search

TheJacksonLaboratory / loinc2hpoAnnotation

loinc2hpo Annotation Data
https://github.com/monarch-initiative/loinc2hpo
Other
24 stars 11 forks source link

Proposed new annotation format #137

Closed pnrobinson closed 2 years ago

pnrobinson commented 2 years ago

Please see README and loinc2hpo-annotations.tsv. I think this format is a little easier to understand. Also, POS and NEG are typically misunderstood and I think it would be clearer to replace them with PRESENT and ABSENT. Thoughts? @iimpulse

callahantiff commented 2 years ago

@pnrobinson - Thanks for doing all of this updating!

The updated format for loinc2hpo-annotations.tsv looks great to me. I had only one small question (for my understanding), what types of concepts will go in the supplementalTermId column?

The request to use PRESENCE/ABSENCE seems completely reasonable to me.

kingmanzhang commented 2 years ago

The new table looks awesome!

POS and NEG were chosen because they were standard concepts in FHIR (along with L, H, N, A). Back then we thought relying on some existing vocabulary rather than inventing new ones will help adoption.

https://www.hl7.org/fhir/v3/ObservationInterpretation/cs.html

If you look at the FHIR vocabulary, it also has similar concepts to PRESENCE/ABSENCE, ND and DET, but they have more strict definition, i.e. below detection limit is NEGATIVE, but is not necessarily equal to ABSENCE. See below.

p.s. I just notice ND/DET are similar to NEG/POS but NOT to PRESENCE/ABSENCE.

Level Code Display Definition
2 NEG Negative An absence finding of the specified component / analyte, organism or clinical sign based on the established threshold of the performed test or procedure.[Note: Negative does not necessarily imply the complete absence of the specified item.]
2 POS Positive A presence finding of the specified component / analyte, organism or clinical sign based on the established threshold of the performed test or procedure.
3 ND Not detected The presence of the specified component / analyte, organism or clinical sign could not be determined within the limit of detection of the performed test or procedure.
3 DET Detected The measurement of the specified component / analyte, organism or clinical sign above the limit of detection of the performed test or procedure.

I do not really get what you meant by saying "POS and NEG are typically misunderstood". I probably would prefer to stick to NEG and POS.

kingmanzhang commented 2 years ago

Looking at the above table, ND/DET could replace NEG/POS?

pnrobinson commented 2 years ago

I think NEG and POS are more general terms. I will finalize the conversion to the new format. The loinc2hpo library now expects that new format as well and runs through without errors.