Thanks for inviting me to your DSCSA Pilot Party this week. I look forward to visiting you at your Silver Spring home so we can catch up on what’s been happing in our lives recently. I’m glad you are thinking more about the Drug Supply Chain Security Act (DSCSA) lately. I am too.
In fact, I’ve been thinking about how nice it would be if you would fix the broken National Drug Code (NDC) as part of the implementation of the DSCSA. It’s really not very hard to do. You already laid out and tested the path that needs to be followed when you implemented the Unique Device Identification (UDI) numbering system for medical devices a few years ago.
Remember how happy that made me? Remember, I called it “revolutionary” (see “FDA Proposed UDI: A Revolution In Number Assignment”), and it was! But before I get to how you could do it for pharmaceuticals, let me explain why you should do it as clearly as I can.
THE PROBLEM WITH THE NATIONAL DRUG CODE
Today, you issue a “labeler code” to companies who plan to submit New Drug Applications (NDA), Abbreviated New Drug Applications (ANDA), Biologics License Applications (BLA) or repackage drugs. These companies use it to define a base NDC that they subsequent register with you as part of the NDA/ANDA/BLA process and once the drug is on the market, they generally use it in all correspondence with you regarding that specific drug. Once you give a company approval to market the drug, you require them to print that number in human readable form and put it into a linear barcode on the drug packages (see “Will The FDA Eliminate The Linear Barcode On Drugs?“). If a recall is issued for that drug, you notify healthcare professionals and the public using that number. The NDC becomes the drug’s identity.
So far so good, but here’s the rub. Remember how your forefathers back in the 1970s tweaked the structure of the NDC? Remember, they lengthened the labeler code and shortened either the product code or the pack size fields, leaving that choice up to the applicant? If you don’t remember (I know you’re younger than that) see “Anatomy Of The National Drug Code”.
Well, when they did that, it confused people who want to reliably extract the constituent parts—labeler code, product code and pack size—from any NDC they were given. Now people cannot tell for sure exactly where the labeler code ends and the product code starts, and where the product code ends and where the pack size starts, without prior knowledge of your entire database of labeler codes. And it turns out that the people who need to extract this data are our friends the Pharmacy Benefit Managers (PBM) and insurance companies. They all want to break every NDC down to those three constituent parts, and they don’t like ambiguity. Weird huh (probably not)?
So years ago they came up with their own fix. They made everybody who gives them an NDC add an extra zero to those numbers, making them all 11-digits long instead of the 10-digits you require. If I were them, I would have called this 11-digit number an “NDC-based reimbursement code”, but unfortunately they didn’t ask me. Instead, guess what they call it? They call it an NDC! That’s right. It’s an 11-digit number that is based on your real 10-digit NDC, and they call it “an NDC”. Confusing, right?
To generate the 11-digit number they add a zero to the beginning of one of the three fields of the real NDC, either the labeler code, the product code or the pack size. They have a logical algorithm for determining which of those three fields gets the extra zero so anyone can do it anytime. Our friends in the Michigan state government have provided us with this little cheat-sheet for that algorithm: https://www.michigan.gov/documents/lara/lara_MAPS_NDC_Guidelines_396516_7.pdf.
And look what happens. For every NDC…er…”NDC-based reimbursement code”, you can always tell which digits are the labeler code, which are the product code, and which are the pack size, because they are always 5 digits, 4 digits and 2 digits respectively. No ambiguity. But too bad our friends in Michigan didn’t provide the algorithm for doing the reverse—converting an 11-digit code back to the 10-digit code. In fact, that is not always possible. If both the product code and the pack size fields of the reimbursement code begin with a zero you can’t tell which zero needs to be removed…unless you have prior knowledge of your entire database of labeler codes, which brings us right back where we started in the first place.
Too bad your forefathers didn’t think of the ambiguity they were introducing back in the 70s. Maybe we would have been able to avoid this unfortunate situation. That is, the situation where drug manufacturers and repackagers are required to use one identifier to reference their products, and at the other end of the supply chain, dispensers are required to refer to those same drugs with a different identifier. The fact that these two identifiers are similar, or related, is of no consequence. They are different. To a computer, they might as well be totally different random numbers. A search for one will not find the other.
