Standard: 2003 NELAC
Section: Appendix B.2.1 and B.2.2
Link to relevant standard
3. Finally there appears to be a highly technical issue and conflict between Sections B 2.1 and B 2.2 of Appendix B in the 2003 NELAC Standard. Section B 2.1 requires the RSD of a method to be less than 50% of the RSD predicted at the Assigned Value of the sample. The NELAC regression equations predict variable standard deviations and RSD across the NELAC concentration ranges and in many instances NELAC criteria require interlaboratory evaluation limits which vary with laboratory population and concentration range. However, good method development procedures require the RSD of a method to be constant across the calibration range which in most cases is not consistent with the NELAC concentration range. The RSD of a method is controlled by the technique of the method and the variability of the instrument not by the NELAC concentration range.
The more important requirement to protect PT sample integrity is in Section B 2.2 and it requires the actual standard deviation of the verification analysis to be within 1.5 times the predicted standard deviation at the Assigned Value of the sample. If a method is capable of insuring that the sample meets the standard deviation requirement of section B 2.2 then it should be considered adequate to meet the requirements of the PT program. If the method is capable of achieving the necessary reliability in terms of meeting the standard deviation requirement of Chapter 2 Appendix B 2.2 then it is fit for use.
TNI FINAL RESPONSE:
(Proficiency Testing Board / NELAP Board, 10-11-09)
Sections B.2.1 and B.2.2 serve different purposes and are not in conflict. The purpose of B.2.1 is to ensure that each analytical method being used is precise enough to effectively detect any bias or inhomogeneity in the sample. Section B.2.2 provides the specific criteria for evaluating the homogeneity of the sample. Both sections must be followed.