Introduction

At the Eurachem/CITAC workshop Measurement uncertainty and traceability: meeting the requirements of ISO/IEC 17025  in Lucerne, June 2002, a workshop session was held on the topic of uncertainty in qualitative analysis and testing. That workshop recommended the formation of a new Eurachem working group to provide guidance on the topic, based on the discussion paper presented at that meeting. The following constitutes the Terms of Reference for the working group.

Terms of reference

Members (at March 2005)

• Steve Ellison (LGC, UK) (Chair)
• Wolfram Bremser (BAM, DE)
• Enzo Ferrara  (IT)
• Leen van Ginkel (NL)
• Rüdiger Kaus (HS Niederheim, DE)
• Ilya Kuselman (Israel, CITAC)
• Bertil Magnusson (SP, SE)
• Ulf Ornemark (SE)
• V Patoprstky (SK)
• Anna-Liisa Pikkarainen (FI)
• Liam Regan (IRE)
• Piotr Robouch (EC-JRC-IRMM)
• Matthias Rösslein (EMPA, CH)
• Rainer Stephany (NL)
• Miloslav Suchanek (CZ)
• Adriaan M.H. van der Veen (NL)
• Maire Walsh (State Laboratory, Eire)
• Roger Wood (Food Standards Agency, UK)

Activity

The Eurachem working group is working on uncertainty issues in qualitative analysis, in response to early concerns about the implementation of measurement uncertainty under ISO 17025 accreditation. Eurachem published a policy paper on the subject in 2000 [i].At the Eurachem/CITAC workshop Measurement uncertainty and traceability: meeting the requirements of ISO/IEC 17025  in Lucerne, June 2002, a workshop session was held on the topic. That workshop recommended the formation of a new Eurachem Qualitative Analysis Working Group  to provide guidance on the topic, based on a discussion paper presented at that meeting. The group met first in May 2003. It is currently working on the topic of the assessment and expression of uncertainty in qualitative analysis and testing. The aim is to provide guidance on issues such as:

• the effort required to obtain sufficiently reliable false response rates (the most common basis for assessing the reliability of qualitative tests)
• methods of determining false response rates, including experimental methods, methods based on databases, and prediction of false response rates from the performance of quantitative methods used in the testing process
• practical methods of expressing the performance of methods, including guidance on the wide range of different terminology in use.

The group is working in cooperation with CITAC, and includes some US representatives as well as European members. The curent draft guidance is divided int two principal sections, one covering policy issues such as the use and reporting of uncertainty in qualitative tests (based on the policy paper above [6]), and the other on technical methods for assessing and expressing uncertainty in qualitative tests. Eurachem currently take the position that measurement uncertainties in the test method control parameters (such as times, temperatures etc.) should be demonstrably under control, that at least the most important false response rate should be assessed and that it is desirable to assess both false positive and false negative rates. Eurachem does not consider that the general requirements covering measurement uncertainty reporting should extend to the provision of probabilistic information in qualitative test reports. This is largely because such estimates are based on probabilties which are themselves very hard to estimate reliably. The latter part of the guidance covers the estimation or experimental determination of false response rates, using experiment, probabilistic modelling, database match rates etc. An earlier published paper describes many of the relevant issues [ii]. The methods are very closely related to procedures appropriate for validation of qualitative methods, a topic which the group sees as important but not yet fully addressed.

[i] Ellison  SLR, Accred. Qual. Assur. 2000, 5, 346-348

[ii] Ellison SLR, Gregory S Hardcastle WA, Analyst 1998, 123, 1155-61