April 29, 2004

Tracking Cattle

Think retinal scanners. Bluetooth wireless data technology. Linux operating system. Radio frequency identification (RFID) tags. Global positioning systems.


Processors use tech to track meat progression
By Michelle Kessler, USA TODAY
To fight mad cow disease, a Colorado slaughterhouse is using a high-tech computer system worthy of a spy movie.

Think retinal scanners. Bluetooth wireless data technology. Linux operating system. Radio frequency identification (RFID) tags. Global positioning systems.

Putting such cutting-edge technology in a wet, messy slaughterhouse might seem a little strange. But Swift & Co., the nation's third-largest meat processor, says it's the best way to make sure steak is safe.
About Managing Technology
Managing Technology is an occasional feature to help companies, business travelers and even consumers make strides through technology. To suggest a topic, e-mail [email protected].

Tainted beef can kill. Mad cow disease, or bovine spongiform encephalopathy, is a deadly animal illness that's passed to humans who eat meat from a sick cow. The usually fatal human version of the disease has infected 155 people worldwide since 1995, the Food and Drug Administration says.

Cattle infected with mad cow usually appear ill. But healthy-looking animals can harbor pathogens such as E.coli O157:H7, a bacteria which usually doesn't hurt cattle but can kill humans. About 60 people die each year of E.coli poisoning, the Centers for Disease Control says.

One of the best ways to prevent deaths is carefully tracking cattle, says Colorado State University food science professor Patricia Kendall. A good tracking system would allow public health officials to quickly contain outbreaks and recall tainted meat.
Practicality of Swift system questioned
The Optibrand cattle-tracking system is so new that Swift & Co. is the first major meat processor to try it. Optibrand, a start-up, has just 10 full-time employees and is on the "cusp of being profitable," says CEO Bruce Golden, a former Colorado State University professor.

But Optibrand has broad aspirations. It hopes to someday track cattle throughout their lives. Often, cattle are bred by one rancher, then passed to another to raise. Once the cattle are mature, they're sent to a third farmer to fatten up before slaughter. Only then are they sent to the slaughterhouse.

A cow that had its retina scanned at birth could easily be identified at a later date. That might prevent situations such as the one that occurred in December, when the first case of mad-cow-tainted beef appeared in the USA. Officials had a very hard time identifying the source of the disease and experts still aren't 100% sure they got the right cow, Golden says.

Some say Golden's vision isn't practical. "It's quite a bit of money," says North Carolina State University food science professor Kevin Keener. "In my opinion, (better slaughterhouse tracking) isn't going to significantly improve things."

Many beef industry experts say it will be very difficult to unite the notoriously segmented industry, which has thousands of small ranchers. "Some of these guys don't even have PCs, aren't even connected to the Internet," says Swift Vice President Gary Acromite.

Colorado State food science professor Patricia Kendall says it will take a national system to make big improvements in food safety. But the system Swift is using "is a huge step in the right direction."

By Michelle Kessler

That doesn't happen today. Most slaughterhouses in the USA process cattle about 20 at a time. Beef is usually identified with that lot number, but no more specific information about which cow it came from. Other countries, including England and Japan, have stricter regulations stemming in part from mad cow outbreaks. No one is yet believed to have caught mad cow in the USA.

If a piece of beef here tests positive for a pathogen, everything in its lot usually must be recalled � often as much as a tractor-trailer load, says North Carolina State University assistant food science professor Kevin Keener. A recall is a tiny percentage of the more than 26 billion pounds of beef U.S. processors produce annually, the Department of Agriculture says. But tracking tainted beef can be a race, since consumers usually eat beef not long after they buy it, Keener says.

Swift hopes its new system, installed in its Greeley, Colo., packing plant in November, will track a steak back to the cow it came from. That could limit the size and improve the speed of recalls, Swift Vice President Gary Acromite says.

