Chair: Rob Hausam, Scribe: Rob Hausam

TermInfo Conference Call Minutes
Wednesday, April 9, 2014- Time: 8:00 AM EasternTime (1200 UTC)

Chair: Rob Hausam, Scribe: Rob Hausam

Attendees:Lisa Nelson, Riki Merrick, Jos Baptist
Regrets:Heather Grain, Daniel Karlsson

Announcements
IHTSDO SNOMED CT ImplementationShowcasecall for submissionsclosesFriday, 2nd May
Call for papers at:
Theme isSNOMED CT - making health records make sense
Consider a submission on the TermInfo project work - volunteers?
Looking at the website – there is a cost associated with the showcase ($99), this should have been included in the announcement
Met with Conformance WG (CGIT) regarding conformance aspects of TermInfo - plan for joint discussion in Phoenix (Have template ID to assert, but how exactly would that work?)
Met with Templates WG (TermInfo co-sponsor) regarding potential for use of templates to manage information model/terminology overlaps - plan for joint discussion in Phoenix

Review and approve April 2nd minutes– move to approve last week’s minutes as distributed – Riki Merrick, Lisa Nelson, no further discussion,abstain: 0, against: 0, in favor: 3
Action items review(if possible, please send your updates in advance)
From January 2014 (San Antonio) WGMTermInfoand Vocab quarters
UpdateTermInfoProject Milestones (Rob) - still pending
Negation
Review and update Core Principles negation guidance (section 6.6 etc.)
Prepare Project Scope Statement to update Core Principles section 6.6 on negation (joint with MnM and Vocab) (Jean Duteau)
Prepare Help Desk article on negation (Lisa)
Prepare Project Scope Statement for Implementation Guide on negation in HL7 products (Jean Duteau)
MnM lead, Vocab co-sponsor
Useful to also have CGIT and SD co-sponsor
Ballot reconciliation forTermInfoDSTU(60 min.)
Continue working through comments for discussion and proposed dispositions that were distributed last week

#154: Copied from last week - Discussion about how to deal with overlap in structure – template designers could use the nullFlavor (existing) when folks want to use post-coordination instead of the structural approach – BUT, if the structured approach is used do not use post-coordination – template designers could also constrain the doubled up concepts out of the template.

Identify the overlap at the time of designing the template and deal with it at this time.

Should have a requirement for an explicit statement of use of structure vs terminology representation, then create the statements of how each should be handled to allow translation between the different approaches.

SNOMED has a lot more nuances for procedure site (direct, indirect etc) – cannot disallow use of SNOMED without loss of meaning

Looking for proposed wording for this section as homework – Rob to draft, Lisa will edit and bring back next week – other suggestions welcome

Today’s discussion: – have reached out to Templates WG and planned a quarter in WGM, but in general we should really work on making the templates we design very explicit – reviewing the value sets for overlap issues and then make explicit SHALL NOT statements to avoid overlap, if so decided – or the other way around, but be explicit about it. Combining with the value set definitions project – that should be also considered there = disallow pre-coordinated concepts / post-coordination with laterality for example.

Deposition comment for persuasive with mod: template design have to explicitly define and declare how the structural and vocabulary elements will work together. If the intention is to use the terminology concept model, then explicit statements shall limit the overlapping use of structural elements in the template. Lisa will draft and bring back next week.

It is not up to TermInfo document about which is better, rather describe the issues and expected solutions for each – if terminology do “X”, if structure do “Y”.

If you have a research project, you may want to use the terminology approach to allow variability, while if you are designing in your hospital system a structure for procedure protocols, you may need to disallow terminology granularity, so need both approaches.

HL7 is in the middle and needs to be able to communicate with both – need to examine the RIM on how it can accommodate the differences in the models of representations.

#157: 2.2.10 – double negation: Negation indicator is on the act, which here is an assertion: so needs to state: “did not observe “finding X not present”” for negation on the act – do not change the words of the concept expressed that is being negated. Also need to show where the negation indicator is applied – If to the value or to the act – should update the example to show the proper way for assertion pattern or value negationIndicator.

“Finding X absent” plus value negationInd = NOT “Finding X absent”

More likely someone wanted to use the Act negationInd = not observed “Finding X absent”

Example in SNOMED of not findings: 168223003^No parasites isolated (finding)

Value negationInd = Not No parasites isolated – this is non-sensical.

ActnegationInd = Not observed No parasites isolated

Act negationInd – why use this on a finding?

