Mappings and Translations

Mappings and Translations

This page is for discussion and download of mappings and data translations made from mappings. Anyone can post a set of mappings, and anyone can post comments on the mappings they see. I hope the discussion will lead to improvement and extension of the mappings.

The page contains a number of zip archives containing Straw Man mappings and message translations. They are currently straw men because, while the mappings are sufficient to drive message translations, very little domain knowledge has yet been put into them, so there are many gaps and issues. Please start commenting on them now.

Sets of mappings and translations currently posted:

• V2 ORU^R01 to CCD Vital Signs Section: V2VSMappings211009.zip.
• V2 OMP^O09 to CCD Medications Section: V2MedMappings211009.zip.
• ASTM CCR to CCD, all sections except Social History: CCRMappings211009.zip.

Each mapping archive contains the following:

• A ReadMe describing the chosen problem (e.g. V2.5.1 Pharmacy to CCD) and explaining the file names
• The mappings in a readable Excel form, output from the tool used to make the mappings
• The same mappings in the native XML form used by the mapping tool – which you can soon use to edit and improve the mappings, and to run your own sample translations from them.
• A set of sample data translations made from the mappings.

You can comment on the mappings without looking at the mappings themselves; you just need to look at the example message translations made from them. Gaps in the translation results usually reflect gaps in the mappings.

The sample message translations have names like A.xml, AB.xml, and you need to know how this naming convention works. File names with one letter before the extension – A.xml and B.xml – are the source messages for translations, got from a variety of places. Typically A.xml is a CCD example, and B.xml is a V2.xml example (V2 bar-hat form is also supplied, in text files).

Then AB.xml is the result of translating from A to B – so AB.xml might be the result of translating from CCD to V2, using the mappings. Similarly, BA.xml is a partial CCD instance got by translation from the V2 message B.xml. Finally ABA.xml is the result of a round-trip going from CCD => V2 => CCD, using the mappings to translate at both steps. Similarly BAB.xml results from a round-trip V2=>CCD=>V2.

Any feature present in the source A.xml, but absent from the round-trip result ABA.xml, is something which could not be translated from CCD into V2 and back again, using the mappings. It may be that a mapping needs to be added to the mapping set, or it may be that V2 simply cannot express the information – so no mapping can be added.

If you find a mapping that needs to be added or changed, you could then look at the Excel form of the mappings and try adding a row for the new mapping. The spreadsheet describes the mapped locations in V2 by XPath expressions, and in CCD by RMIM class names, modified for templates. Or (in due course – see below) you can use the mapping tool to edit the mappings in their native form, and run the translations again.

The mappings were made with the mapping tools available for download at http://gforge.hl7.org/gf/project/v2v3-mapping/frs/ . However, all mappings were made with a CDA-capable version of the tools, which has not yet been posted there; so you’ll need to wait a couple of weeks before being able to refine and use the mappings yourselves.