DSCSA: Saleable Returns Verification

The Healthcare Distribution Alliance (HDA) Traceability Seminar that was held back in early November was so packed with valuable information that I still have a number of topics queued up from that event for RxTrace essays in the future.  Today I want to take a closer look at the results of the Saleable Returns Pilots conducted by the HDA last year to figure out the most efficient way to verify saleable returns.  I discussed the overall project in my report of the Traceability Seminar (see “HDA Delivers Home Run To Record-Breaking Audience”) but today I want to focus in on just two of the approaches piloted.  These are:

  • Manufacturer sends to wholesale distributor product identifiers for only the units purchased by that wholesale distributor, and,
  • Verification Router Service (VRS).

These are the two approaches selected by the HDA as recommendations for drug manufacturers to allow wholesale distributors to solve the saleable returns problem that was created by Section 582(c)(4)(D) of the Drug Supply Chain Security Act (DSCSA), Verification of Saleable Returned Product.  That section requires wholesale distributors to begin “verifying” the DSCSA Product Identifier, including the Standardized Numerical Identifier (SNI) on homogeneous cases or individual saleable units if they are no longer sealed within homogeneous cases, in November of 2019 (see “DQSA: Dancing Around The Returns Problem”, “Drug Verification: EU Vs US”, “The DSCSA Product Identifier On Drug Packages”, “FDA Aligns with GS1 SGTIN For SNDC” and “InBrief: ‘The Smallest Individual Saleable Unit’ In The DSCSA”).

Why don’t the wholesale distributors just scan every serial number (or “SNI”) that leaves their facilities in customer shipments, and then use that data to confirm that they originally received from the manufacturer each saleable return?  The HDA did a pilot with this approach, but so far they have not made public their final analysis of it.  From my experience, there are two reasons this won’t work well.  First, some of the shipments to customers are in the form of full homogeneous cases that were originally packed by the drug manufacturer.  If the wholesale distributor scans the serial number on those cases, but the customer returns the contents after opening that case, the unit serial numbers will not be in their database.  Second, some of the larger wholesale distributors make use of automation systems, like A-Frames, that do not offer the opportunity for every unit to be scanned during order fulfillment.  Although some experiments have been done, adding a step where every unit in an A-Frame is scanned during order fulfillment could lead to the loss of the very efficiency the A-Frame offers to begin with.  

So basically, the HDA and their wholesale distributor members are telling drug manufacturers to “pick one” of the two approaches in my list above, and plan to begin following it before November 27, 2019.  Which approach should you “pick”?  Let’s take a look at both of them.


In this approach, drug manufacturers would need to know exactly which serial numbers are on each homogeneous case and the individual saleable units contained within them, and the loose units in each shipment to wholesale distributors.  That way they can provide the list of those serial numbers to the recipient at the time of the shipment.  How are you going to know that?  You have two choices.  You can either open every case just prior to shipment and scan each of the unit-level serial numbers and the case serial numbers, or, you can build the serial number containment hierarchy as you originally put the units within the homogeneous cases. 

That first choice is unworkable because no one is going to want to re-open sealed cases and re-scan serial numbers in their distribution center or at their 3PL at the time of shipment.  That would be a huge workload and fraught with errors.

You may recognize that second choice as “aggregation data” (see “Does The DQSA Require Manufacturers To Provide Aggregation Data? Survey Says…“).  If you choose to collect aggregation data at the time of case-packing, then later when you assemble a shipment to a wholesale distributor, all you need to do to know exactly which unit-level serial numbers are included in your shipment is to scan the case-level serial numbers and use the aggregation data to build your list.  Just send that list along with the case-level serial numbers to the wholesale distributors.  Easy, right?

Previously some of the larger wholesale distributors said they were planning to require their suppliers to provide them with the full aggregation data for their shipments of serialized product sometime in advance of the November 2019 deadline (see “U.S. Drug Wholesale Distributors Provide Direction To Manufacturers”, and “HDMA DMC Serves Sizzling Steak With A Small Side Of Snake Oil”).  But the 2016 HDA pilots taught them that they technically do not need to receive the actual aggregation data as long as they receive a list of the specific serial numbers that are included within each shipment.  That’s not really a concession because drug manufacturers will be forced to collect aggregation data in order to fulfill that requirement.  They won’t need to send the actual aggregation data if they don’t want to, but they would need to use that data to build the list of serial numbers that the wholesale distributors are expecting.

We know that a large percentage of drug manufacturers are already planning to collect aggregation data prior to 2023 (see “The 2016 RxTrace U.S. Pharma Traceability Survey, Sponsored by Frequentz, Final Report”).  But that percentage is still well below 100%, so those companies who are not should be very interested in the second HDA approach to solving the saleable returns problem.


In this approach, drug manufacturers would need populate a cloud-based Verification Router Service (VRS) with the GS1 Global Trade Item Numbers (GTINs) and a pointer to their own internet-based verification service.  When wholesale distributors, or others, need to verify the serial numbers on saleable returns, they would contact the VRS, provide it with the full DSCSA Product Identifier on the product(s) and the VRS would facilitate their verification.  There are a number of possible designs for the VRS which I’m not going to cover in this essay.  The HDA has already embarked on a project to figure out which design to use.  I’ll report on that later.

