Page 10

eFOODLab_International_03_2014

Figure 2: Equivalence analysis of PT Manganese in green cabbage (BVL). 10 3/14 eFOOD-Lab international Fig 3a: Test results and tolerance limits based on Horwitz and on Q/Hampel Fig 3b: Z scores based on Q/Hampel Fig 3c: Z scores based on Horwitz Turning to Europe and Germany, with respect to international PT schemes carried out by the European Reference Laboratories, an important question concerns the monitoring of laboratories in the country of origin of import. In his presentation, Manfred Stoyke (BVL - Federal Office of Consumer Protection and Food Safety, Germany) explained how a thorough and detailed statistical analysis of PT data can shed light on analytical problems and lead to a harmonization of analytical methods and, in due course, to considerable progress in analytical quality. A prerequisite for such an endeavor is the use of powerful statistical software. The PT-software PROLab Plus is particularly well suited. Steffen Uhlig (QuoData GmbH, www. quodata.de) gave an overview of different approaches to data analysis and showed how effective data analysis is crucial for the success and harmonization of PT programs. He drew special attention to the fact that such an approach to data analysis goes well beyond the requirements of standards such as ISO 17043 and ISO 13528. It is true that the fundamental international standard for the accreditation of analytical laboratories ISO 17025 contains numerous requirements concerning internal and external analytical quality control. However, this does not mean that different internal and external QC programs can be considered comparable or equivalent in terms of resulting analytical quality. Analytical quality control methods may be defined in international standards and examined in audits, but implementation, interpretation and use vary from lab to lab, from sector to sector and from country to country. It is precisely these differences that could be termed differences in Laboratory Quality Culture (LQC). LQC may cause significant differences in analytical quality • between laboratories • between accreditation bodies • between sectors • between countries LQC can thus be assessed by measuring trueness and precision • within laboratories • across laboratories accredited by a specific accreditation body • across laboratories inside a particular sector • across laboratories inside a particular country PT schemes have a central role to play in the harmonization of LQC. Indeed, effective PT data analysis represents a privileged vehicle for achieving the following goals: 1. To facilitate training, learning, networking and best practice sharing in analytical methods 2. To promote harmonized LQC across laboratories and countries 3. To build confidence in LQC across sectors and countries In the following, an outline of the scope of effective data analysis is provided, based on Steffen Uhlig’s talk. An outline of effective data analysis Precision, bias and equivalence of test methods First of all, PT is relevant for staff training within laboratories. PTs can provide Qualit y Management


eFOODLab_International_03_2014
To see the actual publication please follow the link above