdraw
my medical information from all studies
at your institution. And what we’re finding out
is there’s really just two tiers so far,
and this is preliminary. There are institutions that are
making remarkable efforts and investments
to talk individually to patients and take this very seriously and will even talk
to my undergraduate to explain very carefully
what they do. And then there are systems
that are cheating and are just taking information.
There is no consent at all. It is what seems to be
an implied consent if you belong to,
say, these university hospitals that have either
opt-out programs or are performing
this type of research that they don’t call research. So I think what’s needed is some
basic consumer protection laws first that just say you cannot
take people’s tissues or data representation
of their tissues for research without having them
opt in period. And then I’ll leave it to the
people who are real experts because I don’t deal
with patients. And I’m sure there’s a lot
of really great research of how to do that,
and other people can speak. But we’ve got to have a baseline because the institutions
that are reaping… and there’s another word
that sounds like that… the most information
from their patients have nothing in place
in terms of informed consent, and everyone’s jealous of them.
I’m at a state university. They would never
allow me to do that. It’s mostly
private institutions. I was amazed that the university
system in California is able to do this. There’s a lot of money
being poured into these labs, and they’re getting data, I believe,
without informed consent. I’d like to add
something to this. This conversation around
how to get consent and whether consent should be
required for all research is not a new one. It’s existed for a very,
very, very long time. It’s one that we recently
struggled with yet again in revisions to the common rule for federally supported
programs and research. And it’s an issue that certainly
gets dealt with in HIPAA as well. The standard is that you get
informed consent before you use someone’s data
for research purposes. And it’s hard to do well,
as you pointed out, because you oftentimes
get one shot to sort of ask, to try to get that consent
in a very informed way. And John Wilbanks,
who was here earlier, and the materials
that he has developed in his toolkit at Sage, they do it probably the best
of anybody that I have ever seen in terms of trying
to get it to be informed and really sort of
probing people. How informed is it, do you truly
understand what’s going on here. But there’s another issue, which is that we have
these possibilities for consent to be waived
in circumstances where it would essentially not
be possible to do the research because the number of people
involved and having to go back and reconsent them
into a new research project would make it very
difficult to do so and frankly would keep
the research from happening. And so you have
an institutional review board or a privacy board
in the case of HIPAA, but most people use
institutional review boards, which are supposed to have
expertise around ethics to consider whether the value
of the research and the protections
that have been put in to protecting the privacy
of the data, which usually involves removing
some measure of identifiers out of the data,
contractual commitments around reidentification and things of that nature justify doing the research
without consent, without needing to go back and
get the consent of the patient. So it is a weighing
of the importance of the research
questions being asked against what are the
risks from an ethical as well as a privacy standpoint
around the consent pieces. And it doesn’t work perfectly,
and it certainly falls down. But I would be greatly concerned
about an absolute prohibition on doing any research
with any data under any circumstances
without consent because I believe that it tips
the scale in the wrong direction around the ability to use data
for research purposes. I’m not suggesting that
there’s not value to consent and I’m not suggesting that people ought not to have
more control over their data. That’s what we’re all about
at Ciitizen. But at the same time, I don’t think
that where you draw that line is it’s consent or nothing. It also puts a little bit
too much of the burden for protecting people and using their data
appropriately on the individual. It’s frankly really easy to get
people’s consent for things, even if you do it really well
and inform them. A lot of times, and particularly
with sick people, they say yes because
they want to save themselves and they want
to save other people. We need other ways
to ethically review whether some particular use
of data ought to move forward than just relying on
whether says yes or not. Yeah. I’d like to echo
that last point, Deven, that you made with respect
to people saying yes. Sick people say yes to things that I don’t think
they fully understand. I know speaking for myself,
I’m one of those people where I have been asked
right before surgery if I would sign a consent form.
Of course, I’m just like, am I going to live,
am I going to die. Oh, you want me
to sign something? Okay. And I’ll go ahead and sign it.
