Multilevel Interventions in Health Care Conference: Discussant comments by Ernest Hawk, MD, MPH

Uploaded by NIHOD on 05.05.2011

>>>DR. STEPHEN TAPLIN: So we're going to finish now with a
presentation by Dr. Ernie Hawk. And Dr. Hawk is a great
addition to our team. He's the President and Division head for
Cancer Prevention and Population Sciences at the University of
Texas, M.D. Anderson Cancer Center. Dr. Hawk's been
involved in chemo prevention research focused on
non-steroidal, anti-inflammatory drugs and Cox 2 inhibitors.
And he has earned numerous awards, including the NCI
Research Award for Distinguished Achievement in
Cancer Prevention. Dr. Hawk currently serves as the Senior
Editor of Cancer Epidemiology, Biomarkers and Prevention
and Deputy Editor for Cancer Prevention Research.
He earned his medical degree at Wayne State University
and his Master of Public Health degree at
Johns Hopkins University. So we are very happy to have Ernie
involved in this work we look forward to his comments.
>>>DR. ERNIE HAWK: Thank you, Steve. I came at this
less from the perspective of a critique of the papers and
more as an eager learner. Because, indeed with regard to
the first paper, by Dr. Warnecke and his colleagues,
I think we're facing a very similar situation in Houston to
which he faced in Chicago. And he has taken a number of
important steps to try and resolve that disparity. And so
the questions I have really relate to reading the paper and
trying to understand how we might learn from it, to change
practices in Houston. The second paper based on genomic medicine
was much more futuristic in its outlook. And what I attempted
to do is provide a bit more context, some real life
examples of how this is implemented in care in our own
institution. I think we've already heard, particularly in
Dr. Devers' presentation this morning, the tremendous changes
that are underway in American medicine. I've outlined some
of them here. I'll highlight a couple that weren't highlighted
previously. And they're the top two bullets. That is a poor
understanding of cancer and its causes as well as considering
issues of risk which is inherent in all areas of
intervention, including multi-level intervention.
And the poor tolerance for issues of complexity or
personal risk in our population. The explosions in technologies,
the pace of development, have been covered by others.
The transitioning of the medical care system; it is an absolute
earthquake, at least for NCI-designated Cancer Centers
as we think of how we're going to integrate into this new
system. And I think that is tremendously important because
all of the research we're talking about doing is built
upon academic medical centers or at least a lot of it is.
And so how that's going to shake out, I think is going to be
incredibly important. That last bullet there on that
section, the need to demonstrate value is of
critical importance for all cancer centers as well.
We're struggling with how we're going to do that, when indeed
the metrics on either the quality or the cost side are
not well-established currently. In the Warnecke paper, again,
I came at this as an eager learner. They discussed
policies as they're developed at a national or state level
and how they implemented them. We have to remember that
policies are always a little less out of the control of
certainly a research team. They're well intended.
They're hopefully evidence-based. But they
usually represent some compromise between science,
fiscal concerns, and political maneuvering. And so they're
rarely the best that anyone of those disciplines have
to offer. Nevertheless, they gave a very compelling story
of administration implementation of those policies to effect
change, and questioning issues around access, quality and
environmental stress. I thought one of the really interesting
aspects -- I work more in the area of early phase drug
development, and issues of signal transduction are
absolutely essential within a cell. This gave clear evidence
that the role of signal transduction across these
levels of implementation are critical as well.
And that's critical to try to link the good intentions to
effective and efficient actions and ultimately improved
outcomes. They wrestled with these issues of disparities
in breast cancer mortality and considered at least four
possibilities that are on most radar lists I guess facing
a similar situation, poor or limited access to,
in this case, screening mammography, poor quality
of that screening when it's obtained, poor quality of
treatment associated with it downstream of it. Or biological
differences. And their approach to this was quite elegant.
First of all, it sounds as though there was an
evidence base preceding this. But they convened a large
number of individuals, some 102 individuals from seventy-four
organizations, giving some idea of the scope when one attempts
to undertake a multi-level intervention. Secondly, the
task force worked for a period of six months to generate some
thirty-seven specific pragmatic, evidence based
recommendations. I left myself wondering though how are these
identified and agreed upon in this very complex organization.
I think that's critical for progress. What were the
criteria roles and responsibilities of the various
individuals, organizations? And was this ultimately a policy
level intervention alone or did it involve others? These were
questions I had as I read through the piece and attempted
to learn from it for what we might do in Houston.
Next, they had a very effective strategy of taking this forward
to Illinois legislators and pass really groundbreaking
legislation to try to reduce disparities that involved
elimination of co-pays and deductibles for mammography as
well as patient navigation services. However, I was again
wondering whether these were all of the task force
recommendations? Were all of them incorporated? If not,
which ones were selected? How was that process managed?
And then based on the evidence that was generated and
mentioned in the report, did that serve as a foundation for
the action that was taken? Or was it kind of concurrent at
the same time or indeed entirely independent of it? That I think
would benefit at least a reader like me in trying to
understand how this may relate to a situation I'm facing.
And then in terms of follow-up, I was left wanting to know,
it took a lot of great ... had great intentions. Took a number
of very important actions which were just almost
insurmountable, I would have imagine, yet were achieved.
But the real question is have they impacted breast cancer
mortality, their ultimate goal? If not, when will we know
how far downstream does it take to see the effects of
a multilevel intervention if indeed that's what this was.
It would be interesting to know whether all
of the full package was required or specific elements
were particularly critical to the success of the endeavor.
