Mapping Notation Need

Large-scale healthcare applications typically use multiple overlapping models, message formats and data structures:

• National initiatives such as HITECH have adopted several overlapping standards
• Within one SDO, there may be multiple overlapping standards (e.g. HL7 V2, V3 and CDA)
• There is interest in local data formats and Domain Specific Languages with precise semantic relations to standards (e.g. Green CDA, Micro-ITS, RIM ITS, Detailed Clinical Models, Virtual Medical Record)
• Healthcare providers need to interface IT systems with a wide range of different models and data formats.
• The HL7 Enterprise Compliance & Conformance Framework (ECCF) has a Specification Stack with three layers from OMG MDA: CIM, PIM, and PSM. ECCF Conformance Statements are testable relations between items in different layers. They are mappings between layers.

So national programs, SDOs, IT suppliers and healthcare providers frequently need to understand and document the precise relationships between different healthcare models and data formats. Failure to do so carries risk to patient safety. The relationships between different semantic models and data formats are commonly referred to as ‘mappings’, and the activity of defining them is called ‘mapping’ between the formalisms.

While mapping is critical to the integrity of healthcare IT systems and to patient safety, there is no single widely-used notation for mapping, and there are no agreed quality criteria for mappings. Mapping is typically done behind closed doors within implementation projects, with the results buried in code. Mappings are documented informally in Excel spreadsheets, and become shelfware.

A single standard notation for mappings would have clear benefits to the healthcare IT community, and for patient safety:

• Different groups could exchange results in the notation, getting common insights and sharing problems.
• Mappings could be made available for peer review and ballot, increasing their quality and the reliability of the systems that depend on them.
• Mappings in a standard notation are machine-processable, and will create a market for tools
• Mappings support conformance testing
• Mappings expose interoperability problems early – saving costs
• Mappings are more maintainable than code
• A common mapping notation will create a market for mappings in the notation.

Mappings are a critical element of the Healthcare IT eco-system. They will continue to be so, as long as no one healthcare data standard rules the world. Yet mappings have remained a second-class citizen of the Healthcare IT skill-set and toolkit. Failures of mapping have remained hidden, and have contributed to poor system integration.

By adopting one common notation for mappings, HL7 can change this. A standard mapping notation will make mappings a shareable, peer-reviewed resource, and increase their quality.

Return to top page on mapping notations