Question: Where do lab result interpretations go?

Jul 20, 2012

Question:

In an ORU^R01 message, where do you usually put the interpretation of lab result?

Answer

Well, that depends on that kind of interpretation you are talking about. Firstly, Lab results are a combination of data and interpretations, and there is very often no clear division in the way the data should be understood, let alone the report (aside: Doctors and formal modelers continue in the fantasy that lab results are data, not interpretation but this is not the case. Interpretation starts in the lab before any numbers are released, though the main high% tests just go straight out).

So “Interpretation” can be one of the following things:

  • A flag appended to the numerical result (H, L, etc)
  • A comment provided in place of the result
  • An additional comment appended to the result
  • A whole set of pages of content which is interpretation as narrative, that may include some original data

Accordingly, HL7 provides several candidate locations for placing these interpretations:

  • OBX-8: Abnormal Flags : “a table lookup indicating the normalcy status of the result
  • An NTE after the OBX
  • A separate OBX, or set of OBXs

I very much recommend against using the NTE approach - They are used in some areas around the world, but they just don’t carry semantics. So it really comes to, use OBX-8 for flags, and put other interpretations in an OBX of their own. Really, the key to this use is a combination of OBX-3 and OBX-4. OBX-3 is the code that defines what the type of OBX is - it might one of the LOINC codes for comment, such as 8251-1, or a local code. The problem with interpreting lab results mostly comes down to interpreting the oBX-3 codes correctly. OBX-4 may help, in that it can indicate that a particular OBX segment is a child segment of another, but there’s no real consistency between the syntax for using OBX-4, and also quite wqh