Working Group Meeting Book-Gluten

ISPAM Stakeholder Panel/Working Group Minutes August 26, 2018 Final Version

WORKING GROUP DISCUSSIONS: WORKING GROUP ON QUANTITATIVE MICROBIOLOGY METHODS ACCEPTANCE CRITERIA

Patrick Bird (Q Laboratories, Inc.) lead the working group discussions. The current acceptance criteria applied during validations is a 0.5 log difference between the means of the candidate and reference method for each contamination level. The working group agreed to take a deeper dive into how/why these criteria were accepted within the ISO community. The acceptability limits for the accuracy profile were established theoretically, although the existing data was based on the ISO 16140. It was important for the study design to use replicates of the same sample in order to achieve a proper estimate of the level of contamination of the sample and also to report on the repeatability/ reproducibility. The initial process for determining this was done theoretically in which 150 simulations and 37 actual data from previous ILS Ring Trial studies from MicroVal and Afnor; which took it from the theoretical to actual approach. The study design needed the same number of replicates for both the alternative and the reference method with a focus on the repeatability/reproducibility. During the development of the accuracy profile (AP) a proposal was discussed within WG2, that a two-step approach was needed, instead of having just one value for an acceptability limit (AL). The two-step approach was proposed to take into account disapproval of alternative methods due to a large variation between results for the reference method. It was also discussed to set the AL in the first step of 0.3 log units instead of 0.5 log which was not accepted by WG3, so the 0.5 remained. Pool all of the standard deviations of the alternative method across the contamination levels and calculate the bias (difference with the levels) confidence level and evaluate at the 0.5. The question is, can the working group harmonize with ISO 16140-2 and their quantitative statistics due to the workbook that has been validated and build acceptability limits criteria in the two-step approach. The proposal to the working group is to take data generated through the PTM process and run through the workbook calculations to determine whether the results are the same, rejected or more data is accepted and present to the working group during the next call. Paul Wehling will also attend the ISO SC34TC9WG2 on statistics meeting to discuss the development of the AL in more details. VII. NEXT STEPS Erin Crowley and Krystyna McIver (AOAC) discussed next steps, wrapped up all discussions, and answer any additional questions.

Meeting book link – August 26, 2018: https://cld.bz/BwtMsja