OOS — out of specification
An out-of-specification (OOS) result is a test result that falls outside the specification or acceptance criteria — and it stands as valid unless a laboratory investigation identifies an assignable cause.
An out-of-specification (OOS) result is a test result that falls outside the specification or acceptance criteria — and it stands as valid unless a laboratory investigation identifies an assignable cause.
Plain-English explanation, then the primary regulation it comes from.
Understanding oos — out of specification
The default position is the part people find uncomfortable: an OOS result is real until proven otherwise. It cannot be dismissed because it looks unlikely, because the analyst is inexperienced, or because a retest passed. Invalidating a result requires an identified, documented assignable cause — a demonstrable laboratory error, not a hypothesis.
Investigation runs in two phases. Phase I is the laboratory investigation, performed promptly and before the sample preparation is discarded where possible: review the method, calculation, instrument, standards and analyst execution to see whether a laboratory error explains the result. If an assignable cause is found, the result is invalidated, the cause is corrected, and the retest is justified.
If Phase I finds nothing, the investigation must widen. Phase II covers the manufacturing process, batch records, raw materials and any related batches. This is the phase that gets truncated under commercial pressure, and inspectors know it. An OOS closed at Phase I with a vague laboratory explanation and a passing retest is one of the most reliably challenged records on a site.
Retesting and resampling both need pre-defined rules. Testing repeatedly until a passing result appears, then averaging or selecting, is the textbook data integrity finding — and because it is visible in audit trails and sequence logs, it is one of the easiest to detect.
- Valid until invalidated
- An OOS stands unless an assignable laboratory cause is documented.
- Phase I — laboratory
- Method, calculation, instrument, standards, analyst technique. Performed promptly.
- Phase II — full investigation
- Extends into manufacturing, materials and other batches when Phase I finds no cause.
- Retest rules defined in advance
- Number of retests, who authorises, and how results are treated — written before you need them.
- Impact on other batches
- Consider whether the finding affects released or in-process material.
Common failure modes
- Invalidating on a suspected rather than a demonstrated cause.
- Averaging a failing result with passing retests to produce a compliant mean.
- Testing into compliance — repeat injections until one passes, with the earlier ones unexplained.
- Phase II omitted because the batch was commercially needed.
Where this is written down
- European CommissionEudraLex Volume 4 — EU GMP guidelines
Part I, Chapter 6 — Quality Control
- FDAFDA — Current Good Manufacturing Practice (CGMP) regulations
FDA guidance on investigating OOS test results is the most detailed public treatment
- MHRAMHRA — GxP Data Integrity Guidance and Definitions
On audit trails, retesting and record selection
Read next
Looking for a definition rather than an explanation? The GMP glossary covers the abbreviations in one line each.
Knowing the requirement is not the same as closing the gap
If you want to know where your site actually stands against this, the readiness score covers seven quality-system domains in twenty questions, and takes about ten minutes.