Key Topics

Principle 2

Analytical measurements should be made using methods and equipment which have been tested to ensure they are fit for purpose.

Why Is It Important?

• Without knowledge of its performance characteristics, it is impossible to identify a method which can meet an agreed specification (see VAM principle 1).
  
  The performance characteristics are determined during method development. Performance characteristics of published methods which you might consider using should be documented and should be used to make an assessment of the suitability of a method for the application. The performance characteristics of in-house methods should be carefully investigated and documented.
• If you cannot achieve in your laboratory the documented performance characteristics of a method, application to customer samples is a waste of time.
  
  Any laboratory which intends to apply a method must show first of all that it can apply the method to samples which are within the documented scope of the method and obtain reliable results.
• Attempts to apply a method outside its documented scope can lead to serious errors.
  
  You cannot expect reliable results from validated methods unless the measurements are being undertaken within the range and scope (e.g. levels of analyte and types of sample matrix) for which the method has been validated and the equipment is both suitable and properly calibrated.

What To Do

• Review the original sources of the methods currently in use in your laboratory
  
  Your methods will have come from a variety of sources, and the extent of validation data will also vary. You should review the validation data available for the methods which you use for routine analysis. Method documentation should include a statement of the uncertainty associated with the method.
• Provide a procedure which will help analysts to select appropriate methods taking into account the analytical specification and any additional constraints.
  
  The procedure could include a decision tree or a check list to help the analyst. The check list should cover aspects such as: analyte/matrix characteristics, assessment of interferences, concentration range, time, cost, uncertainty. It should also refer back to the analytical specification (see VAM principle 1).
• Ensure that there is a procedure in place which shows that staff can use methods and obtain satisfactory results before application to customer samples.
  
  The procedure should specify that samples of well characterised materials are used in a 'blind' test of performance. The samples should be as similar as possible to the expected customer's samples. If available, certified reference materials could be used as samples.
• Review the procedures which are in place for dealing with non-routine analysis.
  
  You should examine how methods are developed for non-routine samples, and what steps are taken to validate these methods. Assessment should be made of the confidence limits which apply to the results of these methods.
• Ensure that equipment used operates according to a specification appropriate to the requirements of the method.
  
  Instruments used in a method cannot be relied upon unless they have been properly maintained and calibrated.

Additional information

Sources of methods

Careful selection of a method is critical to obtaining reliable data. There are a number of sources you can turn to in identification of a suitable method. These are listed below in order of priority.

• Methods which have been collaboratively studied by several laboratories. Collaboratively studied methods are published by a number of bodies such as AOAC, BSI, ISO, and the Analytical Methods Committee of The Royal Society of Chemistry.
• Methods which have been validated by more than one laboratory, or are recommended by a panel of experts e.g. the Environment Agency Standing Committee of Analysts.
• Methods developed and validated in-house. These would probably have originated from the scientific literature, books or manuals, and have been further developed and validated in your laboratory.

Your analysts will often obtain better results using a method which is familiar to them rather than one which is new to them. Therefore as long as the method complies with the customer's requirement it is sensible to choose the most familiar method.

Method Validation

Validation of a method establishes that the performance characteristics are adequate for the intended use of the analytical result. The key criterion here is that the method as applied in your laboratory provides data which are fit for the intended purpose. This criterion needs to be met before applying the method to analysis of the customer's sample. The amount and type of work needed to validate the method will depend on the previous history of the method and the requirements of the analytical specification. The development of standard methods will have included consideration of all the necessary aspects of the validation. However the responsibility still lies with you to ensure that the validation documented in the method is sufficiently complete to meet your customer's needs. Even if the validation covers your intended application, you will still need to verify that the documented performance can be met in your laboratory.

Important performance characteristics that should be addressed in validation include:

• Selectivity and specificity
• Working range
• Linearity
• Limit of detection
• Limit of quantitation
• Ruggedness
• Precision
• Bias
• Uncertainty (precision and bias)

These parameters must be clearly stated in the documented method so that the suitability of the method can be assessed for its intended application.

Validation does not imply that the method is free from error. It is confirmation that the method can yield acceptable results within the limits of the scope of the method.

The most important point to remember is that you cannot expect reliable results from validated methods unless the measurements are being undertaken within the range and scope for which the method has been validated and the equipment calibrated.

If a method is changed in any way (e.g. if a new instrument is brought into use or the method is applied to a different sample matrix) an appropriate revalidation must be undertaken.

Wherever possible validation should include the use of certified reference materials and calibration standards and should result in a method with known confidence limits (see VAM principle 5).

What happens in the case of a single sample?

In reaching agreement over the appropriate level of effort for a single sample, it is important that both the analyst and customer understand what is required to achieve fitness for purpose (see VAM principle 1). For one-off samples, your analysts should consider assembling a method using as building blocks well known and authenticated steps which may have been proven in similar contexts from previous work. It is almost always possible to run a spiked sample simultaneously with the real sample to confer some degree of confidence in the results obtained. (A spiked sample is a sample whose matrix is very similar, or identical, to that of the samples of interest, to which a known quantity of analyte has been added.)

Examples

Methods should be validated by users in their own laboratories

A dispute arose when an analysis of a food product labelled as being 'free of preservatives' was shown to contain the preservative benzoic acid. The chromatographic method used for the analysis was typically that used, in a recognised proficiency testing scheme, for the analysis for benzoic acid in food products. To resolve the dispute a second laboratory was asked to repeat the analysis. The second laboratory used a different detection system and the results confirmed the product was free of preservatives. The peak which had been attributed to benzoic acid was due to a matrix component. The problem with the first analysis was that the method was being used outside its validated scope. The laboratory staff had not checked if it was suitable for this particular type of food stuff.

Equipment should be properly calibrated before use

Measurements of the amount of alcohol present in a person's breath are now carried out in the UK (and certain other countries) using two types of instrument which depend on different analytical principles. These breath testing instruments are required to detect alcohol concentrations in a subject's deep lung air with high accuracy. When the instruments were first introduced there were significant numbers of challenges in the courts concerning the validity of the results obtained. However, these instruments are now tested to stringent specifications produced by the Home Office, and the instruments are calibrated before and after each use with certified reference materials. This ensures that the results produced are valid. As a consequence the public now has confidence in the validity of the measurements and the expensive blood tests are used only in borderline cases.

Supporting resources

• In-House Method Validation: A Guide for Chemical Laboratories
• mVal: Software for method validation
• Method Validation - Key topics
• Method Validation Training Course