Towards Achieving Semantic Interoperability in eHealth Services - E-Healthcare Systems and Wireless Communications - page 392

based on the use of domain ontologies [Musen,
M.A (1998)] combined with rich terminology and
knowledge driven problem models used at the
heart of a service-based framework to enhance
semantic interoperability in eHealth services.
An example of an eHealth service is EHRs or
data sources that in this service-based framework
become (for example web or grid) services (thus
referred to below as eHealth services or services).
The chapter proposes the use of extensible prob-
lem models enriched with well-defined domain
concepts that can be dynamically translated into
individual data sources (or services). Data sources
or services themselves are semantically described
in terms of the domain ontology, which is dynami-
cally discovered and bound to at run-time. A data
governance ontology is also used to define data
access constraints, that enable enacting the defined
ethical rules and constraints on accessing the clini-
cal data and maintaining patient confidentiality,
thus enabling what is usually referred to as 'local
ownership, local control'.
The following sections first outline service
interoperability and heterogeneity challenges in
eHealth systems, describing related semantic web
and ontology-based approaches to addressing se-
mantic interoperability. Then the chapter presents
our approach to the auto-generation of rich domain
concepts and problem models and their transla-
tion across the architecture to individual data
sources. An outline of a prototype (ePCRN-IIA)
is then presented as an initial implementation of
the approach reflecting on our experiences using
the prototype and identifying requirements and
limitations.

in the arena of service-oriented architectures.
However, the focus in eHealth is on the data itself
and less on the functionality, but in principle these
challenges are similar and include:

• Using common interoperability mecha-

nisms to dynamically discover and in-
tegrate services from a potentially large
number of disparate health data sources, in
this case electronic health record systems
located in health institutions.

• Separating the interoperability of service

functionality from that of communication
mechanisms, in this case the main element
is the heterogeneity of the data that is being
served through the services.

• Enabling dynamic run-time binding to ser-

vices without prior detailed knowledge of
how it will be used that can be computa-
tionally interpreted at run time.

We therefore propose a domain focussed
interoperability representation that utilises ser-
vice-oriented architecture technologies. In other
words, to define the service-oriented interoper-
ability mechanism based on the clinical data that
constitutes the heart of the services. In a complex
and closed domain, such as eHealth, where the
purpose of function of services is defined, the
interoperability mechanism is enabled through
a domain-specific semantic representation or
ontology that provides the required richness and
accuracy. While services provide the uniform
abstract interfaces for communication, their
functional focus include specific definition of a
semantic representation of the data and its access
constraints that can be controlled and supported by
the clinical data providers themselves. This does
not just enable semantic-awareness in services
but also propagates it through the architecture.
Basing the interoperability mechanism within
services, using a domain-specific representation
provides the required extensible richness and com-
plexity needed in the domain, opposed to having

CHALLENGES OF SERVICE
INTEROPERABILITY AND
COMPOSITION IN EHEALTH

The vision of eHealth as enabling flexible and
dynamic connectivity and data integration shares
many of the challenges that are being addressed