DSCSA Uniqueness: SNI vs SGTIN
The FDA and members of the US pharma supply chain have started to trip over some problems that stem from the fact that the National Drug Code (NDC)—the regulatory identifier for drugs in the US—and GS1 GTINs—the practical identifier companies have been using on drug packaging for decades—are similar, but are not the same. In short, the regulatory identifier is not the same as the practical identifier. The GTIN is the “practical” identifier because it is the only one that can be encoded directly into barcodes. And that’s pretty important in today’s modern supply chain.
It’s been this way since the early 1970s when the FDA first invented the NDC, and the Uniform Grocery Code Council (UGCC) defined the Universal Product Code (UPC)—the identifier that eventually evolved into the Global Trade Item Number (GTIN) many years later (see “Anatomy Of The National Drug Code”).
The GTIN and the NDC have common origins, and the needs they fulfill partially overlapped because, like today, over-the-counter (OTC) drugs were sold in grocery stores. So the UGCC decided to reserve ‘3’ in the first digit of the UPC to indicate “drugs”, and the next 10 digits would be the NDC of the drug. The last digit was (and still is) a check digit. This ‘shoehorning’ of the NDC into a UPC worked well throughout globalization and normal technology evolution, eventually leading to the NDC being ‘shoehorned’ into a GTIN structure. Not bad for something very modern created just for the US in the early ‘70s. I wonder if anyone at that time envisioned the problems this would eventually cause. (They probably figured it wouldn't happen until after they retired, and they were right!)
It appears that the FDA had nothing to do with the decision made by the UGCC. Barcodes were not required by the FDA on drug packaging until 2006—three decades later—although by that time, all OTC drug packages and most prescription drug packages already had linear barcodes on them voluntarily. These were mostly UPC barcodes encoding the NDC exactly as originally envisioned by the UGCC. In 2006 FDA finalized what I like to call ‘the linear barcode rule’. That rule required either a UCC/EAN (the former name for GS1) or HIBCC linear barcode encoding the NDC on most drug packages. On their own, the industry chose to use GS1 standards over HIBCC for drugs (see “WAR: GS1 Vs. HIBCC” and “Product Identification And National Registration Codes”).
AND EVERYTHING WAS FINE
And everything was fine…until November 27, 2013. That’s when the Drug Supply Chain Security Act (DSCSA) was enacted.
Well, to be more precise, everything was fine until November 27, 2017, when DSCSA required drug manufacturers to apply a data matrix barcode encoding the DSCSA product identifier, plus human readable text, on all non-exempt drug packages and homogeneous cases (see “The DSCSA Product Identifier On Drug Packages” and “Is Your Drug Exempt From The Federal Drug Supply Chain Security Act?”).
Well, OK, to be even more precise, everything was fine until November 27, 2018 (next month) when the enforcement delay of the DSCSA data matrix barcode requirement will end (see “FDA Delays Enforcement of DSCSA November Deadline: What It Means”).
By definition, the DSCSA product identifier must include the drug’s Standardized Numerical Identifier (SNI), the lot number and expiration date. The SNI had already been defined back in 2010 but had no practical use until the DSCSA (see “FDA Aligns with GS1 SGTIN For SNDC” and “Anatomy Of An FDA SNI”). It was defined as the drug’s NDC plus a unique serial number. According to the SNI guidance, the FDA allows the SNI to be encoded into a GS1 serialized GTIN (SGTIN) for the purposes of putting it into a barcode. The DSCSA requires the product identifier to be applied to the packages of a drug’s lowest saleable unit and to homogeneous cases.
SO WHAT’S WRONG WITH ALL THAT?
OK, if you’re with me so far, let’s look at the problem of uniqueness. First, uniqueness is necessary under the DSCSA because every lowest saleable drug package and homogeneous case must be distinctly identifiable from every other package or case so you can track their histories separately, and so you can “DSCSA verify” individual packages and cases using only an SNI (see “What’s So Hard About Unique Identifier Verification?”).
But the SNI on a drug package or case is not equal to its SGTIN. That simple fact is what leads to trouble—several kinds of trouble, in fact. Here’s why. A homogeneous case always has a different SGTIN than each of the saleable units. That is, every homogeneous case SGTIN is identified uniquely from every other case and every saleable drug package SGTIN. So far, so good. But, that’s not what the DSCSA requires. It requires every homogeneous case SNI to be unique from every other case and saleable package SNI. But remember, the SNI on a drug package or case is not equal to its SGTIN, so just because the SGTINs are unique, doesn’t mean the SNI’s are.
