Towards Achieving Semantic Interoperability in eHealth Services - E-Healthcare Systems and Wireless Communications - page 398

allow specifying the type of the data, or specific
data elements and associate them with the selected
privileges.
The Meta Registry combines the knowledge
about services (i.e. physical resources or health
systems and their descriptions), the domain ontol-
ogy capturing the relevant domain concepts (focus-
ing on primary care) and the mappings between
the two. It also holds the services' corresponding
data governance policies. It acts as a central (do-
main) description service that holds functional
descriptions of the data services themselves, and
descriptions of the underlying health data that
they provide access to, alongside a description
of the access mechanism or rules that governs
access to the data. The Meta Registry is updated
by two mechanisms: self-registration of services
and on-request update of service descriptions.
The self-registration occurs when a new service
is deployed at an organization and the service
registers itself with the meta-registry, declaring
its intention to participate in the ePCRN network,
describing its physical resources, data and data
access rules. For services already in the network,
the registry can request updates of the service
descriptions. Each data service also holds a copy
of its data description and its access rules to en-
able local configuration of changes in the system.
Knowledge Services provide access to custom-
built or publicly available domain support ontolo-
gies or terminologies. In ePCRN, these represent
the available clinical terminologies and vocabu-
laries, some of which are tailored to the primary
care domain. These resources are facilitated by
intuitive interfaces that enable users to search and
find relevant clinical concepts. These interfaces
are linked to the EQAM to enable vocabulary
controlled capture of a semantically annotated
'problem model' or ' user data needs'. Figure 3
illustrates an interface to capture the eligibility
criteria EQAM, while Figure 4 illustrates an inter-
face to find and retrieve clinical concepts from a
knowledge resource. Terminology and knowledge
resources are key to the capture of individual user

needs in a computer-interpretable format and
their translation to individual health systems. In
ePCRN, several knowledge services are used,
including the Enterprise Vocabulary Service [NCI
(2011)] inclusive of the UMLS MetaThesaurus,
and the ePCRN Vocabulary Service inclusive of
local national UK Read Codes and some specific
EHRs' codes. These services have been brought
together by a common service and interface to
enable as rich as possible definitions of clinical
concepts. The definition of concepts includes
annotating each requested concept in as many as
possible available terminologies, creating rich
multi-mappings. These mappings are then stored
in the EQAM at the point of definition for each
instance. The QGS service uses these mappings
and domain data ontology mappings to create a
tailored data request or query format specific to the
designed health system. The richer the knowledge
services, the richer and often more accurate is the
outcome, and the greater is the number of health
systems that can be supported in the network.
In the few cases that we came across where the
health systems uses own-custom terminologies, for
specific part of its data, that are not supported or
provided by the knowledge services, the outcome
of the queried or retrieved data or its quality end
up being limited to the supported data.

DYNAMIC GENERATION OF
SEMANTIC INTEROPERABLE
DOMAIN CONCEPTS AND
DEFINITIONS

To enable semantic translations between services
(or data sources) and the defined problem models, a
mechanism for bridging between the two is needed.
This bridging mechanism includes translations at
two levels: at the problem models level, definition
annotated by the knowledge services' terminolo-
gies, and at the services' levels translated through
the Query Generation Service. Figure 5 illustrates