Wikidata:Property proposal/National Drug Code

National Drug Code (product) edit

Originally proposed at Wikidata:Property proposal/Natural science

DescriptionA code issues by FDA for every drug product on the US market. Includes a labeler code, product code and package code, unique for every drug product.
RepresentsNational Drug Code (Q6972354)
Data typeExternal identifier
DomainMedicine, chemistry
Allowed values[0-9]{4,5}-[0-9]{3,4}
Examplevemurafenib (Q423111) → 50242-090
Sourcehttps://www.nlm.nih.gov/research/umls/rxnorm/docs/rxnormfiles.html
Planned usethis code will be added to all approved drugs and formulations, the latter will be created as needed.
Formatter URLhttps://www.hipaaspace.com/Medical_Billing/Coding/National.Drug.Codes/$1
Motivation

The National Drug Code (NDC), issued by the FDA, identifies every drug approved by the FDA for marketing in the US. Every producer of a drug is uniquely identified, as is the formulation of a certain drug. This property will help to clearly identify drug formulations, combination drugs, etc in Wikidata. Dosage/formulation information should be considered as very important information, especially for drug indications, as many drugs are given to patients at different dosages for different conditions. Overall, this property will help to improve information on drugs in Wikdiata/Wikipedia. Sebotic (talk) 18:34, 8 February 2017 (UTC)[reply]