And what I might be signing is I have no right
to request any of that tissue. It doesn’t fall under
HIPAA protection. The institution that I may have
signed this contract with is able to prevent me
from acquiring that tissue, even if they’re not using it
for anything. But I’m not allowed
to acquire it and put it toward research
at any other institution unless it benefits
them directly. And these are things that
we are often asked to say yes to at the most dire moments
of our lives. I think we have to consider when
it is we’re asking for consent and for what, in addition
to how we’re asking for consent. And that’s not a conversation
I’ve really ever heard brought up in a serious way
in these circles at this time. It’s how you get consent,
but not when you get consent. So I definitely want
to flag that. And then a question, Deven,
for you is I would love to hear if any of the institutions that scored
pretty low on your scorecard were made aware of those scores
and if they responded or if there’s any kind of circle
of feedback where they can be like,
look, we fixed it all. So we emailed
all 51 institutions two days in advance of the scorecard
coming out and let them all know
what their score was, and I got calls from three. Now, having said that… [inaudible] No. One of them
was at the bottom. Two of them were not
so bad actually. They were in the middle. But they were concerned
that they were in the middle. The one at the bottom
was very concerned that they were at the bottom. I wish I had gotten
26 phone calls, but I didn’t. But I want to say
a couple of things about that. One, they were very good
conversations with those three institutions.
Two, we did this, and it’s the first time anybody
has ever done this with a scorecard that names
names and evaluates people. We’re going to do this again for another
I think 250 institutions. And we’re also going
to rescore anybody that we had
more recent requests for. So there’s room for improvement. There’s room also to come down if you’re not consistent
in how you do this. But this is not a one-shot deal.
We’ll do it again and again. And I suspect that over time, if we’re able to sort of keep up
a two-month tick with this, we’ll start getting some more
phone calls. [inaudible] Can I just follow up on that? Why not publish the names of the
institutions with their scores like a consumer reports
type of thing? -I did.
-Oh, you have? It’s on the website. I’m sorry. I missed that. No, it’s okay. Yeah.
It is on the website. It is the first time, I think,
that anybody has done that. Quality measurement 101, right? What gets measured
and publicly reported gets improved over time. So that’s what
we hope to happen. Now you just have to get funding
for little plaques that they could put up
at their reception. We’re a gold group or whatever. Yeah. Woohoo.
That’s a really good idea. Question regarding
the informed consent. It is clearly flawed, but
there are alternatives, right, like dynamic consent and actually ownership
based [inaudible]. I just want
to hear your thoughts on those different
consent models. So having a lot of friends on
the IRB in 2014, ’15, ’16, ’17, the overriding palpable emotion I got from them over the NPRM and the common rule
revision process was abject fear because change is hard and they were concerned
about centralized IRBs. And I think with respect
to consent, I’m not saying anything that people in this room
probably don’t know, but it’s not really
for the participant. It is to mitigate liability
for the institution. And that’s why every consent
form at my institution, at least, is way too
long, way too jargony. And since I was the geneticist,
when I was in the IRB, it was always do we need
[inaudible] language in there. And it was simply kind of
a box checking exercise. And I say that with nothing
but respect and affection for those people because they’re
trying to do this thing that has all of these rules. And again, I think it
comes back to incentive. So I think Deven is absolutely
right that Sage Bionetworks has done heroic work in changing
how we do informed consent and making it a much
more authentic process. But how do you convince an
academic medical center to basically start over? I believe maybe the
new common rule is starting to nudge a
little bit though. Wasn’t it adopted that now
consent forms have to have a summary on the first
page in plain language? -Are you skeptical?
