This wiki has undergone a migration to Confluence found Here

Properties of Requested Act vs properties of Request Act

From HL7Wiki
Jump to navigation Jump to search

Issue

Some properties of a requested-activity that is requested to be performed can't be conveyed. Most likely cause is the fact that an Act in RQO mood is a combination of the request (as an activity, the requesting process) as well as the requested-activity (the activity that an order filler is requested to perform).

Related: Inert vs. descriptive properties of Acts

Use-cases

  • See Order interaction for the creation of a document for the details of a number of use-cases:
    1. How does one, in an order message, convey that the resulting structure should conform to a particular set of templates, or to a particular profileId?
      • Lloyd: At the moment, the only way to do it would be with profileId. Presuming the profileId points to a narrower set of receiver responsibilities with payloads defined to reference narrower static models or templates, you could do it. Beyond that, there's no capability at present. Might be an interesting use-case to toss at SAEAF.
      • René: In other words: if the request identifies profileID A1, then this would imply to a receiver that the response document would have to conform to profileID B26. The link A1-B26 would be part of the definition of profile A1.
      • This same problem also exists for mime-type.
    2. How does one, in an order message, convey that the resulting act should use a specific language? The languageCode of the Act RQO contains the language of the request Act, and has no impact nor implications for any resulting/infulfillmentOf act.
      • (Hans) Within one institution I could only see limited use as practice dictates whether a prime language is used, or in some instances two languages are provided, but when requests for documentation to be used by patients or cross-organizations/country, I would expect some use cases to pop up over time. E.g., "Please re-send discharge summary in English" to a Dutch provider for a patient who presents on vacation skiing in Switzerland.
    3. How would you model: "please create a new document (a new version, in CDA terms) for document with ID X". Or "Please create an appendix of type Z for the document with ID X". Using "pertinentInformation" comes to mind, but we should be able to do better than that.

Discussion

  • (Tom) this matter is reminiscent of earlier discussions about negationInd and other Act attributes. In formal terms, the question is whether the language attribute is descriptive or inert. That determines whether it relates to the act of ordering or the act that is ordered. It might be that we need two different attributes, one for either case (just like we now have actionNegationInd and valueNegationInd). Another reason I'm copying MnM is that I have a growing feeling that we are starting to pay the price for merging two acts into one (the order and what is being ordered). I'm not saying the alternative is superior, but there would at least be merit in some further analysis.
  • (Lloyd) It's reasonable to me that there should be an inert and a non-inert language attribute. Is this something that only applies to observations, or would it make sense anywhere else? Observation is the only one where your event produces net new data, so I suspect it only applies there . . . There might be other characteristics of the observation.value you'd want to constrain too, such as number of repetitions, datatype, etc. We should probably think about what other constraints on the result might be needed and try to solve all of them at once.
  • (Lloyd) Looks like we should be devoting a session to this in May. I wonder if we should have a joint MnM/OO session? Too late to book it officially, but MnM could arrange to just show up in the OO room. It'd be really useful to have Gunther there, as I'm sure he'll have an opinion. (You'll note I've avoided answering your question, largely because I'm not sure of the correct answer)

September 2009 WGM

  • Tuesday Q2: after discussion of the use-cases, suggestion is to use DOCCLIN RQO with a INST act relationship to a DOCCLIN DEF, in human terms: "request for instantiation of a definition". The DOCCLIN DEF contains constraints for attributes such as languageCode and templateId.
    • In a wider context, where definitions have associations with other classes that are part of the definition, there is a wider issue that needs to be resolved. See Act in DEF Mood for details.

Resolution

January 2010 WGM, Sun. Q3

Refer to Act Mood and Inert Vs. Descriptive Properties of Acts