READY TO EAT PRODUCT SAMPLING ALGORITHM UPDATE
Beginning in October, 2016, FSIS is modifying its sampling frequency and targeting in Ready to Eat (RTE) products. (FSIS uses the word algorithm)
FSIS has two routine sampling programs for Ready-to-Eat (RTE) products. They are:
• All RTE products are eligible, and products are randomly selected by inspection program personnel.
• Goal is to verify all RTE products are free of Listeria monocytogenes (Lm) and Salmonella.
• Only post-lethality exposed products are eligible, and products are selected by inspection program personnel using a risk-ranking (fully cooked sliced, hot dogs, salad/spread/pate, etc.).
• Goal is to focus resources to verify riskiest products (as shown by a 2003 FDA-FSIS quantitative risk ranking) are free of Lm and Salmonella.
Samples under both programs are analyzed for Lm and Salmonella.
As a result, FSIS establishment testing has showed a low percentage of positives, which has plateaued. In addition, results show RTEPROD_RISK program has had equal or lower percent positive than RTEPROD_RAND in recent years.
So, the Agency is working on several projects to efficiently target FSIS resources on higher risk facilities and products to further reduce likelihood of Lm contamination in Federal establishments and at retail.
Currently the RTEPROD_RISK algorithm is heavily weighted toward type of product only.
FSIS has concluded that other factors that inform the algorithm (volume, alternative, etc.) may not have been as impactful as intended.
Furthermore, FSIS data show very few Lm positives in the higher risk RTE meat and poultry products (those identified in the 2003 FDA-FSIS quantitative risk ranking) likely because of improved process controls industry has put in place.
FSIS is observing Lm positives in lower risk RTE meat and poultry products, possibly due to less effective controls. For example, an analysis of RTEPROD data from 2010-2013 showed acidified/fermented meat and poultry products have a higher percent positive for Lm than hot dogs (0.46% vs. 0.37% respectively).
Thus, starting with October, 2016, samples the following changes will occur:
• Each eligible RTE establishment will be selected for a random sample at least once every 6 months.
• Any eligible establishment with a positive result (Salmonella or Lm) in an RTEPROD sampling project (random or risk) in the last 6 months is guaranteed to be selected for a random sampling task. In other words, after a positive occurs for an RTEPROD sample, that establishment will be selected for a RTEPROD_RAND sampling task for each of the following 6 months.
• The remaining number of random sampling tasks each month will be randomly assigned to eligible establishments not already selected under criteria 1 or 2.
• Establishments that have at least one post-lethality exposed product in their plant profile and do not already have a random sampling task assigned will be assigned a risk ranking. Establishment selection for risk-based sampling tasks will be based on this risk ranking that considers:
1. The historical percent positive for each product produced at the establishment.
2. The daily production volume of each product at the establishment.
3. The Lm alternative3 used for each product at the establishment.
In addition to revising the algorithm, FSIS also will be allocating sampling resources evenly between the two RTEPROD sampling projects to ensure broad sampling coverage of all RTE products.
Samples had been allocated more heavily to the RTEPROD_RAND program than RISK (25% vs. 75%, respectively).
This change results in approximately 700 samples scheduled each month per project.
There is a limit of 1 RTEPROD sample (either RAND or RISK) per establishment per