This topic was initially discussed at the EAG on November 27, 2023 and a briefing note has been prepared and is available here.
Requirements were developed with lab specialists to meet their needs. There needs to be some more specific use cases to make the effort worth it other than just making the hierarchy more manageable.
Additives are substances that would be incorporated into the specimen, whereas the separator attribute represents something that is not incorporated into the specimen. Also suggested that separators should not also be physical objects, but should be substances.
What is the impact of adding these to other hierarchies that might use these needs to be considered?
There is some use for the HAS INTENDED SPECIMEN even knowing it is not "always and necessarily true".
How should "spray-dried coating" be modeled in this scenario? Is it an additive or just a substance coating on the container? How manufacturers represent the difference between coatings and additives needs to be considered.
Need definitions for each of the attributes that make it clear to modelers what substances are allowed.
Will the level of effort needed to make these SD be worth the results.
2 Comments
Jim Case
Comments from Monique van Berkum
Feedback on “Changes to 706437002 |Container (physical object)| and the physical object MRCM”
2023-12-06
Background
Drivers identified in the briefing document for the proposed the changes include:
Several SNOMED member countries, including Czech Republic, the Netherlands and Sweden, have identified a requirement to specify specimen container types in more detail than is currently available in SNOMED. It should be possible to link these container types both to the specimen they contain or are meant to contain, and to any additives that were added during production of the container or to facilitate analysis of the specimen. This requires a change in, or rather an extension of, the machine-readable concept model.
The lack of a concept model for containers as well as support for specimens in containers make curation difficult and prevents the effective use of these hierarchies for analytics.
Goals of this project which include:
Feedback
The number of concepts to which these attributes can be applied is very limited. For that reason and other reasons outlined below, I do not think that the changes suggested in the briefing will accomplish the Goals identified above.
For Has additive (attribute)
Current proposal:
Summary: This attribute represents a substance that is intended to mix with a sample, e.g. a fluid, granular substance or spray coating on the inside of a container. An additive can be added to a container at time of manufacture, or to a specimen while or after the specimen is added to the container.
Feedback on Has additive (attribute)
On a quick scan of the 1173 descendants of 123038009 |Specimen (specimen)| in the 2023-12-01 Release, I could only find about 12 specimens (below) with some type of “substance” in their FSN. Of those, only the ones in red might include additive substances.
Without a clearer use case, SNOMED may want to consider avoiding going down the path of pre-coordinating Specimens with additives in them. Additionally, I advise against the level of pre-coordination of the concept in green, which pre-coordinates the specimen, the container it is submitted in, and the additive in the container. The permutations of specimens, containers, and substances added to the containers (or to the specimen) are enormous and this might be better handled in an information model, if at all. The few exceptions could simply remain primitive.
Descendants of 123038009 |Specimen (specimen)| in the 2023-12-01 with some type of “substance” in their FSN.
706437002 |Container (physical object)| has only 50 descendants in the 2023-12-01 Release, of which none have an FSN that includes an additive substance.
Also, since specimens can be collected in objects other than “Containers”, the domain for |Has additive (attribute)| would likely need to be a much higher level concept like 49062001 |Device (physical object)|.
Examples of devices, other than “Containers”, in which Specimens might be collected:
706041008 |Device for body fluid and tissue collection/transfer/processing (physical object)| (has 122 subtypes)
Some examples of subtypes, for which |Has additive (attribute)| might apply, include:
Examples of other devices that could be used to “collect/contain/hold” specimens that are not necessarily subtypes of 706437002 |Container (physical object)| or 706041008 |Device for body fluid and tissue collection/transfer/processing (physical object)| include:
For the reasons above, and those in the Conclusion section of this feedback, I would also consider holding off on |Has additive (attribute)| for concepts in the |Physical object| hierarchy until further evaluation reveals that this attribute can be consistently applied across concepts in the |Physical object| hierarchy in a way that it can be relied on for finding duplicates, classifying the hierarchy, analytics, etc..
However, if this attribute is added, consider something more like:
Has additive (attribute)
Summary: This attribute represents an additive substance (e.g., a fluid, granular substance, container coating, anticoagulant, preservative, etc.) that is intended to mix with a specimen/sample and is included with the container at time of manufacture.
Additional Guidance/Exclusions:
Has separator (attribute)
Current proposal:
Summary: This attribute represents a separator, something that is intended to separate parts of the sample. A separator can be a substance, such as a gel separator, but also a mechanical separator, such as a plastic barrier.
Feedback on Has separator (attribute):
I do not think this attribute will accomplish the objective of Goals 1 and 3 (in the Background above) for the Domain: 706437002 |Container (physical object)|.
Of the 50 descendants of 706437002 |Container (physical object)| in the 2023-12-01 Release, none have an FSN that includes a physical separator or a separator substance.
For this reason and reasons similar to those above and in the Conclusion section of this feedback, I would also consider holding off on |Has separator (attribute)| for concepts in the |Physical object| hierarchy until further evaluation reveals that this attribute can be consistently applied across concepts in the |Physical object| hierarchy in a way that it can be relied on for finding duplicates, classifying the hierarchy, analytics, etc..
Conclusion:
In general, devices in which specimens might be collected (containers, etc.) will not necessarily have FSNs that provide information as to whether the device has an internal separator, an additive substance, a separator substance, or is coated with something that might act as an additive versus only as a coating material. While some commonly named collection devices may include this information, many, if not most, may not. The task for modelers of investigating the FSNs of existing physical objects in SNOMED and then trying to look up information to figure out which objects contain separator gels, additives, physical barriers, fixatives, preservatives etc. will be extremely cumbersome.
While it may be possible to make some of these linkages, this does not mean that SNOMED should. In my opinion, at best, these attributes will be sporadically applied making them unreliable for anything (queries, detecting duplicates, classification etc.).
Examples of additives, separators, coating materials (e.g., surfactants) etc., that could be in a collection tube. It is unlikely that FSNs will fully capture this information.
Monique van Berkum
Images and color context have been lost in the above comment. Please see attachment for full feedback.