OHDSI Home | Forums | Wiki | Github

How to implement EDI Code change history

Vocabulary Users
#1

Hello,

I’m looking for help with the idea that enhances EDI capability in OMOP.

Thankfully, EDI is already reviewed by the OHDSI contributors by the publication(Incorporation of Korean Electronic Data Interchange Vocabulary into Observational Medical Outcomes Partnership Vocabulary) .

In order to overcome its limitation described in the paper, we organized drug EDI code history based on our HIS. We think that the application of drug EDI history can contribute to the researches designed using EDI code.

Regarding the update of EDI code history, I think I could use concept_relationship and relationship table for records of its history. however, I’m not pretty sure that it is the optimal way for code history management. Therefore, I’m reaching out for help now.

thank you for your time :slight_smile:
all the best,
ildong lee.

#2

Hi @ildonglee
What do you mean by EDI code history?

If you mean just a usual retrospective research, then the problem is that EDI is not mapped to standard terminology, thus you can’t automatically query the drugs using standard concepts and a hierarchy.
If you want to contriute building the relationships, the vocabulary team is happy to help explaining how to build those.

1 Like
#3

Hello, @Dymshyts. thank you for your attention

we figured out that missing old drug EDI codes makes data loss when researchers building their cohort by using concept set using EDI codes. We’d like to enhance the coverage of drug EDI, We try to map the relationship among the outdated Drug EDI codes and the Current Drug EDI code updated in Athena by adding its expired and valid date information in the concept_relationship table and relation table.

EDI update records are based on our organization’s EDI update records. Hopely it helps researchers who are defining concepts set by using EDI Code in Korea. At the same time, we are also trying to map drug edi code to rxnorm as well.

if you don’t mind could you help us to build relationships among old EDI codes and Current EDI codes?

#4

This is a great comment, and thank you for your efforts to maintain the EDI code. It seems to me that your opinion is to use the relationship table to avoid missing EDI codes from the study. Also, in general, we only use concepts that are standard, and I am wondering if it is correct to argue the opinion of improving the vocabulary to use non-standard EDI code in research.
Is that correct?

In my opinion, if the purpose is to maintenance the history of the EDI code (very important), it seems that you can use the “invalid reasion” column to maintenance it.

In addition, the opinions on mapping EDI code to standard concept are also very good and important. However, it seems necessary to discuss with HIRA, which assigns and manages EDI codes.

1 Like
#5

Appreciate your kind help @Jaehyeong_Cho

Regarding your question, yes it is about using non standard drug EDI code in research. We think that using EDI code facilitates researches based on HIRA Claim data once if it is mapped to standard concept. As a stepping stone, we would like to build relations among this old and new EDI codes.

Certainly, mapping EDI code to standard concept shall be discussed with HIRA. We would like to offer cases that HIRA could consider for their release and governance of EDI. Thank you for your thoughtful consideration.

if you don’t mind i’d like to know a bit more about details to use 'invalid reason column. :slight_smile: thank you for your time!

#6

So you build the links between the old codes and new codes, something like this, where you can see the 'Concept replaced by relationship? That means that the old RxNorm code is repleced by another one.

We have the script, which build RxNorm Extension out of the national drug registries or maps the source drugs to existing RxNorm or RxNorm Extension concepts. We can teach you how to prepare your source data to be ingested by this script.

As a temporary solution (since EDI doesn’t have mappings to standard concepts yet) you can build your cohorts based on source codes.

The invalid_reason indicates whether the concept is active (NULL), replaced by another concept (‘U -
Upgraded’) or deprected without the replacement (‘D - deprecated’).

t