Blog Viewer

I had a chance recently to think about proposed regulation around track-and-trace capabilities within the pharmaceutical supply chain and had presented some thoughts to ASHP that I thought might be worthwhile sharing with this group.

Let me first assert firmly that I am by no means a specialist in supply chain, so most of the questions that would ordinarily be asked regarding those matters belong somewhere besides me. I do, however, have some experience with bar coding and, to the extent that bar coding is a key part of the proposed track-and-trace strategy, have some firm opinions on the infrastructure that would need to be in place to make track-and-trace useful. The good news is that that same infrastructure might well close some of the holes we currently experience in the use of bar coding.

The basic notion of track-and-trace is that the information encoded within a bar code could be scanned and compared to a database of known transactions for that same bar code which would allow a user to review the pedigree of the scanned product, and hopefully discern gaps in that pedigree that might indicate that the scanned items was counterfeit.

For example, if the database for a given instance of a given NDC were to flow through a series of transactions in the New York City area and suddenly get a transaction report that indicated use in California, someone might be concerned that the product scanned in California might be counterfeit.

Note that this strategy would appear to make some rather significant assumptions:

1. The NDC bar code contains a serial number that represents a single instance of the NDC being scanned.

2. There would be a database that knew each and every NDC currently in the supply chain and every transaction that each of those NDC's underwent. This implies that new NDC's would be available to the supply chain before the first item was actually sold.

3. Current inventory systems would be connected to this database and would transmit inventory transactions as scans occurred.

4. All transactions would be reported to this system.

5. The first transactions in the system (those from the manufacturer to wholesalers, etc.) would be legitimate, and that the opportunity for illegitimate transactions would occur subsequently in the supply chain.

Let me propose a few other things that would need to be true.

1. Theoretically there are different NDC's for different package levels of a product. For example, the NDC on a case of 100 vials of an injection is different than the NDC on each of the ten boxes of ten vials within that case, which is , in turn, different from the NDC that is imprinted on each of the vials in that case. Therefore, the meaning of the serial number has to be different depending on the NDC being scanned. For the case, it would be the unique serial number of that case within all cases with that NDC. For the box of ten vials, it would be the unique instance among all boxes of ten. For the vials, it would be a unique serial number among all vials of that product ever produced.

2. Depending on the level of a transaction, a scanning system might scan the entire case, a shelf pack within the case, or an individual unit-of-use container within the shelf pack. In order for traceability to be demonstrated through all these transactions, the track and trace system would need to maintain awareness of the relationship between these individual transactions to traceability could be maintained even though different transactions scanned different NDC's. Continuing with the above example, this would mean that the scan of a case being shipped to a wholesaler from a  manufacturer would also be represented as the same transaction occurring for each of the ten boxes of vials within that case and for each of the vials within each of the shelf packs in the case. This means that a case would have to "know" each of the serialized shelf-pack bar codes it contained as well as each of the serialized bar codes on each of the unit-of-use packages it contained.

Without this level of integration, transactions would only trace based on the NDC and serial number scanned. This means that transactions for a serialized bar code for a case would not list the transactions for the shelf packs or unit-of-use containers within that case. The usefulness of this data under these conditions would be severely constrained.

Assuming this level of integration is in place, a scan of a vial out of our example case could return the transaction in which the case was shipped from the manufacturer to a wholesaler, the transaction in which the case was received by the wholesaler, the transaction in which one of the boxes of 10 vials was shipped to my hospital, the transaction in which that box of 10 was received by my hospital, and the transaction in which one of the vials from that box was about to be used to prepare an IV.

3. Even if the bar code continued to contain only the NDC, its data would become sufficiently great that it could not be contained within a 1-D bar code. A 2-D would be required. 


The good news around all this is that the excuses for not encoding lot and expiration in the bar code would be lost. The process of serializing bar codes would readily permit the inclusion of other information.

The other good news would be that bar code scanners could tell the difference between scanning the same vial 10 times and scanning 10 different vials, which would remove some of the workarounds we know occur in barcoding system.

Knowing as I do that neither commercial databases nor the FDA database maintains timely and complete information about every NDC for every package in the supply chain, I must confess a certain amount of skepticism about how well this is all going to work.

What do you think?

Dennis A. Tribble, Pharm. D., FASHP
Daytona Beach , FL
DATdoc@aol.com