Reference : Using total error as decision criterion in analytical method transfer
Scientific journals : Article
Physical, chemical, mathematical & earth Sciences : Chemistry
Engineering, computing & technology : Computer science
Using total error as decision criterion in analytical method transfer
Dewé, Walthère [Université de Liège - ULg > > Analyse des médicaments >]
Govaerts, Bernadette [Université Catholique de Louvain > Institut de Statistique > > >]
Boulanger, Bruno [Université de Liège - ULg > Département de pharmacie > Analyse des médicaments >]
Rozet, Eric mailto [Université de Liège - ULg > > Chimie analytique >]
Chiap, Patrice mailto [Université de Liège - ULg > Département de pharmacie > Analyse des médicaments >]
Hubert, Philippe mailto [Université de Liège - ULg > > Chimie analytique >]
Chemometrics and Intelligent Laboratory Systems
Elsevier Science Bv
Yes (verified by ORBi)
[en] analytical method ; transfer ; accuracy ; total error ; tolerance interval ; consumer risk ; producer risk
[en] An analytical transfer is a complete process that consists in transferring a validated analytical method from a sending laboratory (called sender) to a receiving laboratory (called receiver) after having experimentally demonstrated that it also masters the method. The decision to transfer an analytical method is usually taken based on the combination of two criteria: one about systematic error and one about random errors. These usual criteria, their combination and their limitations are described in this paper. As it is for method validation, total error-based approaches should be also considered in method transfer. Two approaches are proposed in order to have a total error-based criterion and to take into account the precision of the true value estimate. In the first approach, a beta-expectation tolerance interval similar to the one used in method validation is calculated and then compared to acceptance limits around the estimate of the true value. The second approach, found as slightly more powerful than the first one, consists in estimating the probability to have a result outside these acceptance limits. (c) 2006 Elsevier B.V. All rights reserved.
Researchers ; Professionals

File(s) associated to this reference

Fulltext file(s):

Restricted access
FinalTransferWally.pdfNo commentaryPublisher postprint618.81 kBRequest copy

Bookmark and Share SFX Query

All documents in ORBi are protected by a user license.