Information Technology Reference
In-Depth Information
represent an ideal basis for an intelligent healthcare information system. In the
following, we present some approaches to create a unifiedmodel in the health domain.
One of the biggest computer-supported biomedical knowledge bases is the Unified
Medical Language System (UMLS),
4
a platform that provides a unified vocabulary
for biomedical and health terms. UMLS is maintained by the US National Library of
Medicine and updated on a frequent basis. The knowledge base of UMLS consists of
three components: (a) Biomedical terms from various controlled vocabularies (such
as SNOMED-CT
5
) are defined in a Metathesaurus as semantic concepts. (b) The
semantic relationship between these concepts are defined in a Semantic Network and
the concepts are categorized into broad categories (semantic types). (c) Syntactical,
orthographic, andmorphological information about the biometric terms are defined in
a lexicon. UMLS has a strong focus on the US health market, and hence, multilingual
aspects are not fully supported, making it difficult to apply it to the scenario that is
outlined in this chapter.
Other health-related ontologies include MEDCIN and SNOMED-CT, two classi-
fication systems used to store patient health records. The International Health Termi-
nology Standards Development Organisation (IHTSDO) promotes SMOMED-CT as
a standard for health records. The aforementioned approaches define common vocab-
ularies and their relation to each other for the medical domain, but they define the
terms in a proprietary format that is not machine-readable and shareable. These con-
strains make reusing and sharing of data complicated though. OpenGALEN [
33
,
37
]
is a medical ontology developed in the European GALEN project. OpenGalen offers
a comprehensive knowledge base of medical terms and relations. It provides three
types of ontologies: A high-level ontology, defining general structures, a common
reference model defining the reusable parts intended to be shared between ontolo-
gies, and extensions for subdomains and specific use cases. The GALEN ontology
is available under an open source license.
The presented approaches all contribute to the goal of a common, shareable, and
reusable notion of the medical domain for building health-related applications. Nev-
ertheless, none of these approaches fulfill the requirements to serve as the basis for
a multilingual health information system. In particular, they lack modeling of multi-
lingual descriptions and a connection to related health information is not available.
3.2.3 Identification of User Context
An important feature of online information systems is to adapt information based on
users' individual requirements [
17
]. Requirements can depend on different factors
such as the users' demands and context. In the health domain, contexts such as
health condition of the user and their location plays an important role. For instance,
recommendations for a healthy diet for a person with Type 2 diabetes and a pregnant
Search WWH ::
Custom Search