Multilevel Interventions in Health Care Conference: Presentation by Muin Khoury, MD, PhD

Uploaded by NIHOD on 05.05.2011

>>>DR. STEPHEN TAPLIN: We're going to turn now to
Muin Khoury. And Muin is going to take us back down
to the person level and look very directly at the
person, even below the skin level of what's going on.
Muin is the founding director of CDC's Office of
Public Health Genomics, which assesses the impact of advances
in human genetics and human project on public heath and
disease prevention. And he's a senior consultant to NCI, he's
been working with us now for several years. He's been a
great addition to our staff. Among his many accomplishments,
Dr. Khoury started the Human Genome Epidemiology Network,
HGNET. The evaluation of genomic applications in
practice and prevention initiative, and the genomic
applications and practice and prevention network. He's a
Board certified in medical genetics. Dr. Khoury received
his medical degree from the American University of Beirut,
Lebanon, and a Ph.D. in human genetics and genetic
epidemiology from Johns Hopkins. So we're glad to have
him here as part of the team. We're glad to have a
perspective from the patient and from personalized
medicine and what it means in multi-level.
>>>DR. MUIN KHOURY: Good morning. Thank you for
the introduction. There's a lot to cover in ten minutes,
but as I was sitting at table 7 yesterday and
talking about what we heard, I kept saying to the group, it's
complicated. And now I'm going to get it even more complicated
than it possibly is. I would like to thank all my co-authors
here. I certainly learned a lot from this experience.
Steve Clauser and Mary Fennell and Russ Glasgow were
instrumental in helping this paper. But also Marion Schooner
from the VA and Rand Corporation, and Mark Williams
from the Intermountain Health Care who are both pioneers in
genomic medicine, and Sherry Charlie from DCCPS. So very
briefly I'm going to talk about both the promise and the
challenge of genomic medicine. Talk about the (inaud.)
research agenda in the field, and how multi-level
intervention research will apply or is already being applied.
And then use one example, and then raise a few issues for
discussions. I wanted to start with an anecdote. Actually this
is a paper that Francis Collins wrote back in 1999. Francis was
the current director of NIH, used to be the Director of the
National Human Genome Research Institute at that time. And he
predicted that in the year 2010 this is how medicine will be
practiced. And he used a hypothetical case scenario of a
23 year old man named John who goes to his primary health care
provider with a family history. His father has early heart
disease. And I'm going to add to the mix that his mother has
breast cancer, but that's my own flavor on this. And by the
year 2010, which was last year, John will get a printout of his
personal genome profile. Which, by the way you can do that
right now in 2010. I mean, you can get genome white analysis
of about a million variants for about, $200 to $400 directly on
the internet. Or you can have your whole 3 billion base pairs
sequenced for a price tag that's rapidly dropping. Now
it's about $5,000 to $10,000 and will be very soon under
$1,000. So on the basis of this printout, and these are
obviously fictitious numbers that Francis came up with.
John is at reduced risk of prostate cancer and
Alzheimer's Disease. He probably doesn't need as many
PSA's as others. I don't know. Or elevated risk of heart
disease, colon cancer and lung cancer.
And what Francis went on to say, that John will
have an individualized treatment and prevention and
early detection effort that's tailored to his genes.
A smoking cessation program because John happened to be a
smoker. By the way, there is a test that you can do for
susceptibility to smoking. It's called respiragene. I'm not
endorsing it, but it gives you a printout of your lung cancer
risk based on your genome type. Now, does that mean that if
you don't have that profile you can continue smoking for the
rest of your life? I'm not so sure about that.
Colorectal cancer screening because he's, I think he's at
increased risk of colon cancer. And then a tailored treatment
for his high cholesterol based on pharmacagenome links.
Now, many of these things are in the works, they're not
happening. The numbers are fictitious, although they're
becoming more real by the day based on the science.
But John's mother has breast cancer and there are
tools that are used today to differentiate using based
on the cancer gene expression profiles of the tumor,
whether or not she needs reception, whether or not
gene expression profile would determine her
risk of recurrence, and therefore modulate therapy.
Whether or not she has a pharmacagenome variant for
(inaud.) which will influence the way, which drugs you will
take for the treatment of breast cancer. John's father
on the other hand is heart disease and he's been worked up
for familial (inaud.) single gene disorder. And also he may
be on a regimen to evaluate whether or not he has side
effects, statins, based on a (inaud.) variant. And whether
or not he needs a pharmacagenome test before he
takes Bufferin for anticoagulation because he may
or may not bleed as a result of that. This is not science
fiction folks, this is 2011. Some of this happening, some
of this is not happening. But what we're faced with in
genomics is a large evidentiary gap. I mean, you guys have been
talking for the last two days about things that have been
established, that we move to move into practice. Genomics by
and large is not there yet. But it's rapidly moving in a way
that impinges on all other areas of practice, whether
cancer or otherwise. Many challenges face us, which gets
us to the heart of the onion. I'm exploding, I'm opening up
that onion for you here and I surmise to you that these
personal biologic characteristics on the basis of
the genome type of the people as well as the tumors, and
anything in between, will effect everything we do.
Maybe not today, maybe ten years from now. So how are
we going to enlarge our scope of multi-level intervention.
So we have many genomic applications. By last count,
just last year we've done a survey, 250 new genome based
tests have been added to the market. 2/3 of them are
cancer specific. And yet we have incomplete information on
genome type outcome relationships, gene environment
interaction, gene drug relationships, clinical
utility, whether or not it will improve outcomes.
