OHDSI Home | Forums | Wiki | Github

Question about icd9cm and "Maps To" in v5

Vocabulary Users
#1

Hi,
I noticed something when looking through concepts using the mapping code to snomed. When using the "relationship_id = ‘Maps to’ " in the concept_relationship, I mapped an ICD9CM transplant procedure to a generic “Past history of procedure” concept. For example:

select * from omop_v5.concept src
left join omop_v5.concept_relationship cr on src.concept_id = cr.concept_id_1 and cr.relationship_id = 'Maps to'  and cr.invalid_reason is null
left join omop_v5.concept tar on cr.concept_id_2 = tar.concept_id  and tar.standard_concept = 'S' and 
where src.concept_code = 'V42.82';

Returns concept_id 4215685 - Past History of procedure ( SNOMED 416940007 )

But if I use the “Maps to value” instead of the “Maps to”, I get the procedure concept I am looking for:

select * from omop_v5.concept src
	left join omop_v5.concept_relationship cr on src.concept_id = cr.concept_id_1 and cr.relationship_id = 'Maps to value'  and cr.invalid_reason is null
	left join omop_v5.concept tar on cr.concept_id_2 = tar.concept_id and tar.standard_concept = 'S'
where src.concept_code = 'V42.82';

I am not sure where my mistake is.

Thanks,
Richard

#2

@Richard_Starr:

Nowhere. It’s correct that way. “History of”, “Family history of”, “Need for vaccination” etc. - all those that don’t denote a condition or other event that happened at the time of recording are organized that way: The mapping is to the fact that it is in the past, and the actual condition or event is mapped through “Maps to value” and recorded in the value_as_concept_id field of the OBSERVATION table.

The idea is you can actually send a query “Find me all patient with a history of transplantations”, and it will get back to your the stem cell transplantation, but also kidney transplantations, liver transplantations etc. It is particularly useful for history of malignant disease.

#3

Thanks @Christian_Reich.

I got busy and it took me awhile to circle back around to this question.

Maybe I am just misunderstanding the V* ICD9 codes. For example, mapping with “maps to”, it seems some of the V49.6* map to history ( snomed 41* ) and other map to amputations ( snomed 299* ). Looking at the ICD9 descriptions, I don’t really see the difference between the code definitions.
BTW, V49.64 is the only one that “Maps to value”. V49.63 does not.

'V49.61','44837833','Thumb amputation status','','4198962','Amputated thumb','S','SNOMED','299217009'
'V49.62','44821564','Other finger(s) amputation status','','4184877','Amputated finger','S','SNOMED','299219007'
'V49.63','44830803','Hand amputation status','','4215685','Past history of procedure','S','SNOMED','416940007'
'V49.64','44835482','Wrist amputation status','','4214956','History of clinical finding in subj','S','SNOMED','417662000'
'V49.65','44833111','Below elbow amputation status','','4198961','Amputated below elbow','S','SNOMED','299215001'
'V49.66','44830804','Above elbow amputation status','','4198960','Amputated above elbow','S','SNOMED','299213008'
'V49.67','44825015','Shoulder amputation status','','4184875','Amputated at shoulder','S','SNOMED','299212003'

I hope this makes sense.

Thanks,
Richard

#4

@Richard_Starr:

Yes, they need some polishing. Will do. Thanks for the digging.

t