Guidelines for the determination of the prevalence of Salmonella contamination in consumer poultry at retail level

Abstract

A monitoring system is described to determine the prevalence of Salmonella in poultry meat products at retail level. For the Dutch situation chicken meat samples will have to be collected in 385 shops (confidence level of 95%, accepted error of 5% and estimated prevalence of 50%). The main retail points for poultry meat in the Netherlands are real butchers, regular poulterers, market poulterers and supermarkets. They account for 95% of the chicken sold to Dutch consumers. Based on the market shares of these retailers, 65 real butchers, 44 regular poulterers, 37 market poulterers and 243 supermarkets should be sampled. For sampling it is necessary to divide the chicken meat into four product groups: whole carcasses, parts of leg, parts of breast and other parts. Each of these groups should be sampled in the shops. So there will be 16 (4 shop types * 4 product groups) strata for which an estimate of the prevalence is to be determined. The strata should be sampled by collecting 740 gr. meat per product group per shop and is based on the quantity of chicken a consumer buys at each time. In total about 1550 samples have to be collected. To compare results it will be of importance to examine the samples in an identical way. For example by rinsing the meat samples with buffered pepton water and testing by the fluid for the presence of Salmonella using the ISO 6579 isolation method. Based on the prevalences of the 16 strata, the prevalences for each shop type, product group and the total chicken sold can be calculated. The formula's for calculating these prevalences and the standard deviation are presented. Although, the system is described for the Dutch situation, it can also be applied in other countries where similar data are available.<br>