Understanding Shared Expertise in Communities of Practice
N. Bryan-Kinns and R. Makwana
Department of Computer Science
Queen Mary and Westfield College
University of London
Mile End, London. E1 4NS
England
{nickbk, ranj}@dcs.qmw.ac.uk
Introduction
This chapter discusses a conceptual framework for understanding how shared
expertise is utilised in communities of practice. The perspective taken
here is on how the effort involved in acquiring, transferring, and transforming
knowledge affects the development of community knowledge and in turn the
sharing of expertise. The framework is exemplified through studies of clinicians
undertaking shared care of diabetic patients. Results of these studies
are used to inform the design of artefacts that act as foci around which
expertise can be shared through the growth and development of community
knowledge.
Conceptual Framework
A collaborating group's shared expertise can be viewed as an emergent property,
akin to the gestalt principle of the whole being greater than the sum
of the parts. Individual members of such a group bring with them their
experience, expertise, and knowledge, but the shared expertise they develop
may be more than the aggregation of individual member's expertise.
A community of practice is a collaborating group whose common purpose
and its need to share and develop community knowledge binds it (see [Carotenuto99]
for details of various knowledge communities). Clinicians for example,
share the treatment of patients and supply expertise as well as results
of tests they may carry out on the patients and other relevant information
to one another. A group's work involves the acquisition, translation, and
transfer of knowledge which involves effort. If we are to help support
the group's development of shared experience, then we need to understand
the value of different types of knowledge, and to try to reduce and/or
redistribute the costs in accordance with the group goals and individual
goals; sharing expertise is facilitated by group members' perceptions of
community knowledge.
Our conceptual framework for assessing the value and effort involved
in collaboration has the following four components:
-
Knowledge taxonomy. We distinguish three sorts of knowledge (loosely
based on Johnson-Lenz's knowledge taxonomy [Johnson-Lenz96]). 1) Explicit
knowledge - this is generally knowledge of data and has no ambiguity
as to its meaning or intent. 2) Embodied knowledge - this is the
domain specific knowledge and can be both tacit and explicit. 3) Community
knowledge, which subsumes awareness and knowledge of others and their
actions.
In using the artefacts of a given domain, users acquire domain specific
knowledge from data by applying their understanding - often using data
in the context of other data. This can be tacit, for example, "the high
blood glucose level may lead to complications" or explicit - "the blood
glucose level of over 7 mmol/l is fatal to the patient".
Community knowledge is the common belief system that develops whilst
a group carries out tasks and is generally tacit. It concerns our knowledge
of others, the particular facts we believe they know, their role and expertise
(know-how) and what it is they need to know to complete the task in hand.
-
The value of knowledge. Knowledge becomes value creating when it
is acted on and used, i.e. knowledge has value relative to the context
of the task at hand and the individual and group goals - value with respect
to the group task, and value with respect to the task of the individual
who processes it. For example, a blood test result may have more or less
significance when the overall clinical state of the patient is taken into
account. Or it may have less value by itself than when taken in the context
of a set of such tests showing a trend. Moreover, value is affected by
individual's perception of the knowledge creator's experience.
-
The effort of acquiring and transferring knowledge. Many different
sorts of effort are required for individuals to acquire knowledge locally
and to transfer it to others - via communication channels and shared artefacts
such as documents and databases. We concentrate on interpreting, translating
and educating. In many cases explicit knowledge is embodied in the artefacts
group members use in their normal work. Such knowledge can be used by members
once interpreted (transformed into explicit knowledge) in the context of
the community knowledge. Such transformations require effort which is increased
if the community knowledge is inadequately developed to help with interpretation.
Acquiring knowledge involves such effort as questioning another participant
or interrogating a database or thumbing through a folder of notes. There
is in general a combination of physical effort and the cognitive effort
of searching for salient facts. Similarly, making knowledge available involves
targeting it at another user or group of users - for example, putting it
in a database or writing a letter. Effort may be required to get the recipient's
attention and in transforming the knowledge appropriately. Further effort
may be involved in interpreting knowledge made available by another, or
in educating another in the types of knowledge the recipient needs. Interpreting
knowledge and knowing what types of knowledge a recipient needs relies
on an understanding of the community's knowledge which is not explicitly
represented in the artefacts, but is developed over time by the participants.
-
Trade-offs. After ascertaining value and effort according to the
foregoing analysis in a particular case, we need to go on to understand
the trade-offs between value and effort with respect to both group and
individual goals. These trade-offs help us to understand why users behave
the way they do currently, and will help us to predict the users' response
to changes. Understanding these trade-offs will also help us comprehend
why community knowledge does not develop and/ or does not enhance collaborative
tasks.
Our conceptual framework has parallels in current work. For example, Bannon
and Bødker [Bannon96] discuss the interpretation and packaging of
data by members of one community for other communities which relate to
our notions of interpreting and transformation. Similarly, Star discusses
boundary objects [Star89] which may be needed at the interface between
communities to govern the flow of knowledge between users which relates
to our notion of knowledge of others' needs. The key distinction between
our work and that of others is that we are concerned with the effort involved
in collaboration, and its effect on the development and use of shared expertise
based on community knowledge.
