4 minute read
ISO 22716
By Robert Low, Lead Management System Specialist
Clause 9.5 of ISO 22716 is short — three sub-clauses — but it covers ground that auditors return to repeatedly, because how a manufacturer handles an out-of-specification result says a great deal about whether the quality system is genuinely functioning or just generating paperwork.
This post is part of a series working through ISO 22716 section by section. The series summary with links to every section is available at the end.
Review and investigation. Out-of-specification results need reviewing by authorised personnel and properly investigating. This is not optional, and it is not something that can be skipped because the result “was probably a testing error.”
Justification for re-testing. If re-testing is going to happen, there needs to be sufficient justification recorded for why. This exists specifically to prevent the practice — sometimes called “testing into compliance” — of simply repeating a test until a passing result appears, without any investigation into why the first result failed.
Decision following investigation. After investigating, authorised personnel need to make a decision: deviation, rejection, or pending. The investigation has to come before the decision, not the other way around.
An out-of-specification result is one of the clearest moments in a quality system where genuine root cause investigation either happens or does not. The pressure to move past an OOS result quickly — to keep production moving, to avoid delaying a shipment — is real and constant. The standard’s requirement for documented investigation before any decision is a direct counter to that pressure.
Auditors specifically look for evidence that an investigation considered whether the failure was a genuine product issue or a laboratory error — sample handling, equipment calibration, analyst technique — before concluding either way. A finding that simply states “result was within tolerance on re-test, batch released” without any investigation into the original failure is precisely the gap this clause exists to close.
The most frequent issue is re-testing happening without documented justification. A result fails, the same sample (or a new one) gets tested again, and a passing result is accepted with no record of why re-testing was considered appropriate in the first place. This is one of the most consistently cited findings in cosmetics GMP audits, precisely because it is so easy for an auditor to spot — they ask for the OOS investigation record, and if “we just ran it again” is the only answer, that is a clear nonconformance.
The second common issue is investigations that identify a probable cause but never check whether other batches might share the same root cause. If a calibration drift caused one batch’s OOS result, the same drift was likely present for other batches tested on that same instrument in the same period — and those need checking too.
An OOS investigation that concludes a genuine product issue occurred should flow directly into nonconformance and CAPA management — the OOS result is the trigger, but the corrective action process is where the root cause actually gets addressed. Where the OOS relates to a supplier-provided raw material, it also needs to feed back into supplier monitoring, since a pattern of OOS results from one supplier is exactly the kind of trend a functioning approval system should catch.
This post is part of the Cornerstone ISO 22716 series. See the full series summary with links to all 16 sections.
Back to all
