External quality assessment is how a laboratory proves to someone outside its own walls that its results are right. ISO 15189:2022 requires it for every examination and now brings point-of-care testing into the standard as well. This page is about planning it: which programmes a laboratory needs, how often, and how to combine providers into a portfolio that covers every discipline.
Seven detailed pages go deeper than this overview. Each answers one question a laboratory has to settle on the way to reliable results and accreditation.
Control rules, limits, sigma metrics, third-party controls and the 2025 survey results.
The clauses on QC and EQA, what assessors check, and how CAP accreditation relates.
Analytes, programmes, frequencies and pitfalls for nine disciplines.
Eight EQA providers, portfolio levels, selection criteria and cost drivers.
z-score, SDI, peer groups, root causes and the steps after a failed result.
Five phases from test menu to routine, digital set-up and roles.
Thirty-two QC, EQA and accreditation terms in plain language.
Worse than most laboratories assume. In the 2025 Great Global QC Survey by Westgard QC, one laboratory in three reported being out of control every day or several times a day, up from about one in four in 2021. Most of these alarms are not real analytical errors but false rejections caused by badly chosen control limits — and the usual reaction, repeating the control until it passes, hides the few errors that are real.
of laboratories are out of control daily or several times a day (2021: about 23 %)
simply repeat the control when it fails, instead of looking for the cause first
still apply 2 SD limits to every test — the main source of false rejections
release patient results even though the QC run has failed
Survey results of 1,280+ self-selected laboratories worldwide (Westgard QC, August 2025); Central Asia is not reported separately.
Internal QC catches errors run by run. EQA, the subject of the rest of this page, checks whether the whole system is right compared with everyone else. Rules, limits, sigma metrics and regional survey results are on Internal QC.
Control samples of unknown value are analysed like patient samples, and the results are compared with reference values and with other laboratories using the same method. That comparison — EQA, or proficiency testing (PT) — finds what internal quality control is blind to.
A calibration bias shifts every result in the same direction. Internal QC drifts with it; an outside reference does not.
A patient who moves between institutions needs results that mean the same thing in both. Interlaboratory comparison is how that is shown.
Accreditation bodies read EQA performance as evidence that calibration, reagents and staff work together as intended.
Three layers of one quality system
Clause 7.3.7.3 makes participation in EQA or proficiency testing mandatory for every examination the laboratory performs. The 2022 edition (adopted in Russia as GOST R ISO 15189-2024) puts more weight on risk and, newly, includes POCT. A quality plan has to cover six points:
| 01 · Participation | EQA / PT for every type of examination |
|---|---|
| 02 · Alternatives | Another form of interlaboratory comparison where no programme exists |
| 03 · Analysis | Systematic review of results, including trend detection — not just pass/fail |
| 04 · Corrective action | Documented CAPA for every unacceptable result |
| 05 · Risk | Risk-based prioritisation of controls |
| 06 · POCT | Point-of-care testing inside the quality management system — new in the 2022 edition (formerly ISO 22870) |
Clause by clause, with what assessors look at first: ISO 15189 and accreditation.
Each discipline has its own failure modes, and the programmes are built around them. The frequencies below are per year and describe typical schemes, not a fixed offer. Analytes, sample formats and pitfalls per discipline are on EQA programmes by discipline.
Enzymes, substrates, lipids, electrolytes, HbA1c. The largest share of laboratory test volume falls here — the area where an undetected error reaches the most patients.
CAP chemistry and HbA1c surveys 3–4× · Randox RIQAS up to 24× · Bio-Rad EQAS 12× · INSTAND 2–4× · Labquality 4–6×
Without external control: a systematic haemoglobin bias misclassifies anaemia, wrong platelet counts distort haemostasis, a faulty differential misses blast cells.
CAP haematology surveys 3–4× · Randox RIQAS 12× · UK NEQAS 6× · INSTAND 4×
Pathogen identification and susceptibility testing (MIC, MRSA, ESBL, carbapenemases), mycology, blood cultures, serology. Antimicrobial resistance makes accuracy here a public-health question.
CAP · UK NEQAS · INSTAND (incl. MALDI-TOF) · Labquality · QCMD
The most method-dependent area: results for one analyte differ between platforms, so peer grouping by method and analyser is essential. An hs-cTnI error can mean a missed infarction.
CAP · Randox RIQAS Cardiac · Bio-Rad EQAS · UK NEQAS
Every stage can fail on its own — extraction, amplification, detection, sequencing, bioinformatics, interpretation — and needs its own check.
QCMD · CAP NGS · INSTAND · EMQN · GenQA
INR depends on the thromboplastin ISI, the instrument–reagent-lot calibration and the endpoint method. Peer grouping by instrument and reagent combination is essential.
CAP · Randox RIQAS · Bio-Rad EQAS · UK NEQAS · INSTAND
New in ISO 15189:2022: every POCT examination needs EQA, and identical devices are kept comparable with each other. Devices run by non-laboratory staff, often without LIS connection, are exactly where errors go unseen.
CAP · Randox RIQAS Blood Gas · Bio-Rad EQAS · INSTAND · Labquality
Up to 20 % discordance in HER2 status between laboratories without EQA — and HER2 decides whether targeted therapy is prescribed. NordiQC rates protocols from optimal to poor.
NordiQC · CAP · UK NEQAS
No single provider covers every discipline equally well. The practical answer is a core of broad programmes that carry the routine work, extended by specialists for the high-technology disciplines. Profiles, selection criteria and cost drivers are on EQA providers compared.
CAP (USA) — the broadest coverage and the route to CAP
accreditation.
Randox RIQAS (UK) — up to 24 distributions a year, online
peer comparison.
INSTAND e.V. (Germany) — the DACH standard, simple
logistics for the region.
QCMD — molecular diagnostics and viral loads.
EMQN — molecular genetics (BRCA, CFTR, AML).
NordiQC — immunohistochemistry.
UK NEQAS — specialised tests with deep expertise.
The EQA plan follows from the test menu, and the test menu follows from the equipment. That is why we plan them together: instruments from 68+ European manufacturers, the reagents they run on, and the EQA programmes that prove the combination works. KombiMED is the official CAP distributor for Central Asia and the Caucasus.
Providers, programmes and the full presentation in the buyer portal.
For laboratories outside the US and Canada, six months of CAP PT are the entry ticket. The path from there.
The questionnaire that shows where your laboratory stands today.
