The Recall Notice Arrives on a Friday
A manufacturer withdraws a lot number at four on a Friday afternoon. The question is not whether you stock it — it is which of your locations still holds it, how much left, and when. What a batch trace answers, and precisely where it stops.
The notice is short and it always arrives at an awkward hour. A manufacturer or a regulator withdraws a specific lot number, gives a reason in one sentence, and asks that stock be quarantined immediately and the quantity confirmed.
What happens next in most facilities is a phone call to every store, a person walking shelves with a printed notice, and a number arrived at by asking people to remember. It takes the rest of Friday and a good part of Saturday, and the number at the end of it is an estimate that nobody would want to defend in writing.
The whole of the difficulty is that the question is asked in the manufacturer's unit — a lot number — and most stock records are kept in yours: an item, a quantity, a location. If those two were never connected at the point the goods arrived, no amount of effort on Friday afternoon can connect them now.
The three questions a recall actually asks
They sound like one question and they are not. Each needs a different record, and a facility can be able to answer one and not the others.
| Question | What it needs | Asked by |
|---|---|---|
| How much of this lot is still on our shelves? | Batch-aware stock by location | You, immediately |
| How much of it did we use, and when? | Movements recorded against the batch | The regulator, later |
| Who received it? | A record linking issue to recipient | The clinical side, hardest |
The first is operational and urgent. The second is evidential and arrives in a letter a fortnight later. The third is a clinical question, and it is the one where a supply system has to be honest about its limits rather than imply a capability it does not have.
A recall is asked in the manufacturer's unit and answered in yours. If those were never connected at the receiving door, Friday afternoon is too late to connect them.
What a batch trace returns
Given a batch, the trace produces three things at once, and it is worth knowing the shape of each because they answer different halves of the problem.
-
Every movement, in time order
Each event carries a direction — in or out — a quantity, the warehouse and location it happened at, and when. Read top to bottom, it is the batch's history from receipt onward.
-
Every location still holding it
The locations with stock of that batch remaining, with their quantities. This is the quarantine list, and it is the thing somebody can walk with.
-
The three totals
How much came in, how much went out, and how much remains. These reconcile against each other, which is what makes the figure defensible rather than an estimate.
The reason the totals matter as much as the list is that a recall response gets checked. A quantity that reconciles — this much arrived, this much was used, this much is on the shelf — is a different kind of answer from a number somebody counted on a Friday.
The work is done at receipt, not at recall
None of this is available unless the batch was recorded when the goods arrived. That is thirty seconds of typing per delivery line, on a normal Tuesday, for a benefit that shows up on an afternoon nobody can predict. It is the least popular discipline in a store and it is the one that decides whether a recall takes twenty minutes or two days.
Where the trace stops
It ends at the location the stock left from. The system knows a quantity of that batch was issued from that store on that date. It does not know who received it, because there is no patient in this system and there is not going to be one.
That is a real boundary and it deserves saying plainly rather than being softened. For the operational half of a recall — quarantine what remains, confirm the quantity, report what was used — the trace is sufficient and it is fast. For the clinical half, the record lives in whatever holds your patient data, and the batch trace is the input to that search rather than a substitute for it.
What the trace gives that side of the building is a date range and a location. "This batch was issued from the theatre store between the eleventh and the nineteenth" turns an unbounded search of clinical records into a bounded one, which is usually the difference between a search that gets done and one that gets abandoned.
What AWRA OpsHub does today
- Every movement of a batch in time order — direction, quantity, warehouse, location and when.
- The locations still holding it, with quantities, so quarantine is a list somebody can walk rather than a phone round.
- Totals that reconcile — received, issued, remaining — so the figure you report is defensible.
- Batches carry expiry dates, and a scheduled job runs daily to raise alerts on batches approaching them, on a cadence you set.
- Sales and issues record which batch they drew from, so the outward movements are attributed rather than inferred.
What it does not do
- No recipient. The trace ends at the location the stock left. Who it was administered to is a clinical record this system does not hold, and the trace is the input to that search rather than a replacement for it.
- Stock received without a batch cannot be traced by batch. The record has to exist before it can be searched, and there is no way to reconstruct it afterwards.
- No recall case management. The trace answers the question; tracking the response — who was notified, what was returned to the supplier, what the regulator was told — is a process you run alongside it.
Not ours, by choice
- We will not hold patient data. This is a supply system, and a vendor that offers to store clinical records inside an inventory product is offering you an obligation you did not ask for and a protection standard we would rather integrate with than improvise.
- We will not infer a batch that was never recorded. If receiving did not capture the lot number, the honest answer to a recall is that this stock cannot be traced by batch — and knowing that on Friday is better than a number that looks precise and is not.
Recall case management — a notice raised against a batch, notifications to the locations holding it, and the quarantine and supplier-return steps tracked to completion — is scope rather than a ceiling. The trace, the hold and dispose workflow, and the notification infrastructure all exist and work; assembling them into a case is a written specification and a price.
