03.12.2012 Views

Semantic Web-Based Information Systems: State-of-the-Art ...

Semantic Web-Based Information Systems: State-of-the-Art ...

Semantic Web-Based Information Systems: State-of-the-Art ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

0 E chelberg, Aden, & Thoben<br />

care institutes connected to <strong>the</strong> network could be implemented ra<strong>the</strong>r easily, again<br />

utilising <strong>the</strong> gatekeeper role <strong>of</strong> <strong>the</strong> TTP entity in <strong>the</strong> network. This would assure<br />

that only au<strong>the</strong>nticated entities could participate in <strong>the</strong> protocol ei<strong>the</strong>r as a requestor<br />

or as a record repository and, <strong>the</strong>refore, would provide a minimum level <strong>of</strong> quality<br />

control <strong>of</strong> <strong>the</strong> data communicated within <strong>the</strong> network.<br />

Results<br />

The patient identification protocol establishes a concept that, to our knowledge,<br />

has not been available in <strong>the</strong> health care sector before — an undirected search for<br />

patient records (i.e., a search that does not require a priori information about <strong>the</strong><br />

location <strong>of</strong> <strong>the</strong> records) that does not violate data protection requirements. The<br />

concept matches well with <strong>the</strong> peer-to-peer network structure established in <strong>the</strong><br />

context <strong>of</strong> <strong>the</strong> ARTEMIS project, where many health care institutes may form a<br />

network <strong>of</strong> services in a very dynamic manner. The use <strong>of</strong> semantic annotation allows<br />

one to cope with <strong>the</strong> fact that different institutes identify patients with different<br />

demographics and that no globally unique identifier for patients will be available<br />

for <strong>the</strong> foreseeable future. The possibility <strong>of</strong> including national extensions allows<br />

one to maximise <strong>the</strong> specificity <strong>of</strong> <strong>the</strong> search algorithm in environments that can<br />

make use <strong>of</strong> such extensions, while <strong>the</strong> inclusion <strong>of</strong> phonetic encoding allows one<br />

to improve sensitivity by gracefully handling spelling errors. The integration <strong>of</strong> <strong>the</strong><br />

Retrieve <strong>Information</strong> for Display protocol for <strong>the</strong> second phase (i.e., <strong>the</strong> request and<br />

transmission <strong>of</strong> clinical documents once <strong>the</strong>ir location has been identified) provides<br />

a simple read-only access to clinical documents that easily can be integrated with<br />

<strong>the</strong> existing legacy systems.<br />

An implementation <strong>of</strong> <strong>the</strong> patient identification protocol is currently under development<br />

in <strong>the</strong> context <strong>of</strong> <strong>the</strong> ARTEMIS project. The implementation is based on <strong>the</strong><br />

ARTEMIS middleware, which provides for <strong>the</strong> peer-to-peer infrastructure required<br />

by <strong>the</strong> protocol; that is, message transmission and broadcast facilities based on <strong>the</strong><br />

JXTA peer-to-peer communication platform (Gong, 2002), a policy-based security<br />

architecture for a secure communication over unsecured networks and a pre-selection<br />

<strong>of</strong> record repositories that participate in a query based on geographical location (e.g.,<br />

“we are looking for records located in Germany only”). Pilot applications that will<br />

evaluate <strong>the</strong> protocol in a clinical setting are planned in <strong>the</strong> UK and in Turkey. From<br />

<strong>the</strong>se prototypes, we expect practical experience about use patterns, performance,<br />

and scalability <strong>of</strong> <strong>the</strong> protocol. Krieg, Hense, Lehnert, and Mattauch (2001) report<br />

about an evaluation <strong>of</strong> <strong>the</strong> probabilistic record linkage based on control numbers,<br />

which is used in <strong>the</strong> Münster Cancer Registry in Germany. They use a set <strong>of</strong> 19<br />

control numbers computed from <strong>the</strong> patient’s name (including name <strong>of</strong> birth, prior<br />

name, and phonetic encoding <strong>of</strong> <strong>the</strong> name components) and <strong>the</strong> day <strong>of</strong> birth. The<br />

Copyright © 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission <strong>of</strong><br />

Idea Group Inc. is prohibited.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!