Distribution of aflatoxin in pistachios. 7. Sequential sampling

Sequential sampling for aflatoxin testing in pistachios is evaluated using the aflatoxin distribution and Monte Carlo results previously obtained (J. Agric. Food Chem. 1999, 47, 3771-3775). The sequential protocol is modeled on the current EU test protocol by applying a three-step sampling, using 10, 20, and 30 kg sample averages. An acceptance level of 15 ng/g of total aflatoxin, under consideration for U.S. standards, is applied. Optimization leads to indifference regions of 2-30 ng/g for the first two steps. The resulting OC curve approximates that for a single 50 kg sample. The sequential protocol is applied to the results for a set of 1293 lots of the 1998 crop year, each tested with a single 10 kg sample. Ninety-five percent of the lots would have been accepted on the basis of the single test and 1.5% would have been rejected, whereas 3.5% of the lots would have required retesting.