Normalisation of data from allergens proficiency tests

The problem of allergen analysis using ELISA kits from different commercial products giving significantly different results is widely acknowledged. The effect on proficiency testing results is that different assigned values have to be generated for the different kits used. Some experimental Food Analysis Performance Assessment Scheme (FAPAS) proficiency tests aimed to establish whether the use of a standardised calibrant could be used to normalise the complete data set without recourse to differentiation. Three recent FAPAS proficiency tests (2776 peanut, 2778 soya and 2781 gluten) sent out a second spiked sample, in addition to the usual spiked and unspiked samples. Further analysis of the data was undertaken after the completion of the tests. The ratio of the submitted results for the two spiked samples yielded complete data sets which could be tested for normality of the distribution. Where the raw data for each individual test sample was clearly non-normal and multi-modal, the ratio data yielded a much more normal and symmetrical distribution. The use of one of the test samples as a single-point calibrant has some limitations but the principle of applying a standardisation clearly works. The development of internationally recognised sets of certified reference calibration standards for use by allergens testing laboratories would greatly benefit the analysis.