OHDSI Home | Forums | Wiki | Github

ICD10: A41.01 mapping to SNOMED

Vocabulary Users
#1

We are working on leveraging the OHDSI ICD10 CM to SNOMED mappings.
After downloading the files from Athena and building local tables, one of our terminologists was doing some QA to validate that what I was pulling made sense.
She noticed the following mapping, we confirmed that this is what shows in Athena as well. She is wondering if this is correct?

ICD10: A41.01 Sepsis due to Methicillin susceptible Staphylococcus aureus
mapped to
SNOMED: 448812000 Sepsis due to methicillin resistant Staphylococcus aureus

Thanks,
Lisa

#2

@lms, thanks for reporting!

The mapping of A41.01 is not correct for sure and has to be changed. I suggest using double “Maps to” for the case because SNOMED does not have an equivalent concept. Instead of existing mapping and to prevent data loss, we may use the following combination:
448417001 Sepsis due to Staphylococcus aureus
AND
442073005 Infection by methicillin sensitive Staphylococcus aureus

Anyway, the vocabulary team will fix it as soon as possible.

#3

I recently had to advise someone on which map to take: UMLS or OMOP. I wish the methodology of the OMOP mappings would be better documented. The how-sausage-is-made repo at https://github.com/OHDSI/Vocabulary-v5.0 provides the concepts. But we need how-sausage-is-made repo2 for how the mappings are made. (the sauerkraut with the sausage). I think it is NLP-assisted and later curated. There is no formal validation study looking at random x% sample of the mappings. Can your team @lms share your findings?

#4

+1 @Vojtech_Huser
I’d like to know the methodology of mappings.

#5

@AsiyahFDA, @Vojtech_Huser:

We hear you loud and clear. That needs to be done.

1 Like
#6

I love the sausage making analogy… is there any material available that illustrates that - like a ppt deck or something that i could use/incorporate into my material to try to explain to external stakeholders what’s happening ? thx !

1 Like
t