Multilevel Interventions in Health Care Conference: Conference Closing Remarks

>>>DR. STEPHEN TAPLIN: Let me just do a few things and
acknowledge a few people. First, Heather Edwards who's over
there on the side, who did a lot to work with the authors,
to work with us in keeping the details going.
So I wanted to thank her very much.
>>>[APPLAUSE]
Veronica Chollette who has already left, who left
last night, did a lot of work, worked throughout the time with
Westat. Westat are the people behind us -- contractors led by
Liz Zimmerman who did a huge amount of work to make all the
details possible. So these things don't happen without
people paying a lot of attention to detail.
Irene Prabhu Das who did a lot of work with the authors to
make sure that we were on track. My colleague Steve Clauser,
Mary Fennell who helped incredibly intellectually and
Jane Zapka who also helped with us putting together the
thoughts and organizing this. So my thanks to all of them
for their work to make this possible. I also wanted to
thank the consulting group. There was a group of people,
Tom Vogt, Ernie Hawk, led by Arnie Kaluzny, also
Maria Fernandez. And I'm missing somebody. Mary Fennell was
on that group, Marty Charns who we worked on twice and who
was noticing that we worked on him twice during this
conference to talk. So we appreciate their participation
as well. They also contributed tremendously to our thinking
and will contribute to the supplement. I appreciate their
participation in this work and the contributions that
they made. I also wanted to clarify a couple of things.
The questions came up about why we chose what
we did and what the levels were. We sent out a
paper tiger. We wanted three levels because we wanted to
push things. Because we knew that there were people working
at docs and working at patients. So we set a higher
bar. We set three levels. And we did consciously choose to
have one of those levels be patients. Because most of the
work was in the community and because it was clearly a
conceptual commitment that we thought it was important to be
thinking about multi-level because of the costs, because
of all the things that Otis said today and Otis really
brought us back to the beginning of the motivation for
this meeting. We set out to look at three levels because we
think that we don't do well in health care and we are focused
on health care. And yes, community has a lot of work to
do and a lot to contribute. But we wanted to focus this thing
on health care. And that means we need to make a difference in
people's lives and individuals' lives. So we wanted individual
level outcomes. So that's the perspective we took on setting
this up. We also wanted to be clear that in the opening
comments, I said explicitly expand the foundation because I
grew up as a researcher at the Fred Hutch and at the
University of Washington and knew well the COMMIT and the
people that set the foundation in place for a lot of this kind
of work. And what we want to simply do is build on the work
that they've done. One of the key issues in doing that
expansion is to begin not only to intervene at three levels --
and we set that up as the paper tiger -- but also to measure.
And as I said in the beginning, one of the limitations and why
some of these abstracts are not multi-level interventions by
our original definition was that either they did not have
an intervention at each of the levels and/or they did not
measure an impact at each of those three levels. And it's a
really pretty simple thing, but it's actually not very commonly
reported in papers that you have both of those components.
And that's what we wanted to see in multi-level
interventions. That's why some of the abstracts did not meet
the hard definition we set out. And so I wanted to just point
out that we did talk a lot and we had an active discussion.
And I think the active discussion did exactly what we
wanted to do. It brought up some things. Like the absence
of a discussion on disparities. So all of us thought about it.
It didn't get explicit. And we need to talk about that
explicitly. Because some of what we're doing in terms of
multi-level interventions is affected and does affect
disparities. And we will try to think about how we can
incorporate those issues throughout the supplement.
I think we were open to discussion. I think the
explicit discussion was helpful. I think it is clear
that it's complex. But I also believe that there are ways of
simplifying some of this work and that the complexity is not
a reason not to move forward. It's a challenge. If it was
easy, other people would have done it by now. And we would be
moving forward. I think it also is clear that there are
some leadership issues. You all heard that things change and
people on top and people are skeptical about whether we
should move forward. The question is what does the
science say? What does the problem say? And I think the
problems are demanding that we think about multi-level
interventions. I think there is no way around it. We are living
in a world, in a complex world, of health care. The individual
reductionist approach gets us important information, but it
is not sufficient. We also need to do -- it's not one or the
other, we need to be doing both. And we are the people...
we're thinking about it. I hope more people will be
thinking about it. That's part of what we want to do.
