OHDSI Home | Forums | Wiki | Github

Any suggestions for 3m CRG codes?

Vocabulary Users
#1

I’ve been digging through the vocabularies trying to find a place to put 3M Clinical Risk Group scores. I do not see anything that I can really use.
I am thinking about putting them in the observation table with:

concept_id = 0
observation_source_value = ‘3M CRG’
value as string = ‘5b’
values_as_number = 56000
unit_concept_id = 8509 – risk?
observation_type_concept_id – ??

Has anyone loaded CRG scores before? We are also going to be coming up with our own risk scoring, etc. Is this the proper way to load the unmapped observation data?

Thanks!

#2

@Richard_Starr:

Good try, but no. It makes no sense to come up with your own rules to somehow squish the data in there, because the standardized tools won’t know this.

The right thing to do is to add CRGs to the Standardized Vocabularies. And then the CRG should go into the observation_source_value, and the concept_id in the observation_concept_id.

Are they available publicly? If not, we can bring them in as a proprietary source, or you do it yourself for your own local vocabulary tables.

Let us know.

#3

Thanks! I’ll try them locally and then send out what I end up with.

For example:

observation_concept_id = 12345     -- Congestive Heart Failure and                      
                                      Diabetes Level - 5  
                                      ( vocabulary_id = 3M CRG )
observation_source_value = 61115   -- actual 3M CRG code
value as string = '6'              -- status
values_as_number = 61115           -- full code
unit_concept_id = 0                -- or risk ?
observation_type_concept_id = 123  -- 3M CRG

I would hope 3m is more interested in the licensing of the software that produces the values than the proprietariness of the definitions, but I will research and inquire through our 3m contact. I am not finding a mapping of the detailed codes publicly available. It is in their documentation, which is behind a subscriber wall.

#4

Almost. :smile:

observation_concept_id = 12345 (the concept ID we will assign to hte 3M code)
observatoin_source_value = 61115 (the actual 3M code)
value_as_string = null (you would put something if the 3M codes were e.g. questions and answers in a survey)
value_as_number = null (there is no number here, as there is no measurement or similar)
unit_concept_id = null (there is no unit here, as there is no measurement or similar)
observation_type_concept_id = Pick one of the following for the source of this information. If nothing fits, we can add.

45905771 Observation Recorded from a Survey
38000276 Problem list from EHR
38000277 Lab observation numeric result
38000278 Lab observation text
38000279 Lab observation concept code result
38000280 Observation recorded from EHR
38000281 Observation recorded from EHR with text result
38000282 Chief complaint
43542355 Referral Record
44786633 HRA Observation Numeric Result
44786634 HRA Observation Text
44814721 Patient reported

#5

Sounds great. There are about 1000 of the codes that will need concept_ids.

For the observation type, these values are generated from 18 months of claims and/or EHR data, so I am not sure if any of these fit. I don’t think either of the HRA do. I would think something like a risk score or analytic result.

Thanks for your help!

t