EA-4/21:2026 — new edition of the EA guidelines on small interlaboratory comparisons (up to seven participants): risk assessment, three assigned-value scenarios and EN ISO/IEC 17043:2023 requirements
On 21 April 2026 European co-operation for Accreditation (EA) issued the second edition of EA-4/21 INF, endorsed on 26 March 2026 and applicable from publication. It tells accreditation body assessors when a comparison organised by a few laboratories among themselves can be used to monitor results under EN ISO/IEC 17025 or EN ISO 15189; organisers and participants may use it too.
- A small interlaboratory comparison has at most seven participants, including the organiser. Laboratories use them when proficiency testing is unavailable or inappropriate (ILAC P9:01/2024). EA stresses that the document does not replace the accreditation of a proficiency testing provider under EN ISO/IEC 17043.
- With few results, the uncertainty of the mean and standard deviation grows and the data distribution and outliers cannot be assessed reliably — which is why ISO 13528 recommends that small comparisons should not be evaluated or scored using statistics derived from participants’ results. The new edition introduces a risk assessment by the organiser: is there an independent, traceable reference value and an independent evaluation criterion (SDPA)?
- Three scenarios: (1) a traceable assigned value independent of the data (CRM, reference laboratory, reference method) — zeta or En scores are possible, and z scores with an independent SDPA; (2) another assigned value, e.g. an informative reference material value, an earlier comparison or a participant at a higher metrological level — scores after a risk assessment; (3) no reliable assigned value — scores for information only, with conclusions on repeatability, reproducibility and uncertainties. In the 2018 edition, scenario 2 allowed, in exceptional cases, an assigned value derived from participants’ results.
- The chapter on EN ISO/IEC 17043:2023 requirements lists what to assess: the comparison plan, homogeneity and stability of items, separation of measuring and organising staff and confidentiality of assigned values, and a report with anonymised results. A small comparison must be part of the proficiency testing participation plan, internal audit and management review, and an unsatisfactory result is treated as nonconforming work.