We're not going to stop. This is not going to die away in the
corner because there's no interest in funding at the top.
There is a reason Machiavelli knew it long ago. There are
people out there vested in what we do now and the way we do it
now. And so there's going to be resistance. And they're all
going to say, "Of course, why didn't we think of it before?
Multi-level interventions. Let's fund them." It's not going to
happen. So I think that we are going to keep going and keep
moving. We're going to put the supplement together. And it
does involve some risk. It does involve leadership. Had we hope
that all of you will take part in that. We are here listening.
And we are... "we" meaning the National Cancer Institute is
here listening. There are editors of journals here in the
room that are listening. We have at least three editors in
the room that I know about who we hope will be affected by
some of what they've heard. We also have people from the NCI
besides the people on the committee. And so we wanted to
get a couple of comments from those people in closing to let
you know that there are people listening and to hear their
perspective on what we did. And so we wanted to have Bill Klein
who's the Director of the Behavioral Research Program at
the National Cancer Institute just make a couple of comments
about his perspective on what we've seen here. And then we'll
hear from Rachel Ballard-Barbash in the Applied
Research Program. These are the two people that allowed us
to move forward on this. And without their kind of
leadership and their allowing us to move it and to use money,
we would not be here. So I think it's important to hear
from them and let them share with you where they see
and why they see this work needs to happen.
>>>DR. WILLIAM KLEIN: Thank you, Steve. I've learned a lot
the past two days. I really enjoyed sitting here and
listening. Not only am I leaving here having learned a
lot, but also I have a new team which is the fabulous Table 7.
We had some great discussions and we really got into it
and talked a lot about health disparities and many other
issues. And I think I now have new collaborators,
new friendships with folks who have some common interests.
I'm just going to take a Glasgow minute.
That's what I'm going to call it... because Russ is
very good at keeping people to one minute... to comment on
some connections between what we've heard over the last two
days and some of the values, principles, initiatives, areas
of interest that we think about at NCI. And I think what I'm
going to focus on here, because I only have that short amount
of time, is the various other meetings that we have. Many of
you have attended other NCI meetings like this one. So you
know that that's something we do a lot and hopefully do well
plan meetings around common issues. And I'd like to argue
that this meeting in some ways is the glue that holds together
many of the other meetings that we've had recently. So just to
highlight a few. One was one that Russ mentioned that was
held down in Atlanta a few weeks ago. This was a joint
collaboration between NCI, CDC and the American Cancer
Society. And Otis was there and gave a wonderful opening talk.
And that was focused on cancer communication science. And as
Russ articulated, communication is going to be at every level
of the multi-levels that we've talked about here. And so it's
something that we need to pay attention to. And that meeting
was all about developing a blueprint for cancer
communication moving forward. We also had a meeting last
month on discrimination and health, overseen primarily by
Vickie Shavers who was here and may still be here. She was here
at some point. And that meeting was about how discrimination
can enter into various different pieces of the cancer
continuum and the health continuum more generally. So it
looked at what happens between patients and providers as you
might expect, but also how does the perception of
discrimination, the experience of discrimination lead to
downstream health consequences? And as you can imagine, that
happens at multiple different levels. And so I kept thinking
back to that meeting over the past two days about how the
various concepts, various processes, various outcomes
that we've talked about here can play out in a context where
discrimination might be an issue. We also have hosted
meetings that have less of a content orientation the way
these previous two have and focus a bit more on the kinds
of infrastructure questions you might think about in trying to
do multi-level research. You know about the dissemination
implementation science coming up or at least I hope you do.
And if you don't, ask Russ. There's a lot of great stuff
happening at that meeting. And then we also have a meeting
coming up in April on what we call the Science of Team
Science. This is a new field. It's a field, or at least it
argues to be a new field, a field that focuses on how to
understand team science. How does team science develop?
How can it be nurtured? And how can you get folks to work
together in a way that's productive? Well, Team Science
is something that we need if we're going to do a good job
at approaching research at multiple different levels.
