Feb 6, 2008 (CIDRAP News) – The US Department of Agriculture (USDA) says it is preparing to begin publishing the names of poultry and meat plants that have trouble controlling Salmonella, as the agency extends a set of policy changes designed to reduce the prevalence of the pathogen in meat.
Starting Mar 28, the USDA's Food Safety and Inspection Service (FSIS) will publish online the names and test results for plants where more than 10% of product samples are found to have Salmonella contamination, the agency announced last week. That step will focus first on broiler (young) chicken plants, which have had the most difficulty with Salmonella.
The USDA said the coming changes are the result of a Salmonella-control initiative it launched 2 years ago, after the Centers for Disease Control and Prevention (CDC) reported that Salmonella had become the most common foodborne pathogen, accounting for 38.6% of cases. USDA had seen an upward trend in Salmonella contamination from 2002 through 2005, with about 16% of broiler chicken samples testing positive in 2005.
"The initiative includes concentrating resources at establishments with higher levels of Salmonella and changes to the reporting and use of FSIS Salmonella verification test results," the USDA said in a Jan 28 announcement.
Plants sorted by performance
In the "risk-based" sampling program, poultry and meat plants are sorted into three categories according to their Salmonella test results. The best-performing plants, those in category 1, are sampled by the USDA less often than those in categories 2 and 3, according to FSIS officials.
Laura Reiser, an FSIS spokeswoman in Washington DC, said more plants are being tested under the new program, but the overall number of product samples will stay about the same. Category 1 plants are tested at least once every 2 years and category 2 plants at least once a year, while category 3 plants may be tested "multiple times" annually, she told CIDRAP News. She said the switch to the risk-based sampling approach began shortly after it was announced in February 2006.
The FSIS collects about 75 sets of product samples for Salmonella testing each month, according to the agency news release. A set is a series of samples collected at one site on successive days—51 days in the case of broiler chickens.
The FSIS's current "standard" for Salmonella in tested poultry and meat samples is a 20% contamination rate, based on baseline sampling in the 1990s, according to Reiser. To be in category 1, firms must limit Salmonella to half of that standard (10%) or less. Firms that have Salmonella in more than 10% but fewer than 20% of samples will be in category 2, and those that exceed 20% are in category 3. The names and locations of category 2 and 3 plants will be published on an FSIS Web site starting Mar 28, the agency said.
Chris Waldrop, a spokesman for the Consumer Federation of America, based in Washington, DC, welcomed the move to publish the names. "We have talked to FSIS and encouraged them to post this information on the Web," he told CIDRAP News. "They've taken a while to do it. . . . We're pleased that they're going to post that information on the site. It's good information for consumers to know."
"I think what FSIS came to see is that the Salmonella problem in poultry was bad enough that this was a measure they needed to take in order to drive the industry to change," Waldrop said.
The plan to publish names was also praised by Nancy Donley, president of Safe Tables Our Priority (STOP), an advocacy group based in Northbrook, Ill., that works to prevent illness from foodborne pathogens, but she added some qualifiers.
"If companies can be identified, that's something we hope will spur them on to take measures to make sure they don't get put on that list," she said.
However, Donley observed the plan will not enable consumers in a grocery store to tell whether a product came from a plant with a good record or not. "It's probably not going to be a huge benefit to the consumer. But it's going to make it more apparent to those of us who are watching who the good actors are and who the bad actors are," she said.
10% contamination rate in chicken
The latest Salmonella testing results published by FSIS show that 10.6% of all broiler chicken samples and 9.4% of all turkey samples tested in the third quarter of 2007 had Salmonella. The results varied with plant size. For broilers, they ranged from 8.2% for large plants to 12.5% for small plants and 21.9% for very small ones. For turkey firms, the rates were 9.6% for large plants and 8.7% for small ones; no figures were available for very small firms.
Salmonella contamination rates were much lower in nonpoultry meat products—3.0% in market hogs, 1.3% in cows and bulls, 0.2% in steers and heifers, and 4.1% in ground beef. The highest levels were found in ground chicken, 28.3%, and ground turkey, 16.3%.
