A food recall is usually the end of an investigation rather than the start of one. The work that precedes it runs backwards from people who became ill to a shared point of origin.
Detection starts with clinical samples
Laboratories that identify certain infections report them to public health authorities, along with the genetic fingerprint of the organism involved.
Cases that would otherwise look unrelated are linked when their samples share a fingerprint, even where the patients live far apart and fell ill weeks apart.
A cluster of matching cases is what triggers an investigation, which is why an outbreak is often recognised before any single doctor suspects one.
Interviews look for the common item
Investigators interview cases about everything eaten in the relevant period, which is difficult because the incubation window may have been days or weeks earlier.
Recall of ordinary meals is poor, so shopping loyalty records and card statements are used with consent to reconstruct purchases.
The aim is to find an item eaten by far more cases than the general population would report, which points to a candidate rather than proving it.
Traceback follows the paperwork
Once a food is suspected, investigators work through the supply chain from the shops where cases bought it towards its origin.
Distributors, processors and growers each hold records of what they received and dispatched, and the traceback looks for the point where all the routes converge.
Convergence on a single supplier, batch or growing site is what turns a hypothesis into an actionable finding.
Mixing makes the trail harder
Ingredients from many suppliers are combined in processing, so a single batch of a product can contain material from dozens of sources.
Bagged salads, ground meat, spices and flour all mix widely, and a contaminated input can spread through products that share nothing on the label.
This is why recalls sometimes cover a long list of apparently unrelated items, all of which contain the same ingredient from the same batch.
The recall is a communication problem
Notices identify products by batch codes and dates, because withdrawing a whole brand would be disproportionate and would exhaust public attention.
Retailers remove stock and post notices, and where a product has already reached homes the message has to travel through media and direct customer contact.
Effectiveness is measured by how much of the affected stock is recovered, and it is always partial, which is why recall notices are repeated and why checking a code against a current notice is worth the moment it takes.