31 No. 6
Experimental Requirements for Single-Laboratory Validation
The objective of this recently initiated project is to provide expert guidance on the scope and scale of experiments required for single-laboratory method validation, enabling regulatory agencies to harmonize validation requirements.
This new project will develop guidance on experimental designs suitable for determining method performance characteristics during single laboratory validation experiments. Where possible, the resulting report will include guidance on numerical values for such performance characteristics.
The output is intended to support implementation of the existing IUPAC Harmonized Protocol on Single-Laboratory Validation (PAC 2002, Vol. 74, No. 5, pp. 835–855), which specifies the performance characteristics to be assessed, but currently includes no quantitative guidance on the scale of experimentation required.
It is currently envisaged that the guidance will be informed by statistical power considerations; that is, experimental requirements will be set so as to achieve a particular probability of correctly identifying significant adverse effects (such as a specified upper limit for bias). The guidance will indicate the power appropriate for different situations. For example, verification of performance of an established method, validation of a new method, or validation of a new method intended for critical uses would attract increasingly stringent requirements. It is envisaged that the report will include example simple-experiment designs and associated tables of replicate numbers for each such situation so that regulators can easily specify the level of stringency required and analysts can easily identify the scale of experimentation necessary to meet the requirement.
The requirements will be chosen so that they broadly reflect and harmonize current best practice. The advantage of specifying requirements in terms of test power as well as listing specific experiments and experiment sizes is that it then becomes possible to permit any experiment design that is designed to achieve the necessary confidence. This provides for flexibility in specifying single-laboratory validation requirements, and allows analysts to design improved experiments which have greater efficiency while being able to demonstrate that the stringency is sufficient.
For more information and comments, contact Task Group Chair Steven L.R. Ellison <[email protected]>.
last modified 6 November 2009.
Copyright © 2003-2009 International Union of Pure and
Questions regarding the website, please contact [email protected]