Towards Achieving Semantic Interoperability in eHealth Services - E-Healthcare Systems and Wireless Communications - page 394

SEMANTIC INTEROPERABILITY
AND ONTOLOGIES

same concept and differences in the con-
text in which the data has been captured.
• Chronology : differences in the representa-
tion of attributes and the way values may
change over time.
• Security : differences in data access con-
trols and data governance policies.

Healthcare systems are typical information sys-
tems. Due to the large and complex nature of the
Health Service (the word Service here refers to
the overall provided health and medical services
to the patient, thus HS will be used to remove
confusion) in various countries and the ways they
are provided, a typical health organization can
have several different information systems, rang-
ing from diagnostic medical devices to electronic
health record systems (EHRs), the focus of this
paper. EHRs hold patient health record informa-
tion and support the delivery of day-to-day care
in hospitals in speciality clinics and primary care
clinics.
The concern for the project upon which this
paper is based (ePCRN: electronic Primary Care
Research Network) was to leverage the use of
EHRs, from primary care clinics, to support clini-
cal research. Due to various historic reasons, such
clinics use many different EHR systems with vary-
ing degrees of complexity. This heterogeneity has
resulted in significant differences across EHRs in
terms of their data representations, let alone com-
munication and interoperability between them.
This heterogeneity (and/or interoperability)
occurs in several ways [Sheth, A.P. (1998)., Ja-
kobovits, R. (1997), El-Khatib, H.T. et (2002)]:

Various methods have been suggested to
resolve system and syntactic interoperability
problems, which are often easier to resolve [Sheth,
A.P. (1998), Jakobovits, R. (1997)]. However,
achieving Structural and Semantic interoperabil-
ity, for example, in information interrogation or
interchange between information systems contin-
ues to be a difficult problem. In order to achieve
semantic interoperability in a heterogeneous
environment, the meaning of the information that
is interrogated (or communicated) has to have the
same semantics or meaning across the systems.
In clinical systems, the complexity is in the ways
that different systems represent their medical
information. For example, references to clinical
procedures, problems or diagnosis and they way
each system records or code each concept. Equally,
different systems use different codings to refer
to or represent their data, such as SNOMED-CT,
Read Codes, ICD9 etc.
Therefore the requirements for the clinical
research domain and linkage to EHRs require the
following capabilities:

• System : e.g. numerous heterogeneous
hardware platforms and operating systems;
• Syntactic : e.g. different data formats or
query languages; differences in systems
programmable interfaces or the way data
is accessed; differences in transaction
mechanisms.
• Structural : e.g. different model representa-
tions, composite structures of attributes.
• Semantic : differences in the meaning of
the terms used, different terms used for the

• Provide a coherent view of the data from

different autonomous heterogeneous sys-
tems or data sources.

• Access data securely.

• Enable independent and different data ac-

cess and sharing policies.

• Maintain data and patient privacy at all

times, observing strict ethical rules.

• Enable data source or system changes

independently.