The following e-mail was sent over several external mailers in Fall of 2004: ---
October 7, 2004
CTS II Project
Health Level Seven (HL7), an American National Standards Institute (ANSI) -accredited Standards Developing Organization (SDO) operating in the healthcare arena, recently published a standard for API’s to access clinical terminologies. This specification, which is called the Common Terminology Services (CTS), defines a set of interfaces that can be used to discover and query terminology content. The scope of the CTS specification was limited to (a) read-only use cases and (b) only the use cases that were necessary to satisfy near-term HL7 requirements.
The Vocabulary Technical Committee of HL7 is now initiating a second phase of this specification. Under the rubric CTS II, this specification is intended to be a logical extension of the initial CTS specification with the purpose of accommodating many of the use cases excluded from the initial specification.
The Vocabulary Technical Committee is currently soliciting input from interested parties or organizations regarding the requirements, scope and technology that should be included or excluded from this specification. The TC is interested in input regarding:
Should CTS II specify API’s for:
- Authoring? If so, are there any standards (e.g. DIG, SNOMED-CT) that should be considered and, if so, how should they be integrated? Are there any areas that should specifically be avoided? Should the authoring API’s include description logic classifiers?
- Import/Export? Should these standards be represented as API’s, data structures or both? Does the specification need to include “delta’s” – the ability to exchange a block of update rules?
- Provenance? What sort of provenance information needs to be included in the API’s? Does authentication need to be included in the specification? Does the API need to support role and access constraints?
Are there other areas as well?
The current CTS specification is has OMG IDL, Java and SOAP bindings. Should CTS II use the same approach, or is there a different technology that would make more sense. Has the OMG Model Driven Architecture (MDA) matured to the point that it would be useful? Are other options available?
What technologies will the specification have to interoperate with? (e.g. OWL, OBO)
The CTS II API, at a minimum, needs to support the requirements presented by the Enterprise Healthcare Record (EHR), which includes NHII RHIO and the NCI.
What other requirements does CTS II need to address? What resources should the TC use to discover them?