Multilevel Interventions in Health Care Conference: Discussant comments by Martin Charns, DBA

Uploaded by NIHOD on 05.05.2011

>>>DR. STEPHEN TAPLIN: I wanted to now switch to Marty Charns.
Marty spoke yesterday, he's the Director of the Center for
Organization, Leadership and Management Research, a VA
health services research and development services center of
excellence. He's a professor of health policy and management,
and Director of the program on health care organization
studies at Boston University School of Public Health. He has
over 35 years experience in the field of organization, and he
claims he's the only DBA in the room.
Thank you very much for participating.
>>>MARTY CHARNS: For better or worse he got me twice.
So I'm going to talk about the Yano paper first and then the
Devers paper. So Becky and I, as you probably could tell
hearing us talk, work kind of the same sandbox. I very much
like the work she does, and think the paper makes a number
of very important contributions. I just wanted to
highlight four things that I thought were real strengths.
And those of you that, I don't think people got the whole
papers, but I think it's important to pay attention to
several things in this paper. First off, there is support
through the various presentation and discussion
around the importance of multi-level studies.
Another important point is that the paper distinguishes between
efficacy and effectiveness. And it's not the only paper
obviously that does that, but it so much is an important
point and puts so much of the emphasis on how critical the
implementation process is, and the importance of context.
The examples in the paper which Becky cited, were a central
part of what they saw as their objectives in putting the paper
together. I think really are a major contribution. I'll talk
in a minute about some things I'd like to see them do in
their final version that would extend that, but they're very
good examples. They're very informative. And in the paper,
they also talk about the theories that were used in
those different examples. And that's something we mentioned
yesterday as being important in some of our discussion, and as
something that often is lacking, being the explicit
identification of the theories that are being used.
Some things that I'd like to see greater emphasis on. I didn't
see a lot of discussion, and I actually saw a little bit of
interchange of words, around adoption and implementation.
So an organization can decide to adopt a new practice, but
that doesn't mean they actually get it implemented so it works.
And that whole process of implementation is largely what
was talked about in Becky's presentation. But just to make
a little bit more distinction between those two different
things. Sustainability is probably the biggest bone I
have to pick with her paper, and perhaps we just look at it
as an honest disagreement. So I agree that you don't want to
sustain something forever because then you never move
onto better stuff. On the other hand, if we don't have
some degree of sustainability, if not the exact practice,
at least the gains we have achieved through the
implementation process. If we don't have that, then at best
we've wasted energy and we've slipped backwards. And at worst
what we've done is raised people's expectations that
things were better but now we've lost it. So the next time
we try to implement something they're further disillusions
about the change process. So I'd like to see a little bit
more balance to my colleague around the issue of
sustainability. Also discussion of spread. I think it's
important to differentiate between spread within
organizations and spread among organizations. And it's the
same concept, but I think some of the issues that come into
play, because they're at different levels may be
different. And we might bring in the idea of organizational
feel that Dick Scott raised yesterday when we're talking
about the spread across organizations. And then a
little bit more, of course you didn't all read the paper,
but a little bit more of the descriptions of how the
mechanisms work. You've got such a rich set of examples
here, and you'll hit against the page limit for the paper.
But that's something that I think we could benefit from.
Raised a couple of questions for me. The first one is the
levels in the studies. Are they really all multi-level?
Certainly they're two level, but are they all three level?
I maybe missed it and saw it, missed it in the paper but
Becky cited it in the presentation, that there was a
practice level intervention in the Harvard Vanguard example.
And this again raises the question that I raised
yesterday in our presentation as to whether patient and
provider represent two levels or not. And I know from an
analytical perspective when we do hierarchical linear
modeling, we often cluster patients by provider and then
providers within practice, if we can cluster as many levels
as we wish we could. So that's one argument to say they should
be separate levels. But there's another argument which really
said providers and patients really are both individuals at
the same level. So I think that's a debate for us. How do
we identify levels and might we describe more about how the
interventions interact, so this particular paper within
organization. So if we look at the VA, you know, the VA is an
organization. So is that it? Is that a level? Well, this
paper went beyond it to say medical centers in the VA are
nested hierarchally within VISN, Veterans Integrated
Service Networks, or regional networks. And what was
happening at the network level, the incentives, the measures,
the leadership, are important for what happens at the next
lower level of the medical centers. So this was a good
example of where within what we've called kind of loosely
one level in our multi-level diagram, where it's important
here to differentiate among those levels. But I think the
important question for us as we move forward in this area is,
how do we distinguish among levels. The question is, what
makes a level. I don't have the answer to that and it may be
totally dependant upon the specific situation being
studied. It's triggered another idea. Becky talked a lot about
implementation, about PSA cycles. And so I thought one of
the things that might be useful for us to think about is that
sometimes what we're trying to do, if you look at the left
side of my continuum here, is implement an evidence based
practice. And in some ways it's a push strategy. We have
evidence, we have a practice, we want to push it out into the
field. Now, of course, we know that you have to adapt the
practice to the particular setting. And one of the
concerns is as you move further and further towards the right
of the continuum as you're adapting it, at what point have
you lost what the original practice was that you were
trying to implement. So there's this question about maintaining
fidelity that you have to look at. Coming from the other
side, often organizations start with quality improvement.