In the same quarter, 74% of broiler plants (139 of 189) and 84% of turkey plants (31 of 37) achieved category 1 (a 10% or lower Salmonella rate), FSIS figures show. Twenty-three percent of broiler plants were in category 2 and 3% in category 3. For turkey plants, 16% were in the second category, leaving none in category 3.
At first glance, the 10.6% contamination rate for broiler chickens, for example, indicates an improvement from the 16% contamination rate seen in broiler chickens in 2005. However, the FSIS says that because of the changes in the sampling program, comparisons with figures from earlier years, or even with the previous quarter, are not valid. The reason is that, with the risk-based program, the plants chosen for testing are no longer a representative random sample, according to the FSIS's Reiser.
The FSIS is working on nationwide studies of the prevalence of various foodborne pathogens in order to provide a baseline for future comparisons and trend spotting, according to the quarterly Salmonella report. A 12-month study of pathogens in broiler chickens is under way, and more studies are being planned, the report says.
Shifts in sampling
As part of the switch to risk-based inspections, the FSIS is cutting back on sampling at ground beef plants producing less than 1,000 pounds a day, according to the agency announcement. Officials said those plants account for a large share of samples but a very small percentage of the ground beef supply. For efficiency, the agency will sample those plants less often but will run Salmonella tests on samples that are already collected for Escherichia coli O157 testing.
Also under the new program, category 1 plants have the option to "test new procedures, equipment, or processing techniques to facilitate improvements" in Salmonella control, the FSIS said. One thing that means is that they can increase their production line speeds, according to Reiser. As part of the deal, they must collect and test samples from each production line on every shift.
"These additional samples will provide the agency with key microbial data on attribution of human illness to FSIS-regulated products," the FSIS release stated.
The agency said it is also taking steps to uncover links between foodborne Salmonella and human illnesses. Under a plan developed last year, the USDA's Agricultural Research Service is comparing the pulsed-field gel electrophoresis (PFGE) patterns of Salmonella isolates collected by the FSIS with those of strains collected from human patients and provided through the CDC's PulseNet system.
The USDA just recently began submitting Salmonella PFGE patterns to PulseNet, according to Reiser. "Now, all will be added and clusters of illness will be looked for," she said.
Consumer groups' concerns
While welcoming the FSIS plan to reveal the names of plants with Salmonella problems, Waldrop and Donley expressed concern about some other aspects of the agency's Salmonella program.
Donley suggested that moving to a risk-based sampling program now is premature. "There's just not enough good data to support going in a whole risk-based direction," she said. "There's not a good data infrastructure."
She pointed to last year's massive ground beef recall by Topps Meat Co. as an example of potential problems with risk-based inspection. The company recalled 21.7 million pounds of ground beef because of possible E coli contamination.
If the same risk-based system had been in use for E coli, Topps Meat would have been a category 1 facility, with reduced inspections, Donley said, adding, "So it goes to show you there's a hole in that."
Both Donley and Waldrop had reservations about the possibility of faster production lines in meat and poultry plants. "If something like that happens, you'd want to make really sure that plants were still maintaining food safety," said Waldrop. "In the past you could link increased line speeds to reduced sanitation and increased problems in terms of worker safety."
"As speed picks up, it's harder and harder for inspectors to do their jobs," commented Donley. "There's certainly a worker safety issue too—those birds are whizzing by."
She and Waldrop also wondered if the FSIS policies will eventually lead to unwarranted complacency about Salmonella. They speculated that could happen if all or nearly all plants eventually meet the 10% standard for category 1.
"Eventually all the plants will be able to meet it and you won't have any further gains," said Waldrop. "The ideal would be to move it [the cutoff percentage] down on a regular basis."
Donley agreed: "If you see that type of movement, that's great, but then are the performance standards too low? There needs to be continuous improvement."
Jan 28 FSIS News release:
Mar 6, 2006, CIDRAP News story on Salmonella monitoring changes
FSIS Salmonella test results for Jul-Sep 2007
FSIS Salmonella sampling results for 1998-2006
FSIS prototype form identifying plants with Salmonella problems