Stream: implementers
Topic: v2 OBX segments in FHIR DiagnosticReport
Gaston Fiore (Apr 28 2016 at 20:58):
I'm trying to represent to following v2 messages in FHIR using DiagnosticReport.
For a positive finding of the HLA-B*57:01 allele:
OBX|1|..|50956-2^HLA-B*57:01^LN ||10828004^Positive^SCT~LA6576-8^Positive^LN-ANS|…
And, for a negative finding of the HLA-B*57:01 allele:
OBX|1|..|50956-2^HLA-B*57:01^LN ||260385009^Negative^SCT~LA6577-6^Negative^LN-ANS|…
In v2, 50956-2 would populate the OBX-3 field of the OBX segment. In DiagnosticReport, is 50956-2 the name/code for the diagnostic report, or just an "Observation" within "contained" (like the Positive/Negative code is)? If it is another "Observation" within "contained", what LOINC code would be the appropriate one for the name/code of the DiagnosticReport?
Lloyd McKenzie (Apr 28 2016 at 21:39):
Typically OBXs would map to Observations. Also check out the Genetic Observation profile. (Click on Profiles from the Observation resource)
Eric Haas (Apr 28 2016 at 22:28):
In general the OBX-3 maps to Observation.code and the OBR-4 maps to DiagnosticReport.code. Although many implementers seem to skip the DiagnosticReport altogether and put the code in OBR-4 in Observation.code and If OBR-4 = OBX-3 and only a single OBX you could map to a single Observation. Otherwise could use .Observation.related to for panel components. This second approach is lossy since only some of the stuff in OBR can be mapped to Observation.
Gaston Fiore (Apr 29 2016 at 15:26):
Thank you very much for the explanation @Lloyd McKenzie and @Eric Haas ! I'm assuming OBX-5 also maps to Observation.code, correct? I definitely want a DiagnosticReport in this case since I'm trying to represent a genetic test result report from a lab. I'll point to two Observations from the DiagnosticReport corresponding to the OBX-3 and OBX-5 fields. Do you see something wrong with this representation?
Grahame Grieve (Apr 29 2016 at 16:15):
OBX-5 maps to Observation.value
Grahame Grieve (Apr 29 2016 at 16:15):
have you looked at
Grahame Grieve (Apr 29 2016 at 16:16):
http://hl7.org/fhir/observation-mappings.html#v2
Gaston Fiore (Apr 29 2016 at 16:27):
That makes more sense to me, thanks @Grahame Grieve . I was going to take a look at how to handle LOINC answer codes because I wasn't understanding how they'd link to the LOINC code they refer to. The link you suggest plus reading some more about codes so I know how to include the SNOMED-CT codes for Positive and Negative as well (in addition to the LA codes) should solve my problem. Thanks!
Eric Haas (May 01 2016 at 21:13):
@Gaston Fiore take a look at the Observation Examples several use coded result value like Snomed for example this one
Eric Haas (May 01 2016 at 21:14):
note I'm pointing to DSTU2
Last updated: Apr 12 2022 at 19:14 UTC