It will also help Swift track the best suppliers, he says. Today, "I really couldn't tell you why I get really great steaks ... one week, and the next week, I don't," Acromite says. Swift hopes to someday put a special sticker on tracked beef, indicating quality. "Customers will pay a premium for that," Acromite says.

Tagged, scanned, photographed

Beef tracking is harder than it might sound. Slaughterhouses are bloody, damp places not conducive to computer systems. Cows don't come in a standard size and are cut into pieces before they're sold.

Swift is overcoming those challenges using a new handheld scanning device from Fort Collins, Colo., start-up Optibrand. The scanning device runs on the Linux operating system. It looks similar to a bar-code reader you might see in a supermarket checkout lane, with a base station and scanning wand. The device is packed with features, including:

�Retinal scanner. Workers use the wand to scan a cow's eye. The wand takes an image of the blood vessels, then uses a computer program to turn the pattern into a unique ID number.

�Digital camera. The retinal scanner turns into a regular digital camera with the push of a button, letting ranchers and processors snap pictures of their herds.

�Bar-code reader. An attachment that clips on to the wand turns the camera into a reader for the bar codes many ranchers put on plastic tags in cows' ears.

�Radio frequency identification tag reader. An optional second wand can read RFID tags, computer chips containing information that can be automatically detected by nearby sensors. Cows are occasionally implanted with RFID tags, and that's likely to become more common.

�Global positioning system. The base station is equipped with GPS, which uses a satellite feed to pinpoint its location. Optibrand's GPS system can identify the location of almost anyone on the planet, to within about 50 feet.

�Bluetooth. All data collected by the base station can be sent to a nearby computer via Bluetooth wireless data technology. Bluetooth uses radio waves to send information short distances. It is most commonly used to sync electronic devices, such as personal digital assistants, or PDAs, with computers.

Cows that come into the Swift plant might or might not have some form of ID from the ranch. Swift starts by retinally scanning every cow, so each has a unique ID number. One plant worker scans the eye of each carcass as it zips by. The plant worker also scans in bar codes, RFID tags or other information.

All that data is wirelessly transmitted to a computer in Swift's back office. That eliminates the need for wires that might not last in the rough environment, Acromite says.

Tracking gets trickier when cows are prepared for butchering. When a cow's head is cut off, Swift attaches a plastic tag with its ID number from the retinal scan to the carcass. That tag stays on until the carcass is cut into pieces.

It's too hard to attach a tag to every piece of meat, Acromite says. Swift uses probability to match a steak with the ID number of the cow it came from. If Swift is butchering one cow every minute, the company knows that steaks cut between 2:15 and 2:16 probably come from a certain cow. While not perfect, probability can narrow a steak's origin to a handful of cows, Acromite says. "Rather than having it come from a lot of 20 animals, I've now got it down to two or three," he says.

Hamburger is trickier. Typically, when cows are butchered, odd cuts of meat are tossed into a big bin, then ground into hamburger. A burger might have meat from dozens of cows. Swift is working on a probability system for hamburger, but so far hasn't come up with anything feasible, Acromite says.

Swift's system cost "hundreds of thousands of dollars," Acromite says. He won't give details, but Optibrand says a scanner costs about $1,800. Back-end computers can cost thousands more. Smaller ranchers might purchase only the handheld scanner, then upload the data to a PC as one uploads photos from a digital camera, Optibrand CEO Bruce Golden says.

Optibrand collects a fee for each cow scanned. Infrequent users pay about 90 cents a scan, while bigger ones might negotiate volume discounts, Golden says.

Swift says the money was well spent. In addition to quality controls, the system might put Swift ahead of new tracking regulations expected from federal regulators concerned about mad cow. "In the event that the government mandates RFID tags for individual animals, this technology will be able to capture it, and whatever else is coming at us," Acromite says.

USATODAY.com - Processors use tech to track meat progression

Posted by Craig at April 29, 2004 04:30 PM