search

FDA / openfda

openFDA is a research project to provide open APIs, raw data downloads, documentation and examples, and a developer community for an important collection of FDA public datasets.
https://open.fda.gov
Creative Commons Zero v1.0 Universal
567 stars 131 forks source link

Definition of package descriptions #181

Closed beckyconning closed 2 years ago

beckyconning commented 2 years ago

Summary

Package descriptions that contain * don't seem to make sense.

Two examples which seem at odds with each other

One

2 KIT in 1 CARTON (50474-700-61) > 1 KIT in 1 KIT * 1 mL in 1 VIAL, GLASS * 1 mL in 1 VIAL, GLASS

The above seems to say that the package contains two kits and that each kit contains two 1ml vials (note that this doesn't say whether these vials contain the same or different drugs or strengths).

so we can break this up into

2 KIT in 1 CARTON (50474-700-61) > 1 KIT in 1 KIT > 1 mL in 1 VIAL, GLASS

2 1 1 ml

and

2 KIT in 1 CARTON (50474-700-61) > 1 KIT in 1 KIT > 1 mL in 1 VIAL, GLASS

2 1 1 = 2 ml

This seems correct

Two

24 BLISTER PACK in 1 CARTON (41226-627-24) > 1 KIT in 1 BLISTER PACK * 2 BLISTER PACK in 1 CARTON > 4 CAPSULE, LIQUID FILLED in 1 BLISTER PACK * 16 CAPSULE, LIQUID FILLED in 1 BLISTER PACK

However the above seems to say that there are twenty four kits (24 * 1) and that each kit contains a two blister packs of four capsules and one blister pack of sixteen capsules.

So we break that up like

24 BLISTER PACK in 1 CARTON (41226-627-24) > 1 KIT in 1 BLISTER PACK > 2 BLISTER PACK in 1 CARTON > 4 CAPSULE, LIQUID FILLED in 1 BLISTER PACK

24 1 2 * 4 = 192 capsules

and

24 BLISTER PACK in 1 CARTON (41226-627-24) > 1 KIT in 1 BLISTER PACK > 16 CAPSULE, LIQUID FILLED in 1 BLISTER PACK

24 1 16 = 384 capsules

However i'm pretty sure that in reality this kit only contains 8 of one type of capsule and 16 of the other. So the initial 24 refers to the total number of capsules (not sure why thats relevant info as they are different drugs with different strengths).

Conclusion and next steps

So how are we meant to make sense of this? Perhaps we could try ignoring everything in the first element (as split by * apart from the number of kits?

Is there some documentation somewhere? I can't seem to find it. It seems unfortunate as the ones that don't contain * tend to be quite clear and consistent even if they contain >.

beckyconning commented 2 years ago 
24 BLISTER PACK in 1 CARTON (41226-627-24) > 1 KIT in 1 BLISTER PACK * 2 BLISTER PACK in 1 CARTON > 4 CAPSULE, LIQUID FILLED in 1 BLISTER PACK * 16 CAPSULE, LIQUID FILLED in 1 BLISTER PACK

The above description and those descriptions which are similar are certainly erroneous and we will have to live with it being wrong rather than trying to special case for it. Ignoring the split on ">" elements prior to the first element that contains "KIT" would fix the above but we can't use this methodology due to correct descriptions such as below.

10 CARTON in 1 CARTON (0944-5001-10)  > 1 KIT in 1 CARTON (0944-5001-01)  *  1 mL in 1 VIAL, GLASS (0944-5011-01)  *  1 mL in 1 SYRINGE, GLASS (0944-0011-01)
beckyconning commented 2 years ago

As an aside there are many kit package descriptions (i.e. array_length(split(package_description,"*")) > 1) which don't include NDCs on some or all of their kit components. This is quite useless as how can we tell what these quantities are referring to?

Ok this kit contains 8 capsules and 16 capsules but which drugs are these?

I think it would be valuable for the NDC to have a clearer and more descriptive standard for kit package descriptions (i.e. array_length(split(package_description,"*")) > 1) and a stronger validation process for such applications.

dkrylovsb commented 2 years ago

Unfortunately, the openFDA Team does not maintain or manage the National Drug Code Directory -- we merely process the NDC downloadable files and expose the data via an API "as is". I think you have valid concerns, but this Github repo wouldn't be the best place to report them simply because it is not actively monitored by the FDA CDER team (i.e. the team that actually manages the information in NDC).

I would recommend emailing your concerns to this address -- edrls@fda.hhs.gov -- which can also be found on the https://www.fda.gov/drugs/electronic-drug-registration-and-listing-system-edrls/points-contact-drug-registration-and-listing page.

Sorry I could not be more helpful.

beckyconning commented 2 years ago

@dkrylovsb Thank you, this is very helpful.

As a closing remark to help anyone else looking into this here is an explanation of my best guess of the meaning of such package descriptions.

When split by * the first element represents the kit quantity. Subsequent elements are things contained within the kit.

As such

2 KIT in 1 CARTON (50474-700-61) > 1 KIT in 1 KIT 1 mL in 1 VIAL, GLASS 1 mL in 1 VIAL, GLASS

The above seems to say that the package contains two kits and that each kit contains two 1ml vials (note that this doesn't say whether these vials contain the same or different drugs or strengths).

so we can break this up into

2 KIT in 1 CARTON (50474-700-61) > 1 KIT in 1 KIT > 1 mL in 1 VIAL, GLASS

total = 2 1 1 ml = 2ml

and

2 KIT in 1 CARTON (50474-700-61) > 1 KIT in 1 KIT > 1 mL in 1 VIAL, GLASS

total = 2 1 1 = 2 ml

See the above comments for other gotchas and concerns.