-Two cheers. You reminded me. That is consent forms, at
least for clinical trials, are supposed to be made public. But I don’t know when that’s
going to take effect. But I think once we can hold
these 35-page documents up to the light, that maybe like what
Deven did with HIPAA, it will be possible
to shame people. Can I just make a
real quick plea? We have a pretty
diverse audience. Can someone describe or tell
us what common rule is and why it’s important? Before you do that, I also
wanted to make a comment. A lot of the things that
we’re talking about today don’t necessarily fall under
the common rule or HIPAA. And there might be other
things that are relevant that anybody on the
stage could bring up. I think everybody
would appreciate that. Well, I was just reviewing
before I got up. I just got a Department
of Justice grant to use genealogical,
forensic genealogy to help people get out of prison who are innocent by finding
the real perpetrators when we have a little
DNA from a crime scene. And they just put out a
temporary statement about forensic
genealogy that is, I don’t know,
seven pages long, and I can’t figure out if the
money they just gave me, if I can use it for what I
said I’m going to do. Every group seems to have
new rules coming out because of what’s going on. But what we haven’t seen and
what I haven’t heard much of today is actionable
legal ramifications. Like what needs to happen is
the lawyers need to fire up. There have to be lawsuits. People have to be able to show they’ve been damaged
by these things. And then we’ll have some
meaningful discussions. But as it is now, every shop
seems to make up their own. I write my own consent forms. I’m at Boise State University. How many forms does my IRB see? How aware are they? They’re professors like
me on that committee. But what I found is basically
like with the Vanderbilt story and I haven’t followed up
with them in several years. When I asked how many
would you have gotten. They got 100,000 biobanking
samples at the time I talked to them. How many would you have
gotten if it was opt-in? And she said 50,000. She had a number, which means
they did some research. And so what it seems to me
is if I’m bringing grandma to the hospital because
she broke her finger and I’m saying, Grandma, do you want to give your
DNA, it’s just like when the retirement person
comes around to me and says you got to put your
money in somewhere in this 401. And all they’re trying
to determine is how risk averse am I. If I’m someone who throws
caution to the wind and is all for giving
humanity my DNA, that might be a green,
right, on a form. And there might be a yellow
and there might be a red because really what we need
to determine from patients is what is their risk aversion in
a very simple calculation. We don’t have to explain
to them what they’re doing. But if somebody is like, I don’t
care what you do with my DNA, they could check off the
form that says give my DNA. But that’s really what, to me,
the clinicians need to determine in my experience
with sick people, is you need to
assess this person. Is this a person who wants
to give their stuff and doesn’t care about risks? Are they really risk averse, in which case you need
to care for them? Or are they on the middle ground and you got to spend more
time talking with them. But by treating everyone
as a semi-expert… I mean, it frightens me when you say there has to
be a frontpage summary of a consent form. That means it’s a
multiple page form. No one reads anything
beyond one page, right? I mean, we’ve all signed
mortgage documents. There’s 50 pages. I’ll
look at the first page. So I really think this kind of
measurement of risk aversion is probably what we should go to when we can obtain
personal consent. An interesting thought. To get back to the baseline
question around level setting. So the common rule is a set of
requirements on human subjects research which applies when
you’re doing research on humans and when you are doing research
on identifiable data about human beings. And it needs to be followed when you are doing
federally funded research or if you are doing
research that itself might not be federally funded but you are in an institution
that receives federal funding and you have agreed to
do all your research in compliance with
the common rule. -Christi can correct me.
-No. I would just add that what
the common rule requires, just the basics, is it does require that
these research activities be reviewed by an
institutional review board before they begin who will
evaluate the protocol, the risk and benefits
of the protocol, and make sure that the
risks are reasonable in relation to any benefits
that might be received. And they also are charged
with ensuring that informed consent is received
or obtained, I should say. And there are detailed
provisions about what needs to be in the
informed consent form. So we’ve had the common rule. It’s called the common
rule because over a dozen federal agencies
have signed onto it. So we call it the common rule. But it’s been place
for over 20 years, but about six years ago
it got another look because it was
recognized that research and technologies have
changed a lot. The original common rule was
probably not as protective or was missing some things. And so it went through I think
it was a four-year long process to revise it. There was an advanced notice
of proposed rulemaking, a notice of proposed rulemaking, and then it finally
was finalized. And it went into effect
in early 2018, I believe. It went back and forth for
so long I don’t remember. January 2018. But it generally is our
touchstone for research ethics. And I will add, although the
common rule covers generally federally
funded research, some private researchers
are still they are still
covered by some federal research
subject protections to the extent that
their activities are regulated by the FDA. So the FDA has adopted its own
research subject protections, which are not identical,
but are close to it. Right. And then there’s HIPAA which also has rules
around research uses of data for entities
covered by HIPAA. So a very short legal primer. But it is correct that
it doesn’t apply to everyone collecting
data in this space. I just want to say it
becomes an alphabet soup. And something that I’m
really worried about is sort of different
definitions and standards. So the common rule talks about
data being not identifiable and the HIPAA talks about
it being de-identified. And then the 21st Century
Cures Act was enacted, and it’s concerned about
sensitive private information. And those definitions, they
are not talking to each other. And I think that further
confuses this landscape. So if you have the impression
that the common rule is like the Old Testament and we
are Talmudic scholars or presumptive
Talmudic scholars, you’re absolutely
right because there are all of these kind of
metaphysical weird things. For example, Greg knowingly or otherwise
alluded to this earlier. If you’re doing
quality improvements, then your work is not
subject to the common rule. And so there’s an incentive if
you don’t want to go through the IRB gauntlet to frame
what you’re doing as quality improvement. On the other hand, we’ve seen
counterexamples most famously or infamously Peter Pronovost’s
checklist at Johns Hopkins that Atul Gawande
wrote about where they were trying to
just put up a checklist. Wash your hands
before you do surgery. And they were studying that. And the fact that they were
studying that in the eyes of the Office of Human Research
Protections made it research and required full IRB review. So there was another
question online that I think would be relevant. They asked or commented that
there has been a lot of discussions about challenges and ethical considerations
of sharing data, issues with
interpretation, privacy, and clear communication of the
limitations of the privacy. Do panelists have any
suggestions on paths forward or next steps? -Repeat it?
