How Many Samples Do I Take?

Abstract

It is sometimes falsely assumed that microbes are evenly distributed throughout the entire food mass. This is rarely true. Sporadic contamination of food production lots is a very common phenomenon requiring increased sampling to gain some measure of confidence in results. Sample composites are often used to reduce the testing burden. However, before using a compositing scheme, one should verify that it is valid for the food matrix, the microbe of interest, and the assay used. Focused sampling and testing of the portion of a product in which the microbe is most likely to be is one approach which can be considered. Often “positive” results are questioned. Verification of “positive” samples through retesting is not an effective or appropriate means to determine whether or not the testing laboratory made an error. Nevertheless, there are diagnostic questions that can be addressed to the testing laboratory before accepting a “positive” result. Food processors should establish appropriate specifications for their incoming ingredients. Certain key elements should be included in any microbiological specification. Sampling plan inadequacies usually have to do with taking an insufficient number of samples without regard to accepted statistical sampling plans. Sampling plans to consider include attributes plans and variables plans as described in ICMSF plans and others.