SOP: Out-of-Specification (OOS) Result Investigation

This is a ready-to-use SOP. Replace every <<FILL: ...>> placeholder with your own specifics, set your document numbers and dates, and route it through your normal document control, review, and approval. A worked filled specimen follows the template so you can see how a completed version reads. Verify each cited regulation against the current source before you rely on it.

Document control header

1. Purpose

This procedure defines how <<FILL: COMPANY NAME>> investigates out-of-specification (OOS) laboratory results so that every failing reportable result is explained on evidence, not invalidated for convenience. The objective is to determine whether an OOS arose from a specific, documented laboratory error or reflects a true failure of the material, and to reach a disposition decision that holds up in front of an inspector.

2. Scope

This procedure applies to all OOS results generated by GxP laboratories at the sites listed in the header, covering chemical and physical testing, microbiological testing, and biological assays, for raw materials, in-process samples, finished product, stability samples, and any other regulated test. It applies whether or not the affected batch has already been distributed. It covers the handling of out-of-trend (OOT) and aberrant results to the point of distinguishing them from an OOS. It does not govern manufacturing deviations or general nonconformances, which are handled under <<FILL: SOP-ID for deviation management>>, nor does it govern stability OOS trend evaluation beyond first triage, handled under <<FILL: SOP-ID for stability program>>.

3. Responsibilities

4. Definitions

• Out-of-specification (OOS): a reportable result that falls outside an established acceptance criterion (a registered specification, a compendial monograph limit, or a defined in-house limit) for the test performed.
• Reportable result: the final value the analytical method defines for comparison to the specification, for example the mean of a defined number of preparations, not each individual injection, weighing, or titration.
• Out-of-trend (OOT): a result inside specification but statistically atypical relative to the product’s own history. An internal alert signal, not a registered specification.
• Aberrant result: an atypical pattern in the raw data (a suspect single injection, an anomalous prep) that is investigated and accounted for, never silently deleted, and is not by itself an OOS.
• Assignable cause: a specific, identifiable error in execution, supported by contemporaneous evidence, that explains the result. The only valid basis to invalidate a result.
• Retest: analysis of additional preparations taken from the original, retained sample, under a pre-approved plan.
• Resample: drawing a new sample from the batch, justified only by documented evidence the original sample was unrepresentative.
• Additional testing: testing that characterizes the batch further without erasing the original OOS, which remains a real data point.

5. Procedure

5.1 Triage at the moment of the result

The analyst takes these actions immediately, before anyone reruns or reprocesses anything:

1. Confirm the value is a reportable result outside the acceptance criterion, not an aberrant single injection or an OOT. If it is aberrant or OOT, handle under section 5.7 and do not open an OOS in error.
2. Preserve everything: the original sample preparation, standards, mobile phase, the column where practical, and all instrument data and sequences. Do not discard solutions, do not reinject the vial “to confirm”, do not reintegrate, and do not adjust the system clock.
3. Notify the QC supervisor and QA the same working day and log the event in <<FILL: tracking system / log>> with a unique OOS identifier.

A reinjection or reintegration performed before review now sits in the record and must be accounted for forever. The single fastest way to corrupt an OOS investigation is to rerun on instinct.

5.2 Phase 1: laboratory assessment

Phase 1 is performed by the analyst and the QC supervisor under QA oversight, and is completed before Phase 2. It evaluates execution only. It does not involve additional testing of the original sample beyond what is needed to evaluate that execution. Document each of the following as checked, with the specific evidence reviewed:

1. Method: the correct, current, approved method and version was used, and the right reportable-value formula applied. Confirm against the controlled document, not a bench printout.
2. Instrument: inside its calibration and qualification window, with system suitability (SST) met before and during the sequence. An SST failure captured in the run is a clean assignable cause.
3. Sample preparation: weighed and diluted per the method, correct diluent and dilution factor, final concentration inside the validated range. Check the balance printout, glassware class, and the dilution chain.
4. Reagents and standards: in date, stored correctly, traceable to a certified source (USP, EP, JP, or a qualified working standard bridged to one), and prepared correctly. Check whether other samples in the run trend the same way.
5. Calculation: re-run the arithmetic independently. Check the dilution factor, potency, and response factor. For spreadsheets, confirm the formula was validated and not edited mid-run.
6. Analyst qualification: trained and qualified on this method and instrument. This is supporting evidence only, never a standalone assignable cause.
7. Raw data: open the actual chromatogram or trace. Look for co-elution, abnormal peak shape, baseline disturbance, wrong integration, detector saturation, or a missed injection. Confirm integration was per method and not adjusted to move the number, and that any reprocessing is shown in the audit trail.