-Yeah. Would you mind? Sure. So there has been a lot of
discussions about challenges and ethical considerations
regarding the sharing of data, issues with
interpretation, privacy, and clear communication of the
limitations of the privacy. Do panelists have any
suggestions on paths forward or next steps? I do think that there is — there’s a lot of
models out there, and there are more
popping up every day. And I actually don’t think
that’s necessarily a bad thing because I don’t know
that there is any one way to do data sharing for research, just it’s this way or no
way at all, right? We’ve had lots of examples here of more patient
controlled models, like the one that we’re
trying to build at Ciitizen where people get their data and
push it out to the researchers in the ways that they want to. We’ve had other examples of sort
of ClinVar and other databases where there’s an attempt to
really create a public resource. Corey and her question
and some other folks that I was talking to at the
break were talking about, all right, we’re
talking about data, but what do we do about tissue. Patients might have the ability
to go get their data, but they don’t have the ability
necessarily to get their tissue and make sure that it is
used for research purposes. So I think there’s lots of
sort of potential out there. There are research
consortia that exist around pediatrics in particular. They tend to do a really good
job actually at coming together to share data for
research purposes. So I think there’s lots of
promising models out there. Maybe we don’t do
enough to study the different aspects
of those models and why they work
to build trust, to facilitate sharing, to do so in a way that
produces results that then can be replicated. We report the outcomes of that
research or we hope that we do, and we don’t do
enough to study maybe why some of those
research networks and approaches work or
don’t work so that we can learn the mechanics of how we do this
as well as getting the results. Maybe we ought to be able to
trademark our genomes and trademark our person
while we’re alive. And I don’t know what happens
once people are dead with the common rule. Are samples from
deceased people covered? No. So the moment you
die, that’s it? [inaudible] You’re not human.
HIPAA survives though. For 50 years, but there are
special provisions around research with deceased
person’s information. It’s a little less stringent if
you can justify that you need the deceased person’s data
for this particular project. So that might open
up a whole field. One small suggestion. And that is we need more
genetic counselors for the interpretation piece. They are so important. They are more important
now than ever. And I would love to see them
embedded within any systems that are providing information
back to patients to make them available. [inaudible] maybe somewhat
related questions. Christi, on your slide you
mentioned that you were kind of concerned about the policy
and the regulation aspects. Some things slipping
through the cracks. Maybe in my mind a
related question, which is the
relationship between personal utility and
clinical utility. It seems like the regulation may
impact one but not the other. Well, actually the HIPAA access
right says that you cannot… I don’t know if it’s the
interpretation or the actual regulation, but in terms of
personal utility, you can have access
to anything you want so long as it qualifies as protected health
information contained within the
designated record set. So if you want your DNA
to make wallpaper out of it, that’s fine. And your healthcare
provider is not allowed to deny you
access on those grounds or any grounds and really
is not even supposed to ask why you want it. So personal utility is
respected whatever utility you might be able
to gain from it, and that includes enjoyment. And I think that is
a potential benefit. Individuals can
enjoy, for instance, participating in
scientific discovery to the extent
they’re doing that or just enjoy learning new
information about themselves. So I’ve been hesitating
to ask this question. I’m Josh Waterfall from
the Institut Curie. I’ve been hesitating
to ask this because it’s really for Richard Milne who isn’t with us anymore, but maybe some of you
can give some insight. If I saw some of
his slides correctly, one of the questions
was they distinguished whether people were willing
to let their data be analyzed by academic researchers within their own country
versus other countries. When I deposit data, this is
not a difference I can make [inaudible] has
to check the country that the research is from. Is this really a
direction that consent and data sharing is moving in? What can you say about that? A couple things
to say about that. One is when we’re talking
about genomic data, there are actually national
laws in Russia, India. Many parts of the
world are passing laws in order to prevent
biocolonialism. They’re not comfortable just
sending their data to an out-of-country depository. So there are national laws. There’s a thicket of them
that have to be negotiated. And so that’s a reality. The other thing,
it is interesting. When Mary Majumder has done a
survey of people’s concerns about sharing data
