OHDSI Home | Forums | Wiki | Github

Claim Detail Position in Type Concepts

(Nadya Zvyagina) #1

Hi all! I would like to bring up type_concept_id topic again. What’s the real purpose of those
Carrier claim detail - 1st position, Carrier claim detail - 2nd position, etc.
Who really cares about these? Is there anyone who uses these?
@Dymshyts, @aostropolets, @Christian_Reich, @Gowtham_Rao

(Gowtham Rao) #2

Thank you for bringing this up @nzvyagina . This needs someone to champion, build consensus and standardize.

(Nadya Zvyagina) #3

I can prepare an initial proposal next week.

(Jen Duryea) #4

The “carrier claim” type concept ids are supposed to represent information coming from physician claims (a.k.a. claims billed on HCFA 1500 forms, a.k.a. Medicare part B physician services). The “position” part of these type concept ids are to represent the position on the claim where the diagnosis or procedure code comes from. We use these type concept ids for storing Medicare claims information.

(Seng Chan You) #5

@nzvyagina We’re using ‘Carrier claim detail - 1st position, Carrier claim detail - 2nd position, Carrier claim detail - 3rd position’, and etc.
Korean national claim database system stores upto 10th position of the claim for the diagnosis. When defining the phenotype, we usually care whether the condition is in the 1st or 2nd position.

But there is redundancy or ambiguity of condition type concept ids. for example,

45756835 OMOP generated Carrier claim header - 1st position


45756843 OMOP generated Carrier claim detail - 1st position


44786627 OMOP generated Primary Condition

Usually first-position diagnosis is primary condition. So we do need to standardize how to store order of the conditions and priority of the condition.

(Nadya Zvyagina) #6

Hi Jennifer! Thank you for your answer. Yes, I understand what they are used for during the conversion. But I’m mostly interested in if anyone uses these type concepts in further data analysis? Are there any studies involving the exact position number?

(Mark Danese) #7

Yes. Many algorithms require a person to have a diagnosis code in the first or primary position for the inpatient setting. For example, myocardial infarction is typically limited to inpatient diagnoses (inpatient facility usually). There isn’t as much need in the outpatient facility or provider setting. The information is useful for provenance and traceability from the source data set.

(Seng Chan You) #8

@Mark_Danese Yes, and that is why we need to standardize the condition_type_concept_ids between ‘position’(1st, 2nd,…) and ‘priority’ (primary, secondary) among OHDSI network.

(Gowtham Rao) #9

In general, (in USA claims) there is a need to distinguish between primary and secondary diagnoses. Among secondary diagnoses, there is no much value in knowing if the the secondary diagnoses is in 3rd position or 20th position. I.e., There is not significant difference in the interpretation of the different position numbers within secondary diagnosis.