It is interesting to note what happens in the middle of the supply chain. At some point, someone has to receive products with the FDA-mandated number, and then switch to the dispenser-mandated number. That “point” is known as, wholesale distributors. This topic recently came up on an industry call and I noted that each of the three big wholesale distributors appear to handle this situation a little differently. Apparently one of them converts the number immediately and all internal operations are based on the “NDC-based reimbursement code”. They refer to the 11-digit number as “the NDC”. The other two appear to run their internal operations on the FDA-mandated (real) NDC and only convert it to the dispenser-mandated code on invoices and other communications that their customers see. They both refer to the 10-digit number as “the NDC” and the 11-digit number as “the 11-digit NDC”.
I know you haven’t had much experience working with wholesale distributors, but historically they have been a quiet, compliant bunch who never complain much. If suppliers formatted barcodes improperly, fine, they just made internal adjustments to accommodate the bad barcodes rather than let the offender know they were doing it wrong. If suppliers were willing to adopt Electronic Data Exchange (EDI) for B2B commerce, great. If not, well, that was OK too. Best not rock the boat. That is, until the last few years. With the approach of deadlines for serialization and standardized transaction documentation mandated by the DSCSA, they have become a little more (OK, a lot more) interested in driving their suppliers toward EDI and more accurate, readable and standardized labeling. That’s a good thing because in the long run, it makes their work more efficient and therefore more profitable.
This discrepancy between the product identifier that the manufacturer uses and the one their customers use has always been accommodated internally by the wholesale distributors, without complaint, right along with those other technology accommodations. But as you know, one of the reasons Congress enacted the DSCSA was to standardize the identification of drugs at the unit and case level, and standardize the transaction documentation throughout the supply chain. The law requires everyone to use the NDC in the product barcodes and in the official transaction documentation. That includes the dispensers. You should be ready for that to uncover some problems since none of their system are set up to use that number. The real FDA NDC is a foreign number to dispensers. You will either be expected to allow the wholesalers to switch from one numbering scheme in their documentation to another, like they have been doing for years now, or you will need to cram the 10-digit NDC down the throats of the dispensers, the PBMs and the insurance companies. Good luck with that latter approach.
HERE’S HOW TO FIX IT
Don’t do either. Just fix it. It’s easy. All you need to do is expand UDI to include pharmaceuticals over a span of time that is comparable to the UDI roleout. That’s it. You’re done. The moniker “UDI” would then stand for “Unique Drug/Device Identification”. Or you can duplicate the entire UDI program under a new regulation for drugs and call it the “Unique Pharmaceutical Identification” (UPI) program. Either way, you would allow drugs to be identified with any FDA-accredited number issuing agency. Of course, this would include GS1 US, HIBCC and ICCBBA, the three current agencies accredited under UDI. Don’t worry, no drug manufacturer will use ICCBBA or HIBCC standards to identify their drugs because I predict the wholesale distributors would immediately announce that they will only buy drugs that are identified with numbers issued by GS1 US. This would maintain the existing situation where all drugs in the U.S. supply chain are marked with NDCs encoded into GS1 GTIN-12s or GTIN-14s and depicted in GS1 barcodes (see “Depicting An NDC Within A GTIN” and “The DSCSA, the NDC, Inventory Management, GS1 GTINs…and Turkeys”).
Look how easy the transition would be. Because GS1 US would already be a valid issuing agency, all existing NDCs would be grandfathered in as valid UPIs when they are properly depicted as GTIN-14s, a 14-digit number. That conversion is almost as simple as the conversion from the NDC to the “NDC-based reimbursement code”. And the rules for extracting the GS1 Company Prefix (the equivalent of the NDC “labeler code”), product code and package size are well established globally and in multiple industries.
Yes, dispensers, PBMs and insurance companies would need to change their systems to make use of the new UPI instead of their customized “NDC-based reimbursement code”, but guess what? They have already done that because of UDI! Your path is now clear! UPI would allow these organizations to merge their current medical device claims software with their pharma claims software. That could enable benefits like the future use of GS1’s Global Data Synchronization Network (GDSN) for receiving standardized pharmaceutical master data published by the actual manufacturer. Or FDA could take that on with a pharma equivalent to the GUDID. As I understand it, these organizations are seeing great benefits from UDI, they just need to be convinced that they will see the same kind of benefits with drugs.
Think of it, everyone using the same number for the same product! Think of the efficiencies you will enable and the errors you will prevent—just like UDI. Ask your brothers and sisters in the devices group how they view it (see “Benefits of a UDI system”). Let’s get started. The sooner we move forward with a plan like this the sooner things will begin operating the way we all want them to.
Sincerely, Your Friend,