Multilevel Interventions in Health Care Conference: Presentation by Elizabeth Yano, PhD

>>>DR. STEPHEN TAPLIN: Next we're going to talk with
Elizabeth Yano who's a co-Director and Research career
scientist at the VA, Greater Los Angeles Health Services
Research and Development Center of Excellence,
and adjunct professor of health services at the UCLA School of
Public Health. Dr. Yano's research involves evaluating
organizational influences on quality of care. She has been
the chief architect of a series of national organizational
surveys spanning VA's quality transformation,
as well as organizational initiatives in primary care,
women's health, HIV care, and in-patient quality. She has
also evaluated impacts of system reorganization and
practiced (inaud.) quality interventions. Dr. Yano has
been in the thick of it for a number of years and has an
incredible amount of experience to build on and to share
with us. She won't share all of it with us in these
ten minutes, but she will point to some directions we
can go and think about. So, Becky, welcome.
>>>[APPLAUSE]
>>>Dr. Yano: Thank you very much. So I have the interesting
task of trying to pull together everything we've been
talking about multi-level interventions in the last
day and a half, and thinking about what it really
takes to implement and spread them into real,
routine practice. In ten minutes. So with that,
I want to acknowledge my co-authors Larry Green,
Karen Glands, Johnny Yeeyan, Brian Mitman who's here,
our NCI connection Veronica Cholette,
and Lisa Rubenstein. So as we've already been discussing,
scientific evidence about what works in health and health care
tends to take decades to move into the routine care you and I
might experience in the doctor's office or the bedside.
And most of that evidence is single site and single level.
And as a result, that evidence from our perspective is,
unfortunately, quite flawed because it's been tested under
highly controlled and homogenized circumstances in
settings that are very rarely those in which the vast
majority of Americans obtain their health care. So when
applied to real world settings, we really shouldn't be very
surprised when we see a voltage drop in the kind of outcomes
that we can expect once we try and take it to scale.
That's led to greater recognition of the contextual
influences underlying both intervention success or
failure, which has motivated interventions that target those
context levels moving forward. The dilemma is that there are
very few multi-level interventions that have been
conducted along the cancer care continuum and very fewer that
have actually been implemented into practice. So how is
implementation different from everything we've been talking
about so far? Well, in this situation we are no longer
talking about testing the original efficacious
intervention no matter how many levels it's been designed for.
We're testing a set of strategies for deploying the
multi-level intervention into practice,
which requires adapting them to different contexts and focusing
on activities that facilitate the update,
adoption and implementation of attributes of each intervention
at each level that you want to work within.
That, as folks before me have already mentioned,
requires engagement and involvement of a wide range of
stakeholders, partners, and implementation at each level.
And this is harder for us researchers. And I think one of
the reasons so many of these studies are at the patient and
the provider level is, that is where we have our illusion of
control. When we begin to go to other levels,
the researcher's capacity to influence adoption and
implementation is acutely determined by the handoffs and
the support that's constructed through the partners who live
in that other realm where you as a researcher don't
typically reside. Now as if implementation wasn't hard
enough, trying to spread multi-level interventions in a
way that achieves a universal and permanent new ways of doing
business is a huge challenge. So our process in this regard
was to identify a series of cancer and non-cancer
multi-level intervention exemplars that had been
implemented into practice or police that span different
levels and different stages of the care continuum.
Now there's little time to give you the details of each of one
of these interventions. And many of them probably missed being
picked up by the other literature reviews that the
other groups did because commonly in their publications
they only reported as least two levels. But did plenty of work
at the third and fourth and sometimes fifth levels.
Which again shows the limitations of our publication
possibilities. We include the pool cool diffusion trial which
was a skin cancer prevention program,
the choice cancer education program,
improving systems for colorectal cancer screening at
the Harvard Vanguard Medical Associates Integrated Medical
Group. The best practices comprehensive tobacco control
programs led by the CDC once the California and the
Massachusetts programs were adopted in terms of best
practices and then disseminated to other states. We also look
at the Tides depression collaborative care models.
And we include this as a non-cancer exemplar because
it's one of the only ones that actually spans all of
the levels of the onion. And then the VA colorectal
cancer collaborative or C4. This is just to give
you an idea of how these different exemplars span
some of the levels that we've been talking
about over the last day and a half. The other key thing here
is that the interventions that are within each one of these
levels are many of the usual suspects that you're familiar
with, reminders among patients, reminders among providers.
But as you get to the higher levels,
they tend to be evidence based quality improvement approaches
to actually change manager behavior,
to change leadership activities. So in our efforts
to basically do what you might call a cross case analysis
across these exemplars, we came together with a series of
lessons learned that I'll go through in the time that I
have. The first one, which I think is a theme you've heard
about again in the last day and a half,
is the importance of the combinations as well as the
phases of multi-level intervention implementation.
And that's attention to stakeholders at each level,
and to understand how those levels may interact. And where
possible, to create inner dependencies across those
levels. For example, in the tobacco control programs,
local funding was based on mapping what you were going to
do in a local community to what the state's priorities were.
And that allowed for a certain amount of homogeneity across
those levels. It's also important to determine the
quality of evidence that you actually have for interventions
at each level and where, in lieu of evidence,
blending experience and expert opinion to fill in the gaps.
