Show simple item record

dc.contributor.advisorHauswrith, Manfred
dc.contributor.advisorPolleres, Axel
dc.contributor.authorSahay, Ratnesh
dc.description.abstractHealthcare applications are complex in the way data and schemas are organised in their internal systems. Healthcare standards have been proposed to reduce heterogeneities between healthcare applications. Widely deployed healthcare standards such as the Health Level Seven (HL7) Version 2 specifications are designed as flexible schemas which allow several optionalities while constructing clinical messages. Such flexibility and optionalities offer choices to meet clinician messaging requirements. However, these optionalities result in a quadratic number of interfaces and alignments between Version 2 applications. To overcome the alignment problem HL7 has proposed Version 3, where a centrally consistent information model controls terminologies and concepts shared by Version 3 applications. However, it is impossible to predefine all clinical possibilities in a central information model. Consequently, Version 3 allows the creation of local vocabularies and constraints, thus further impeding interoperability of HL7 applications. The goal of this thesis is to reduce the integration burden between HL7 applications. Our hypothesis which we aim to prove with the contributions summarised in this thesis is that "An ontology-based integration framework can improve healthcare data interoperability by reducing: (i) the integration burden (per message) between heterogeneous healthcare applications; and (ii) the number of alignments between heterogeneous healthcare applications". We address three main research challenges in the development of an ontology-based integration framework (i) how to build HL7 ontologies by reusing domain artefacts reside in separate global and local spaces; (ii) how to automate alignment of HL7 ontologies with greater accuracy of correspondences retrieved; and finally (iii) how to resolve inconsistencies caused by context-sensitive (or local) healthcare policies while mediating heterogeneous HL7 messages. We provide (i) a semi-automatic ontology building methodology for the HL7 standard; (ii) a semi-automatic ontology alignment methodology for HL7 ontologies; and (iii) a detailed investigation and a solution path towards realising context-awareness and modularity for local healthcare policies. This thesis offers an ontology-based integration framework called Plug and Play Electronic Patient Records (PPEPR) that reduces the number of alignments and the integration burden between HL7 applications. Further, the PPEPR framework enables interoperation of HL7 Version 2 and Version 3 applications. HL7 ontologies and their alignments are successfully deployed and evaluated under the PPEPR framework.en_US
dc.rightsAttribution-NonCommercial-NoDerivs 3.0 Ireland
dc.subjectHealth Level Seven (HL7)en_US
dc.subjectSemantic Weben_US
dc.subjectSemantic Interoperabilityen_US
dc.subjectOntology Building Methodologyen_US
dc.subjectOntology Alignmenten_US
dc.subjectHealthcare Policyen_US
dc.subjectOntology-based Integration Frameworken_US
dc.titleAn Ontological Framework for Interoperability of Health Level Seven (HL7) Applications: The PPEPR Methodology and Systemen_US
dc.contributor.funderScience Foundation Ireland Grant No. SFI/02/CE1/I131 (Lion-1)en_US
dc.contributor.funderScience Foundation Ireland Grant No. SFI/08/CE/I1380 (Lion-2)en_US
dc.contributor.funderEnterprise Ireland Grant No. CFTD 2005 INF 224en_US
dc.local.noteHealthcare applications are complex in the way data and schemas are organised in their internal systems. This thesis offers an ontology-based integration framework called Plug and Play Electronic Patient Records (PPEPR) that improves interoperability between Health Level Seven (HL7) applications.en_US

Files in this item


This item appears in the following Collection(s)

Show simple item record

Attribution-NonCommercial-NoDerivs 3.0 Ireland
Except where otherwise noted, this item's license is described as Attribution-NonCommercial-NoDerivs 3.0 Ireland