Stream: terminology
Topic: Gestational Age
Michelle (Moseman) Miller (Aug 22 2017 at 13:05):
We are trying to understand if there are specific codes that imply whether the gestational age being observed is for mom's fetus (when she is pregnant) or mom (when she was a baby herself). We started looking at LOINC 11884-4 (Gestational age Estimated) and SNOMED 412726003 Length of gestation at birth (observable entity), but after dissecting the codes (especially SNOMED), we got twisted because gestation is defined as "the process of carrying or being carried in the womb between conception and birth"
The issue is when a single patient's chart has her own gestational age at birth documented as well as her child's gestational age (at the time of delivery).
I believe the FHIR extension, http://build.fhir.org/extension-observation-focal-subject.html, helps clarify the subject when subject (fetus) differs from patient. I'm just curious whether terminology can also be used to differentiate these 2 observations (mom's gestational age at birth when mom was a baby -- versus -- mom's fetus' gestational age when mom was pregnant)?
Swapna Abhyankar (Aug 22 2017 at 14:15):
I'm just curious whether terminology can also be used to differentiate these 2 observations (mom's gestational age at birth when mom was a baby -- versus -- mom's fetus' gestational age when mom was pregnant)?
In LOINC, you can tell who the subject of the observation is by looking at the System Part. ^Patient means the observation is about the patient and the result is stored in the patient's record, ^Fetus means the observation is about the fetus when the fetus isn't the immediate patient (i.e., the pregnant mom is having the ultrasound and the result is stored in the mom's record), and ^Mother means the observation is about the patient's mother and the result is stored in the patient's (usually an infant or child) record.
So for example,
85724-3 Age^at delivery Time Pt ^Mother Qn = the patient's mother's age at delivery of the patient
43993-5 Age^at delivery Time Pt ^Patient Qn = the patient's age at which she delivered
76516-4 Gestational age^at birth Time Pt ^Patient Qn = the patient's gestational age when the patient was born
18185-9 Gestational age Time Pt ^Fetus Qn = the patient's fetus's age when the patient is pregnant
Hope this helps!
Michelle (Moseman) Miller (Aug 22 2017 at 16:54):
Thanks, the last couple LOINCs might work. I think we got twisted looking at the SNOMED code, but maybe LOINC has better options.
Robert McClure (Aug 24 2017 at 15:00):
@Michelle (Moseman) Miller I agree, that's is confusing wording for SNOMED. Where did you find that definition? In any case, the default attribution is "patient" for SNOMED hence unless changed, the gestational age would be for the person whose record it's in. I'll bet that would still be confusing and since this is an observable - go with LOINC.
Michelle (Moseman) Miller (Aug 24 2017 at 20:20):
@Robert McClure an internal terminologist provided the definition (using a medical dictionary) - so it wasn't tied directly to SNOMED from what I understand....
In any case, I failed to clearly articulate the question earlier. My original question was from the perspective of a pop health platform that aggregates data from disparate source systems, so we are at the mercy of how the source system codifies gestational age. We were trying to assess whether or not the focal subject was known from the SNOMED code. I agree LOINC is much cleaner (when LOINC is provided). Based on your response, though, it sounds like we can assume that the SNOMED code has a focal subject of patient. Thanks -- I appreciate the input.
Rob Hausam (Aug 25 2017 at 12:26):
@Michelle (Moseman) Miller SNOMED CT 412726003 |Length of gestation at birth (observable entity)| is a primitive concept with only the "Is a" relationship to "Fetal gestational age", and the default context is essentially "present in the patient of record", as @Robert McClure said. But that context can potentially be overridden by how the concept is used in the information model. If that same SNOMED concept is used in a pregnant mother's record in an observation that specifically indicates the fetus as the subject, then it would be interpreted as the gestational age of the fetus. So, back to your original question, looking at only the SNOMED CT code itself is probably not going to be sufficient to differentiate between mom's gestational age at birth when mom was a baby vs. mom's fetus' gestational age when mom was pregnant. There's some further information about this in the SNOMED CT Terminology Services Guide here: (https://confluence.ihtsdotools.org/display/DOCTSG/12.3.21+Normal+Forms+That+Take+Account+of+the+Information+Model), which is written in the context of transforming SNOMED CT expressions (which includes precoordinated concepts) to normal forms.
Michaela Ziegler (Jul 15 2020 at 13:44):
We have created a profile for the time of gestation in weeks and days. We want to combine the value with weeks and days, for example 30 weeks & 5 days. Therefore we have represented the element valueQuantity as a component.
Now we have found that LOINC code 49052-4 (Gestational age in days) may not say exactly what we want the component to represent.
How do you make the specification? Do you convert the weeks and days into a total of days? Or are there other suggestions?
Lloyd McKenzie (Jul 15 2020 at 17:43):
FHIR doesn't really support what you're trying to do. You can express a value in weeks or days, but not "X weeks + Y days". The LOINC code you identified would be total days, not 'remainder' days. Converting a total number of days into a weeks + days thing would typically be a presentation/user-interface function. The exchanged data would always be in days.
Michaela Ziegler (Jul 16 2020 at 06:06):
Thanks @Lloyd McKenzie :+1:
Last updated: Apr 12 2022 at 19:14 UTC
