Deviation vs OOS
A departure from procedure vs a failing test result — handled differently.
What a comparison is not
A comparison is SPEQ’s reading of how two published documents differ. Neither is the right answer, it is not a determination of which applies to you, and neither is summarised in a way that replaces reading it.
A deviation is any departure from an approved instruction, procedure, or standard — planned or unplanned. An out-of-specification (OOS) result is a specific event: a laboratory test result that falls outside an established specification. Every OOS is investigated, but OOS investigations follow a distinct, well-defined path (laboratory phase, then full-scale) that differs from general deviation handling.
| ASPECT | DEVIATION | OOS |
|---|---|---|
| What it is | Any departure from a procedure, spec, or instruction | A test result outside an established acceptance criterion |
| Where it arises | Anywhere in operations (manufacturing, warehouse, lab) | The laboratory (analytical testing) |
| Investigation path | Deviation investigation → root cause → CAPA if warranted | Phase I (lab investigation) → Phase II (full-scale, incl. manufacturing) if lab cause not found |
| Key reference | GMP deviation-management expectations | FDA OOS guidance; the “Barr” logic; 21 CFR 211.192 |
| Invalidating a result | N/A | Only with a documented, investigated assignable cause — no “testing into compliance” |
| Outcome | Corrected; batch/impact assessed; CAPA as needed | Result confirmed or invalidated; batch impact and other-batch scope assessed |
Log a deviation whenever something did not follow the approved procedure or standard — the focus is understanding what happened, its impact, and whether a CAPA is needed.
Run an OOS investigation whenever a test result fails specification — the focus is a disciplined laboratory phase first, then a full-scale investigation (including manufacturing) if no lab cause is found, before any decision to invalidate.
An OOS is a special, tightly-governed kind of event; a deviation is the general category of “we didn’t follow the procedure.” The classic enforcement trap is treating an OOS like a routine deviation — closing it as “lab error” without the required investigation, or retesting until you get a pass. Follow the OOS-specific logic (documented assignable cause before invalidation; scope to other batches) and use deviation handling for everything else.
Deviation vs OOS: frequently asked questions
Common questions on how Deviation and OOS differ and when each applies.
Is an OOS a type of deviation?
Broadly, an OOS is a kind of event that departs from the expected outcome, but it is handled under its own specific investigation framework (FDA OOS guidance, 21 CFR 211.192) rather than general deviation procedures. Many firms track them as separate record types for exactly this reason.
Can I retest an OOS result?
Only within a defined, justified investigation — never to “test into compliance.” Retesting is permitted when the laboratory investigation identifies a possible assignable cause and your procedure defines the retest scheme in advance. You cannot average a passing retest with a failing original to make a batch conform.
What is OOT vs OOS?
Out-of-Trend (OOT) is a result that is within specification but atypical relative to historical data or expected trend. It can be an early signal of a developing problem and is investigated, even though the spec was technically met. OOS is a hard specification failure.
When does an OOS become a full-scale investigation?
When the laboratory-phase (Phase I) investigation does not find a clear, assignable laboratory cause, the investigation expands (Phase II) to include manufacturing and other potential causes, and assesses impact on the batch and other batches.