Sampling Results for FSIS Regulated Products

Overview

Depending on the data available, calculations will either be percent positives, volume-weighted percent positive, or prevalence.[4] When possible, the calculations will include upper and lower confidence bounds. Follow-up sampling is specific to an individual establishment and the samples are only intended to gauge whether the establishment's corrective actions worked. Therefore, confidence bounds are not provided for such projects.

Percent positive (PP) – The percentage of samples of a specific FSIS-regulated product with a specific pathogen. These calculations may vary significantly from the actual prevalence of the population, but FSIS cannot make a more precise calculation with the data currently available.
Volume weighed percent positive (VWPP) – The VWPP for a specific product-pathogen pair is calculated by combining the production volumes (production volume can be defined as the amount (weight or numbers of head) of product produced per year) for establishments with their sampling results. This provides a more accurate estimation than a percent positive, but the data do not meet the statistical requirements to make a prevalence estimate. VWPP is more public health focused than PP, in that weighting results by volume reflects exposure of the public to specific pathogens.
Prevalence – The estimated proportion, nationally, of a specific FSIS-regulated product with a specific pathogen.

[1] Current calculations are national estimates for all establishment sizes. FSIS does not believe that the data support calculations based on smaller geographic regions or establishment size at this time.

[2] This evaluation, Use of FSIS Regulatory Verification Sampling to Generate Prevalence Estimates, is available on the FSIS website. Use of FSIS Regulatory Verification Sampling to Generate Prevalence Estimates

[3] FSIS has made modifications to several of its sampling projects since this evaluation was published. These modifications were made both to improve verification and to allow more precise calculations. Details of these changes can be found in the FSIS annual sampling plans available on the FSIS website.

[4] For each product pathogen pair, FSIS chose the most statistically valid method. Factors that impacted this decision are: the dates of sampling results; the number of samples available; the sampling design, or some other property of the sampling project that limits the conclusions that can be drawn from the aggregated sampling results. The different types of calculations should not be compared.