What Is Proficiency Testing? A PT and ILC Guide for Labs

Proficiency testing (PT) is the evaluation of a laboratory's performance against pre-established criteria by means of interlaboratory comparisons. An interlaboratory comparison (ILC) is the organisation, performance and evaluation of measurements or tests on the same or similar items by two or more laboratories under predetermined conditions. Every PT is an ILC, but an ILC is not PT unless it includes a performance evaluation.

This guide explains why PT is needed, how it works, how z-scores, En numbers and zeta scores are calculated and interpreted with numerical examples, what to do after an unsatisfactory result and how to choose a provider.

Why is proficiency testing needed?

• Monitoring the validity of results: ISO/IEC 17025 requires laboratories to monitor the validity of their results and, where available and appropriate, to take part in PT or ILCs.
• Accreditation: accreditation bodies expect regular, planned PT/ILC participation for the technical areas in scope. For frequency and planning rules, refer to your accreditation body's current guidance and, internationally, to the ILAC P9 policy.
• Independent assurance: internal quality control shows a laboratory is consistent with itself; PT shows its results agree with other laboratories and the assigned value.
• Early warning: problems with methods, instruments or staff are found before they become customer complaints.

How does proficiency testing work?

1. The provider publishes the scheme and technical protocol: items, parameters, dates and evaluation method.
2. The laboratory registers and receives the item.
3. The item is measured with the laboratory's routine method by its routine staff. PT items get no special treatment.
4. The result, and the uncertainty where required, is submitted to the provider by the deadline.
5. The provider evaluates the results statistically and issues a report.
6. The laboratory reviews the report, interprets its result and starts corrective action where needed.

Types of schemes

How is the assigned value determined?

The laboratory's result is compared with the assigned value, the value taken as correct. Approaches described in ISO 13528 include:

• Known formulation (if the item was prepared from known quantities)
• The value of a certified reference material
• A value determined by an expert reference laboratory (common in calibration)
• Consensus of expert laboratories
• Consensus of participants (using robust statistics resistant to outliers)

Performance statistics: z-score, En and zeta

z-score

z = (x − X) / σ_pt, where x is the laboratory's result, X the assigned value and σ_pt the standard deviation for proficiency assessment.

Example: assigned value 10.00 mg/L, σ_pt = 0.20 mg/L, laboratory result 10.46 mg/L: z = (10.46 − 10.00) / 0.20 = 2.3. The result is questionable; on its own it is not a failure, but the cause should be investigated and monitored in later rounds.

Some providers define criteria differently; the criteria in the scheme protocol apply.

En number (normalised error)

The most common statistic in calibration comparisons, taking the uncertainty of both sides into account:

E_n = (x − X) / √(U_lab² + U_ref²), where U are expanded uncertainties. |E_n| ≤ 1 is satisfactory and |E_n| > 1 unsatisfactory.

Example: the reference laboratory measured a gauge block as 100.0000 mm with U_ref = 0.0005 mm; a participant found 100.0012 mm with U_lab = 0.0008 mm. E_n = 0.0012 / √(0.0008² + 0.0005²) = 0.0012 / 0.00094 = 1.27. The result is unsatisfactory. The problem is not only the size of the deviation but that the laboratory's stated uncertainty does not cover it. Underestimated uncertainty is one of the most common reasons for a failed En number; see our measurement uncertainty guide.

Zeta (ζ) score

ζ = (x − X) / √(u_lab² + u_X²). Similar to En but calculated with standard uncertainties and interpreted like a z-score. It shows whether the laboratory's stated uncertainty is consistent with its result.

What to do after an unsatisfactory or questionable result

1. Check the data first: look for unit, transcription, calculation or reporting errors. Many failures come from transfer errors rather than the measurement itself.
2. Analyse the root cause: review method, instruments, calibration, reference materials, sample preparation, staff and environmental conditions in turn.
3. Assess the impact on past results: could the same problem have affected results already issued to customers?
4. Implement and record corrective action.
5. Verify effectiveness: show the fix worked through internal quality control, re-measurement or the next PT round.

Questionable results should also be recorded and trended across rounds; repeated questionable results in the same direction may indicate a systematic problem.

How to choose a PT provider

• Providers accredited to ISO/IEC 17043 are preferred.
• The scheme should cover the parameters, matrix and measuring range in the laboratory's scope.
• The number of participants and the way the assigned value is set affect the reliability of the evaluation.
• Dates and reporting time should fit the laboratory's plan.
• If no suitable scheme exists, bilateral or multilateral comparisons with other laboratories can be planned.

How to plan PT/ILC participation

Laboratories usually prepare a multi-year PT/ILC plan covering the accreditation cycle. It shows which scheme will cover each technical area or sub-discipline in which year, together with previous results and risk assessment. Areas with failed results or newly added areas are planned more often.

For organisers of PT and ILC schemes

Running a scheme is a long process: protocol, applications, quotations, sample circulation, result collection, statistical evaluation and reporting. Our YT-LAK software manages proficiency testing and ILC programmes in a shared structure: with sequential circulation and parallel distribution you can see which laboratory holds the item at any moment, laboratories submit results from their own accounts, and evaluation uses z-score, En, deviation or consensus depending on the scheme. On the laboratory side, KalDataV2 can open a corrective action directly from a proficiency test result.

Frequently Asked Questions

What is proficiency testing?
The evaluation of a laboratory's performance against pre-established criteria by means of interlaboratory comparisons. It shows whether the laboratory's results agree with other laboratories and the assigned value.

What is the difference between PT and an ILC?
An ILC is laboratories measuring the same or similar items and comparing results. PT turns ILC results into a performance evaluation against pre-established criteria.

How is a z-score interpreted?
Generally |z| ≤ 2 is satisfactory, between 2 and 3 questionable, and |z| ≥ 3 unsatisfactory. The criteria in the scheme protocol apply.

What does an En number above 1 mean?
The difference between the laboratory result and the assigned value cannot be explained by the stated uncertainties of both sides, so the result is unsatisfactory. Check the deviation and whether uncertainty was underestimated.

How often should a laboratory take part in PT?
Frequency depends on the accreditation body's rules, the technical field and risk assessment. Refer to your accreditation body's current guidance and the ILAC P9 policy.

Should a PT provider be accredited?
Providers accredited to ISO/IEC 17043 are preferred. If no suitable accredited scheme exists, other options can be considered with justification.

What should be done after a failed PT result?
Check data and calculations first, analyse the root cause, assess the impact on customer results, implement corrective action and verify its effectiveness.

What is ISO/IEC 17043?
The international standard setting general requirements for the competence of proficiency testing providers. The current edition was published in 2023.