A testing or calibration laboratory lives on confidence in its results. That confidence is not asserted, it is demonstrated: with validated methods, equipment calibrated with metrological traceability and a declared measurement uncertainty. The international reference framework for all of this is the standard EN ISO/IEC 17025:2017, which sets out the general requirements for the technical competence and impartiality of laboratories. In this guide we explain what the standard requires, how a method is validated, what calibrating with traceability means and how accreditation is achieved.
ISO/IEC 17025: technical competence, not just management
Unlike ISO 9001, which certifies that a quality management system works, ISO/IEC 17025 accredits that the laboratory is technically competent to produce valid results. This is a crucial difference: a certification speaks of processes; an accreditation speaks of demonstrated capability for a specific scope of tests. The 2017 revision reorganised the requirements into structure, resources, processes and management, and introduced risk-based thinking and an approach to impartiality and confidentiality.
The resource requirements cover competent personnel with defined functions, facilities with controlled and recorded environmental conditions (temperature, humidity, vibration depending on the test), equipment with a maintenance and calibration plan, and metrological traceability. The process requirements cover the review of requests, the selection and validation of methods, sampling, the handling of test items, ensuring the validity of results and the report.
Method validation and verification
Before applying a method, the laboratory must demonstrate that it is fit for its intended use. When the method is standardised and applied without modification, it is enough to verify it: confirm that the laboratory can run it with the expected performance. When the method is in-house, modified or used outside its intended scope, it must be validated in full.
Validation characterises a series of performance parameters:
- Accuracy: degree of agreement with the true value, assessed with certified reference materials.
- Precision: repeatability (same operator, same day) and intermediate reproducibility (different days, operators, equipment).
- Linearity and working range: the range where the response is proportional to the analyte.
- Limit of detection (LOD) and of quantification (LOQ): the minimum quantity reliably detectable and quantifiable.
- Selectivity and robustness: the ability to measure the analyte without interference and tolerance to small variations in conditions.
Each parameter is documented with experimental data and acceptance criteria defined in advance. A method validated without prior acceptance criteria is not a validation; it is a collection of data with no conclusion.
Calibration and metrological traceability
Calibrating a piece of equipment means comparing its readings with a reference standard to determine the deviation and, where appropriate, correct it. The key is metrological traceability: an unbroken chain of documented calibrations, each with its uncertainty, that links the laboratory's instrument to a national or international standard and, ultimately, to the units of the International System. In Spain, the national standards are kept by the Spanish Metrology Centre, and the chain usually passes through accredited calibration laboratories.
A thermometer that reads 100.2 °C when the standard shows 100.0 °C has an error of +0.2 °C; that figure, with its uncertainty, goes onto the calibration certificate and must be applied or taken into account in every subsequent measurement. Critical equipment carries a calibration plan with a periodicity justified by its historical drift, not by arbitrary habit. Between calibrations, intermediate checks with working standards detect drift before it invalidates results.
Measurement uncertainty
No measurement result is an exact number: it is a value accompanied by an uncertainty that quantifies the reasonable doubt. ISO/IEC 17025 requires estimating and declaring measurement uncertainty in calibrations and, when it is relevant to validity or interpretation, in tests too. The reference methodology is the Guide to the Expression of Uncertainty in Measurement (GUM).
The process identifies all sources of uncertainty (repeatability, equipment resolution, the standard's uncertainty, environmental conditions, drift), combines them through a law of propagation and expresses them as an expanded uncertainty with a coverage factor, typically k=2 for a confidence level close to 95%. Declaring a result of 100.2 ± 0.3 °C (k=2) is radically different from declaring only 100.2 °C: the former lets you decide whether the value meets a specification; the latter does not.
Table: certification versus accreditation
| Aspect | ISO 9001 (certification) | ISO/IEC 17025 (accreditation) |
|---|---|---|
| What it assesses | Quality management system | Technical competence for specific tests |
| Granted by | Certification body | Accreditation body (ENAC in Spain) |
| Scope | The organisation's processes | Specific tests or calibrations |
| Metrological traceability | Not explicitly required | Central requirement |
| Uncertainty | Not applicable | Mandatory in calibration |
Managing non-conformities and corrective actions
However rigorous a laboratory is, sooner or later it will detect a result that does not meet the criteria: an out-of-tolerance calibration, a poorly preserved sample, an unsatisfactory proficiency test. ISO/IEC 17025 requires a documented procedure for managing this nonconforming work. The procedure must define responsibilities, assess the significance of the non-conformity, decide on the acceptability of the affected work and, where appropriate, notify the client and recall the reports already issued.
