How would you map DRUG_NAME/PLACEBO (CLINICAL TRIAL)?

Implementers
1

There are quite some of these records in our database

GLIPIZIDE/PLACEBO (CLINICAL TRIAL),2.5MG

I guess it mostly means a double blinded study that we have no idea the patient was given the drug or some placebo, because that record will never be updated or undlinded

How would you map this? Do you map it to GLIPIZIDE or do you just map it to 0?

Depending on the drug itself, the number could be quite big - not sure if it would be significant.

2

Hi @lychenus13
What an interesting question indeed! So it’s unknown if an actual drug or placebo was administered. In this case, I would create a 2bil custom concept for this drug just to be able to perform analysis on concept_id rather than drug_source_value.

3

@lychenus13 and the Clinical Trial folks:

Not sure you want to use the OMOP CDM for the primary data analysis of clinical trials. There is CDISC, and that has all the mechanisms for randomization, visits, missing values, edit checks and monitoring, etc.

OMOP is for secondary use. What can you do with the data once the trial is over. At that point, the exposure should have been unblinded. From an OMOP perspective it makes no sense to have a field filled with “don’t know what it is.” That, by definition, is concept_id=0.

4

It is not that we are using OMOP CDM for the primary data analysis of clinical trials, but the data we have for very general OHDSI studies include records of clinical trials. And in our system, the record is never updated. All we know is that some patient had gone through that procedure and something is taken in.

We are not sure if that drug would cause some significant factor in any study.

Well that said if the general opinion is to map it to zero then lets do it with zero for the moment then. Thanks

5

Use placebo. Legitimate RxNorm ingredient.