Studies of Clinicians
An interesting aspect of the clinical situation studied is that the set
of members working to care for a single patient have diverse sets of individual
expertise. Typically each member directly collaborates with one other member
at a time. This means that each pair of members develops its own set of
knowledge, which reflects the particular aspect of the patient's case they
are working on. Difficulties arise when members need to transfer knowledge
to and from different contexts i.e. for different collaborations. Moreover,
there are difficulties in the development of shared expertise for the group
as a whole.
Problems encountered are briefly recounted in the following points in
terms of the conceptual framework. The problems typically affect the efficiency
and quality of the group work. Often the problems result in the duplication
of effort by clinicians; tests are repeated unnecessarily, and patients
are asked similar questions by different clinicians. Also, procedures are
sometimes omitted altogether, and development of shared expertise is frustrated
by poor knowledge of the current state of the collaboration and members'
roles.
-
Transfer of domain specific knowledge - either embodied in artefacts
such as the medical record or via some communications channel: The timely
transfer of knowledge between clinicians in the shared care of diabetic
patients is a hit-and-miss affair; some knowledge may be transferred, others
may not. Furthermore the clinicians have no way of knowing whether they
have missed knowledge that has been pushed for them, or whether the knowledge
was not pushed in the first place. In our studies we have found that clinicians
are well aware of this problem, indeed, they do not even expect acknowledgements
of attempts to push knowledge, and often devise ad-hoc methods of tackling
the problem.
-
Community knowledge: Often clinicians construct expectations that
their colleagues will perform certain tasks or impart certain information
to the patient. These expectations are formed from beliefs about the other's
roles, expertise, and knowledge, but are often not successfully conveyed
to others. This indicates that community knowledge has not been successfully
developed and utilised.
-
The value of knowledge: Our studies have included preliminary investigations
into the clinicians' perceived value of knowledge in terms of individual
value and the value based on perceived expertise. One such value judgement
arises in the use of blood test values. Two kinds of tests of blood sugar
levels are used: 'finger-prick' tests and 'lab' tests. Although finger-prick
tests are easier to obtain, the lab tests give more reliable results and
so are more highly valued.
-
Trade-offs: Our initial investigations of clinicians' perceptions of trade-offs
between value and effort in group work have highlighted one of the reasons
for the prevalent duplication of effort. Often tests (e.g. blood tests)
and enquiries (e.g. details of a patient's diabetic history) are duplicated
by clinicians because the information is valuable to the task at hand,
but the effort required to obtain the knowledge from other clinicians is
higher than finding it out by themselves. This completes an individual's
task, but the duplication reduces the collaboration's efficiency.
Design Implications
We have designed a shared workspace (see [Kindberg96] for a discussion
of workspaces and their implementation) in order to address the issues
of efficiency and quality of group work in terms of duplication and omission
of effort, delays in communication, inconsistencies, development of community
knowledge, and effectiveness of expertise-sharing highlighted in the previous
section. From the holistic view of shared diabetic patient care our design
addresses the problems of collaboration and sharing of data and knowledge
in unusual cases. These are cases where clinicians need to seek advice,
or collaborate over and above the usual sharing of data - which may be
automated for a core set of data. The clinicians therefore need to understand
and determine which information is needed, or should be made available.
The design provides two main artefacts around and through which collaboration
can take place, and community knowledge can develop.
The first is the 'timeline' which provides a graphically laid out view
of all events and communication acts pertaining to a particular patient
by the clinicians involved. This aims to provide awareness of others' actions
as well as some notion of the context in which actions have taken place.
The second artefact provides a shared medium through which ongoing discussions
can be developed. Moreover it supports dynamic referencing of clinical
data which allows clinicians to support their perspective or interpretation
with concrete information.
Summary
This paper discussed a framework for analysing group work in terms of types
of knowledge, its value, and the effort involved in acquiring, transferring,
and transforming it. The framework was used in this paper to highlight
the problems in the development and maintenance of community knowledge,
in particular the community knowledge developed in shared care of diabetic
patients. Furthermore, the framework was used to inform the design of artefacts
around which community knowledge could develop. We are currently developing
the framework in terms of its evaluative power, and will be using it to
evaluate our designs in field trials with clinicians as part of our ongoing
iterative development process.
References
Bannon, L., and Bødker, S. (1997), Constructing Common Information
Spaces. Proceedings of ECSCW '97, pp.81-96.
Carotenuto, L., Etienne, W., et.al. (1999), Towards flexible
support for voluntary knowledge communities. Proceedings of Changing Places
workshop, April 1999, Queen Mary and Westfield college, University of London.
Also available at: [http://www.dcs.qmw.ac.uk/research/distrib/Mushroom/workshop/prelimProgramme3.html]
Johnson-Lenz, P. & T., (1998), Community of Inquiry and Practice
Report. Available at: [http://www.awaken.com/AWAKETECH/Wisdom.nsf/By+Title/FrameSet?OpenDocument]
Kindberg, T. Mushroom (1996), A framework for collaboration and interaction
across the Internet. Proc. CSCW & the Web, 5th ERCIM workshop, Busbach,
U., Kerr, D., and Sikkel, K. (eds), (Bonn, 1996), GMD, pp.43-53.
Star, S.L., (1989), The structure of ill-structured solutions: boundary
objects and heterogeneous distributed problem solving, in Distributed Artificial
Intelligence, vol. 2, Gasser, L., and Huhns, M. (eds), Pitman, London,
1989, pp.37-54.