Concept id 4180886 (Phototherapy of skin) has several descendant concepts visible in the Hierarchy displayed in ATHENA, but when I search for this code in concept_ancestor.ancestor_concept_id, 0 rows are returned. Any idea what’s up?
Thanks!
Actually, I seem to have figured out the root cause - all of the descendant concepts are either ICD10PCS or OPCS4 vocabularies, which don’t appear in the concept table at all in my database. Maybe an issue with our implementation? I’ll follow up with an engineer here - but would be helpful if someone could confirm this is just my internal issue vs a global one 
thanks!
Hi Katy! Welcome to the wonderful world of OMOP vocabularies. 
What kind of data are you working with? It looks like the concept is a lot less frequently coded in primary care data. In my data, I see the following counts for Concept ID = 4180886 (Phototherapy of skin):
- In one large claims data set: 216,681 pts
- In a pre-adjudicated (open) claims data set: 111,917 pts
- In hospital EHR data: 58,298 pts
- In a primary care EHR: 7,275 pts
- In another primary care site: 546 pts
I see a lot more coverage of the parent concepts in my primary care EHR data:
If you are investigating your ETL process, @MNairn or @areckard may have some queries they could share to help you perform diagnostics on your ETL.
Hope this helps!
Best,
Kristin
Hello Kristin,
Thanks for your reply! I’m actually referring just to the OMOP CDM tables, not the clinical data I’m working with (EHR data).
Those descendant concepts with ICD10PCS and OPCS4 vocabulary do not appear in my concept table at all, and when I search for those vocabulary_ids in the concept table, I get 0 rows. Given that they do show up for you and are in Athena, I think it’s an issue with our implementation of the CDM. We must not have downloaded these specific vocabularies (still following up with our engineers to confirm this, but I suspect it’s the case).
Anyhow, looks like (as you point out) these child concepts are quite rare and also none of them have any other non-standard concepts that map to them. So I don’t think I’m missing out on anything super important for the moment 
Thanks again!
Katy
Hi again!
So I did get confirmation from engineering that we did not download those 2 vocabularies. They only downloaded the vocabularies that were pre-checked-off on the download page.
I’m curious - why were those specific vocabularies pre-checked and the others left off? I assume it’s because the pre-checked ones are the most common and the rest are more “niche” (as it seems the ones in my example are)? Are there other unchecked vocabularies that are used in standard concepts?
Our engineering team is looking into adding more of the vocabularies but I’d like to understand how high-priority this task is for them. I don’t think we need every single vocabulary as we’re in the early stages of our implementation, but I do want to make sure we’re not missing anything important.
Thank you!
Katy
Yeah. That’s the idea. It used to be that “standard vocabularies” are prechecked, and the others are not. However, this system doesn’t really work, since vocabularies are not standard, but concepts. We need a better UI to guide the users what to click.
Look. You should routinely download the vocabularies, at least during each refresh. Because things change in there, and if you are on an old version chances are that you could run into problems with Network Studies. If your IT department doesn’t want to deal with the details tell them to click on everything that’s free. Athena will remember the choice. It’s not that much data, and you never have to worry about this again.
Thanks for your reply! Makes sense about the checkbox system; agree that the UI could be clarified a bit.
We went ahead and downloaded all of the additional free ones and will continue to do so moving forward