Today's post provides a current assessment of the integration of serialization into common pharma manufacturing processes. Certainly not a new topic and one that has been covered in numerous articles from colleagues and vendors. As a point of reference- nearly 18 months ago a great overview was provided by David Colombo of KPMG in Pharmaceutical Online (Read it here).
How far have we come since then? Did the industry heed the message or are we still treating serialization 'in a vacuum' and missing integration into key operations?
Let''s dive into two key manufacturing processes where ensuring proper serialization integration is critical.
Serial Number Reconciliation- No pharma manufacturer (following GMP) would release product without quantity reconciliations being performed and documented in batch records. Quantity of items packed, quantity of items sampled, quantity of items damaged/destroyed- this is the norm of any packaging operation. So why should it be any different with serialization?
With serialization an entirely new data set exists (serial numbers) which must align with the quantities documented in a batch record- So it stands to reason that checking the quantity of SNs against the quantities in batch records each and every time seems like a painfully obvious concept- many will assume it's being done. But if you are a manufacturer and have ever experienced mismatches in your serialization data then quite clearly reconciliation was missed somewhere along the way. (Comment and let me know your experiences!!)
So, what's preventing serial number reconciliations from being a pervasive concept? In many cases it comes down to limitations of serialization solutions. Many vendor solutions don't support the ability to do serial number reconciliations in two main areas:
- Packaging site/CMO solutions which only allow the collation and communication of serialization data to manufacturers at the time of shipment. Following good practices- if a manufacturer only receives the serialization data for a batch after the product has been loaded onto a truck it is way too late in the process. When designing packaging site/CMO integrations identify trigger points during the batch process which will support reconciliation (such as upon lot closure). If either your CMO's system or your system can only support integration at time of shipment- be prepared to justify the risk of not being able to perform serial number-to-batch record reconciliation until 1) after you've had to provide batch release and 2) after your product is already in transit. What's the preferred capability in this area? Some solutions are able to take in batch record inputs (even better if they are electronic batch records!!) and automatically compare against processed serialized data- providing alerts when quantities do not reconcile.
- Solutions which do not support capture/communication of 'end-of-life' data (end-of-life covers any scenario where a serial number is no longer continuing in the commercial supply chain). This has long been a heated debate within the industry with organizations deeply rooted on both sides of the fence- in regards to DSCSA some choose to follow the regulations and only capture data for the 'good' items that get shipped/sold while others will require tracking of all allocated serial numbers. For what it's worth- I've always been firmly planted on the latter side. I summarize the debate like this- there is no right or wrong answer, but if you are only tracking the 'good' items then the extent of your organization's capability is serialization compliance, not true traceability. Again, for many organizations there is nothing wrong with that as they're likely still 'checking the box', it just means additional areas will have to be addressed before being able to take advantage of other benefits of true traceability (e.g. brand protection, process/quality efficiency monitoring, etc.)
Serialized Shipments- For many organizations, capturing serialized shipping information is still a future endeavor. But as 2019 requirements (both regulatory and industry driven) start to surface the need to track items from manufacturing through distribution arises. One practice, however, that appears time and again is the concept of 'Ship by Lot'. In the old, quantity-based world this concept had merit as it was an easy way to update quantities of items in ERP systems. In a serialized world, however, continuing this practice brings significant risk. The principle is simple- You don't ship lots, you ship shipments.
What allows this practice to continue in a serialized world is that often a shipment equals a lot. Again, in the old, quantity-based world it was easy to say "Ship Lot 123" and then enter the quantity actually being shipped. In a serialized world, performing the same action "Ship Lot 123" makes a big assumption that all serial numbers which have been previously associated to that lot are now being shipped. The problem is that organizations are not going through and actually scanning each serial number as part of an explicit shipping step and therefore systems have to 'infer' which items are now being shipped. The serialized world relies on inference in situations where aggregations exist (physical product packed and sealed in larger physical containers) but shouldn't rely on inference for creating significant traceability steps (e.g. shipping).
This practice is most commonly seen at companies who have chosen not to aggregate for the time being but are often 'forced' into generating a shipping step by either their vendor or a downstream customer. Additionally, certain vendors have made this practice too accessible by making it a standard (and even recommended) capability in their platforms.
My summary: Wait until you are aggregating before you start capturing serialized shipping information unless your volumes are so small you can warrant doing an explicit shipping step. Moreover, once you are aggregating there is no reason to continue a practice like 'Ship by Lot'. Instead perform a true shipping step which includes scanning of parent level containers.