And I think a very important issue any time you're affecting
something as important as public policy with all of its
broad implications. Typically, at least in Texas, when you do
something, it's typically a zero sum game. When you do
something, you give up something else. And so very
important questions of how is that political and financial
resources, what were the implications of focusing more
on breast cancer mortality? Was anything left off that would be
similarly important? I've listed some of the things of
varying importance here -- childhood vaccination, obesity
programs in the population, down to something that would
perhaps be less concerning, a public prostate cancer
screening program. So I think it's very important to learn
from examples such as this about how we think about
prioritizing among opportunities. The second
paper by Khoury, as I mentioned, was more futuristic
in its vision. I think it's incredibly powerful and
important. And Dr. Khoury and his colleagues have done a
great deal to advance this field. So I was a humble
reviewer. What they point out is extremely rapid pace of
discovery, the long and complex translational paths, to first
of all establish validity and efficacy, much less approaching
effectiveness. And that ultimately, of course, we're
interested in hard outcomes of real interest which is clinical
benefit. But there's a lot of investment in the T0, T1, the
early phases of exploration and relatively little in later
phases. Nevertheless, those later phases are no less
critical to the ultimate success that's intended.
So the opportunities far exceed investments. Fundamentally,
as I read this paper, it looked exceedingly familiar to me to
the complex issue of biomarkers and our difficulty in advancing
those. And so if we broaden the discussion now from genomics to
this broader concept of biomarkers, we can see there
are at least ... and here I've detailed six different ways
that these are used clinically in a population of individuals.
And our track record in applying biomarkers is not
without at least controversy. And so I highlighted here two
papers in prostate cancer screening, using a protein
based approach, not genomic, but protein based approach.
And the disconnect that we have within our process of
scientific discovery through regulatory approval and
application in the public, particularly in the realm of
biomarkers where the definitive proof that's required before
approval is validity, reliability of the assay,
rather than the overt connection to clinical benefit,
different than when developing a drug. In the realm of risk
biomarkers, we don't require that demonstration of clinical
benefit before approval and dissemination, adoption broadly
in the public. And so we're left with a conundrum such as
is exemplified with PSA screening where we have some
studies showing harm or at least no benefit and others
showing benefit and less harm. And we're left with a real
conundrum with regard to what to do. And I fear that genomics
is taking us in a similar path. And hopefully we'll be able to
learn from history and adjust the path. Well, there are
already many levels of evidence for developing biomarkers or
genomic markers. So these have been codified in the literature
and reasonably well adhered to. The challenge is getting
anything down that stream of evidence development is
extremely challenging. We also -- before adoption into the
clinic - have, as a body of investigators, have come up
with some clearly laid out rules, intended clinical use
that's delineated. The magnitude of outcomes is clear.
The estimate of the magnitude is accurate, reliable and
validated. So those are done. But the fact is in the face of
desperate patients, different clinical decisions are made
rather than referring back to these criterion levels of
evidence. And already at Anderson, what we've seen in
the last couple of years is remarkable. So these are tests,
things like KRAS evaluation of colorectal cancer, but a
variety of other markers that are applied in patients with
the very best of intentions, trying to bring the best
medicine to bear on the patient sitting before you. So there's
an explosion in these sorts of biomarkers requests. Few of
them are reimbursed. And so we're spending literally
millions of dollars on these tests and hopefully improving
the outcome of some patients. Because these are used for very
valid reasons to try to improve patient selection, target
identification to match agents to disease as well as get
better assessments of the context in which cancer is
occurring. But the challenge is we don't ... the evidence has
not caught up in that case to the medical practice.
Back when the TRWG was mentioned yesterday, when we
led that, we identified a number of critical infrastructures
that are necessary to transform our scientific discoveries
into clinical advances. I'll note that few of these exist.
I'll not spend any time on that. So, in conclusion, with regard
to the Khoury paper, one of the cases that I got out of the
paper was that because of the complexity, speed, potential
for misunderstanding, mis- communication as well as the
insufficient, in my view at least, regulatory oversight
process that we're in at the moment, particularly applied to
genomic technologies and biomarkers more broadly and the
lack of adequate preparation among practitioners, we had the
real, very real potential, while advancing our
technologies to actually increase harms, worries,
litigation, cost and disparities, all without
improving outcomes, at least uniformly. And I think ...
this is a sobering message ... but in my view at least it's a
part of the reality. And I'd love to end on a hopeful note.
I think there's no other way forward than to boldly move
into that future. But I see a lot of challenges here.
And I think the paper does a nice job of stating them.
At least perhaps not emphasizing them in quite the way I would,
but nevertheless bringing that reality to the fore. I think
there are tremendous reasons for hope. And one of the
things I think we could do is try to link at least within
government areas that are under our control. The communication
is NCI putting forth initiatives that would answer
critical questions for CMS or for FDA. Is FDA using the
evidence that's created by NCI and NIH, et cetera?
This implies a linearity that isn't reality. It's much more
complex, of course. But at least within the realm of
government, I think this is absolutely critical to model
the change we're looking for in the broader field. So I'll end
on a note of hope. And this is actually by Dr. Fineberg.
And he stated many years ago, nature is probabilistic,
information incomplete. Outcomes are valued,
resources limited. Decisions unavoidable. So, to me it's
that notion of full steam ahead, regardless of the challenges
and how daunting they may appear. Because there's really
no path in any other direction that will be fruitful.
Thank you for your attention.