Phase 1 closes with exactly one of two conclusions:

• An assignable laboratory cause was found, documented with evidence, source identified. A retest of the original sample, or recalculation of the existing data, is justified to obtain a valid result.
• No assignable laboratory cause was found. The result stands as a confirmed laboratory finding and Phase 2 begins.

“No error found, reanalysis requested” without the underlying documented review is not a Phase 1. There is no third conclusion such as “probably an error” or “result looks wrong”.

5.3 Phase 2: full investigation

Phase 2 starts when Phase 1 finds no assignable laboratory cause. QA leads it; manufacturing, process, and where relevant supplier quality and validation contribute. The question is now whether the result is telling you something true about the batch. Examine at least:

1. Production review: read the batch manufacturing record. Were all steps performed as written and in sequence, all parameters inside their proven acceptable ranges? Were deviations open during manufacture? A blend-uniformity OOS next to a deviation about a mixer running short is the root cause, not a coincidence.
2. Related batches and history: plot the result against the historical distribution for this product and process, adjacent lots, and any shift after a raw material change. A genuine product failure usually has company.
3. Raw materials and components: from approved suppliers and released lots, with no borderline or trending incoming results.
4. Process validation: does the validation package show this process reliably yields conforming product under the conditions actually used? An OOS inside a supposedly validated space is a stronger product signal than one tied to a clear excursion.

Use a formal root-cause method (fishbone across man, machine, method, material, measurement, environment, then 5-Whys or fault-tree to drive each branch to ground). Rule each candidate in with evidence or out with a reason. The output is a documented root-cause assessment that supports a defensible disposition.

5.4 Retest, resample, and additional testing

These are not interchangeable. Confusing them is a fast route to a data integrity finding.

• Retest is allowed only with a sound scientific reason and advance QA approval against a predefined plan. The plan states, and is signed before any new result is seen: who authorized it, the scientific rationale, the number of retests, who runs them (often a second qualified analyst), and how the result set will be evaluated. Deciding the acceptance logic after the numbers come back is testing into compliance and is prohibited.
• Resample is allowed only when the investigation identifies a specific, documented reason to believe the original sample was unrepresentative (a sampling error, evidence of non-homogeneity). The basis must be in the record before the new sample is pulled. Random resampling in hope of a better outcome is prohibited.
• Additional testing characterizes the batch; it does not erase the original OOS, which stays in the evaluation.

All results, including the original OOS and any outlier, are carried into the final statistical evaluation. An outlier test may be applied only where prospectively defined and scientifically justified, typically for higher-variability biological assays; it may never be used to invalidate an OOS for a chemical assay where the failing value is the result of interest.

5.5 Assignable cause and invalidation

A result may be invalidated only when an assignable cause is identified with documented, contemporaneous evidence. Defensible causes include a documented preparation error confirmed by balance printouts or raw data, an equipment malfunction or SST failure recorded at the time, a contamination of a standard confirmed by reanalysis, or a demonstrable integration or software error where reprocessing per the method (with the audit trail showing what changed and by whom) yields the corrected value. The following are not assignable causes: “the result does not fit our expectations”, “the analyst is new” with no specific error shown, “this instrument has given us trouble before” with no evidence for this run, an outlier designation built on the OOS value itself, or a passing retest where the original failure was never explained.

The analyst never invalidates their own result. QA owns the invalidation decision. You invalidate a result by explaining it, not by outvoting it.

5.6 Disposition and scope beyond the batch

1. QA determines disposition (release, reject, or further characterize) from the evidence: whether invalidation is genuinely defensible, what additional testing showed read in full, whether the OOS is isolated or part of a pattern, and the product type and patient risk. For a confirmed OOS with no assignable cause, rejection is the expected outcome in most cases.
2. Where the cause could have affected other lots, extend the investigation to every potentially affected batch, including distributed product, per 21 CFR 211.192. Escalate to field action or recall per <<FILL: SOP-ID for recall / field action>> where distributed product is implicated.
3. For investigational product, notify the sponsor’s quality function and evaluate impact on subject safety and study data; raise urgency where dosing is ongoing.
4. A rejected batch is segregated, labeled, and quarantined under control, and its disposition is documented in the batch record.

5.7 OOT and aberrant results

An OOT result is inside specification but statistically unusual against the product’s own history; investigate it under <<FILL: SOP-ID for OOT>> as an early-warning signal and feed it to periodic review, but do not log it as an OOS. An aberrant single injection or prep is investigated and accounted for in the raw data record, never silently deleted, and is not an OOS unless the reportable result itself is out of specification.

