The U.S. Department of Agriculture, in charge of keeping the potentially lethal E. coli bacteria out of your food, has concluded its own testing process for ground beef is flawed and may be missing bacteria during tests.

These findings come from an audit released Monday by the Inspector General’s office of the USDA. It warns that the current sampling method “is not designed to yield the statistical precision that is reasonable for food safety or to verify that plant controls or interventions are working as intended.”

The audit makes four recommendations for improving inspections of the nearly 4 billion pounds of ground beef produced annually in the United States. These recommendations include developing a redesigned sampling program to provide “higher confidence” in the testing regime.

The audit was done at the request of Rep. Rosa DeLauro, D-Conn.

Most of the ground beef consumed domestically is made of beef trim, the various bits left on a carcass after the choice cuts have been butchered. A device similar to a carrot peeler is used to slice roughly 4-inch pieces of trim off the cow, which are then stacked into large bins and sent off for testing before being ground. Under the current N-60 method, inspectors test 60 of these slices for E. coli. The process is overseen by the Food Safety and Inspection Service (FSIS).

Phyllis K. Fong, the USDA Inspector General, has warned that, in situations where E. coli is present in 1 percent of the inspected bin, the current screening method would miss it over half the time. Or, as the report puts it, “if the contamination level is very low, FSIS is more likely to miss contamination than to detect it.”

At the same time, she expressed confidence that planned audits into the testing of beef trim should help bolster public confidence that FSIS’ tests are accurately identifying E. coli and ultimately preventing contaminated meat from being distributed and consumed.

The report acknowledges that increasing the number of samples taken by inspectors at processing plants is “impractical,” given the time and cost such sampling requires. Instead, it recommends that FSIS move towards a system that would allow them to identify the highest risk plants, focusing inspection resources where they are most needed. 

DeLauro issued a statement saying she was very disturbed by the report’s findings and that “the report’s conclusion represents a significant public health threat.”

Activists in the food safety world have argued for years that the current E. coli testing method is superficial. “The problem is you’re just checking the top, not the bottom or the middle,” says Tony Corbo, a senior lobbyist with Food & Water Watch, a food watchdog group. “These bins usually hold 2,000 pounds of meat product, and they’re only checking 60 pieces.”

Given the testing method, Corbo argued, statistics showing that the industry has reduced E. coli in recent years are flawed. “They simply can’t make that claim, based on how they’re collecting the samples,” he said.

Other experts caution that the findings may be overblown. Dr. Keith Belk, a professor at Colorado State University’s College of Agricultural Sciences, said the findings do not relate directly to the redundancy measures in place, such as a series of E. coli checks that occur after the trim has been ground — something FSIS points out in their response to the IG’s findings. To assume this means the system has failed because of these findings would be incorrect, according to Belk, and he pointed to the fact that industry and FSIS spends more money testing “than any other two countries combined” as proof that all parties have a vested interest in keeping meat safe.

Escherichia coli is a bacteria that takes various forms, but is best known for its O157:H7 strain that can lead to painful, and possibly deadly, food poisoning. It is often found in the intestines of cattle and other livestock, for whom the bacteria is harmless, but can be passed on to humans when feces on the animal’s skin get into the meat during the butchering. Fecal contamination can also spread if the cow’s gut is punctured during the process. The bacteria can also be found in milk products such as cheese, as the contamination can pass through the udder of the cow.

There have been few large outbreaks of E. coli in recent years. The Centers for Disease Control lists 10 outbreaks going back to 2006. The most recent outbreak from ground beef contamination occurred in November 2009, as 26 people across eight states became ill. Nineteen people were hospitalized, and two died.

Overall, the CDC estimates there are roughly 70,000 cases of food poisoning related to E. coli every year, at the cost of $405 million to the economy.

Activists in the food safety world have argued for years that the current E. coli testing method is superficial, despite FSIS numbers showing a year-by-year drop in the potentially harmful bacteria. Between 2008 and 2009, E. coli samples found in federally inspected plants dropped from 0.47 to 0.30 percent of the tested material. (2010 year end numbers are not yet available from FSIS.) The American Meat Institute cites a 45 percent reduction in E. coli tests since 2000. But the AMI also said that no sampling test can truly eradicate the risk of the bacteria. “Individual test results do not—and cannot— guarantee that a product is ‘free’ of E. coli O157:H7,” they write.

AMI’s Executive Vice President James H. Hodges released a statement saying the report hit on “a central dilemma” faced by the industry. “The more we reduce E. coli O157:H7 in beef trim, the more tests we will need to conduct to find it.”

Despite USDA’s pledge, in the wake of the findings, to explore new methods of inspecting for E. coli, Belk agreed that the meat industry will never be 100 percent safe. “Zero risk does not exist,” he said. “Only with politicians and political science does zero risk exist.”