The issue… There are many things we know...
lower left of Stacey... Examples: early ambulation after hip surgery,
early extubation after heart surgery, aspirin after MI.... In a large
organization or health system, these are done in some places well
and in other places not at all. Replication of improvement is an important
issue. In general, people are confused by the often failure to replicate
a logically sound improvement idea. Complexity theory could provide
important sense-making and strategy-aides. The question we should
be seeking to answer for the leaders out there struggling with replication
of improvement is: What should I do different from what I am currently
doing?
Caution… There is a need for thoughtfulness
in deciding what is needed in a situation. Yes, taking an area of
certainty and driving it home is the problem. But also yes, creating
conditions for adaptation and creativity is the problem.
Complexity Concepts: Fitness Landscapes, Coupling,
Context… Kauffman's "rugged fitness landscapes" may help us understand
what is happening here... In a smooth landscape, optimization of whole
system is easy... On rugged landscape, optimization of whole system
is hard... On smooth landscapes, all can see the optimum point and
all roads lead there... On rugged landscapes, one can be trapped on
a low "optimum" where any change leads to degradation and where it
is hard to see that there are even higher optimum points elsewhere.
Tight coupling between practices and the context of
those practices (organizational environment factors) leads to rugged
landscapes. So, we need to understand how strongly coupled the "practice"
is to the "context" before we simply try to replicate it. If the practice
is tightly coupled to a particular context, then it will be harder
to replicate into another context. Everett Rogers talks about this
in his book Diffusion of Innovation. There he observes that groups
often need to "rediscover" the innovation. This has the effect of
breaking the coupling with the distant context and creating useful
coupling to the local context. A social system that facilitates rediscovery
by valuing diversity, collaboration, risk taking, etc. might be more
able to accept good practices initially discovered elsewhere. (Another
reference is a paper by Mark Granovetter, "The Strength of Weak Ties."
Glenda Eoyang's book contains a chapter on coupling.)
For successful emergence, a complex system requires
loose coupling in order to consider new ideas and tight coupling to
engage them locally. Both the describer of the "best practice" and
the receiver of the practice need to be conscious of this need to
continuously play with coupling. (Jim Howard pointed out that Grey
Elrodt did a study of diffusion of clinical pathways within VHA and
found this loose-tight coupling phenomenon.)
Rogers also points out that innovation often begins
in the 2% "first adopters." These folks are often not listened to
in their home organizations (loose coupling). The next 10% of "adopters"
are key. They are respected in their organizations (tighter coupling)
but they will also listen to the odd-ball early adopters.
The idea of weak and tight coupling might also be
related to the concept of legitimate and shadow organizations. Tight
coupling in the legitimate and weaker coupling in the shadow. Diffusion
of innovation might be easier in organizations where there is some
degree of harmony among these aspects of the organization.
Plsek post … "My question is: How would you
know if a practice was tightly coupled to its context? What would
you be able to observe that would indicate to you that simply replicating
the idea back at your place was not going to be a simple matter?
(The answers to these questions could be a great contribution to
the "arts" of benchmarking and rapid cycle improvement based on
change concepts.)"
Levine post … I think the answer depends
on a better understanding of the human relationships within the
organization under review. It seems to me that process as a whole
as well as the component parts of the process have differing levels
of dependence on individuals. If we had a way to "measure" these
individual relationships, we would have some sense of how well prepared
another institution was to incorporate the benchmark institution's
process.
Dooley post… Remember Shortell's important
recent research re hospitals and TQM: midstream and late adopters
(70% of the total) of TQM do so because they think they must in
order to remain "legitimate"; the adoption process has little to
do with perceived or expected benefit. Therefore, organizational
readiness or context--while I admit that it *should* matter--is
largely ignored by organizational decision-makers.
Complexity Concepts: Tune the CAS… Recall one
of the "emerging principles of complexity:" *Tune* your place to the
edge by *tuning* info flow, diversity, and so on. Tuning implies that
sometimes you need more info, diversity, etc and sometimes you really
do need less. Great opportunity to overlap with traditional QI approaches
around the PDSA cycle of learning. Tune the info, diversity, etc *up*
and reflect on what happens, tune it *down* and reflect. Continuously
learning about the CAS.
