FHIR Chat · Non-numeric GenLab result with UoM · implementers

Stream: implementers

Topic: Non-numeric GenLab result with UoM


view this post on Zulip Maria Hu (Sep 23 2019 at 23:34):

ResourceObservation.value can have datatype of valueQuantity or valueString etc.
valueQuantity contains Units whereas valueString doesn't.
If I were to send a GenLab result which is text but contains UoM, what data type should I use? and how?
e.g. GenLab result for a Hep B Viral Load test, which can have a result 'Undetectable' and Unit = 'copies/mL'.
Please advise, thanks.

view this post on Zulip Eric Haas (Sep 23 2019 at 23:46):

can use a comparator to indicate less than a certain value as your undetectable. Strings and codes typically do not have UOM. Instead consider using http://hl7.org/fhir/observation-definitions.html#Observation.dataAbsentReason for these exceptional values. How do you currently do this?

view this post on Zulip Maria Hu (Sep 23 2019 at 23:53):

can use a comparator to indicate less than a certain value as your undetectable. Strings and codes typically do not have UOM. Instead consider using http://hl7.org/fhir/observation-definitions.html#Observation.dataAbsentReason for these exceptional values. How do you currently do this?

Thanks. Currently use valueQuantity but stuck with text result + UoM, that's why I'm inquiring how.
There are also other text result, 'Undetectable' is just one.

view this post on Zulip Lloyd McKenzie (Sep 24 2019 at 14:40):

You can use the originalText extension (http://build.fhir.org/extension-originaltext.html) to convey a free text version of any complex data type.

view this post on Zulip Eric Haas (Sep 24 2019 at 17:31):

How do you do this in V2 messaging? I don't recall seeing the units on "Undetectable" or "Indeterminate" or Others.... @Riki Merrick ?

Also can define the Observation using ObservationDefinition and define the UOM there. So maybe you won't need to send for these exceptional values.

I am not sure what Original text buys you. unless you provide a numeric results and "Undetectable" or "Indeterminate", etc. I personally think the DAR is better.

view this post on Zulip Maria Hu (Sep 24 2019 at 18:28):

ResourceObservation.value can have datatype of valueQuantity or valueString etc.
valueQuantity contains Units whereas valueString doesn't.
If I were to send a GenLab result which is text but contains UoM, what data type should I use? and how?
e.g. GenLab result for a Hep B Viral Load test, which can have a result 'Undetectable' and Unit = 'copies/mL'.
Please advise, thanks.

Do we have any suggestions how to map lab result in text with UoM?

view this post on Zulip Lloyd McKenzie (Sep 24 2019 at 18:34):

Fill in the 'unit' and put the full result in valueQuantity originalText extension

view this post on Zulip Maria Hu (Sep 24 2019 at 18:36):

Fill in the 'unit' and put the full result in valueQuantity originalText extension

Ah ok thanks Lloyd !

view this post on Zulip Eric Haas (Sep 24 2019 at 20:19):

@Lloyd McKenzie ... thats not a bad idea. but is totally new to me. Since this issue comes up in many quant lab tests. I think if we want to promote a single way to this in FHIR that it should be documented in the spec like we did for Observation.DAR. I think should doc this nugget in the Quant datatype and even consider making it an element. Right now the only way to learn about Lloyd's bag of tricks is here> :-)

view this post on Zulip Lloyd McKenzie (Sep 24 2019 at 20:32):

The question is how common this (text results + computable units) is in practice and whether it's worth surfacing this as common guidance.

view this post on Zulip Eric Haas (Sep 24 2019 at 21:55):

I will bring in OO where we have lots of lab represented

view this post on Zulip Grahame Grieve (Sep 24 2019 at 22:00):

I've not seen both represented before


Last updated: Apr 12 2022 at 19:14 UTC