across borders. And it turns out that the trust
level drops off as soon as you add the going to
a foreign country. It drops off even more
actually than to a company. So there’s something going on
that hasn’t been really probed. I don’t know that
that’s showing up in informed consent
documents yet. But it’s one of the
things that we’ve observed as a phenomenon that’s
beginning to emerge. There’s lot of friction in
international data sharing. I’ve seen it show up in
the data security context. So I’ve never seen it in an
informed consent document to deal with individuals
participating in research or individuals consenting to have their data
used for something. But when entities
consider whether they will share
with another entity, there is a lot of concern
across cross-border transfers of data because of the
differences in the laws. And then also it comes up
also in data storage and the willingness to
use offshore providers, particularly in countries where there is some
uncertainty regarding whether the rights
that the entity feels they have in that
data will be respected. Along those same lines, for
those of with rare cancers, sharing genomic data across the
borders may be the only way we get a large enough sample
size to come to any conclusion. The dynamics of
sharing rare disease data are so unique and so important. And you’re absolutely right. It doesn’t mean that it’s not
an obstacle unfortunately. I’m wondering if you could
talk a little bit about underrepresented communities,
not just in the researchers but also some of the
issues that came up I believe it was Havasupai some
years ago in Arizona. Finding out their ancestry
and forensics about that and populations that came over. That at the definitional
level and what is ethical and how do you
share information. But then also just communities
that have a well-earned mistrust of medical establishment,
of participation, and also just simply of
at the point of diagnosis, is this really when you need and when you should be
asking someone to think about these expansive big issues when they’ve just received a
terminal diagnosis of some sort. Bless your heart for
asking that question because I realize I should
have put that in my talk. I really apologize. There are a bunch of issues
that do bear on indigenous communities in
the United States and Canada and all over the world. Many of the tribes have actually
— the Navajo had a moratorium on genetics research. They’re beginning
to rethink that now. There’s a whole group of the
National Congress of American Indians that is thinking about
this and their policies are focused on retaining
[inaudible], which is very serious
and has real legal meaning. And a history of distrust. My own university is
the source of that distrust in part because of
the Havasupai case. At least at ASU, we have decided we just have to
grab that by its reins and confront it directly
and be quite honest that we know that we screwed up and
let’s not do that again. And we know that we can’t be
setting the health priorities for an indigenous community. That said, we do have a
really serious problem. And think about the example
that I gave of BRCA mutations. If you’re an Ashkenazi
Jew, you have a good data set and a lot of people
that have been tested. One of the numbers
that was on my slide was a study that
came out in 2016, three years ago, was done
in China on sequential people who came in with
breast and ovarian cancer. 41 percent of the
variants that were discovered were not in public data sets
that they compared that to. We’re going to find that over
and over again in Africa, in Asia, in Latin America, and in every indigenous
population for certain genes, like late onset dementias
and things like cancer, late onset cancers. We’re going to be finding
founder mutations in populations that have not been tested. We’re going to find variants
that we don’t know how to interpret
until we actually get those communities engaged
in our health system and include them in our systems
of gathering information. And that’s one of the
things that all of this is trying to solve, right? We’re trying to diversity
the base of people that are in the tent. So again, we’ve
got these problems that are almost
antithetical to one another. One is not enough stuff going on and well justified distrust
of the research system. And we have to solve both of
those problems simultaneously for indigenous communities too. I’ll just talk about the
forensic side of things. The criminal database we have
for DNA is enriched for minority populations for
all the socioeconomic reasons you might imagine. However, there’s
been a bit of a flip. And where this became
important is in something we call familial searching. It’s not genealogy. It’s when we look at
the criminal database. And some states will
allow you to look for a relative of evidence. So if we have semen, as I spoke
about in that earlier case, I was trying to get our state to
look for a brother, cousin, uncle of the semen donor in the
convicted offender database. Idaho wouldn’t
allow me to do it. Some other states do that. But that, obviously, makes you
more likely to find a match in enriched minority
populations. So there’s significant questions