Finally, we do a variety of different meetings and other
initiatives in other areas that haven't really been touched
on or talked about in too much detail over the past two days,
but are clearly relevant. And so, I'll just mention one.
And that is the focus on tobacco. We have an entire
branch within our behavioral research program focused
on tobacco. And they don't just focus on cessation or addiction
which is often the way people think about tobacco. But they
focus on the entire continuum. They talk about policy,
everything in between the individual decision to smoke
and decisions at a policy level to tax cigarettes, for example.
And so, trying to think about tobacco at multiple different
levels, that's something that people have been doing in
the tobacco field for years, decades in fact.
And so we can learn a great deal from them in trying
to move forward in the kinds of things we're talking
about here. The same would be true of food and obesity,
physical activity. Rachel can tell you about the
work that she's doing with NCCOR, the National Childhood
Collaborative and Obesity Research. They take a
multi-level approach as well. So essentially, I guess the
bottom line here is that multi-level in my view is not a
content area. It's not a scientific discipline. It's a
value. It's a principle. It's something that is inherent in
just about everything we do and everything we think about.
And so it's great to have so many people at the table
thinking about it in so many different ways. So thank you
for being here. I thank the two Steves and their big team
for putting this together. And I hope you have enjoyed it.
>>>[APPLAUSE]
>>>DR. RACHEL BALLARD-BARBASH: Well, Steve Clauser suggested I
had five minutes. Steve Taplin said I had one. I'm not sure if
that's their perspective of what I might contribute. But I
think your point about apathy, ignorance and greed is a big
problem related to our health care. I think we've actually
addressed the issues of apathy and ignorance. I think we have
done a very poor job of acknowledging how much greed is
driving the problem in U.S. health care. And we follow the
principle that free market and competition is going to get us
there. We in fact followed that same principle in how we setup
our research, right? I mean, that's what's driving how we
move forward in determining how research should happen. We do a
limited amount of directed research. And as our budget
declines, we'll probably do even less directed research.
So until we probably acknowledge that we have to
consider other models in a free market economy as a way to
solve this issue, I think we, in fact, might make quite
limited progress. In this area, I think that we've heard a lot
about the fact that we need a lot more development. We need
much more understanding of measures and methods at each of
these levels. And having just recently come from the review
of the CABIG program where NCI lept way far ahead of where the
evidence suggested we would make gains, I think we have to
be cautious about saying we're ready to institute big
multi-level interventions at this point in time. But I think
we can make some great gains by really improving the science
about measures in these areas. I'm also, this area of
multi-level analysis, as Bill said is ubiquitous in
everything that's being debated at NIH right now, whether it's
basic science and talking about how the cell operates in its
macro environment or whether it's related to prevention or
whether it's related to this area. So I think the important
thing is not to spend too much energy in defining what we mean
by a multi-level intervention and whether or not it needs to
include the patient and which other levels, but rather that
we do collect information at these different levels and
understand how they interact and how they're influencing
each other. And I'm not convinced from what I've heard
today that we necessarily always have to have the
information on the patient to make progress. Because we have,
as was described in some other areas, made progress if we have
measures across several levels without necessarily measures at
the individual level. There is, as Steve Clauser noted,
there have been a number of existing infrastructures that
can help us make more rapid progress in this area.
The Cancer Research Network is one. Dr. Collins is moving
forward with an HMO RN collaborator to make this open
to a broader array of institutes to look at issues for non-major
diseases or diseases like obesity or tobacco that affect
health issues for many disease end points. So I think an
important thing as we move forward in this area is to
understand that we're not going to be able to move it out of
just NCI. It's going to have to be across not just NCI or NIH,
but other groups like AHRQ. That was one of the things that
drove the NCCOR initiative. So I'll comment on it briefly
because I think there are a couple of decisions we made
about priorities for that initiative that might be
relevant here. For people who don't know about it, the NCCOR
initiative stands for the National Collaboration on
Childhood Obesity Research. And the focus of that effort as a
way to bring together a lot of activities related to childhood
obesity by Robert Wood Johnson, CDC and NIH, was to say what
might we have in common across these groups that we could help
move forward more rapidly. And it also came as the research
community recognized that we needed to move beyond
individual work to understanding how environment
and policy was influencing what we were seeing in this
country. So the focus of that effort is really on environment
and policy. We felt the first need was to really enhance
knowledge about measures and knowledge about data available
on those measures at the state, local and national level.
