Clinical Semantics - Computational Intelligence in Healthcare: Advanced Methodologies - page 7

Fig. 1. A framework for health care and life sciences (adapted from [1])
information, to migrate existing domain ontologies and coding systems (e.g., from
HL7, UMLS) and diverse, vendor-specific electronic health record systems to support
the fully structured approach. Furthermore, apart from the initial construction effort,
the approach would necessitate the establishment of a centrally-planned co-ordination
effort on a massive scale to account for the inevitable (and hopefully welcomed!)
evolution of both information and semantics in the domain. This issue has been prac-
tically highlighted by the authors in a system migration evaluation study carried out in
co-operation with an industrial partner [2] as discussed later.
The question arising from this discussion is that might there be a way to avoid the
complexity of construction and maintenance inherent to the Semantic Web and related
approaches? A discussion on some possible answers to this question is the topic of
this chapter.
1.2 Open Information Management
Earlier work by the authors has probed into the question of human abstraction level
information management and has developed an approach coined as 'open information
management'. For a general review of the approach see [3] and for discussion in the
scope of health care see [4], [5] and [6]. A textbook edited by the authors on the topic
is also available [7].
The fundamentals of this approach can be summarized under three observations:
- The expressive power of natural language
- The associative power of unrestricted linkage
- The descriptive power of direct observation
The basic claim of the approach is that these observations lay the foundation for the
development of information management tools capable of learning on the human
abstraction level with a principally unbounded case space. In practice, the learning
process results from user interactions with expressivity similar to natural language
supported by the associative and descriptive powers of the latter two constructs. The
implication is that semantics is automatically and natively constructed through a