about civil rights that can be brought up there. The flip is that it’s generally
Caucasians who participated in the genealogical studies that
are the publicly available databases like GEDmatch to the
point where before… GEDmatch just changed its rules. Now you have to opt in to
being criminal searches. But before that, they
had enough in there, 1.2 million or something, where there was a 65 percent
chance that any Caucasian in this room who left DNA, we could find your last name through this publicly
available database by swabbing your
comp or whatever. And so that just shows you how
these different demographic groups have concerns that are
sometimes kind of unpredictable. Another issue is if we ever get an equal rights amendment
in this country, there’s going to be a
huge concern about the fact that we have male specific
markers in forensics. So the Y chromosome DNA
is a big part of my field that’s because 95-plus percent
of the rapes are by males. And so we look at
male specific markers. There is no equivalent
female specific markers. And so again, you have a
database bias where one group is represented much more
for good reasons. Anyway, I think all of this
starts becoming very complicated as there’s more and
more data being shared. Who is it that’s participating
in medical studies? Who has to worry about the risk
of their information being shared versus people who are not
reached by the medical system? So there’s a lot of social
issues that we’re dealing with. I just have one point
I want to bring up with respect to consent. It relates more to the
commercial area and it was from ten years ago. Might help inform figuring
out the questions you want. At the time, we were
talking about privacy and personal information and shipping things
to people, orders placed over the internet. A woman succinctly
put it quite well. She said there’s great concern
inside our population for retaining personal
information, not releasing it, making sure that
everything is kept private. And they’re willing to give that
all away for free shipping. [inaudible] motivation
for the people coming in, you might find [inaudible]. Maybe they said
yes to free shipping, but they didn’t agree to then to have that data
subsequently used for about 90 other
purposes, right? Yeah. [inaudible] Okay. So Christi, you spoke
of third-party tools that interface with
DTC for consumers to use. And this morning
during session two, there were a lot
of concerns with DTCs and in particular the
fact that individuals are not empowered with the
resources to really understand and use the information
provided by DTCs. So the question is, are
there third-party tools that can help
address this issue? If not, how do we incentive the
development of the resources that people actually need
to really understand DTCs? Yeah. I think that many of these tools are
a response to the limited information that the DTC genetic
testing companies are providing or can provide under
FDA regulations. So my understanding is since you
have access to your raw data, that was the response. Let’s build these tools to
help understand those data. So… Yeah. I mean, I guess I’m not
understanding the question. I think we are seeing the
universe somewhat expand. And I will say Promethease was
just purchased by MyHeritage, which is a major
direct-to-consumer genetic testing firm. And so I’m very interested on
what they are going to do and whether this might set some sort of precedent for
DTC genetic testing companies offering more sort of
ancillary interpretive tools along the lines to the bridge
to the literature services. And that’s what
they call themselves, that they’re not providing
diagnostic information. They’re just taking that
information from another source and organizing it and
presenting it to users. So Promethease pulled its
information from SNPedia, which is a crowdsourced wiki. But other third-party tools are
pulling from other data sources. So there was another
question which actually has gotten some likes,
several likes online. And the question is, what would
be the funding mechanism or motivation for researchers
for the extra work in summarizing research data and sharing research
data back to the patient while all this work
may not lead to a visible publication. So it’s getting I guess
back to how to motivate people and how do we incentivize
researchers to do this. So the analogy I use is
open access publishing. So when Harold Varmus
and Richard Roberts and Michael Eisen, et al. hatched this whatever
18 years ago, there was a lot of resistance. There was a lot of resistance from commercial
science publishers. There was a lot of
resistance from societies. This is our revenue source. And to his enormous credit, Harold Varmus said
we’re going to do this. So I think it could… somebody
using the bully pulpit could help. But I think it has to become
sort of a cultural norm. It has to be this kind of thing. It’s looked upon as a good deed, maybe something you
can put on your CV the same way NSF
rewards public outreach, public engagement. And that takes a long time. And I’ll just add
to that real quick. Andy Faucett from Geisinger. One of the things we
did early in ClinGen is we created a wall of fame and a wall of shame like
