Negation Requirements Project Minutes 1 June 2016

From HL7Wiki
Jump to navigation Jump to search

Back to Negation Minutes


Meeting Information

HL7 PC-CIMI-POC Meeting Minutes

Location: PC call line

Date: 2016-06-01
Time: 11:00-12:00 ET
Facilitator Jay Lyle Note taker(s) Jay Lyle
Attendee Name Affiliation

y Jay Lyle JP Systems
y Richard Esmond
Gerard Freriks
y Rob Hausam
y Susan Barber
y Jim Case NLM
y Karl Poterack


Agenda Topics

  1. Discussion of principles
  2. Domain coverage


Minutes/Conclusions Reached:

  1. Principles
    1. We agree that avoiding counterintuitive semantics & intractability both suggest that avoiding Boolean types in information modeling is advisable, except where the domain property is actually and explicitly Boolean. I.e., even "yes/no" questions are better modeled with "yes/no" value sets. If a DL implementer wishes to convert such data into logical negation, that can be done on a case-by-case basis.
    2. We agree that consistency is important; which is to say, all else being equal, similar content should be represented with similar structures.
      1. The difficulty is in determining how similar things must be before we consider all else equal. E.g., is "no known allergies" the same as recording the absence of any other condition, or is the allergy case different enough procedurally to merit its own pattern?
    3. We agree with the CIMI maximally coordinated principle, being that the structure of a complex value should be modeled in order to support use of the parts by systems that may not support parsing of description logic expressions.
      1. This principle has a floor: where such properties are not used, and precoordination is the norm, precoordination is acceptable (e.g., LOINC).
      2. Once again, the difficulty is in the determination of what specific cases belong on the respective sides of that distinction.
      3. The suggestion to "avoid negation in unary code patterns" another example of the difficulty in drawing that boundary: postcoordinate in the information model, except where you don't.
      4. How prevalent must a precoordinated pattern be before it is deemed acceptable?
  2. Note that while "No allergy to [X]" is a negative, for which a specific substance has been tested, "No known [drug] allergy" is actually a null, not a negative. It means "No allergy to the substances to which I've been exposed, be we don't know anything else."
  3. Example of menses:
    1. "No menses" a member of the pattern of "absent normal function"; whether this should be a function + absent or a precoordinated deficit is TBD.
    2. "Date of last menstruation" may be null (or in distant past), but should not be expected to communicate "none" or any other reason for its own state; that's a separate question.

Meeting Outcomes

  • review Use Cases for completeness, classification for accuracy (all)

Next Meeting/Preliminary Agenda Items
  • Continue review of requirements

© 2012 Health Level Seven® International. All rights reserved.