What clause 7.7 requires
Clause 7.7, “Ensuring the validity of results”, has two complementary parts: what you can check yourself at any time, and what you can only see by comparison with others.
| Part | Content |
|---|---|
| 7.7.1 Internal monitoring | Ongoing measures, where possible and appropriate: reference materials or quality control materials, control samples, substitute or check instruments, functional checks of measuring and test equipment, control charts, intermediate checks, replicate tests using the same or a different method, retesting of retained items, correlation of results for different characteristics, review of reported results, intralaboratory comparisons, testing of blind samples |
| 7.7.2 External comparison | Performance monitoring by comparison with results of other laboratories, where available and appropriate: participation in proficiency testing and/or in interlaboratory comparisons other than proficiency testing |
| 7.7.3 Evaluation | Monitoring data is analysed and used to control and, where applicable, improve the laboratory's activities. If results fall outside pre-defined criteria, appropriate action is taken to prevent incorrect results from being reported |
Note the link to clause 8.5 (risks and opportunities): the monitoring programme is not a fixed list. It should reflect risk, test volume and the importance of the method, and be planned and reviewed.
Proficiency test or interlaboratory comparison: the difference
The terms are often used interchangeably, but the standard separates them. A proficiency test is an interlaboratory comparison whose organisation and evaluation meet formal requirements; a comparison can also run informally between laboratories.
| Type | Characteristics |
|---|---|
| Proficiency testing (PT) | Run by a provider, ideally accredited to ISO/IEC 17043. Homogeneous, stable samples, an assigned value, statistical evaluation per ISO 13528, anonymised results. Recognised by accreditation bodies as evidence |
| Interlaboratory comparison (ILC) | Several laboratories measure the same or similar items and compare results. Effort and weight of evidence depend on how well the items, assigned value and evaluation are defined. Useful when no suitable PT exists for your field |
| Comparison with a reference laboratory | Your sample or standard is measured by a laboratory with higher accuracy. More a traceability proof than a performance comparison, but often accepted as evidence for 7.7.2 when justified |
Evaluating results: z-score and En number
The provider normally evaluates your result. You should still understand the figures, because the required reaction depends on them.
| Statistic | Meaning and evaluation |
|---|---|
| z-score = (x − X) / σ | Deviation of your result x from the assigned value X, relative to the standard deviation for proficiency assessment σ. |z| ≤ 2: satisfactory. 2 < |z| < 3: questionable (warning signal). |z| ≥ 3: unsatisfactory (action signal) |
| En number = (x − X) / √(U²lab + U²ref) | Takes into account the expanded measurement uncertainty of your laboratory and of the assigned value. |En| ≤ 1: satisfactory, |En| > 1: unsatisfactory. Common for calibration laboratories and quantities reported with an uncertainty |
| Zeta score | Like the z-score, but using the standard uncertainty of your laboratory and the assigned value instead of σ. Shows whether your stated uncertainty budget fits the observed deviation |
A single unremarkable value proves little. Monitoring becomes meaningful through the history: if several consecutive rounds show z-scores with the same sign, that points to a systematic bias even if each value is below 2. Keep a per-method overview for that reason.
Planning and carrying out participation
- 1
Group your scope into sub-disciplines
Cluster methods with similar technique and measured quantity. Each group should have a participation within the cycle — not every single method.
- 2
Clarify your accreditation body's requirement
Read its policy on ILAC P9 and participation frequency. It defines how many participations per cycle and sub-discipline are expected.
- 3
Find a suitable provider
Provider databases (e.g. EPTIS) list schemes by quantity and matrix. Prefer providers accredited to ISO/IEC 17043. If none exists, justify that and plan an interlaboratory comparison or a reference laboratory comparison.
- 4
Create an annual plan with owners
Date, scheme, method, responsible person, budget. Link the plan to the audit programme so gaps become visible.
- 5
Treat the sample like a routine sample
Same methods, equipment and personnel as in daily work. Do not measure repeatedly and report the best result unless routine work does the same — otherwise the participation shows something other than your performance.
- 6
Evaluate and record the result
File the provider's report, transfer statistic and assessment per method into the overview, review the uncertainty and sign off the evaluation.
- 7
On a flagged result, find causes and make actions effective
See the next section. Also assess earlier results produced with the same method, equipment or personnel.
When the result is flagged
An unsatisfactory result is not a scandal; it is the purpose of the exercise: it found an error before a customer did. What counts is the reaction.
- Rule out transcription and reporting errors first (unit, rounding, swapped sample).
- Then work through the cause chain: sample handling, instrument and its calibration status, method and evaluation, environmental conditions, personnel, uncertainty calculation.
- Check whether the deviation was also visible in internal monitoring (7.7.1), and why not.
- Assess the impact on reported results: which test reports came from the same method and instrument since the last unremarkable monitoring result? Where necessary inform customers and amend reports (clauses 7.10 and 7.8.8).
- Define and implement a corrective action and verify its effectiveness — for example by a repeat on retained material or participation in the next round.
- Record everything: cause, decision, actions, effectiveness evidence. Assessors read this file first.
Typical assessment findings
- No plan covering the scope of accreditation; participations happen ad hoc or by budget.
- Participation without a documented evaluation — the provider's report is in the folder, but nobody signed it off.
- Questionable or unsatisfactory results without a root-cause analysis.
- Actions without an effectiveness check.
- No internal monitoring between external rounds; for methods without PT, no justified alternative.
- The sample was treated differently from routine samples, for example through replicate measurements or especially experienced staff.
Here too, plan, participation, evaluation, deviation and action are usually scattered. Keeping them in the same place as the method and the affected test orders lets you determine the effect of a flagged result on earlier work in minutes rather than days.