The Snapper Platform is an Eclipse-based application designed to support describing the meaning of terms in existing clinical terminologies using concepts or expressions from SNOMED CT. However Snapper is not bound to SNOMED CT, and other ontologies can be used as an alternate "base" ontology.
Creating extensions for SNOMED CT™
The SNOMED CT ontology has been identified by many Health Departments world-wide as the standard clinical terminology to be used when collecting clinical information about a patients treatment.
Whilst SNOMED CT is vast, its coverage can be patchy, particularly in specialist areas. SNOMED CT is a living ontology with releases twice a year. However this delay can be problematic for adoption. Furthermore, there is no guarantee the concepts you need will make it into the international standard. There are also vast amounts of legacy ad-hoc terminology data, and a slow and often simplistic reaction to the problem of SNOMED CT support by the large commercial software vendors.
A solution to these problems is to create your own local SNOMED CT extension. To do this, you must be able to explain the relationship between your concepts, and those in the base SNOMED CT ontology. This is achieved though the authoring of expressions and this is where Snapper comes in.
Snapper allows you to import a list of ad-hoc terminology source terms from an existing data set and then map these terms to SNOMED CT using SNOMED's post-coordination expression syntax (see screenshot above). Realizing from the outset that post-coordination of concepts is difficult, an expression editor was designed with features similar to a smart program editor, such as colour-coded syntax highlighting, automatic description completion and a templating engine driven by the SNOMED CT concept model.
Usability issues addressed
- The Automap feature in Snapper attempts to automatically generate the simplest of mappign expressions - a one to one mapping. This can cut down the mapping workload considerably. Additionally, users can write a script to perform some of their own mapping, allowing for handling of common abbreviations or other features specific to the source terminology.
- Drag and drop functionality of concepts into the expression editor from the search results view and the graphical visualization was added as a first step to supporting a direct manipulation approach to composing expressions.
- Many people interact with SNOMED CT terms and not the underlying concept identifiers. Concept identifiers feature heavily in our expression editor, and whilst this conforms to the official SNOMED CT documentation, alternative visual representations supporting user interfaces metaphors such as direct manipulation will better cater for this group of users.
- Many users have no experience with the SNOMED CT compositional grammar and the syntax is somewhat obscure and unintuitive. Ideally the editor should handle these details of the grammar and support the user's goals of identifying the focus concept for mapping and then mediate its subsequent refinement based on the Concept Model constraints.
- There are general mapping patterns, akin to software design patterns, that represent a set of best practices in mapping, developed over time. To assist in the authoring of mappings with reproducible quality and ease repetitive mapping tasks, the tool should support user-specified workflows and templates to capture these patterns.
