Proficiency testing under ISO 17025: when is participation mandatory?
Proficiency testing (PT) is a study in which several laboratories receive the same samples or measurement objects, and compare their results with each other. The RvA (Dutch Accreditation Council) requires accredited laboratories to take part in PT studies. This also applies to calibration laboratories, as the RvA states on its calibration page . The requirement is in section 7.7 of ISO/IEC 17025, on ensuring the validity of results. Also read the scope of your accreditation and the checklist for your internal controls.
The RvA is strict about this. According to its explanatory notes (INF002, section 7.7.2), you must take part in a PT or an interlaboratory comparison (ILC) for each activity or calibration. An ILC is a comparison that laboratories organise among themselves. Anyone applying for or extending accreditation must already have taken part. The results must then have been evaluated and meet the requirements.
The purpose is simple. When your laboratory gives a result, the client wants to know whether that result is correct. A proficiency test shows whether your outcome falls within the expected spread of other laboratories. The RvA publishes an overview of PT study organisers .
What is the difference with internal quality control?
Internal quality control looks at your own work, every day. Proficiency testing looks at your work compared with others, periodically. Both are needed, but they answer different questions.
| Question | Internal control | Proficiency testing |
|---|---|---|
| Who sets the reference value? | Your own laboratory, with a control sample or standard | An organiser, based on the joint results or a set value |
| How often? | With each batch or according to your procedure | Periodically, according to a schedule you choose in advance |
| What does it show? | Whether your work is stable today | Whether your work is comparable with that of other laboratories |
| What do you do with a deviation? | Correct it directly, and redo the work if needed | Investigate the cause, improve, and sometimes repeat |
| Which material? | Reference or quality control material is required, but it does not have to be certified | Samples from the organiser |
What the auditor wants to see
The auditor does not expect a perfect score. They want to see that you take part, that you read the results, and that you act on a deviating outcome.
Taking part in one proficiency test and then doing nothing further is not enough. The auditor looks for a continuous pattern: participation, analysis and follow-up.
Evidence an auditor usually asks for:
- A list of the proficiency tests you took part in, with date and subject.
- The organiser’s report, with your result and the reference value.
- Your analysis of the result: within or outside the permitted spread, and why.
- If there was a deviation: the root cause analysis and the action you took.
- Where possible, a look at trends over time. The RvA asks for this when evaluating PT and ILC results (INF002, 7.7.3).
Fictional example: a water laboratory in the second round
A fictional laboratory for drinking water analysis takes part in a proficiency test for nitrate every year. In the first year, the result falls just outside the spread. The laboratory does the following.
| Step | What the laboratory does (fictional) | Evidence |
|---|---|---|
| 1. Detect | The deviation is visible in the proficiency testing report | Report with z-score |
| 2. Investigate | Comparison of the calibration and processing process | Investigation note |
| 3. Cause | Outdated calibration series in the spectrophotometer | Analysis with date |
| 4. Action | New calibration series, extra check with each batch | Revised procedure and log |
| 5. Follow-up | The next proficiency test falls within the spread | Report from the following year |
The result is not the action on its own. The auditor reads the chain from the signal to the improvement.
How do you read a proficiency testing report?
A proficiency testing report is often long and full of numbers. You do not need to understand everything to use it well. Focus on four parts.
| Term | What it means | What to look at |
|---|---|---|
| Reference value | The value against which the results are compared, set by the organiser | How was that value determined, and is it reliable? |
| Your result | What your laboratory reported | Does the number match what you put in your report? |
| Spread | How far apart the results of all participants lie | How wide is the permitted band? |
| Deviation score | A measure of how far your result lies from the reference. Many organisers use a z-score | Does the score fall within the limit set in advance? |
A score outside the limit is not the end. It is the starting point of your investigation. See the steps under nonconformities.
If you are a calibration laboratory, the measurement uncertainty of the individual laboratories is included in the evaluation. This is set out in INF002 under section 7.7.3. A result with a small uncertainty therefore calls for a stricter look than a result with a large one.
Check the unit and the range. A result that looks good may relate to a different concentration range from your accredited method. Always compare the scope of the study with your own scope first.
How do you plan proficiency testing under ISO 17025?
Not every proficiency test fits your scope. Choose studies that match your accredited methods and ranges. A study on a different matrix or a different concentration range says little about your own work.
- Make a list of your accredited methods and ranges.
- Find organisers that run studies for those methods, for example through the RvA list.
- Record for each activity which study you take part in and when. This is your PT plan. The RvA asks for this plan with the application (ASR008 describes the components).
- Note why the study fits your scope.
If you have several laboratories within one organisation that perform the same tests, each laboratory must take part periodically. The same applies to work at client sites or in a mobile laboratory. For calibration laboratories, the RvA uses the HCS codes from the scope as the basis for participation (ASR106, section 4.1).
A fictional example: a laboratory with one site and a mobile measuring van calibrates mass and pressure. It plans a study for both working areas and also takes part separately for the measuring van. In this way, the plan covers both the fixed laboratory and the location at the client.
Common mistakes
- Only taking part when it suits you. Laboratories that only join studies they are good at stand out in the planning.
- Ignoring the score. A report with a deviating score that is not discussed in the management review stands out.
- No link with the scope. Taking part in a study of a method you have not been accredited for provides little evidence for your accreditation.
- No participation before the application. The RvA expects you to have taken part and assessed the results when you apply.
- Cost as the only criterion. Cheap studies are not automatically suitable. Check the quality of the organiser and the relevance to your method.
Frequently asked questions
Must I take part every year? The RvA speaks of periodic participation. Record the frequency and the justification in your PT plan, and if in doubt, ask your assessor for their view.
What if no proficiency test is available for my method? The standard also accepts an ILC with other laboratories. Organise one yourself and record the choice. Reference material belongs to your internal quality control and does not replace comparison with other laboratories.
Does a proficiency test count as evidence for my accreditation? It is part of the evidence. It does not replace the validation of your method or your internal control.