Information Technology Reference
In-Depth Information
transfer” (eliciting knowledge from human experts and transferring elicited knowl-
edge into knowledge-based systems) in the early eighties to “knowledge modeling”
(creating computer models to solve particular problems in the area of interest) more
recently. In the model-oriented KE methodology, the knowledge and problem solv-
ing techniques of the domain experts are explicitly modeled. This modeling process
requires a specification of
what
the system should do, a plan describing
how
the
system should solve specific problems, and an explanation
why
the system solves
the problem in a particular way.
Knowledge Management (KM), on the other hand, is often defined as an area of
business administration. Thus, KM views knowledge as a key asset and resource in
an organization. KM concentrates on three aspects: people who create and commu-
nicate knowledge (
who
), knowledge communication process (
how
,
where
,
when
),
and knowledge itself (
what
). KM is concerned with the entire KM process in the
context of business enterprise. Thus, KM encompasses knowledge creation, acqui-
sition, validation, presentation, distribution, and application. KM in medicine and
healthcare faces many additional challenges such as extreme complexity of biomed-
ical systems, high costs of errors, and fast-growing body of knowledge [9].
In broader context, KE and KM address all fundamental questions regarding
knowledge: who, what, how, when, where and why. These questions are answered
by various types of knowledge: declarative (what), procedural (how), causal (why),
and organizational (when, where, who). One of the specific areas of interest for both
disciplines is decision making. However, KE and KM have, to some extent, dissim-
ilar approaches to decision making. Whereas KE has a long history of a practical
approach to constructing computerized decision support systems, KM has been con-
cerned more with organizational aspects of decision making. In spite of historical
differences, both disciplines are closely interrelated and could benefit from collabo-
ration. This need for interdisciplinary approach is especially visible in knowledge-
rich applications, such as medicine, in which KE is strongly intertwined with each
step of the KM process: knowledge acquisition, creation, validation, presentation,
distribution, and application. Specifically, computer-based systems, which support
clinical decisions, such as clinical decision support systems (CDSS), must explicitly
represent the knowledge and its source, maintain the currency of the knowledge, and
create guidelines for the application of knowledge in specific clinical settings [2].
The rest of the paper is structured as follows. Section 14.2 defines the key terms:
expert system, decision support system, and clinical decision support system. Sec-
tion 14.3 describes the Knowledge Base (KB), the most important component of
CDSS. Furthermore, it lists user requirements regarding the medical KB. Section
14.4 discusses the verification and validation of KB using the data warehousing and
data mining approach. Section 14.5 provides the conclusions and future work.
14.2
Decision Support Systems in Medicine
This section describes the Clinical Decision Support Systems (CDDS) in the context
of Expert Systems (ES) and decision support systems (DSS). We define ES, DSS,
