Practical considerations for the interpretation of microbial testing results based on small numbers of samples.

While adequate, statistically designed sampling plans should be used whenever feasible, inference about the presence of pathogens in food occasionally has to be made based on smaller numbers of samples. To help the interpretation of such results, we reviewed the impact of small sample sizes on pathogen detection and prevalence estimation. In particular, we evaluated four situations commonly encountered in practice. The first two examples evaluate the combined impact of sample size and pathogen prevalence (i.e., fraction of contaminated food items in a given lot) on pathogen detection and prevalence estimation. The latter two examples extend the previous example to consider the impact of pathogen concentration and imperfect test sensitivity. The provided examples highlight the difficulties of making inference based on small numbers of samples, and emphasize the importance of using appropriate statistical sampling designs whenever possible.