Difference between revisions of "Simplification in the UK"
|Line 29:||Line 29:|
== Simplification Inputs and Process ==
== Simplification Inputs and Process ==
Message simplifications are defined using mapping and simplification tools, which are an updated version of the tools posted on the HL7 GForge.
Message simplifications are defined using mapping and simplification tools, which are an updated version of the tools posted on the HL7 GForge
Revision as of 17:42, 20 February 2011
Simplification of V3 and CDA in the UK: Background
The NHS have defined a number of Version 3 and CDA message exchanges and content, and published them in the Message Implementation Manual (MIM). Many of these are now in widespread use across the UK. Currently the NHS is defining the second release of its Integration Toolkit (ITK), to provide easly implementable standards for local integration, meeting the stated requirements of UK providers of health and social care. It is likely that the ITK will include CDA and V3 messages; for instance Release 1 of the ITK includes a CDA discharge sumnmary, which has been implemented by seversal suppliers; and Release 2 will include V3 messages for the national data spine.
NHS is considering whether to use message simplification techniques in ITK. It is in any case possible for suppliers and providers to use simplification as an implementation technique to read or write conformant full CDA and V3 messages, as defined by NHS, on the wire.
A project is under way to produce simplified forms of the CDA documents defined by NHS as part of the Common Assessment Framework (CAF) for health and social care. The intent of CAF is for caregivers in health and social care to be able to share assessments and care plans. The technical work of that simplification project is described below.
Outputs of the CAF Simplification Project
There are five CDA CAF documents which describe the assessments of patients and citizens, and their care plans. Some of these overlap strongly with the CDA discharge message defined in the MIM. The simplification project will produce the following outputs (for each CAF document, unless stated otherwise)
- XML schema for the simplified message
- Example instances of the simplified message
- Transforms in both directions between full and simplified messages (as XSLT, or via a Java transform engine)
- Results of round-trip tests of the transforms (Full=>Simplified=>Full)
- Mappings between simplified and full CDA forms, in the Neutral Mapping notation
- A single simplified class model, supporting all the CAF messages.
The mappings fully define the semantics of the simplified messages, in terms of RIM-based CDA semantics.
It is intended that as long as the information carried by the simplfied and full messages is a closed set (e.g no unlimited recursion, no undefined data types), the round trip betweeen full and simplified messages will preserve 100% of the information. This means that systems using the simplified messages will be able to ineroperate freely with systems using the full messages.
Rendered text in the CDAs is passed through the transforms unchanged.
Apart from rendered text, the simplified messages are typically about three times smaller than the full messages, in terms of bytes or number of distinct node types. In terms of simplicity of the XML, a simplified message should be essentially 'as good as it gets': if, for some defined domain, the information could be conveyed more simply, you can re-run the simplification process to define that simpler message.
Simplification Inputs and Process
Message simplifications are defined using mapping and simplification tools, which are an updated version of the tools posted on the [GForge].
To define a simplified message, you navigate the full CDA or V3 message tree structure and mark up, using the tools:
- which leaf nodes are to be retained in the simplified message
- business names for those leaf nodes
- business names for internal nodes above the retained leaf nodes
- which internal nodes (typically those with multiplicity 1..1) are to be flattened, to give shallower XML
Once these have been marked up, production of the simplification outputs (XML schema, transforms, mappings, simplfied class model) is done automatically by the tools. There is tool support for applying common simplifications across different parts of one message, or across different messages.