Deven did in a sense I think publicizing the labs that are contributing their
data on a regular basis and the labs that are
not contributing their data. And then also going
to the institutions that are ordering testing and saying are you using labs
that aren’t sharing their data. We actually have
a policy at Geisinger that we only use
labs that share data. And if you don’t share data,
people go somewhere else. So I think you have to… But again, it’s recognizing
those who are doing what we think they
should and recognizing those who are not doing what
we think they should and being willing to
stick it out there. And if I could just
enter one thing. I’m going to sound Pollyanna
here, and I apologize for that. I think most people, even in the
hardest core molecular biology labs that are doing biomedical
research actually do want to reach the end patient. That’s why they’re doing
what they’re doing. And we just need to make
it a little bit easier and have a mild reward set. And we don’t have that. It’s a culture of
that’s not real science. And I think we need to
broaden that definition. Certainly I see as part of my
task broadening the definition of what good science
is to include the people that you’re doing
the science for. We haven’t built
that into our culture, but I think that’s
what’s going to happen. Millennials will bring that. We need to work hard
to make it easier. I also think it’s the culture
of human subjects research at academic medical centers. My friend Carl Elliott
compares getting IRB approval, it’s a combination of a tax
audit and a trip to the DMV. And so people are like,
if I can just do this. It’s like what a lot of
my students say to me. What do you want,
Professor Angrist. Just tell me what you want,
and I’ll give it to you. And that’s not the way
to have a sort of engaged, reflective this is
a good thing to do that I should want
to do rather than what do I have to do to get
this in my rearview mirror. So if I could also
make a comment on this. I don’t have a question.
I just want to add in. I do think that there’s
something seriously wrong with our reward mechanism
because they do have to publish in order to keep
their jobs, right? We’ve really done a bad job of
setting up the environment for clinical researchers and
pre-clinical researchers. But also, there is a movement
afoot to change that. Unfortunately, it’s not coming
from maybe the people with the most money that would
wield the most influence there. But there are small efforts. There are foundations who are
starting to write that into their grant guidelines. If you’re not sharing the data, then you don’t get
this pot of money, right? And so I think we just need to
look at that on a broader scale. And certainly NCI could make a
huge difference with that if they were to change
some of their policies and put a little
more carrot there and a little less stick. But there are some
efforts out there. One’s going on right now. I’ll just give a little shout
out to researchsymbionts.org. It is actually an award that
people can apply for if you’re sharing your data. They want to recognize that. So please do. There haven’t been a lot of
applications right now, and the deadline is approaching. So if you know somebody that’s
sharing their data, please encourage
them to apply to that because we want to reward
that kind of behavior. I was wondering if
anyone had any familiarity with any case law that’s actually where
patients have shown successfully through the court
systems and won damages. We’ve talked a little bit about how long these
consent forms are, in part just to protect the
institution and that lawyers all have different
interpretations with all these
different institutions. But what are they drawing
from in terms of real fear? I know there was the recent case
that was just filed this summer. A gentleman thought that his
data had been turned over from some Chicago hospital
inappropriately to Google for some research. But I was wondering if anyone
had sort of a more current understanding what’s actually
happened in the court system. That’s a good question, but I haven’t really
studied that in a while. There’s the issue around the
oxygenation rates for preemies and some of the issues that
happened there with respect to the research that
was done there. That wouldn’t surprise me if
there was some case law on that. But I’m not familiar with it. And I’m not familiar either with
any case law around harms from — physical harms. However, Bob just reminded me of
an Alaska case that I believe is still working its way
through the court system. So Alaska is one
of a handful of states that has a genetic property law. It’s a state law that says an
individual’s genetic information is that person’s property, and
it cannot be disclosed, et cetera, without permission. And there was an individual
in Alaska who was tested by ancestry.com and was receiving
all kinds of spam email and got online to see
if his email address had been posted
somewhere online. And that’s when he found that
actually his genetic data had been posted online. So he sued Gene By Gene under
that statutory provision. And it was going back and forth. Gene By Gene claimed it was
unconstitutionally vague, and that didn’t go anywhere. -Bob, what’s the latest?
