Start from the result, not from a ticket
Out-of-specification (OOS) work begins when a result does not meet the approved specification. The quality event should be born from that result — with sample, method, instrument, analyst, and specification version already attached.
When OOS is a generic ticket in another system, investigators spend the first day rebuilding context. That delay is not only inefficient. It is how the laboratory record and the quality record diverge.
OOS is not the same as a deviation
OOS is a result against a limit. A deviation is a departure from an approved process. They often meet — a method step skipped, then a failing value — but they should remain distinguishable. Inspectors ask whether you can tell a laboratory failure from a process failure.
Out-of-trend (OOT) is a third category: within limits, historically strange. If your LIMS only has a single “issue” object, you will encode those distinctions in titles. Titles are not a quality system.
Investigation needs evidence objects
A usable workflow lets QA attach retests, instrument status, chromatograms or files, and second-person review without leaving the event. Phase-appropriate steps (lab hypothesis, full investigation) should be visible as status, not as folklore.
- Link to the triggering result and specification version
- Record hypothesis and findings as attributable text
- Capture retest decisions and outcomes
- Keep the original failing value; do not overwrite it to “make the CoA clean”
CAPA is closure with a memory
Corrective and preventive action exists because investigations that only explain the past will repeat. CAPA should state the action, the owner, the due date, and the effectiveness check. Closing a CAPA because the meeting ended is not effectiveness.
Lab-originated CAPA belongs next to the samples and methods that produced it. Some organizations also operate an enterprise QMS. That is fine — if the boundary is explicit. Two unrelated CAPA numbers for one failing assay is how findings are born.
What “good” looks like in a demo
Ask the vendor to fail a result, open OOS, record a deviation if process error is involved, launch CAPA, and show the sample’s history still listing the original value. Then ask how a CoA is blocked or annotated while the event is open.
Aureon LIMS treats QC/SPC, OOS, deviation, and CAPA as core quality operations — bound to the laboratory record — not as a separate “quality app” you integrate someday.
Metrics QA should actually watch
Volume of OOS by method, time to close, recurring root causes, and CAPA effectiveness failures are more useful than a vanity “quality score.” If analytics exist, they should be assistive. They should not auto-close events or auto-release product.