I have just started learning OMOP. I am super excited to learn it and attend the upcoming symposium.
I came across concept_id 45576364 that links to two standard concepts, both are valid (without valid_end_date) and have same valid_start_date. Not sure which one should I use.
This is the query I am using:
select c_scr., '’ as Divider, cr., '**’ as Divider, c_tar.
from concept c_scr
join concept_relationship cr on c_scr.concept_id = cr.concept_id_1 and cr.relationship_id = ‘Maps to’
join concept c_tar on cr.concept_id_2 = c_tar.concept_id and c_tar.standard_concept = ‘S’
where c_scr.vocabulary_id in (‘ICD10’,‘ICD10CM’)
and c_scr.concept_id = 45576364
I am sorry, the query did not paste properly in the original post
This is the correct one:
select c_scr., cr., c_tar.*
from Concept c_scr
join concept_relationship cr on c_scr.concept_id = cr.concept_id_1 and cr.relationship_id = ‘Maps to’
join concept c_tar on cr.concept_id_2 = c_tar.concept_id and c_tar.standard_concept = ‘S’
where c_scr.vocabulary_id in (‘ICD10’,‘ICD10CM’)
and c_scr.concept_id = 45576364
Yes, you would map them to both, which means you will end up with two OMOP condition_occurrence records for the 1 source record. This can happen because ICD codes are not always describing just one diagnosis. Sometimes it combines multiple diagnosis into one.
For this ICD term, it is both Malignant lymphoma of lymph nodes of multiple sites and Non-Hodgkin’s lymphoma.
Yes, this is a complex ICD10: Non-Hodgkin lymphoma of multiple sites of lymph nodes. There is no exact equivalent in SNOMED, and hence it got mapped to the first half and the second half: Nothing else we can do today, but we are planning to introduce those combo-concepts as SNOMED Extensions and then line them up correctly into the existing SNOMED hierarchy. This will fix this problem. But we are now talking to SNOMED about the pros and cons of doing something like that (vs having them do it).
Till then, please map to both Concepts and write them both into the CONDITION_OCCURRENCE table.