If you are evaluating traceability, the question that separates real from decorative is whether outward movements are attributed to a batch. Plenty of systems record a batch at receipt and then issue stock without saying which batch it came from, which gives you an arrival record and no trace at all.
The receiving discipline this depends on
Everything above is downstream of one habit, and it is worth being specific about what that habit costs and where it can reasonably be relaxed.
What to insist on at the receiving door
- Lot number captured on every delivery line for anything with a lot number. Not most lines — every line, because the one you skipped is the one that gets recalled.
- Expiry date captured at the same moment. It is printed on the same box and it costs nothing extra to type, and it is what makes expiry alerting work at all.
- One delivery, one batch record per lot. Splitting the same lot across two records because two people received it produces two traces and one wrong total.
- Batch recorded before the stock is put away. Once it is on a shelf among identical-looking cartons, the lot number is gone.
- A written rule about what is exempt. Consumables with no lot marking, stationery, anything where batch tracking is theatre — decide the list deliberately so the exemption does not spread.
The last one matters more than it looks. Batch discipline dies from ambiguity rather than from refusal. If nobody has said which items need a lot number, a busy storekeeper will make the decision item by item at four in the afternoon, and within a month the answer is "the ones that were easy".
Expiry is the same machinery, used earlier
A recall is an unplanned version of a question you should be asking on a schedule anyway, and the batch record answers both.
Batches carry expiry dates, and a job runs every day to raise alerts on the ones approaching them, at whatever reminder cadence you have set. That is worth naming specifically because expiry alerting is a claim made frequently and scheduled rarely — plenty of systems have an expiry field, a report that reads it, and nothing that ever opens the report.
The practical connection is that a facility with good expiry discipline already has good recall discipline, because both are built on the same thirty seconds at the receiving door. If you are trying to justify the receiving habit to a store team, expiry is the easier argument — it saves money every month, visibly, rather than saving a Friday once every few years.
The expiry rhythm is worked through in pharmacy expiry management, the quarantine and destruction side in hold first, destroy second, and the custody discipline for controlled items in controlled medicines.
Our take
The recall is decided at the receiving door, months before it happens. Capture the lot number and the expiry on every line that has one, and a withdrawal notice becomes a twenty-minute exercise producing a reconciled figure rather than two days producing an estimate. Be clear about the boundary too: the trace tells you which locations still hold it, how much left and when, and it stops at the door. Who received it is a clinical record, and what the trace gives that search is a bounded date range and a location — which is usually the difference between a search that happens and one that does not.
See a batch traced end to end
Every movement in time order, the locations still holding stock, totals that reconcile, and expiry alerting on a daily schedule rather than a report nobody opens.
Explore traceabilityFrequently asked questions
How quickly can we answer a recall notice?
If the lot number was captured at receipt, immediately — the trace returns every location still holding that batch with its quantity, every movement of it in time order, and totals for received, issued and remaining that reconcile against each other. That last part is what makes the figure defensible rather than an estimate. If the lot number was not captured at receipt, there is no way to reconstruct it, and the honest answer is that this stock cannot be traced by batch.
Does the trace tell us which patients received the stock?
No. It ends at the location the stock was issued from — the system knows a quantity of that batch left that store on that date, and it does not know who received it, because there is no patient record here and there is not going to be one. What it gives the clinical search is a bounded date range and a specific location, which turns an unbounded review of records into a targeted one. That is genuinely useful, and it is not the same as answering the question.
What has to happen at receiving for this to work?
The lot number and the expiry date captured on the delivery line, before the stock is put away. It is roughly thirty seconds per line, on an ordinary day, and once cartons are on a shelf among identical-looking ones the lot number is gone. The part most facilities get wrong is not refusing to do it — it is never writing down which items require it, so a busy storekeeper decides case by case at four in the afternoon and within a month the answer has become "the ones that were easy".
Do outward movements record which batch they came from?
Yes, and this is the question worth asking of any system claiming traceability. Sales and issues record the batch they drew from, so the outward movements are attributed rather than inferred. A system that captures a batch at receipt and then issues stock without saying which batch it came from gives you an arrival record and no trace — you know it came in, and nothing about where it went.
Does it manage the recall process itself?
No. It answers the question; running the response is a process alongside it — notifying the locations, quarantining what remains, returning to the supplier, reporting to the regulator. The hold and disposal steps exist as inventory operations with a reason and a name attached, so the individual actions are recorded, but they are not assembled into a case with a status. If you handle recalls often enough that case tracking matters, that is worth scoping deliberately rather than assuming.
Is expiry handled by the same records?
Yes, and it is the easier argument for the receiving discipline. Batches carry an expiry date and a scheduled job runs daily to raise alerts on the ones approaching it, at a cadence you set. Worth checking specifically in any evaluation, because an expiry field and a report that reads it are common, while something actually scheduled to run and tell somebody is much less so. A facility with good expiry discipline already has good recall discipline, because both rest on the same thirty seconds at the door.