The thing to notice about the VRS approach is that, unlike the “Manufacturer Sends Serial Numbers With Shipments” approach outlined above, there is no need for drug manufacturers to know or even care where they shipped each unit and case.  Where drugs were shipped does not have any role in the VRS concept.  That comes from the DSCSA definition of the phrase “Verification Or Verify”.  Here is Section 581(28):


The term ‘verification’ or ‘verify’ means determining whether the product identifier affixed to, or imprinted upon, a package or homogeneous case corresponds to the standardized numerical identifier or lot number and expiration date assigned to the product by the manufacturer or the repackager, as applicable in accordance with section 582.”

So because there is nothing in the definition of “Verification Or Verify” about where a drug product was shipped, you can verify that drug product by just determining whether or not its product identifier corresponds to one that the manufacturer assigned.  That’s it.  So if a drug manufacturer makes all of the DSCSA Product Identifiers (see “The DSCSA Product Identifier On Drug Packages“) they ever assign to drug packages and homogeneous cases accessible to their verification service that is called through the VRS, then anyone with the proper access—including wholesale distributors—can verify drugs they have in their possession.  That is, they can “DSCSA Verify” them.

This elimination of the need to care about where each serial number was shipped by the manufacturer greatly simplifies the work those manufacturers need to go through to provide wholesale distributors with the ability to verify saleable returns.  Now, all they have to do is make sure their verification service is aware of all DSCSA Product Identifiers sold into the U.S. supply chain.  That’s pretty simple.  And for wholesale distributors to verify each unit- or case-level SNI received as a saleable return, all they have to do is exercise the VRS appropriately with each identifier.

But for this to work, drug manufacturers have to be willing to respond to verification requests provided to them through a third-party, cloud-based repository.  That means it needs to be very secure.  And, the design of the VRS needs to provide screamingly fast performance, high availability (HA) with disaster recovery (DR) so the wholesale distributors are never delayed in processing their returns.  This is a perfect need for a centralized service.

Some argue that a distributed system is better for the industry than a centralized system.  They claim that putting everything into a single, central repository leads to a dreaded “single point of failure”.  But they fail to realize that a solution with a single point of failure is much better than one that has many points of failure.  With a single point of failure, there is no question where that point is, so you can design a solution (with plenty of security, HA and DR) to protect it from a lot more possible causes of failure than you can protect each of the many points of a distributed system (see “InBrief: ePedigree Models and Points of Failure”).  So an industry-funded central VRS service must be designed with that single-point vulnerability in mind.  You need to call in experts who have lots of experience protecting large financial systems.

Notice that the VRS is starting to sound a little bit like the European Medicines Verification Organization (EMVO)—the “European Hub” anyway.  The EMVO is funded by the industry so that companies serving the E.U. market can meet the Falsified Medicines Directive (FMD) and Delegated Regulation (EUDR).  That law requires the industry to develop and fund a system of repositories that can be used by pharmacies to verify the authenticity of drugs at the point of dispense (PoD) (see “EU FMD: Aggregation Is Not Mandated, But It Will Be Necessary”).  I noted this similarity and explored some of the implications in “Product Identifier Authentication (PIA)”.

Does the VRS concept sound scary to you?  Of the two approaches recommended by the HDA, this is the one that manufacturers will get the most benefit from, but only if they demand that this approach be tied in with reverse chargebacks so it is worth their investment and effort (see “Aggregation –> Chargeback Accuracy –> ROI“).

The VRS could start out being used for wholesale distributor verification of product identifiers, grow into dispenser verification of product identifiers, and then eventually be morphed into the full “secure, interoperable, electronic” system (known in the DSCSA as the “Enhanced Drug Distribution Security”, or “EDDS” phase) that the DSCSA mandates in late 2023.  It could, if its initial design takes that path and end-point into consideration.


One of the wholesalers speaking at the HDA Traceability Seminar said that they realize that drug manufacturers will also need to provide a way of manually verifying an SNI for the times when all else fails, but that cannot be the only way a given manufacturer offersfor verification. 

So which of these two approaches described above are you going to pick?  In my view, it’s a little too early to pick, but it is time to recognize that you will need to pick one or the other in the near future, and the window for your choice will be fairly short.  The time to get behind one approach or the other will happen as soon as the HDA is able to make clear that the VRS will definitely be implemented and when.  (I hope they are prepared to take that pressure.)

If you choose to provide the serial numbers in each shipment to your wholesale distributor customers, you will need to begin doing that no later than about one year from now so they can build up a list of serial numbers in advance of the returns they will expect after November of 2019.  If you choose to go with the VRS, you won’t have to start interacting with it until just before the November 2019 deadline, but at that time you will need to be able to verify product identifiers shipped into the supply chain starting about one year from now.  The longer you delay the decision of which approach you plan to take, the greater the odds you will be forced to take the VRS approach.  (For an early discussion of this choice, see “The Aggregation Hoax and PIA“.)

The best place to get compete and factual information about this choice and the two alternatives is at the upcoming HDA Distribution Management Conference and Expo, March 5 to 8 in Palm Desert, California.  I’ll be there as usual.  Maybe you should too.


[NOTE:  This essay was updated on January 15, 2017 to more accurately reflect the architecture of the VRS that the HDA is currently planning. –Dirk.]