The difference between a laboratory that improves and one that repeats errors lies in root-cause analysis. Correcting the symptom (repeating the test) without understanding why it happened guarantees recurrence. Tools such as the Ishikawa (cause-and-effect) diagram, the five-whys method or failure-mode analysis help reach the real cause: was it a miscalibrated instrument, a poorly written method, insufficient training, an uncontrolled environmental condition? The corrective action attacks that cause, and its effectiveness is verified afterwards; closing a corrective action without confirming that the problem does not recur is a closure only on paper.
The standard also distinguishes between correction (what is done now to contain the problem) and corrective action (what is done so that it does not happen again). Confusing the two is a common audit finding and reflects a system that puts out fires instead of preventing them.
Staff qualification and the chain of custody
Results are produced by people, and the standard recognises this by requiring demonstrated competence for each function. This goes beyond a qualification: it means nominally authorising each analyst for the specific tests they run, keeping records of their training, supervising those still in the process of qualification and periodically reassessing competence. An analyst qualified for chromatography is not automatically authorised for spectrometry; each technique has its own authorisation.
The chain of custody guarantees the integrity and traceability of the test item from receipt through to the issuing of the report and its final disposal. Each transfer is recorded: who received the sample, in what condition, where it was stored, what treatment it received. In sectors where the result may end up before a court—forensic analysis, doping control, food safety—a broken chain of custody invalidates the result, however technically correct it is. The unequivocal identification of each item, normally through coding that prevents the loss of traceability, is the foundation on which all subsequent confidence rests.
Ensuring the validity of results
The standard requires monitoring that results remain valid over time. The usual tools are proficiency testing, in which independent laboratories analyse the same sample and the results are compared; interlaboratory comparisons; reference materials analysed periodically; and control charts that monitor the process drift. Satisfactory participation in proficiency tests is, moreover, one of the pieces of evidence the accreditation body requests.
Common mistakes in the laboratory
The first is declaring results without uncertainty, which prevents the client from deciding conformity against a specification. The second is calibrating by calendar without justifying the periodicity with drift data, which leads to calibrating too often or, worse, too seldom. The third is confusing verification with validation and treating a modified method as validated without characterising its performance. The fourth is environmental control without records: conditions that are monitored but not documented are worthless in an audit. The fifth is a broken chain of custody, which invalidates the traceability of the test item from sampling to the report.
Frequently asked questions
Is accrediting a laboratory the same as certifying it? No. ISO 9001 certification validates the management system; ISO/IEC 17025 accreditation recognises technical competence for specific tests or calibrations, with traceability and uncertainty. A client that demands legally defensible results asks for accreditation.
Who accredits in Spain? The Spanish National Accreditation Body (ENAC), a signatory of the multilateral mutual-recognition agreements, which allows the reports to be accepted internationally.
How often is a piece of equipment calibrated? There is no universal interval. It is set according to the equipment's historical drift, its criticality and the manufacturer's recommendations, and adjusted with intermediate checks.
Why is declaring the uncertainty mandatory? Because without it a result cannot be compared with a specification limit or with another laboratory. Uncertainty turns a number into a defensible statement.
Conclusion
A laboratory's competence is not measured by the sophistication of its equipment, but by its ability to demonstrate, to any auditor or client, that every result is traceable, that the method is validated for its use and that the uncertainty is quantified. ISO/IEC 17025 articulates that demonstration in concrete requirements: methods validated with acceptance criteria, equipment calibrated with metrological traceability, uncertainty estimated according to the GUM and validity ensured through proficiency testing. Achieving ENAC accreditation is not a decorative seal, but proof that an independent third party has verified that competence for a defined scope. At Summum Marketing we guide laboratories through implementing these requirements and preparing for the accreditation audit, focusing on the evidence that is actually examined.