Method Validation: The 5 Essential ISO 17025 Clause 7.2 Records

Method validation is the laboratory’s proof that a method does what the customer needs it to do, and ISO/IEC 17025:2017 clause 7.2 is precise about when proof is owed and what it must contain. Clause 7.2.1 governs the selection and verification of methods: the laboratory uses appropriate methods, keeps them current, and verifies that it can perform a standard method properly before introducing it. Clause 7.2.2 governs validation: non-standard methods, laboratory-developed methods, standard methods used outside their intended scope, and modified standard methods must be validated, and the validation must be “as extensive as is necessary to meet the needs of the given application”.

The line between the two — verify a standard method, validate everything else — is the one most laboratories draw in the wrong place, and the records clause 7.2.2.4 requires are the ones assessors ask for first. This guide sets out the verification–validation boundary, the performance characteristics a validation establishes, the five records the clause requires, what happens when a validated method changes, and the six method validation failures assessors find.

Verify or validate: the 7.2 boundary

The boundary follows the method’s authority, not its difficulty. A published method run exactly as written on the matrix and range it was written for needs verification; the moment the laboratory changes the method, the matrix, the range or the purpose, it owns the method and owes a validation. Our guide to ISO 17025 covers where laboratories lose findings; 7.2 is high on the list.

What a method validation establishes

Clause 7.2.2.3 lists the performance characteristics a validation can include — the choice depends on the method and the application — and a validation plan states which apply and the acceptance criterion for each.

The uncertainty row is the bridge to clause 7.6: validation data is the main input to the budget, and a validation that establishes precision without capturing the other contributions produces a budget built from repeatability alone. Our guide to measurement uncertainty covers that trap.

The five records clause 7.2.2.4 requires

1. The validation procedure used — the plan: what was tested, how, by whom, against what acceptance criteria.
2. The specification of the requirements — what the customer or application needs the method to achieve: the range, the uncertainty, the detection limit, the matrices.
3. The determination of the performance characteristics — the results for each characteristic in the plan.
4. The results obtained — the raw and processed data behind the characteristics.
5. A statement on the validity of the method, detailing its fitness for the intended use — the conclusion, signed by someone authorised, that the requirements in record 2 are met by the characteristics in record 3.

Record 2 is the one most often missing. Without a specification of requirements, the validity statement has nothing to be valid against, and an assessor reading “the method is validated” will ask “for what”.

Method development and change

Clause 7.2.1 requires method development to be a planned activity, assigned to competent personnel, with periodic review to confirm the customer’s needs are still being met, and with modifications approved and authorised. Clause 7.2.2.2 requires that when changes are made to a validated method, the influence of the changes is determined and, where it affects the original validation, a new validation is performed. The practical mechanism is a method register with a validation status per method and a change control that routes every modification through an impact assessment: instrument replaced, reagent supplier changed, software updated, matrix added. Our guide to proficiency testing covers the ongoing evidence that a validated method stays valid.

Six method validation failures assessors find

• Verification recorded as validation, or nothing recorded at all, for a standard method that has been run for years.
• A standard method run on a matrix it was not written for, with only the original verification behind it.
• No specification of requirements, so the validity statement is unsupported.
• Characteristics chosen because they were easy, not because the application needed them — precision without LOD for a trace method.
• Changes without impact assessment: a new instrument, a new column, a software version, and the validation date unchanged.
• Validation by one analyst on one day, so intermediate precision — the characteristic that predicts real performance — is unknown.

Running method validation well

1. Classify every method on the schedule as verify or validate, and record the reason.
2. Write the requirements specification first, from the customer’s decision: what range, what uncertainty, what detection limit.
3. Plan the characteristics and criteria before generating data; a criterion set after the result is not a criterion.
4. Use the validation data for the uncertainty budget and for the decision rule discussion under 7.1.3.
5. Keep the method register current with status, date, responsible person and change history.

Frequently asked questions

What is method validation under ISO 17025?
The confirmation, required by clause 7.2.2 of ISO/IEC 17025:2017, that a non-standard, laboratory-developed, out-of-scope or modified standard method meets the requirements of its intended use, established through performance characteristics such as range, accuracy, uncertainty, detection and quantification limits, selectivity, linearity, precision, robustness and bias, and recorded with a statement of validity.

What is the difference between verification and validation?
Verification (7.2.1) confirms the laboratory can perform a standard method as published and achieve its stated performance; validation (7.2.2) establishes the performance of a method the laboratory has developed, modified or is using outside its published scope.

Which performance characteristics are required?
Clause 7.2.2.3 lists candidates — range, accuracy, uncertainty, LOD and LOQ, selectivity, linearity, repeatability and reproducibility, robustness, bias — and the laboratory selects those the application needs, stating the acceptance criterion for each.

What records must be kept?
Clause 7.2.2.4: the validation procedure, the specification of requirements, the determination of performance characteristics, the results obtained, and a statement on the validity of the method for its intended use.

When must a method be re-validated?
When changes are made to a validated method and the assessment of their influence shows they affect the original validation — a new instrument, matrix, reagent source or software version are the usual triggers.

Where this leaves you

Draw the method validation line where clause 7.2 draws it — verify what you run as published, validate what you own — write the requirements before the data, choose characteristics the application needs, keep the five records, and route every change through an impact assessment. The assessor’s first question is not whether the method works; it is what you required of it and how you proved it.

References

• ISO/IEC 17025:2017 — General requirements for the competence of testing and calibration laboratories — Clause 7.2.1 selection, verification and development of methods; 7.2.2 validation, performance characteristics and records; 7.6 measurement uncertainty.

More on ISO 17025

• Method validation — you are here
• ISO 17025: accreditation, scope and findings
• Measurement uncertainty
• Proficiency testing
• Decision rules and statements of conformity
• ISO 17025 vs ISO 9001

The Method Selection and Verification Procedure, the Method Validation Procedure, the Laboratory Developed Method Control Procedure and the Method Register and Validation Status are in the ISO 17025 Toolkit, or start with the free templates.