And therefore there is uneven coverage and reimbursement
of these things. Very little regulatory oversight.
As I said, you can buy your genome online.
Direct to consumer advertisement and non-outcomes, and uncertain
quality of lab testings, and a few more things. So, where
are we? I don't want to get into the weeds here about my model
of translation, the T1 to T4 model. Russ and I keep arguing
about this and I'd love to argue some more. But I use it
only as a concept to show where cancer genomics is in terms of
the research. So you discover a few genes and a lot of more
biologic insight, and then you develop promising application.
That's your T1 bench to bedside model. There is a lot of that
going on right now in cancer and other areas. T2, there is
very little of that going on. This is the area where you take
promising applications, you evaluate them clinically in
trials, and then on the basis of that you develop evidence
based guidelines. And there is very little of that in genomics
right now. And then there is even less about implementation
science and the kind of multi-level intervention
research. And then even less of outcomes research that shows
the reduction of the burden of disease. So a couple of years
ago we did a portfolio analysis of the NCI funding in cancer
genomics. We just published this paper. There was, of 1,000
genomics grants that NCI funded in FY20007, the vast majority
were discovery research. (inaud.) of bench to bedside and
then it tapers off. T2 and beyond is less than 2%. This is
where your evaluation, intervention, outcome research,
health services. There was only one lone grant that looked at
outcomes related to be RCA one. We also did a literature
analysis at the time. It matters the same finding.
Less than 2% of the published cancer genetics literature,
actually less than 1% is T2 and beyond. This is the field that
we need more of in order to move these promising genomic
applications into practice. So when writing the paper it was
really a very hard paper to write because most of it
doesn't fit. We're still more upstream, and here we're
talking about more downstream implementation. A few things
came to mind in the onion. Especially with the clinical
and lab interface. That's very crucial for genomics.
What people call in clear terms, the pre-analytic,
analytic, and post-analytic. So many others can happen,
and you ordered the wrong test for the wrong people
based on rapidly moving evidence. The behavioral social
science and communication issues between patients,
patient's families, patient provider diodes are very,
very important. And then health care organizations.
Some of them adopt, some of them don't adopt. And there is no
framework for genomic medicine. The public health system is by
and large absent in this debate, but there is a little
bit of discussion about regulatory oversight. So in
brief, folks, this is an area that is completely new and
there is very little going on. Just one quick example just to
illustrate the complexity of it. So four years ago our
office at CDC started an evidentiary panel, which is an
independent panel, dedicated completely to genomics. And we
modeled it after the U.S. Preventive Service Task Force.
Some of you may have heard about it. This group has been
developing methods, adapting methodologies, looking at
outcomes, and they have evaluated about ten genomic
applications. More than half of them are related to cancer.
The only positive recommendation they came up
with is this one. So this is a rare genetic disease that
effects about 3-5% of all new colorectal cancer, and it's
(inaud.). And you get colorectal cancer before the
age of recommendation for screening, which is 50 and
beyond. So what the working group found, and this is going
to be a hell of an implementation in practice.
So they found that there is sufficient evidence to
recommend offering genetic testing for lymph syndrome to
all individuals with newly diagnosed colon cancer, to
reduce morbidity and mortality in their relatives.
So basically every year we have about 150,000 new cases of
colon cancer in the U.S., and about 3% of them or more than
4,000 have lymph syndrome. So the recommendation said that
all new cases, regardless of family history, will have to be
tested for lymph syndrome. Why? Because once you find
those 4,000, those needles in the haystack, you're going to
effect maybe about 12,000 of their relatives.
These relatives have to be on surveillance for colorectal
cancer long before the recommended 50 and beyond.
So this is something new. This is genomic medicine involving
family members. And we held an implementation meeting last
September at the CDC where we brought people together to
discuss how this is possibly going to be done.
Obviously there are multiple levels that could be used,
but there are challenges. Providers don't know
about this. Why are we testing cases for something genetic
without their informed consent. Who's going to cover for it
because basically you're not dealing with people in the
same health care system. You have a cancer in one place
and then the siblings will be another system or not
insured that all, who's responsible for informing
relatives. And there are infrastructure needs and
then testing limitations. So suffice it to say there is,
obviously with any new recommendations that come out,
debates, points and counterpoints, to screen or not
to screen, that is the question. And so I want to
leave you all with a few issues for discussions. Obviously this
is a new area that is going to effect everything we do in
medicine, but not in the hyper way that people have been
hyping it up, but maybe more incrementally over time.
An evolution rather than a revolution. But the pace of the
technology is moving quickly. So is there anything unique about
it or should we just subject it to the same way we've tackled
everything else in medicine. And I'm mixed minded. I don't
believe in genetic exceptionism, but there are
areas of both quantity and quality to differences we need
to deal with. How do we deal with the lack of evidence
because you're not going to be doing randomized clinical trial
on everything that comes your way, and it will change very
quickly. How are we going to manage the push and pull forces
of translation. Obviously if you can buy your genome online
and people are pushing it more one way. So the question is,
based on our review of the literature, there is really
very little implementation research or T2, T3, T4 done,
let alone it's multi-level. And you see it a little bit more in
the rare single gene disorders like BRC1. There is a heavy
literature focus on breast cancer genetics. So how are we
going to develop that agenda that's actually informed by the
other fields that we've been talking about. And perhaps
maybe genomic medicine can inform these other fields given
that it's an out of the box type thinking. Thank you very much.
>>>DR. STEPHEN TAPLIN: Thank you Muin, so here's an area where
we can do some work and anticipate multi-level effects
because the actual delivery is not where we think it is.