In Brazil, the National Agency of Sanitary Surveillance (ANVISA) has built their pharma serialization regulation around GS1 standards. They embrace the GS1 Datamatrix and GS1-128, both encoded with GS1 Application Identifiers (AI) and using GS1 Human Readable Interpretation (HRI) (see my previous essay, “The ANVISA Unique Medicine Identifier (IUM) on Drug Packages”, for my thoughts on HRI), the GS1 Global Trade Item Number (GTIN) and the GS1 Serial Shipping Container Code (SSCC) to be specific. But there is one GS1 standard they steer clear of: the GS1 serial number. Why is that?
I’m not sure, but I think I have a pretty good theory. See if you agree.
But first, let’s review how GS1’s serial numbers are defined (see the GS1 General Specifications for more details and specifics). GS1 serial numbers can be any length, up to 20 characters. The allowed character set includes 82 distinct characters, including digits 0-9, the upper and lower case English alphabet, and a set of 20 symbols and punctuation characters. Most applications, trading partners and regulations limit the serial numbers used within their closed domain to something shorter and often limit the allowed character set by written agreement or guidance. But the feature of GS1 serial numbers most pertinent to this discussion is that they must always refer to a specific GS1 GTIN. That is, any reference to a GS1 serial number without including its associated GTIN is, by definition, totally ambiguous and meaningless.
WHY DOES THAT NOT APPEAL TO ANVISA?
ANVISA wants the serial number (held within AI 21 in the barcode) to be unique across all GTINs registered by the registrant, not just the one GTIN referenced along with it on the product. See the difference? GS1 says the serial number must be unique only for that GTIN, but ANVISA says serial numbers used in their supply chain must be unique to the manufacturer (more accurately, the registration holder). That really complicates things for larger manufacturers, inside and outside Brazil, because it means they have to coordinate the serial number assignment across all of their products that are targeted for the Brazilian market. For large multi-national corporations, that could involve dozens—perhaps hundreds—of packaging facilities around the globe and include a wide diversity of products.
I originally thought this was just a simple mis-interpretation of GS1’s General Specifications by ANVISA, but now that I better understand their regulation, it makes better sense. The serial number assigned to the package is not relative to the GTIN, despite the use of AI 21. (Notice that the GTIN is not part of the Unique Medication Identifier, or IUM.) Instead, the serial number (which is part of the IUM) is relative to the holder of the ANVISA Medication Registry Number (AMRN) for the drug (also part of the IUM).
The AMRN, by its definition, will indicate to ANVISA which company and product the serial number is referencing, much like the GTIN would, but the company referenced in the registration might not be the manufacturer. It could be an importer instead (see RDC No- 54, 10 December 2013, Chapter I, Article 2). And if ANVISA regulations allow, multiple importers might register and import the same product with the same manufacturer’s GTIN. To keep straight which company is responsible for each set of imported packages, the serial number is unique across all products registered with ANVISA by the specific registration-holder.
Of course, most multi-national companies import their own product so this will likely be a technical hardship for those companies when assigning serial numbers to their drugs packaged for the Brazil market. But, perhaps a bigger hardship for those companies might be that this technical approach seems to enable parallel import of pharmaceuticals, if ANVISA regulations allow it.
Who knew GS1 serial numbers could be so confining!
If my theory does not hold up, let me know in a comment below.
2 thoughts on “Why Does ANVISA Embrace GS1 Standards, Except The Serial Number?”
The only way I can see this working is if ANVISA issues a unique registration for each potential labeler. While this is not contained in any of the documentation I have seen, the translations are not forthcoming (English) and have had to rely on 3rd parties (not ideal, to say the least). Nothing that I have read implies this is the case; ANVISA only references the manufacturer. Perhaps our Brazilian colleagus can shed some light; my Brazilian team seems as confused than I am…
I can not see your point why only this would enable parallel importing. From my point of view having several importers of the same product would also be possible with a uniqueness per GTIN. As taking parts from the whole production amount and spread it to different ANVISA Medication Registry Number (related to registration holder and the product) would still end up in a uniqueness of serial number per AMRN (for sure with “gaps”(if you will call it like this in a randomized environment in serial numbers which are used by the other parallel importers).
Or differently said: split an amount of Products where you have a bijective assignment to a serialnumber in subsets will end up in a mapping which is still bijective.
Or did I get your point wrong Drik?
Comments are closed.