Because of the way the GS1 System differentiates trade items and groupings of trade items (which are also trade items), and the way the UGCC decided to encode NDCs within UPCs (now GTINs), the differentiating elements of the GTINs of the unit-level trade items, and the cases that contain them, exist outside of the NDC element that is encoded within the GTINs. You can see this from the figure below (click image to enlarge).
I’ve underlined the NDC encoded within each of the GTINs shown for three different package types of the same drug (30, 60 and 90 count), and the groupings that might be created (a bundle and a homogeneous case of 24 bottles). Notice that the NDC encoded within the GTINs for each of the three products do not change. The same NDC is encoded within the bundle and the case GTINs for the same bottle size. That is, the differentiating elements of these GTINs—the things that make them unique from each other—are outside of the NDC element. The NDC for the bottle of each size (30, 60, 90) is the same as the NDC contained in the GTIN for its corresponding bundle and case (see “How To Properly Define GTINs For Your NDCs”).
I know what you’re thinking right now. You’re thinking, “But the serial number portion of the SNI will make them unique, even if the NDCs are the same”. And you’re right…but you’re also wrong. If the serial numbers for the units were taken from the same pool of numbers as those of the homogeneous cases, then the serial numbers would be unique and you would be right. But in practice, the serial number pools are more likely unique across GTINs, not across the NDC element of the GTIN. What this means is that there is no protection from assigning the identical serial numbers to packages and cases. That’s fine, if you need uniqueness by GTINs, but because of the DSCSA, we need it by NDCs. The problem is, some (most?) serial number management solutions manage serial number uniqueness by GTINs, which means serial number pools for units and cases will normally NOT be the same. And consequently, you can have duplicate serial numbers between units and cases.
If serial number pools are defined based on GTINs rather than NDCs, duplicate SNIs between units and homogeneous cases are possible (click on the image to enlarge). If the serial numbers are randomized, the odds of duplicates would be small. But the DSCSA does not require randomization. For sequentially defined serial numbers, the odds of duplicates are 100%, and if the sequential serial number pools for units and cases both start at the same serial number, then 100% of your case SNIs will have a matching unit SNI.
Big deal, you say. And you would be right…until November 27, 2018, because on that date, drug manufacturers have to be ready to respond to verification request based on the SNI. So if your unit SNIs and your homogeneous case SNIs have duplicates, how are you going to tell the difference when you receive a request for verification? The entire DSCSA product identifier—SNI, lot number and expiration date—are going to be the same between some of your units and cases. And, to make matters worse, most of the matching units are probably not going to be inside of the homogeneous cases that match their SNIs.
Now, it appears that this situation may not be as bad as I have made it sound. The reason for that is that GS1 and the industry have been treating the “SNI” requirement in the DSCSA to actually be an “SGTIN” requirement. That is, wherever the DSCSA talks about an “SNI”, everybody has been reading it as an “SGTIN”. Despite the fact that these identifiers are not the same, if you always read “SGTIN” whenever the text of the DSCSA shows “SNI”, then you will probably be OK. The reason is that when you read that the DSCSA requires manufacturers to “verify units and homogeneous cases at the SNI level”, you will mentally change that to “verify units and homogeneous cases at the SGTIN level”. And when you do that, you will ensure that there are no ill effects of this SNI uniqueness problem. There is no SGTIN uniqueness problem.
You can see the difference between the way the industry thinks of the package and case identifiers and the way the FDA does in the current human readable debate (see “FDA’s Late Recommendation On Human Readable”). The FDA still thinks the identifier that is important is the NDC, while the industry has never thought that. Rather, to the industry, the GTIN is the only identifier that identifies each drug. In a serialized world, that sentence would read: FDA sees SNI, while the industry only sees SGTIN as important.
This clash is pivotal and now that the FDA is seeking input in to the future format of the NDC, whichever side wins the argument will determine how drugs are identified in the future (see “FDA Seeks Input On The Future Format of the National Drug Code”). We must come out of this debate with a single identifier for drugs in the US. Otherwise patient safety and/or supply chain efficiency will suffer.
- 504 views
- 1 version
- 0 replies
- 1 follower
- Posted By:
- Dirk Rodgers
- October 7, 2018
About this forum
- 195 views
- 5 topics
- 2 followers