ABSTRACT
Because of the paucity of quantitative data on numbers of other enteric pathogens in food, the reliability of the examination of processed foods for Enterobacteriaceae was estimated taking Salmonella as a model. For this purpose an assessment was carried out of the risk of accepting Salmonella contaminated consignments of foods, despite a negative outcome of (i) examination of 1.5 kg samples for Salmonella; (ii) examination of one or two 1 g samples for Enterobacteriaceae; (iii) simultaneous application of both tests. The computations were based on the results of the examination of 6830 samples of dried foods, processed for safety, out of a total of 18170 samples.Only 69 samples permitted the exact calculation of the epsilon-factor, defined as c.f.u./g of Enterobacteriaceae/c.f.u./g of Salmonella; 4642 were positive for the former group but ;free' from Salmonella, and the rest were negative in both tests. Numbers of c.f.u./g for both groups, and hence the epsilon-factors, varied widely between commodities and also between different consignments of the same food product. The average for epsilon amounted to 5.8 x 10(3), far from the base-line value of 0.75 x 10(3) assessed earlier. In only 0.1% of samples did the Enterobacteriaceae test fail to achieve the required consumer protection.This investigation therefore substantiates that testing foods processed for safety by examining accurately chosen quantities for ecologically well selected and taxonomically thoroughly defined index organisms is a most effective procedure in terms both of consumer protection and simplicity of examination without compelling the food industry to achieve hardly attainable microbiological quality standards.
