ISO 17025 proficiency testing: interlaboratory comparison, the PT plan and acting on results

What does ISO 17025 proficiency testing require?

ISO/IEC 17025:2017 does not have a clause called proficiency testing. The obligation lives inside clause 7.7, Ensuring the validity of results, which asks a laboratory to monitor its own performance and, where available and appropriate, to compare its results with those of other laboratories. A note to that requirement adds that providers meeting ISO/IEC 17043 are considered competent, which is why most laboratories look first for a scheme run by an accredited provider.

Two terms need to be kept apart. Proficiency testing, as ISO/IEC 17043 defines it, is the evaluation of a participant's performance against criteria fixed in advance, by means of interlaboratory comparisons. An interlaboratory comparison is the wider activity: several laboratories measure the same or similar items and the results are compared.

The PT provider is also an external provider under clause 6.6, Externally provided products and services. Choosing it, checking its competence and recording that evaluation belong to the same system that governs your calibration suppliers and reference material producers.

Proficiency testing or interlaboratory comparison: which one?

Start with the scheme catalogues of accredited providers. A scheme is a better fit when its items match your matrix, its levels fall within your working range and its round dates let you respond before the next assessment. ISO/IEC 17043 describes several designs in its Annex A, including simultaneous, sequential, split-level and sampling schemes, so ask which design a scheme uses before you register.

Some methods have no scheme at all: a rare matrix, a customer-specific test, an unusual calibration range. Here the "where available and appropriate" wording matters, and Annex A of ISO/IEC 17043 recognises alternative interlaboratory comparisons. A defensible comparison has a documented design, an agreed way of setting the reference or assigned value, performance criteria written before any results are seen, and at least one other competent laboratory taking part.

Where neither route is possible, say so in the plan and show how the other monitoring activities under clause 7.7, such as control charts, cover the risk instead. A real gap, explained, is better than a comparison that proves nothing.

Building a proficiency testing plan

A plan turns ad hoc registrations into a programme you can defend. It lists the method families in the scope of accreditation, the scheme or comparison chosen for each, the planned frequency, the provider and the evidence of its competence. Group methods sensibly: where one technique, one analyst team and one set of equipment serve several analytes, a scheme covering the most demanding analyte can speak for the group, provided the plan explains the reasoning.

Frequency is set partly by risk and partly by the accreditation body. Bodies publish their own policies. ANAB, for example, states that its accredited laboratories must take part in proficiency testing every year, and the NABL FAQ asks applicant laboratories to have participated successfully in at least one programme before accreditation is granted. Read your own body's policy before fixing the plan.

Schedule the rounds so that at least one result for each family is back and reviewed before the on-site assessment; round dates follow the provider's calendar, not yours, which is one reason the hub gives six to twelve months for a first accreditation.

Reading a PT report and the performance scores

A report issued under ISO/IEC 17043 has to state, among other things, the assigned value and how it was determined, its traceability where relevant, the standard deviation for proficiency assessment, the statistical procedures used and each participant's performance.

The statistical reference most providers use is ISO 13528:2022. It sets out several routes to an assigned value: formulation, a certified reference material, results from one laboratory, a consensus of expert laboratories or a consensus of participant results. It also describes the performance statistics you will meet, namely z, z′, zeta and En scores. Which one appears, and the limits that separate acceptable from unacceptable, is the provider's choice and must be stated in the report. For calibration schemes, ISO/IEC 17043 requires the assigned value to be metrologically traceable.

ISO/IEC 17043 lists the validation of uncertainty claims among the purposes of proficiency testing. A laboratory whose results agree with the assigned value but whose reported uncertainty looks implausibly small should review its budget; the measurement uncertainty guide covers how.

When a result fails: investigation and corrective action

An unsatisfactory result means a reported value may have been wrong, so treat it as nonconforming work and find the cause: was the item handled, stored and prepared as the instructions said; were units, dilution factors and transcription correct; was the equipment in calibration; was the analyst authorised; does the failure point to the method itself?

The clause routes for this are nonconforming work (7.10) and corrective actions (8.7 under Option A). How findings are raised and closed during the assessment is covered in the nonconformities guide; the internal audit guide explains how to test the PT trail yourself.

The record assessors ask for

Assessors tend to ask for the plan, pick one method from the scope and follow it through: the registration, the report, the signed review and, if the result was poor, the investigation and corrective action. They also check that the provider was evaluated as an external provider and that the plan has actually been followed.

• The PT plan, dated and approved, with coverage mapped to the scope of accreditation.
• Provider evaluations, including evidence of ISO/IEC 17043 accreditation and its scope.
• Every report received, with a documented review by an authorised person.
• Investigations, impact assessments and corrective actions for unsatisfactory results.
• A summary of PT performance taken into management review.

ISO 17025 proficiency testing: questions buyers ask