Is there a drug classification based on the Conditions it is used to treat- So something like an ICD code or diagnosis/condition and tthe list of recommended drugs to treat this dx/condition/ICD code.
@Christian_Reich @aostropolets
@PriyaDesai , there is no complete open-source vocabulary mapping between drugs and their indications. But at the OHDSI Symposium, I encourage you to take a look at the work by Yilu Fang, because that’s exactly the problem she’s aiming to solve by creating a novel workflow using GPT to parse FDA structured product labels and create an indication ontology.
1 Like
@Patrick_Ryan I see she is part of Chunhua’s lab- can you point me to any relevant papers ?
thanks
No papers yet, @PriyaDesai. The Symposium is the first time this is coming out. Will be very interesting what Yilu has come up with.
I also would be cautious and tune my expectation. Yilu has a special use case. But I would not expect that you can have a comprehensive mapping from each drug to all of the diseases, even if we had resources to create such a thing. it doesn’t work. For example, doctors prescribe sleeping pills for all sorts of situations, not just insomnia, but also anxiety, pain, pre-operative, etc.
Dear Christian, I admire your knowledge - and I apologize for stupid question. Anyway 
we would like to build the database based on electronic prescription records here in Czech Republic - so we have in cloud all Rx prescriptions linked to each patient - we would like to help physicians to see the probable diseases each patient suffer. It would require link drug code or ATC 3-5 code to simplified indication and/or disease - the best would be ICD 11 which would be widely implemented.
I understand you recent answer to @PriyaDesai - about the “sleeping pills” - but it is OK - we can list all the indications/diseases with label: approved/off label.
Of course we can get the indications from SmPCs or PILs but usually it is “too long” or “not standardised”. So we wanted to get physicians just briefly informed - this patient has/had “prostate cancer”, “hypertension”, “diabetes type II” “allergy” etc.
Once again thanks for any help
Petr