The use of social marketing for intervention meshes was also a
key factor in the success of these implementation
activities. And something, unfortunately,
we're not as good at as researchers in terms of getting
our message across to diverse groups that don't necessarily
have the same values and plans that we do. The use of rapid
cycle improvement or plan to study act pilots to test within
and across levels were key. And consideration of staged
approaches, and giving each level adequate time to learn,
to consume and to consider the changes into their normal
processes of care as you're giving them new evidence to
consider and to enact, is absolutely key. It's also
important to consider the top down and bottom up
implementation aspects. Not all of these,
coming from the VA people assume for health care reform
and other kinds of quality transformation efforts,
that Dr. Kaiser in Washington said,
this will be. And everyone kind of rolled over. And it's
changed. In reality the VA's quality transformation was a
bottom up development where primary care was actually
enacted two years before Dr. Kaiser came into town.
Now there was an important substraight there that allowed
for the top down to work for that point. But each one of
these exemplars provides examples where there are people
in the front lines that you might not expect to be
champions, that you need to acknowledge and recognize their
value. As well as the importance of leadership buy in.
Again, that same issue about partnerships within and
across levels. Now, I don't know how many of you are used
to working within your own office. Many of you I'll see
here at a meeting like this and I won't even see back at my
home shop. I haven't seen Brian in months, right?
And it's one of those things where we end up not
coming out of our own silos if you will. And so the exemplars
really showed us the importance,
the essential aspects of research clinical partnerships.
Not just for implementation, but also sometimes co-funding
those implementation activities. Our lack of control
over implementation leaves us in a situation where we're
outside of our comfort zone very often. I don't know how
many of you got your Ph.D.'s or your M.D.'s,
assuming you're going to go have to figure out how to talk
to legislators. Or that you're going to have to figure out how
to talk to a CEO who's really not interested in the evidence
that you're providing. This requires a lot of shared
knowledge, trust, and specification of roles at each
level, team building before, during and after
implementation, the continual identification of stakeholders
in the network because these people change just like people
in your own teams change, and their interests change
over time. We found strong leadership support at each
level was absolutely essential. And again,
you can't come in and do your initial launch meeting and then
go away for three years, then come back and say this is what
happened. This is a continual set of relationships. You have
to use their leadership support to elucidate other key players.
They are the ones that I think someone spoke about yesterday
that are ensuring that people at these different levels are
accountable, and they have roles and coalition building.
And we were just talking about electronic health records.
Most of these practices did benefit from electronic medical
records, but not all. And those that did indicated that if you
don't have your favorite IT person in your back pocket,
you're in a bit of trouble. We also talked about
implementation facilitators and barriers. And I think that
Arnie talked about the other day the absolutely critical
importance of organizational supports. Whether it's direct
grants or special allocations for resources,
or simply the protected time to do quality improvement and
implementation. Some of those resources may be centralized,
they had in the tobacco control program. Local places needed a
media campaign, but the state was really able to do that with
their resources and spread it locally. Or shared resources,
for example with EMR support, where one organization may
create a template, and that's distributed throughout
organizations. The barriers though in this kind of work are
not insubstantial. They require interdisciplinary cooperation,
and that may be met with resistance. There are turf
issues, especially if there's competition for resources.
And I mentioned before this problem with silos where we're
used to speaking to ourselves and not to people
with different disciplines and backgrounds.
The perceived value of the interventions is
always balanced with competing demands
among busy members at each level. So we have to also
understand the importance of the policy context,
the fiscal climate, the performance incentives that are
critical to understanding what's surrounding players at
each implementation level. Now that gives me two examples in
the no time that I have left. Harvard Vanguard had a
situation where NCQA had just introduced a new (inaud.)
measure. Two of the four major health insurers in
Massachusetts were participating in that field test.
They had pay for performance integration into
the contracts in their state, and a state wide monitoring
program, all at the time that they were trying to do their
project. A beautiful perfect storm. And the master
settlement agreement with the tobacco industry ensured that
there were actually resources that could go to state and
local communities to enact that evidence. The determinates that
spread are very similar to those of implementation,
with the key issue here of the explication of handoffs.
You're not going to be in charge of your intervention
anymore. You have to find partners who are going to own
it, want it, and make it theirs. And when they do,
they're going to try and change it. And you're going to have to
learn how to work through that process. So in conclusion,
we found despite those challenges,
implementation and spread of multi-level interventions into
routine practice and policy is feasible and effective.
But attention is needed within and across those levels.
We talked about some, the current mismatch between the
reviews and the reality of doing this work. And I want to leave
you with the notion though that sustainability in this work is
indeed a myth. We hear this from our industrial engineering
colleagues. You always will have new evidence,
new stakeholders, and a new context that will be
continually changing. But we think the investment will pay
important dividends. Thank you.
>>>[APPLAUSE]
>>>DR. STEPHEN TAPLIN: Thank you, Becky.
I realize that the time constraints are difficult.
What we're trying to do is start a dialogue,
open up the papers, begin to have people understand some of
the issues. And I think the reason that we wanted Becky
involved here was her vast experience in implementation,
and to begin to see that some of the issues you deal
with in implementation are things that we can
think about in research. What levels, what are the
influences, how do we move it. How do we anticipate
some of the issues ahead of time.