So the issue is, they begin with a problem or a process that
they want to improve. It's a pull strategy in terms of
trying to bring in evidence to help do the improvement.
And there's much more flexibility in terms of what's
going to happen. Well, I think the examples provide some
mix of that. And I think as we think about implementation,
because I know organizations use both of these ends of the
continuum and blends in the middle. That we think about
what is being used and what are its characteristics. Slight
plug. Actually, I'll build off of this in a minute. This the
organizational transformation model that my colleagues and I
developed to look at organizational improvement.
I just want to highlight a couple of things in it.
If you're really interested in it, the URL is OTM for
organizational transformation model at BU,
for Boston University, .edu.
You can find out everything you want about the model.
There's a journal article also. So if we look at
this you'll see on this diagram two things I want to point out.
One is leadership. And what we found in our work is that
leadership is important not only at the top of an
organization, but right down through frontline supervisors
to the front line. A multi-level concept if you will.
And also you see a vertical arrow on the left side
of the diagram called alignment. And that also is a
multi-level concept and it's something what, while Becky
didn't refer to it explicitly, she did refer to it implicitly
in the paper. How at different levels of the organization are
things supporting rather than providing barriers for the next
lower level. I'm going to move on. Importance of context for
implementation. You know, I think that as we look at the
state of our science and we're learning a lot about
implementation. I think one of the things that's still,
I'm pushing the same concept here. That still we're lacking
is a lot of knowledge around sustainability. And we know the
importance of things at the organizational level of culture
and leadership support both for implementation and
sustainability. So to pay attention to those things as
we're looking at multi-levels. Moving onto the health care
reform. Obviously this is probably the biggest change
I've seen in my career in this field. One of the things I
think this does is present a wonderful opportunity for
research. And the reason is there is going to be not only
change to look at over time, which is one aspect, but also
as Kelly described it, there is going to be variation in the
way organizations implement their specific changes. Now, a
suggestion for this particular paper. Or if it's not going to
go in the paper then I think it's something we need to do.
Is look specifically, do a very detailed analysis of how health
care reform and the different parts of it are going to
directly effect the different levels in the multi-level
model. Kelly referred to effects at different levels.
I don't know if I'm smart enough to know what those effects
might be. But I think it would be useful to do that, to think
that through in a detailed level in advance. Similarly,
look at what happens at the interactions or analyze what we
think would happen. And what will happen, so look at the
multi levels and look at the different types of care along
the continuum, so that we have a more detailed analysis of
what we anticipate the impacts would be. Those could give us
some propositions to test. I agree for the need for big
science because I did go to business school I guess.
I think it's important that we have a balanced portfolio.
And right now the part of that portfolio that's unbalanced is
we don't have enough big science for addressing this.
So I think it is important to do. Not that we would do no
individual investigator initiated work. I think it's
necessary because it's an opportunity that is going to
pass. And if we can't move quickly to do research on it we
will miss this opportunity to learn. We need some common
measures. It's been talked about before. The electronic
health record may help. Organizations that have those
already in place have a wonderful tool to use for
research. But organizations that are just starting to
implement, we know the implementation is going to take
too long to be able to have the data available for us the way
those that already have it in place would be able to do.
The importance of multi-method work, testing propositions,
identifying contextual factors, which largely is probably going
to be qualitative work, and identifying factors for
additional study. I want to urge us to look at this over
time. It's going to unfold. It's a wonderful opportunity to
study the processes of implementation. It's not a
steady state where we're taking a snapshot, but can we look at
the dynamics. And finally, my final plug, don't forget the
history. You know, that all organizations are not starting
at the same place as they implement their ACO or their
patient centered medical home. Thank you.
>>>DR. STEPHEN TAPLIN: Thank you Marty.