You could negate the act of culturing a sample, but should we use it on the finding?

Have to limit the use of negationInd – the only things to allow negation for is the Act, since this is for Act.negationInd – but this is the old form, that can do both – maybe we have to have an example on how both functions are covered here (both Code and Values are descriptive attributes – need to add explanation of what this means first using real examples and then apply the negationInd in both ways – once we have the new text, then adjust the disposition as to where to go.

Status of spreadsheet at end of call:

Call adjourned 9:30 AM ET – NOT further discussed:

Status of jointTermInfo/Vocab/MnM negation "guide"
Vote on the IG PSS planned forThurs. April 10
Need to fill in additional facilitator roles - final review on call today
Continue planning discussion for DSTU re-ballot(if time permits - likely defer to WGM)
Sept. 2014 or Jan. 2015 target?
If Sept., need NIB after May WGM
Include LOINC and current RIM

Agendaitems for next call

Keep moving forward for discussion at a later time:

#49: Did we deal with what branch in SNOMED we should be drawing from for coding units? Add <258666001^Unit (qualifier value) as applicable hierarchy for use in the translation code? But informal units are not necessarily covered her – then what?

CDA taskforce is working on example of units that are not in UCUM – may be have 3 patterns:

If unit, but not in UCUM, use Translation with SNOMED, otherwise use {}? Discuss next week

For consideration here is existing text for section 2.2.13.2 with example:

2. In the case of “informal” units, which have no standard UCUM representation, a SNOMED CT concept identifier MAY be used in the translation sub-element of the unit element.

oExamples of “informal” units may include counted items (such as tablets or capsules). Items such as these are frequently encountered in the role of a "unit", but are not true units at all (as they include dose form or other information). The use of these “informal” units, although common, is discouraged. The nature of the counted items should be captured in the appropriate information model or terminology structures. In these cases the unit "1" (the UCUM symbol meaning "the unity") SHOULD be used and the SNOMED CT representation of the nature of the counted unit MAY then be used in the translation sub-element of the unit element.

Examples:

Use of the unitary unit=”1” combined with a translation code from SNOMED from <258666001| units (qualifier =value)|

<translation code="428673006 "

displayName="Tablet - unit of product usage (qualifier value)"

codeSystem="2.16.840.1.113883.6.96"

codeSystemName="SNOMED CT"/>

</value>

1.) is not okay because translation must express the same concept as the enclosing PQ (An alternative representation of the same physical quantity expressed in a different unit, of a different unit code system and possibly with a different value.) And a tablet is not the same as "1" (but check the exact SNOMED CT definition of "unit of product usage (qualifier value)"...)

units are NOT in UCUM, and NOT in SNOMED CT:

2.) is formally okay, but "cycle times" is a property - and not a quantity. How many cycle times? 21.3 ! So a better solution would be a suitable LOINC / SNOMED code in the enclosing Observation.code

3. Use of a nullFlavor=”OTH” and then putting the non-standard units in the originalText as shown below.

<originalTextTHOUS/MCL</originalText

</translation>

</value>

The updated ballot spreadsheet is on gForge a copy is here:

3.) The UCUM part is correct and sufficient.Which information does thetranslation part add? Maybe it is not needed here.Moreover, I am not sure if the strategy with nullFlavor "OTH" works here in the same way as in CD - I would argue that in PQR the coded concept is the union of value and code, therefore it makes no sense to give a value but no code. If this information isreally required, I think it would be more correct to write the following?

originalText125THOUS/MCL</originalText

</translation>

</value>

There was so much discussion again and again about these "tablets" and other countable entities in the past... For me, it is very clear that this is unneccessary and redundant. A prescription could be

500 mg Yxocilin = (500 mg) of (Yxocilin)

or it could be

2 tablets Yxocilin250 = (2) of (Yxocilin250 tablet)

In the first example, the amount of ingredient is expressed, in the second example the amount of product is expressed.

In both cases, there will be some information in some of the enclosing structures that tell as whether the PQ measures an ingredient, or a product consisting of countable entities, or the observation of "cycle times"...

Maybe, examples should be extended to more complete SubstanceAdministrations or Observations in order to understand what is going on.

Just a few more remarks:

Note that in UCUM frormally "10^" and "10*" are considered unit expressions, the exponentiation follows exactly the same rules as in m3 or mol-1 etc.

The Webaddress not valid anymore, please use