And so, this February, in early February, we released the
catalogue of surveillance systems which has seventy-five
surveillance systems around the U.S. that provide information
related to how we're doing health monitoring in the area
of diet, weight, physical activity. That is a searchable
data system. For example, when one of our fellows wanted to
understand what information we had related to a parameter, I
think they were looking at some measure of physical activity.
In an effort that would have taken that person probably
a week to several weeks to do, she did it in ten minutes using
this system. So I think we do need to use the power of things
like the Web to bring things forward more rapidly to the
research community, we're releasing in early April, an
effort to summarize and synopsize all of the data
that's been published on measures at the environmental
and individual level related to diet and physical activity.
Again, it will be searchable. We will be continually adding
onto those kinds of information systems. And I certainly heard
from many investigators that it would be useful to have some
sort of Web resource that could provide some sort of similar
information about what kinds of measures and metrics were
available. I think Dick Scott's point about the need to
really understand the social environment and how it was
changing over time is very important. Ernie Hawk said
today that irrespective of whether there's a research
initiative in this area, cancer centers around the
country are now trying to figure out how to measure
quality, how to look across levels. This is happening and
we need to figure out and provide, I think, the effort to
understand what are the rigorous ways to really study
this area and move it forward. One of the things that I have
been struck in our recent discussions about how to engage
the more basic and clinical community in getting interested
in this issue of multi-level analysis, comparative
effectiveness research, quality improvement, is to make it
relevant to the issues that they think are going to drive
the future of progress in medicine. And at NIH and at
NCI, where we hear excitement in that area is in the area of
genetic characterization that may help better identify those
treatments that will improve outcomes and reduce side
effects. In the area of prevention, because we don't
have those kinds of biomarker at this point in time, there's
much more conversation about how we might build risk models
that are better at predicting differential risk across
subgroups of the population and so that we're moving our
interventions forward in that way. And that's partly
predicated I think by -- maybe less so from some of our
screening efforts which I think we've gotten some good results
out of -- but certainly many of our prevention efforts which
have been quite globally focused have not led to the
promise that they were initially thought to hold.
And I think that's been not true in diabetes and heart disease.
But it has been very true for some of our prevention efforts
I think outside of tobacco perhaps in cancer. And it's in
part because we launch those efforts before we probably had
a sufficient assessment of how to characterize and select
people based on the risk. So I know this is the end of a long
day. And I also wanted to comment that there are a number
of other meetings that are relevant to this, in fact are
relevant to this area of work. So about a month ago, there was
a meeting on CER and methodologic research.
There's a huge overlap between CER and this area of research.
And the issue of how to really develop standards and enhance
our criteria related to external validity was a big comment that
came up in that meeting. And the issue of how to come up
with the right research design for the question that you're
asking came up at that meeting as it did in this meeting.
Other meetings that are moving forward the HMORN is having a
large meeting, as they do every year, up in Boston. And one of
the sessions for that, people were asking is there much
effort to link health care delivery systems to their
communities? So one of the efforts in that meeting is a
workshop on using GIS methodology to characterize
food and physical activities environments for individuals
and actually link that to the data on their BMIs. We don't
have data in most systems on any other parameters that
influence weight in these systems for kids as well as
looking at it related to diabetes management.
So I think there are nascent efforts to try to link health
care into the community. And I think looking across a number
of these levels without applying too limiting a definition of
what we mean by multi-level analysis will be important as
we move forward. It's been a great meeting, not in a great
venue. And I hope that we will take lessons from that.
>>>[APPLAUSE]
>>>DR. STEPHEN TAPLIN: All right. Thank you all. It's been
great having you here. We appreciate your activity.
Enjoy the day. Thank you.
>>>[APPLAUSE]