The Australian Health System is currently grappling with the transition to a fully electronic system which will allow health data to be accessed when and as required. The Health Information Environment project tackles issues related to this National goal.
The long-term goal of this project is to develop tools and techniques to support the use and analysis of health data in electronic form including use in decision support systems.
Since the key to such use is the semantic integration of data from multiple sources and NEHTA has identified SNOMED CT as the pre-eminent clinical terminology for the exchange of clinical data, this project aims to exploit and extend the formal description logic foundation of SNOMED CT to both enable the faster uptake of SNOMED CT in the health system for capture of data and to develop the tools and know-how required for the use of the semantics captured in this data.
Research topics that will be addressed include:
- Concepts that are not representable using clinical terms from an existing ontology
- The large size of the existing ontology
- Errors and inconsistencies in the existing ontology
- Evolution of the ontology over time
- Management of concrete domains - how to incorporate numeric values into SNOMED CT
The tools described below are aimed at enabling the use of the SNOMED CT terminology in the near term, with tools for the development of further content, mapping of existing terminologies to the SNOMED CT terminology and access to content from a single server.
Other projects within the Health Information Environment are making use of these tools.
The Australian e-Health Research Centre is also researching other ways of using clinical terminologies. These include:
- Improving the classification of free text document by mapping words and phrases to concepts in the ontology
- Using ontologies to improve the process of collecting data
Snorocket™
A classification engine that has been licensed to the IHTSDO for use in the SNOMED Workbench for the ongoing maintenance of SNOMED CT internationally. Snorocket is fast, able to classify SNOMED CT at least an order of magnitude faster than other known classifiers.
Snapper Platform
The mapping of existing data to expressions using an ontology involves:
- Normalising the expressions, which may contain multiple codes from an ontology, for consistency purposes
- Calculating subsumption relationships, so that the relationship between expressions is understood
- Finding efficient ways to process the existing ontology extended with new terms and relationships
This work aims to make it possible to fully describe data which is already collected using existing ontologies. The outcomes of the project can then be used in building better information systems with data which is already collected.
Ontoserver
An ontology server designed to support the requirements of SNOMED tooling including the ability to manage multiple SNOMED extensions, subsumption queries, and SNOMED Reference Sets.
