When is a Test Method Valid?

Well in short: When it has, through rigorous experiment, been proven to reliably deliver results fit for the intended purpose!

It is stated in paragraph 5.4.5.1 of ISO 17025:2005 General requirements for the competence of testing and calibration laboratories that “Validation is the confirmation by examination and the provision of objective evidence that the particular requirements for a specific intended use are fulfilled.”

A valid test method then will be a test method

• that has been subjected to rigorous experimental examination in the laboratory; and
• has, following an in-depth evaluation of the results of the examination, been confirmed fit-for-purpose for an intended use.

Before starting a validation study it is crucial to establish and document the customer’s requirement (also referred to as the Analytical Requirement, also keep in mind you may the "customer" setting the requirement). The customer’s requirement is the target you are trying to hit. It is a set of performance criteria that the customer expects from the method, e.g. it must be able selectively and accurately determine the presence of Aflatoxin M1 in milk at levels ranging from 0.01-0.10 µg/kg +/- 0.01µg/kg using LC-MS/MS. If a literature search of your proposed method clearly shows that it is unable to provide the Selectivity, Accuracy and Measurement of Uncertainty in the specified Range, then clearly the method is not fulfilling the requirements of the intended use. When a new, unpublished method is considered, a series of pre-validation experiments will have to be done to determine if the method has potential to meet the customer’s requirements.

Once a specific method, that ought to meet the customer’s requirement, has been identified, a clear Method Validation Plan (MVP) must be established (documented). The MPV serves as a work instruction detailing how the experimental examination of the method will proceed, i.e. what experiments must be conducted to yield adequate and suitable data to serve as objective evidence that the method is fit-for-purpose. Thorough planning at this stage will reduce the overall time-and cost budget of the method validation exercise.

Method Performance Parameters that may be addressed during method validation of e.g. the Aflatoxin M1 method in the example include (in no set order):

·        Limit-of-Detection (LOD) & Limit-of-Quantitation (LOQ)

·        Linearity & Working Range

·        Precision; Intermediate-Precision & Repeatability

·        Recovery & Accuracy

·        Robustness/Ruggedness

·        Uncertainty-of-Measurement

The MVP must clearly stipulate experiments that will generate sufficient and relevant results that can serve as objective evidence of the method’s performance in each of the parameter evaluated. The MVP should also include equipment and materials that will be needed, human resources required, a project time-line & a project-budget. It is also useful to at this stage stipulate the exact statistical methodology that will be followed to evaluate the result for quantitative parameter description (as relevant).

Once the MVP has been drafted and approved, the actual experimental phase of the validation study start. Having a clear plan and strategy ensure that the validation study stays on track and progress as planned. However, in real-life it often necessary that, led by experimental results, the MVP be adapted during validation. This type of experimental feed-back driven changes are totally in order as long you keep your eyes on the desired outcome, time-line and budget. With increased experience and more in-depth planning, one would expect that adapting the set MVP mid-way through a validation study will be kept to a minimum.

Experimental results obtained during the method validation study must be subjected to statistical evaluation to yield quantitative measure of each parameter investigated during the study. Metrics for the LOD, LOG, Repeatability, Accuracy and Uncertainty-of-Measurement must be calculated and documented in the Method Validation Report.

The Method Validation Report typically include:

• The Method Name
• The Scope Application
• The Analytical Requirements
• A reference to the documented MPV that was followed (Attached as Annexure)
• Reporting on each of the Validation Parameters evaluated - obtained metrics
• A specific reference to the Uncertainty-of-Measurement of the Method
• A concise, tabulated summary of the Obtained Metric vs the Analytical Requirement

… and if the Obtained Metrics are meeting or exceeding the Analytical Requirement…

• A formal statement that the method was validated and found fit-for-purpose
• Signature(s) and Dates of parties involved in the validation and making the declaration of fitness-for-purpose
• References used
• Reference to the location of the raw supporting data, calculations and other relevant data supporting the validation study (Attached as an Annexure if possible)

In the event that a method doesn’t validate, i.e. after careful experimentation it was found that the method simply cannot deliver the LOQ required by the customer (as set in the Analytical Requirement), this must be documented and statements to that regard must be included in the Method Validation Report. It may turn out at a later stage another customer will have need for what the method has to offer – and of course it is good validation practice to document both successes and failures.

Of course a careful literature study, intelligent pre-validation and realistic Analytical Requirements all goes a long way to ensure that a costly validation exercise indeed pays off first time.

Once a method has been declared valid for use in the laboratory is must be taken up and published in the laboratory’s Documented Quality Manual as a Standard Operating Procedure (SOP). Following publication of a new SOP employees must be trained in the use of the SOP, with Documented Training Records archived appropriately.

Ongoing validation of a validated method is… (Sorry time’s up – will come back to this topic J)

PS: My son reminded me that in on-line gaming the MVP refers to Most Valuable Player; don't get confused.

(Not proof read - welcome to comment and point out mistakes - EZ)

END