5.8 Timelines and documentation

1. Open and complete Phase 1 within <<FILL: target, commonly 1 business day>>.
2. Close the full investigation within <<FILL: target, commonly 30 calendar days>>. Any extension requires a documented justification and an interim status update before the target date. Late closure with no rationale is itself a finding.
3. Record the entire investigation on the OOS investigation record (the paired record template), signed and dated by the reviewer at each phase and approved by QA.
4. Retain the record per the records retention schedule, for not less than <<FILL: retention period>>.

6. Acceptance criteria

An OOS investigation is acceptable when all of the following are true:

• Triage preserved the original data and no rerun, reintegration, or clock change occurred before review.
• Phase 1 documents every assessment area as checked with specific evidence and reaches one of the two valid conclusions, with no in-between.
• Any retest or resample was pre-approved against a written plan, and all results, including the original OOS, are in the final evaluation.
• Invalidation, where claimed, rests on a documented assignable cause a reader who was not there could verify; the analyst did not invalidate their own result.
• The disposition follows from the findings, the investigation extended beyond the batch where the cause warranted it, and the record is complete, signed, dated, and closed within the committed timeline.

7. References

21 CFR 211.192 (production record review and investigation of discrepancies and failures). 21 CFR 211.165(f) (rejection of batches failing to meet specifications). 21 CFR 211.194 (laboratory records). FDA Guidance for Industry, Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production (originally October 2006; revised Level 2, May 2022). ICH Q7, section 8.15 (deviations) and section 2.2 (responsibilities of the quality unit(s), which cover review and approval of deviations and investigations) for active pharmaceutical ingredients. ICH Q10 (pharmaceutical quality system) for the systemic handling of quality events. EU GMP Chapter 1 and Chapter 6 (quality control); MHRA GxP Data Integrity Guidance and Definitions. 21 CFR 820.90 (control of nonconforming product) for finished-device testing; under the QMSR (effective February 2, 2026) this maps to ISO 13485:2016 clause 8.3, incorporated by reference.

Confirm the current version and clause numbers of each reference before issue.

8. Record generated: OOS investigation record (summary)

The full per-phase record is in the paired OOS investigation record template. The summary fields captured for tracking are:

9. Revision history

10. Approvals

Filled specimen

The following shows the summary record completed for an example finished-product assay, so you can see the level of detail an inspector expects. The company, system, and numbers are illustrative; replace them with your own.

In this example the supervisor and analyst reviewed the run: SST passed (standard replicate RSD 0.4 percent), the standard was in date, the balance log and worksheet checked out. The chromatogram showed the main peak shoulder merged with a late-eluting peak, and the integration baseline had been dropped manually so part of the analyte area was excluded. Reprocessing per the method gave 99.1 percent, and the audit trail confirmed who applied the manual integration and when. The original 93.4 percent was documented and retained, the corrected value supported release, and a CAPA addressed integration practice and analyst retraining. Finding to documented cause to corrected value to retained original is exactly the sequence a reviewer is expected to demonstrate.

Common inspection findings this SOP prevents

• OOS results invalidated without a specific, documented assignable cause (the signature finding of data manipulation).
• Reinjecting or reintegrating the original sample before Phase 1 is closed, contaminating the investigation.
• A passing retest used to release a batch while the original failure sits unexplained, that is, retesting into compliance.
• Phase 1 recorded as “no error found” with no evidence that each assessment area was genuinely checked.
• The original OOS dropped from the final evaluation, or an outlier test used to discard a failing assay value.
• Root cause stated as “human error” naming the analyst but never the actual error.
• The investigation evaluated in isolation, never extended to related batches or distributed product when the cause warranted it.
• The investigation closed past the SOP timeline with no documented justification.

How to adapt this SOP

1. Set your document number, owner, and effective date in the header, and your tracking system in section 5.1.
2. Point the cross-references in sections 2, 5.6, and 5.7 to your real deviation, recall, stability, and OOT procedures.
3. Set your Phase 1 and full-investigation timelines in section 5.8 to whatever your quality system commits to, and hold to it; an inspector holds you to your own SOP.
4. Define “reportable result” for each method in your method documents so analysts know exactly what is compared to specification.
5. If you use predefined outlier procedures for biological assays, reference each validated procedure and the assay types it covers, and state plainly that it never applies to chemical assays.
6. Confirm every regulation in section 7 against the current published version before issue.