gms | German Medical Science

GMDS 2013: 58. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS)

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie

01. - 05.09.2013, Lübeck

Using the i2b2-Web Frontend to Query Custom Medical Data Repositories: Emulation of a Virtual i2b2 Server

Meeting Abstract

Suche in Medline nach

  • Raphael W. Majeed - Justus-Liebig-Universität, Gießen, DE
  • Rainer Röhrig - Justus-Liebig-Universität, Gießen, DE

GMDS 2013. 58. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e.V. (GMDS). Lübeck, 01.-05.09.2013. Düsseldorf: German Medical Science GMS Publishing House; 2013. DocAbstr.260

doi: 10.3205/13gmds138, urn:nbn:de:0183-13gmds1380

Veröffentlicht: 27. August 2013

© 2013 Majeed et al.
Dieser Artikel ist ein Open Access-Artikel und steht unter den Creative Commons Lizenzbedingungen (http://creativecommons.org/licenses/by-nc-nd/3.0/deed.de). Er darf vervielfältigt, verbreitet und öffentlich zugänglich gemacht werden, vorausgesetzt dass Autor und Quelle genannt werden.


Gliederung

Text

Since its release in 2004, the i2b2 software has become a valuable tool for clinical researchers. Physicians can use its browser-based query frontend intuitively without additional training or reading through documentation. Yet, a meeting of the European i2b2 academic user group revealed several reasons hindering hospitals from using i2b2 productively – most commonly difficulties with the process of extracting, transforming and loading (ETL) external data into i2b2, as well as a complicated installation/configuration procedure which led to different simplification approaches [1], [2].

Since many hospitals already have clinical data repositories, we propose a solution which provides the easy to use i2b2 web frontend while avoiding the complexities of the i2b2 server: A platform and database independent application which allows connections from the i2b2 web frontend by emulating basic REST interfaces of i2b2’s core cells. A feasibility study was conducted by developing a software prototype.

Methods: To avoid additional dependencies and the overhead of application servers, an OSGi centred Java architecture was chosen for the software prototype. The necessary HTTP-REST web service interfaces of i2b2’s core modules were determined by examining the XML message history of the message log, which is available from the public i2b2 web client demo application. The public i2b2 demo uses five web services: Project management (PM), Ontology (ONT), Clinical data repository (CRC), User/project workplace (WORK) and File repository (FR). Only the required functionality was implemented. A pre-existing PostgreSQL database was used as backend for the CRC cell. It contained 3.8 million facts for 11000 patients. The database schema was similar but not identical to i2b2’s database schema. The original ontology cell uses a monohierarchical tree structure stored in one or more database tables. To examine a more powerful and flexible approach, we used SPARQL to access SKOS ontologies stored in RDF triple stores, as proposed by Löbe 2010 [3].

Results: The i2b2 web client connects to the prototype server without any errors or warnings. Simple queries can be performed and return correct patient counts. Query performance was similar to i2b2. A multidimensional SKOS ontology is used by exposing a flattened view to the i2b2 web client. Concept keys are replaced by RDF ids.

Discussion: We were able to provide a prototype which enables users to use i2b2’s established web client to access their clinical data without the difficulties and shortcomings of the i2b2 server: direct access to the native data repository (1) removes the need for ETL and (2) avoids license costs and database administration overhead. (3)Installation overhead is reduced to a minimum: The prototype server can be installed and run with a single click. The full installation package fits into a 10–20MB zip archive. While the prototype application can still be run in a Java application server like Tomcat or JBoss, the strict adherence to the OSGi4.3 framework allowed us to reduce dependencies to a minimum and still keep the application highly modular. The prototype is part of the HIStream [4] project and allows clinical researchers to use the i2b2 web client to perform queries on a real time data warehouse [5].


References

1.
Mate S. Evaluation von i2b2 am Universitätsklinikum Erlangen [Studienarbeit]. Lehrstuhl für Medizinische Informatik. Friedrich-Alexander-Universität Erlangen-Nürnberg; 2009.
2.
Arnold G, Hart G, Key D, Eastman D. PS1-01: Infrastructure Setup for the Informatics for Integrating Biology and the Bedside Framework. Clinical Medicine & Research. 2011; 9(3-4):172–3.
3.
Löbe M, Stäubert S. Bringing Semantics to the i2b2 Framework. In: Fähnrich K, Franczyk B, editors. Informatik 2010. Lecture Notes in Informatics. Bonn: Ges. für Informatik; 2010. p. 734–8.
4.
Majeed RW. HIStream - A realtime HL7 stream based clinical decision support system; 2012 [cited 2013 Apr 15]. Available from: URL:https://sourceforge.net/projects/histream/. Externer Link
5.
Majeed RW, Röhrig R. Automated realtime data import for the i2b2 clinical data warehouse: introducing the HL7 ETL cell. Stud Health Technol Inform. 2012; 180:270–4.