  Notified participants of WikiProject Chemistry

  Notified participants of WikiProject Medicine

Discussion
  •   Question The code given in the example stands for what? Pure substance? Formulation? Product? Or maybe even specific package type? Correct me if I'm wrong: if there are 10 manufacturers, each of them produce XXX-drug in two different strengths and two different packages, there would be 400 NDCs? I searched for ascorbic acid and I'm a bit shocked, because you want to create items for formulations of drugs ("this code will be added to all approved drugs and formulations, the latter will be created as needed") and creating a WD item for "75854-303-30 | prenate essential | - ascorbic acid, cholecalciferol, .alpha.-tocopherol, dl-, pyridoxine hydrochloride, folic acid, cyanocobalamin, biotin, calcium carbonate, ferrous fumarate, potassium iodide, magnesium oxide, doconexent and icosapent | TABLET, COATED | 90 mg/1, 220 [iU]/1, 10 [iU]/1, 26 mg/1, 1 mg/1, 13 ug/1, 280 ug/1, 145 mg/1, 29 mg/1, 150 ug/1, 50 mg/1, 300 mg/1, 40 mg/1" and for other 300 formulations of ascorbic acid is a pure nonsense for me. Wostr (talk) 19:03, 8 February 2017 (UTC)[reply]
We do not need to include the packing information (e.g. if in one pack there are 20 or 40 pills), so the digits after the last hyphen could be left away. But the dosage per e.g. pill or vial, encoded in the second set of digits is important [1]. Compounds and dosing can go into WD statements, marketing names goes into the label, the data is all structured, so that's not a problem. Sebotic (talk) 10:29, 9 February 2017 (UTC)[reply]
  • So   Oppose. At least until I see a detailed plan on how this property will be used, which parts of the NDC will be imported, what will be the relations between active ingredients (pure compounds), formulations and their strengths/packages, and see some estimates regarding possible number of new entries. Doing something like that in a hurry, without a detailed, discussed plan seems to be unreasonable, especially seeing what kind of data is included in this database and knowing that we still have trouble with managing existing data. Wostr (talk) 12:00, 9 February 2017 (UTC)[reply]
OK, seems like I did not lay out the picture clearly enough (and you seem to have missed my responses to users questions below). First, I adjusted the proposal for product only, not for packaging. So the structured data are available from this FDA website (product.txt). This file has all data which is needed to establish Wikidata drug formulation items, including brand name, active compound, dosage. For every formulation (NOT package size) a new WD item will be created (113K in total, that's the formulation count in this file). Each item will have a link to the active compound(s), salts, filling/carrier substances (if available) using has part(s) (P527) and quantity (P1114) as a qualifier to indicate the quantity. Furthermore, route of administration and indication will be added for each compound. Here, it becomes clear why we need formulation WD items: Many drugs are used in different formuations for different diseases and routes of administration. This is very important, as it affects the side effect profile of a certain drug.
Regarding your statement on issues in Wikidata: Some parts of WD have issues, but these are not connected to what should happen here. In case you are referring to issues with chemical data: I did a very detailed analysis of these issues and I would really appreciate if you could contribute to resolving these. Sebotic (talk) 22:33, 9 February 2017 (UTC)[reply]
@Wostr: Are you okay with the added explanation by Sebotic? ChristianKl (talk) 15:49, 16 February 2017 (UTC)[reply]
@ChristianKl:. As a chemist, I don't see any reason why we should have elements for formulations (ultimately, an infinite number of such elements; 113k for only one country at this moment; and unusable for any Wikipedia infoboxes), but I won't have any objections as long as this mess is kept away from the pure compounds elements. Wostr (talk) 16:11, 16 February 2017 (UTC) PS quantity (P1114) to indicate the concentration of a substance in a mixture? Another nonsense... Wostr (talk) 16:11, 16 February 2017 (UTC)[reply]
  •   Oppose Not convinced about the general interest of this dataset in WD. I don't think WD should become the unique database of everything. What would be interesting is the part of the code which represents the active substance but the formulatio or the producer are not interesting. Snipre (talk) 20:01, 8 February 2017 (UTC)[reply]
Why do you oppose Wikidata becoming the unique database for everything? Isn't "free access to the sum of all human knowledge" the mission? But even when it comes to the information about drug producers, having a way to list all drugs that a certain company produces is useful. The ability to add a list to a Wikipedia page for a drug that lists all the companies that produce the drug can also be useful. ChristianKl (talk) 22:15, 8 February 2017 (UTC)[reply]
As I discussed below, the best way is to create formulation items, which contain the info which compounds are in a drug at what dosing. Compound information can be added via 'has part'. The only unique id for formulations is a NDC. Furthermore, NDCs are important, because every patient sees them on the packaging every day. From a encyclopedic perspective, this is more important than many drug identifiers we have right now. Sebotic (talk) 10:29, 9 February 2017 (UTC)[reply]
As stated below, this was just a quick example, drug formulations should be their own items, having marketing names, etc. Sebotic (talk) 10:30, 9 February 2017 (UTC)[reply]
  •   Support Dhx1 (talk) 09:25, 9 February 2017 (UTC)[reply]
  •   Support Good idea but not as a property of the compound. I think a better way would be to create a new page for a specific formulation, add it's regulatory numbers (multiple if it's shared between countries) and other information such as URL to the monograph, name, the (active) ingredients and their concentrations, the route of administration, dosage form, appearance, scheduling, etc. Basically make an open database similar to Health Canada or LexiComp. I can think of a few useful applications for this: pill identification (Lexi has this in their app), finding all formulations containing a specific drug and it's quantity (see Wikidata:Property_proposal/strength), comparing regulatory status of formulations around the world, want to find the names of medications with a specific compound (especially in a foreign place you can't speak the language), etc. Most countries have their own regulatory number. For example DIN in Canada for which a database is maintained by Health Canada (eg DIN 02272849 for a formulation of Ibuprofen). Ecuador, Chile, and Bolivia all have unique numbering systems too (Reg. Sanit. No. and Registro I.S.P. No.). I believe you cannot have too much information in WD, so long as it's well organized. Devon Fyson (talk) 09:30, 9 February 2017 (UTC)[reply]
Yes, I only put it on a chemical compound for a quick example, it makes sense to create a new item for a formulation and also to link the prescription information to the formulation, not to the chemical compound. Compound and formulation can be linked via 'has part' Sebotic (talk) 10:29, 9 February 2017 (UTC)[reply]
I added it as an instance of pharmaceutical code (Q7180748) and changed the description slightly. Need there be this pharmaceutical code proposal for every country? For example I now propose the Canadian DIN be added. Maybe there should a generic proposal for adding pharmaceutical codes as they are needed. Devon Fyson (talk) 04:43, 24 May 2017 (UTC)[reply]