Stream: Orders and Observation WG
Topic: Specimen Cardinality
Patrick Werner (May 04 2019 at 20:07):
Currently the specimen has a cardinality of 0..1, in the genomics use case we have the need to have multiple specimen. If you have an Observation on germline/somatic mutation you always look at least at 2 specimen, sometimes even 3.
Eric Haas (May 04 2019 at 22:28):
specimen can be split into multiple sub-specimen. I would have a single parent specimen with multiple childen @Riki Merrick @Lorraine Constable can you comment
Patrick Werner (May 04 2019 at 22:44):
but these specimens doesn't have a relationship. One specimen would be the tumor, the other blood from the patient or blood from a relative.
Eric Haas (May 05 2019 at 00:04):
what is the relationship of observation to specimen? are there three observations that get combined into a calculated result.
Lloyd McKenzie (May 05 2019 at 00:16):
Example - Person A is definitely a father/mother of Person B on the basis of the analysis of specimen A1 and specimen B1
Riki Merrick (May 05 2019 at 12:46):
Yes you will need to be able to track the multiple children - and children's children, when you are talking about anatomic pathology - if you can reference the imediate parent, you can daisey chain them together - if you can reference all specimen IDs, whithout a hierachy, that could work, but keeping the lineage intact would be better.
Patrick Werner (May 05 2019 at 13:37):
what is the relationship of observation to specimen? are there three observations that get combined into a calculated result.
The relation is that i have an observation which is based on two specimen. I need multiple specimen to compare the DNA of these. This could be 3 Observations, but when reporting genomic cancer variants, you would expect a single observation pointing to multiple specimen.
Eric Haas (May 05 2019 at 15:21):
can you provide a specific example ie a reference to a method or something to illustrate this. do you stick three disparate example in a blender and then run the analysis on the blenderized sample? or do you really have three separate observation each with the sample and then should maybe ref thes observation? Otherwise I think this is an edge case and an extension like 'aggregate sample' may be a better option.
Patrick Werner (May 05 2019 at 17:37):
we discussed the issue in CG and came to the same conclusion. The "correct" usage would be 3 different Observations, as this is too much overhead for us we will go with an extension.
Brendan Keeler (May 05 2019 at 18:29):
Isn't this a possible use case with anatomic pathology as well?
Brendan Keeler (May 05 2019 at 18:34):
Nevermind, answer is listed above.
Patrick Werner (May 20 2019 at 15:44):
We will define an Extension for 0..* "related" Specimen for the clinical genomics use-case. Does O&O prefer to have this Extension in FHIR-Core, or should we do our own extension?
Eric Haas (May 20 2019 at 18:37):
but can add as genomics extension, OO has not had input on genomics trackers to OO resource. (Maybe we should). Otherwise OO tracker.
Bob Milius (May 20 2019 at 18:51):
But adding it to core wouldn't take effect until R5, right? If we add it to the IG, then it could be used as soon as the IG is published (this summer?). Once it becomes published as part of core in R5, then a new IG can point to that.
Andrea Pitkus, PhD, MLS(ASCP)CM, CSM (Jun 11 2019 at 13:26):
late here, but another great example is transplant testing, and transfusion type and cross match (very common). In both cases, you have a donor specimen and recipient specimen. It's vital to keep them separate. They are collected separately, but used in the final report typically on the recipient (as that is where billing, the order/request originates, etc.)
It's a bit different than the use case with derived specimens as described by @Riki Merrick and common in AP, microbiology and the like.
Riki Merrick (Jun 11 2019 at 13:46):
I am even later than Andrea ... So it sounds like what is happening in CG is that each specimen is analyzed and then a conclusion is formed, tking these observations into account; I think the extension is a good idea. We are working on speciment resource this cycle - calls are Tuesdays 3 -4 PM EDT on the OO main line (Online Meeting Link: https://join.freeconferencecall.com/ord - Online Meeting ID: ord) - please feel free to join us on one of those).
Andrea Pitkus, PhD, MLS(ASCP)CM, CSM (Apr 29 2020 at 13:05):
also cultures as folks know....
Last updated: Apr 12 2022 at 19:14 UTC
