OOS / OOT

Out of Specification / Out of Trend

The governing principle comes from the 1993 Barr Laboratories judgment and is now codified in FDA guidance: an OOS result is presumed valid until an investigation demonstrates otherwise. You may not discard a failing result because a passing retest exists. The prohibited behaviour has a name — testing into compliance — and it means repeating analysis until a favourable number appears, then releasing on that.

The investigation runs in two phases. Phase I is the laboratory assessment, conducted before any retesting: the analyst and supervisor review the data, calculations, instrument performance, standards, and technique to see whether an assignable laboratory error explains the result. Only a confirmed, documented laboratory error justifies invalidating it. If Phase I finds no assignable cause, Phase II opens a full investigation extending into production — batch record, process parameters, equipment, materials, and other batches potentially affected.

Several practices are tightly constrained. Retesting must follow a pre-defined, scientifically justified plan, not be repeated at will. Resampling is only appropriate where the original sample is shown to be unsuitable. Averaging a failing result with passing ones can mask a genuine problem and is generally not acceptable for release decisions. Outlier tests have limited application and cannot be used to reject a chemical-assay result on statistical grounds alone. OOT results — inside specification but off the expected trend — carry no automatic reject, but they are the early signal a robust system investigates.

• —The original result stands until an investigation shows an assignable cause — no testing into compliance.
• —Phase I: laboratory investigation before any retest; only a documented lab error can invalidate a result.
• —Phase II: full-scale investigation into production, materials, equipment, and other affected batches.
• —Retesting needs a pre-defined, justified plan; resampling requires the original sample to be unsuitable.
• —Averaging a failing result with passing ones can mask the problem; outlier tests have limited use.
• —OOT is inside specification but off-trend — the early warning, and a maturity indicator when acted on.

FDA guidance, Investigating Out-of-Specification (OOS) Test Results for Pharmaceutical Production (2006; revised 2022); 21 CFR 211.192 (production record review) and 211.160 (laboratory controls); EU GMP Part I Chapter 6; MHRA guidance on OOS/OOT investigation; the principles trace to United States v. Barr Laboratories (1993).