Tom Petzinger's Forklift Company example illustrates
the use of tuned up info flow and tuned down power differential. They
made information available as opposed to saying thou shalt do it this
way. Netscape is getting improvement in its browser by sharing information
about its source code with a diverse and increasingly connected collection
of programmers on the internet.
Complexity Concepts:
Stacey Matrix… The Stacey matrix is a wonderful map for plotting issues.
The questions are: Where are we on this diagram? What direction do
we wish the system to go now (towards more adaptation or towards more
certainty)? And what actions on my part, as a member of the CAS, are
appropriate?
Don't settle down forever in the lower left of Stacey.
Yes, we might be certain and agree *now* about something, but we should
also from time to time purposefully pull up into the mid zone and
see if some adaptation is appropriate. A useful bit of language to
remind us about this is the phrase: "current best practice for the
time being."
CAS theory gives us guidance on how to "play" the
system on the Stacey diagram. To move toward the mid-zone: provide
more information, stress divergence and diversity of thought, loosen
the couplings, and so on. Vice versa to move the other direction.
A key piece of information we need for adaptation
is our true position on the Stacey diagram.
Plsek post... I wish that we had a "stacey-o-meter"
that we could dip into a CAS and read where it truly is. One practical
way to do this would be to simply have some dialogue where everyone
gets a chance to put a dot on the diagram where they think the system
is on the particular issue under discussion. This would provide
feedback to the system and provide a grounding for further discussion
about options (maybe multiple options). Such simple reflection might
have helped the Xerox situation described by Steve. Such simple
reflection might keep many organizational "leaders" from wasting
lots of time forcing clockware OR swarmware onto a situation when
one or the other really isn't appropriate. What would our organizations
be like if *everyone* in the organization understood and routinely
used this one simple aide from complexity thinking?
A potentially nice demonstration project… Teach
the Stacey diagram to everyone in the organization... use it for extensively
for reflection... see if the organizations capacity for successful
action and adaptation is improved.
Complexity Concepts: Holland's Framework… Another
potentially rich set of ideas for understanding diffusion of good
practice comes from John Holland's book Hidden Order. See Steve Larned
3/9/98 post for more information about how tagging, agents, internal
models, aggregation, catalyzation, non-linearity, flows, and diversity
might help us think through the issues around diffusion.
Thoughts About Demonstration Projects… A natural
place to demonstrate new thinking based on the complexity sciences
would be in the midst of current efforts around EBM, clinical paths,
guidelines, etc. that people are naturally finding frustrating. This
current frustration would make for a natural resonance with new thinking
from complexity.
Does the current trend toward consolidation into
larger organizations in healthcare make diffusion of innovation easier
or harder?…
(Post from Mark Levine) Consolidation in
and of itself does add to the coupling challenge. I think it is
imperative to remember that consolidation causes disruption to former
relationships while at the same time causing opportunity to develop
new relationships. Focusing on currently existing generative relationships
and actively working to maintain those relationships, it would seem
to me, is the key. Encouraging a new identity and structure in the
consolidated organization that fosters development of new generative
relationships is the other key. (An interesting article about medical
systems can be found in the Feb 15, 1998 Annals of Internal Medicine
Nolan, T Understanding Medical Systems p. 293.)
Post from Al Herzog... Paul, in theory, these
consolidations should make the diffusion of best practice into other
organizations-from those with the best practice for `X`, much easier.
In practice, this has not happened; certainly not in my system,
and, as I listen to and read people like Brent James, neither in
his. All of that provides some comfort, but ultimately much frustration.
I think much of this has to do with trust--the level or degree of
it between/among different people but there may be other, perhaps
even stronger reasons. I struggle with this one but am so far without
a good explanation.
Open questions...
- Can we use complexity principles and Aides (eg.
tune to the edge, coupling, fitness landscapes, shadow organizations,
and so on) to build on and supplement existing thinking about diffusion
of improvement?
- Can these additional complexity-inspired approaches
help us even more as we move toward consolidation and ever-larger
organizations in healthcare?
Thoughts About Demonstration Projects… Need
to do more thinking about this.
Ultimate Goal of This Line of Thinking…
Better diffusion of improvement knowledge, evidence, best practice,
and results of benchmarking studies within today's healthcare organizations.
What can leaders do that is different from what they are currently
doing (which is leading today to lots of frustrations)?
