If only we had the same expectations for food safety that we have for road safety. To reduce traffic accidents, we practice defence in depth. We make all drivers pass a test. We ask for cars that are crash-resistant and have air bags and seat belts. We have stop signs, speed bumps, flashing amber lights. We engineer highways for a clear, long view. We have lots of police presence and random spot inspections.But when it comes to reducing accidents in the food industry, what we have is the equivalent of one police officer in every city chasing down cars that run red lights.
From this lack of deep defences comes the mad cow scare now jeopardizing the $7.5-billion-a-year Canadian beef industry. The sampling techniques used by inspectors watching over this business make focus groups look comprehensive. In recent years, Canadian governments have done annual tests on 3,400 head of cattle from among a herd of 3.5 million. U.S. governments test 19,990 out of 96 million cattle. Marijuana growers could only pray for levels of inspection like that.
In both countries, it's only diseased, fallen ("downer") cattle that get examined in the sample. Well, what if an emerging trend can't be spotted in such a small sample group, and what if cattle carry the disease for six to eight years before they get really sick? This is why most European countries examine all fallen cattle and conduct tests on samples of others as well. That can't be done in Canada, a vet with the Canadian Food Inspection Agency once said. "You can't test them all because there are too many.'
Then there's the type-two error - the key to understanding the shoddiness of food regulation across North America. It goes something like this: until one case of mad cow was discovered in Alberta in mid-May, Canadian food inspectors were spending a lot of time on the lookout for Chronic Wasting Disease among western elk, farmed in the West for velvety horns sold to the Asian folk medicine and penis-enlargement trade.
The type-two error is the regulatory equivalent of falling for the sucker punch, swerving to avoid a jab from the right - an outbreak of mad elk disease, for instance - while leaving yourself defenceless against a knockout punch from another direction.
From 1996 to 2002, 35,740 elk and deer, including all that died of disease, were tested for Chronic Wasting Disease, the elk and deer equivalent of BSE or mad cow. In the same period, 9,026 beef cattle, including some whose deaths seemed mysterious, were inspected for mad cow disease. One of these inspections, completed four months after a sickly steer was slaughtered and rendered into pet, chicken and pig feed, determined that mad cow disease had entered Canada.
Mistakes of the type-two variety are usually made when the people in charge think they've done a damn fine job of avoiding the standard, straight-ahead type-one errors, such as forgetting to sterilize equipment. Hubris, it's called, as in pride cometh before a fall. Canada's food regulators were in just that state of mind, getting ready to launch their new agricultural policy framework designed to brand our food exports as the safest in the world. Perhaps that launch will be postponed.
A few weeks before the mad cow discovery in Alberta, the Canadian government joined the United States to denounce Europe's restrictions on genetically engineered imports. Europe's food regulations, the North Americans insisted, are not science-based, as they are in Canada, where every reasonable effort is made to prevent an outbreak of mad elk disease.
Mad cow is "a European disease, and that's about it. We're free of this disease," the CFIA point man for mad cow said three years ago. And just to prove it wasn't a problem, they didn't spend much time looking for it, which is just how type-two errors are made. This is the brain wasting disease in our food system that we need to worry about.
The methods used for meat extraction - mechanically deboned meat and advanced meat recovery, they're called - stretch the meaning of meat and the limits of safety. Both methods take muscle, connective tissue, nerve fibre, bone fragments, marrow and often bits of spinal cord that make some hot dogs and hamburgers chewy and affordable. These body parts are the freeway for prions travelling up to the brains of diseased cattle.
Then the government safety folks allow the corpses of sick animals to be ground up for pet, chicken and pig feed because pets, chicken and pigs don't get mad cow disease. Of course, type-two-error experts would say, "What if some farmers feed their cattle chicken feed by mistake? What if the assembly line in the feed mill for cattle is contaminated by the previous run of chicken feed? What if the rendered chicken, pigs and pets that can be sold as part of cattle feed still carry the rogue proteins that cause mad cow?" Europeans outlaw any use of animal corpses for animal feed.
The damage-control position on mad cow is that the government needs to hire more inspectors and inspect more carefully. True enough. But the real madness is the type-two error, which can't be corrected by adding more staff. That will take a conversion experience on the part of North American food regulators that will lead them to develop protective systems based on defence in depth and the precautionary principle.