MicroscopyAll participants were requested to determine the presence or absence of land animal and/or fish, to indicate the type of material found and the method used.Incorrect positive results (positive deviations) were expressed in a specificity score and incorrect negative results (negative deviations) were expressed in a sensitivity score. An optimal score is 1.0.The results are analysed in two ways: numbers below threshold (between 1 and 5 particles per determination cycle inclusive) have been considered positive and as alternative considered as negative. By comparing both ways of analysis it is possible on one hand to compare the results with those from previous ring trials (there these numbers were considered positive based on the legal principle of zero tolerance), and on the other hand to compare it to the official method (where numbers between 1 and 5 are considered negative).Most of the specificity and sensitivity scores for microscopy were at good or reasonable levels. In the combination of fish meal (0.1%) and ruminant MBM (0.1%) the detection of fish material was sub optimal. Considering numbers of particles below threshold as negative, the sensitivity scores show a considerable drop. The results indicate that the overall performance of the microscopic method is satisfactory, but applicants of the microscopic method could benefit from good and effective training and documentation in order to achieve a higher reliability in identifying particles. PCRThe specificity (samples A and B) and sensitivity (samples C and D) scores for PCR were between 0.87 and 0.91. Combined scenarios for microscopy and PCR.Several participants applied incorrect numbers of determination cycles, either too many or too less.The scenarios for correct combination of microscopy and/or PCR are published in an accompanying SOP. Several deviations from these scenarios were applied, such as examination by PCR of the terrestrial animal feeds, reporting the presence of fish in the fish feeds, and the absence of a final conclusion combining the results of both methods where appropriate. It is, however, a very good situation to have all (intermediate) results reported for each of the two methods in order to have a good documentation in the framework of a ring test.

No abstract