GIFSQ Tools · Module 14 — LIMS & Quality · Tool 271 of 500
Work through an out-of-specification result: how far the original sits over the limit, what the two retests say, and whether the failure is confirmed.
Try sample values — fills the form with the worked example below.
–
Original deviation (%)
–
Retest mean
–
Retest deviation (%)
–
Retest vs original (%)
Equation: Original deviation (%) = (original result − limit) ÷ limit × 100. Retest mean = (retest 1 + retest 2) ÷ 2. Retest deviation (%) = (retest mean − limit) ÷ limit × 100. Difference (%) = |retest mean − original result| ÷ original result × 100. The failure is confirmed when the retest mean is still above the limit.
Method: Out-of-specification investigation per ISO/IEC 17025:2017, clause 7.10 (nonconforming work): an out-of-specification result is investigated, retested where the procedure allows, and the confirmed outcome — not the first number — decides the batch.
Assumptions: the retests are fresh determinations on the same sample under the same method; the limit is the upper specification limit for this parameter.
Limitations: two agreeing retests confirm the result, but they do not find the cause — a confirmed failure still needs a root-cause investigation before the batch is dispositioned.
Rounding: deviations to 2 decimal places; retest mean to 2 decimal places; intermediates unrounded.
This is a calculation aid, not a laboratory report — confirm against your accredited method before citing a value. Records are kept only in this browser on this device (localStorage); this is a device-only record aid, not a shared multi-user LIMS, and nothing is sent to any server.
Method: Out-of-specification investigation per ISO/IEC 17025:2017, clause 7.10 (nonconforming work): an out-of-specification result is investigated, retested where the procedure allows, and the confirmed outcome — not the first number — decides the batch.
Assumptions: the retests are fresh determinations on the same sample under the same method; the limit is the upper specification limit for this parameter.
Limitations: two agreeing retests confirm the result, but they do not find the cause — a confirmed failure still needs a root-cause investigation before the batch is dispositioned.
Rounding: deviations to 2 decimal places; retest mean to 2 decimal places; intermediates unrounded.
This is a calculation aid, not a laboratory report — confirm against your accredited method before citing a value. Records are kept only in this browser on this device (localStorage); this is a device-only record aid, not a shared multi-user LIMS, and nothing is sent to any server.
Worked example: Original 12.0 against a 10.0 limit, retests 11.8 and 12.2: original deviation 20.00%, retest mean 12.00, retest deviation 20.00% — the failure is confirmed.
Enter values above to see results.
Never discard the original result or average it away: the investigation record keeps the original, both retests and the decision. Tool 271 of 500 — GIFSQ
