2020-03-16 - OBSERVABLE Meeting

Observables and Evaluation procedures

Since Observables were introduced into SNOMED CT the demarcation between Observables and Evaluation procedures has been a potential source of variability of SNOMED CT use. In order to progress this issue the group need to agree on what knowledge is needed to inform any future decision.

2019-11-18:

<<404684003 |Clinical finding (finding)| : 363714003 |Interprets (attribute)| = << 386053000 |Evaluation procedure| ==> 7148 concepts

...but also

1350 hits for <<386053000 | Evaluation procedure (procedure) |: [2..*] { 260686004 |Method (attribute)| = <<129265001 |Evaluation - action (qualifier value)| }

3 hits for <<386053000 | Evaluation procedure (procedure) |: [2..*] { 260686004 |Method (attribute)| = <<129266000 |Measurement - action (qualifier value)| }

611 hits for <<386053000 | Evaluation procedure (procedure) |: [2..*] { 246093002 |Component (attribute)| = * }

2019-12-16:

Some more exploration of existing Evaluation procedure content:

(<<386053000 |Evaluation procedure (procedure)| : [2..*] { * = * }) MINUS <<363679005 |Imaging (procedure)|

10 000-ish Evaluation procedures have 2 or more role groups...

<<386053000 |Evaluation procedure (procedure)| : [2..*] { 246093002 |Component| = * }

611 have 2 or more Components, indicating that it could be a panel. However, many examples are just strangely modeled and likely not intended to be panels. Thus, it's not straight forward to identify panel-like procedures among the Evaluation procedure. Sometimes, e.g. 442553005 | Measurement of chloride in peritoneal fluid specimen (procedure) |, additional Components are inherited from ancestors.

(<<386053000 |Evaluation procedure (procedure)| : [2..*] { 246093002 |Component| = * }) MINUS <<122869004 |Measurement procedure|

Even non Measurement procedures have multiple Components...

<<386053000 |Evaluation procedure (procedure)| : [3..*] { 246093002 |Component| = * }

A few have 3 or more components, many of which are in fact panels, but again likely not all of them. E.g. see <67899004 | Complement component, classic pathway (substance) |.

Briefing note should include:

• Current status in regard to overlap
  • Actual concepts
  • Concept models
• Current status in regard to modeling of Evaluation procedures
• Orders and results use of SNOMED CT
  • Actual use of Evaluation procedure concepts for reporting results
  • Actual use of Evaluation procedure concepts for ordering
• Panels vs. single-measurement observables/procedures

2020-01-28:

The issue of panel concepts (sets of distinct Observable entities) was discussed. The current Observables model, by design, does not allow representation of panels.

Many examples of panels are not well-defined, e.g. a Complete Blood Count may contain different observables in different laboratories. The benefit of adding content on the international level is likely limited. However, SNOMED CT should provide the means to represent such panels when there is agreement on meaning on a national, regional or local level.

Several Evaluation procedure concepts contain multiple groups, but all are not correctly modelled.

2020-02-24:

Aim to get feedback from members about usage of Evaluation procedures.

Further aim to identify candidates for Procedure-to-Observables transfer, i.e. Procedure concepts which represent observations which has a single, distinct result/value.

Some gaps exist in creating observables corresponding to those procedures, most notably techniques are missing.

2020-03-16:

The topic has been lifted in the Modeling Advisory Group.

Panels (aka observables without values): Examples of types of panels are groupers for lists, screening and function tests. Naming today is insufficient to distinguish from e.g. totals (e.g. amino acid panel vs. total amino acid etc.).

Definition and naming of panels (observables OR evaluation procedures).

Piper Allyn Ranallo is  doing a mental health clean up based on Mental & Behavioral CRG work - she has submitted some requests for editing and applying the observable entity model to existing content: 285231000 |Mental function (observable entity)|, 8373002 |Nervous system function (observable entity)|, 4065008 |Affect, function (observable entity)|311465003 |Cognitive functions (observable entity)|. Need advice on applying the model to these function observables.

Hello Daniel and Suzanne

I have now received a reply from Piper to Daniel's questions:
https://prod-request.ihtsdotools.org/#/requests/preview/745434?kb
The new comment is:
Hi Nicki, Regarding the relationship between functions --

• both mental and cognitive functions are specific kinds of central nervous system (CNS) functions.
• Motor functions and sensory functions are specific kinds of nervous system functions, with some motor and sensory functions inhering only in the PNS (peripheral nervous system). others inhering in only the CNS, and others inhering in both the PNS and CNS.
• A cognitive function is generally thought to be a more specific kind of mental function, although I don't know whether there's a clear consensus among scientists about how to unambiguously define the line between a mental and a cognitive function. In general, most scientists describe a cognitive functions as higher level, or more complex, mental function -- functions related to things like consciousness.
• 'Executive cognitive functions' are an even more specific (i.e., more complex, higher level) kind of cognitive function that includes reasoning, memory, attention, etc.
• The relationship between cognitive and intellectual function is unclear to me. An intellectual function may be synonymous w executive cognitive functions. I just don't know.
• Regarding the body structure in which the functions inhere - we'll need to look in more detail at the literature and consult with cognitive neuroscience experts.
• Regarding the process that the functions realize, we will need to work on defining that more specifically.
• In terms of modeling these concepts, one thing that is unclear to us is the distinction between processes in the qualifier hierarchy and processes in the observable hierarchy.
• Another question we have relates more generally to different needs for modeling mental and behavioral phenomena (v. biological phenomena). While biological phenomena can be meaningfully defined in terms of body structures and physiologic processes, most mental, behavioral, and social phenomena are more appropriately defined in terms of functional systems that may or may not have clearly defined biological underpinnings. This is a topic that MBH-CRG would like to explore with the folks working on the Observable and Investigation Model Project. Thanks you, Piper

Evaluate the need for a hierarchy of processes that could be used to define “ability” observables and would allow for useful taxonomic groupings.

SNOMED Clinical findings may be defined in terms of "ability" observables and use values of | has interpretation | to indicate ability, volition, and dependence.

Where the use of the Observables model in SNOMED CT have been highly limited, we are currently seeing increase of activity among members (UK primarily) and also in the international edition (pathology, nutrition, mental health). This seems like a good moment to review where we are and where we like to be in terms of:

• Foreseeable use cases (around the corner: dentistry)
• Modeling decisions made, and to be made
• Documentation update

Next meeting is 2020-04-20 20.00UTC