-You know? I don’t know. I just know about the case. You asked if there’s case law. There probably will be. My go to on this kind of stuff
is Andy’s colleague Jen Wagner at Geisinger who really
tracks that space closely. I’m sorry if I put
her on the spot. My question is about IRB
approval for consent forms. At my institution, we’re seeing
more and more referral to another IRB for a
multisite study. And some of
these things come in, and our IRB never
would have approved, but this is
getting handed to us. Now we have the single IRB
initiative coming in next year. I’m worried that researchers
are going to be going to look, because they don’t want to go
through the tax audit procedure, looking for where is
it easiest to get approved and who is the
lowest denominator. And that’s what we’re
going to be getting in. I really think some of
them are extremely subpar. I’d like to hear what
you think of that. Well, I was just going to say. It’s really interesting. The history of the IRB system
is you go back to… it was actually created by NIH under threat of legislation
from the Senate. And all of us have gotten
used to these exemptions. All those exemptions were
written in 24 hours just before the regulation was posted
to the Federal Register. Charlie McCarthy has been
threatening to write a book about this forever. So the principles
were really clear. You need to get
informed consent, and we need an independent audit of whether the research
makes sense enough that people should
volunteer to get into it. There was agreement
on that general principle coming out of
the National Commission. But it took ten years
to get the common rule through the first time. It actually took ten years to
get the common rule revised. And so it’s a really,
really inertial process. But one of the reasons that it’s
so inertial is all the authority was left at the
institutional level. There is no appeals process. There’s a federal office that
kind of oversees the process, but there is no appeals process. The way the law usually works is
you have a district court that has an appeals court
that has a supreme court, and you get some harmonization
over time through the case law. We don’t have that
capacity for IRBs. So, yes, there is going to be
inconsistency among IRBs for the foreseeable future
because there’s no mechanism for harmonization and
building case law. It’s just a flaw in how the
system was set up because it was set up really quickly
under pressure. I don’t think — you’re right.
Could there be form shopping? Yeah. There could. And the other problem
is there are IRBs that are ridiculously
anal and painful. And that’s an equal problem. And we’ve got both problems in
the real world at the same time. We have one last question
before we end for the day from the online audience. Yeah. So… Is there a role for policy
makers to provide tools and/or guidance to
research participants to help them interpret
their genetic data or should we focus
more on facilitating access to genetic
counselors and other experts? I don’t know if I would want
the policy makers educating the public on this. But certainly Christi’s call for
more genetic counselors feels like an issue where if the
policy makers put some funding behind that in order to help
those programs flourish and to encourage
people to pursue that as a career is
one possible start. Quickly related to that, there’s never been
federal funding to support genetic counseling, one of the fields that
got in a little bit late. So yes. And just a comment
kind of in general, which is I think let
1,000 flowers bloom. We’re experimenting right now
with a bunch of new technologies that are generating oodles and
oodles and oodles of new data that we’ve never
had to deal with, and we don’t have the system for helping people
interpret that yet. The online support groups
can be incredibly powerful for doing that, and
they’re asking for help. We have to build that bridge. The genetic counselors are
really, really important. There aren’t enough of them, and we don’t have a
system for paying for them. That’s a priority for them right
now is get the states to say, hey, yeah, we’re actually
health professionals, please pay us other
than through a doctor. So there’s a bunch of stuff. And I think we just need to let
the system kind of adjust and we need to be paying
attention to it. And, yes, policy makers damn
well better be paying attention to it because they’re the
ones who control the money. So before I forget, I
just want to say that I am a very poor
person’s Michelle McGowan. I wish she were here. But I would loudly echo
what has been said about the need for more
genetic counselors. And I would go further. I don’t know if it’s still true, but there used to
be more active astronauts in NASA than certified
adult geneticists. So we have unleashed this
torrent of data on ourselves and each other, but we no longer have a
PhD medical genetics program. And as a clinical geneticist, one said to me I
could have stopped after I completed my residency
in internal medicine, but I decided to do a
fellowship in clinical genetics so I could make less. Quick comment on that. We need not only
geneticists and counselors, but we need
molecular tumor boards. For people who have
genomically driven cancers, you can’t go to
a genetics counselor. You have to go to a
molecular tumor board, and there are very few of them. So I think we’ll
end on that note. This is a nice
transition to tomorrow, which is a more clinically
focused session. So please join me in
thanking the panelists and all the speakers for today.

Leave a Reply

Your email address will not be published. Required fields are marked *