Variant CJD burst on to the public and political scene at 31 minutes past three on the afternoon of March 20, 1996 when the then Health Minister, Stephen Dorrell, made his infamous announcement in the House of Commons of 10 cases of an unusual kind of CJD in young people, and their possible link with BSE. The price of British beef was not the only thing that collapsed that day. So did the trust of the public in government ministers, government experts, and government pronouncements. A heavy price had been paid for a decade of reassurance that beef was safe. To prevent panic, there had long been a policy of sedating the public and of "leaning into the wind", a process described to the BSE Inquiry by a civil servant as making "more reassuring sounding statements than might ideally have been said". At the end of the day its effect had been the exact opposite of its intention.

Many other mistakes were made during the handling of BSE. The BSE Inquiry concluded that although the right policy decisions were taken, they were often marred by the time taken to reach them, and the lack of vigour in considering how to bring them about and in their implementation, enforcement and monitoring. Nevertheless, it has to be acknowledged that policy makers faced about as difficult a challenge as could be imagined. The infectious agent that caused BSE was new, so assuming that its properties would closely resemble its relatives ? prions, like the one that causes scrapie in sheep ? was a gamble. But testing this notion was very difficult. Prions neither stimulate the production of antibodies nor circulate freely around the body, so there are no blood tests for them. Their infectivity resists many disinfectants and survives heating in hospital autoclaves under conditions that kill bacterial spores ? the traditional test objects ? with room to spare. They grow very slowly. This means that tests for infectivity are also very slow ? many months at the quickest, and often years. It also means that it takes a very long time to find out whether control policies are working.

The average incubation period of vCJD has been estimated to be 16.7 years. This puts an almost insuperable obstacle in the way of finding out how vCJD cases became infected. Even in ordinary food poisoning outbreaks where the longest interval between exposure and illness is a few days it is often difficult for victims to remember what they ate at the crucial time, vital information for the investigators trying to track the poisoned food. So it is not surprising that epidemiologists have failed to find factors pointing unequivocally to the routes of transmission of infection. This makes it very difficult to predict how many cases of vCJD there are going to be.

The epidemic of BSE in cattle in Britain is nearly over. The number of cases peaked in January 1993. Operation of the over 30 months rule (preventing older cattle being eaten) and the removal at slaughter of the tissues and organs where prions are found in infected animals (specified bovine materials) means that the likelihood of prions entering the food chain is vanishingly small. But there is another problem. It is the transmission of CJD from human to human by medical procedures.

This is not a hypothetical risk. So far more people world-wide have died of classical CJD contracted in this way than from vCJD, the majority either after treatment for short stature with growth hormone extracted from human pituitary glands (although this stopped in 1985 when hormones prepared by genetic engineers became available, the long incubation period means that recipients are still dying ? four in Britain in 2003) or dura mater (the tough membrane lining the brain) grafts. However the number of patients treated with these materials pales into insignificance in comparison with those who routinely receive blood transfusions. The experience of the transmission of other infections this way has been a bitter one. Blood donations are screened for syphilis, hepatitis B, hepatitis C, human immunodeficiency virus and the T-cell lymphotropic viruses because of it. What about CJD? Although experiments in rodents show that prions can be found in the blood early in infection, searches for infectivity in the blood of CJD patients using the most sensitive tests have given negative results. Epidemiological studies have shown that previous transfusions do not increase the risk of developing CJD. No cases of CJD have been reported in haemophiliacs, who receive more blood products than any other group. So the risk remains hypothetical. But most of these studies relate to classical CJD. vCJD is different.

In the light of these uncertainties the government invoked the precautionary principle ? that a lack of full scientific certainty should not be used as a reason to avoid taking preventive action. It had received scientific advice that the removal of white cells from donated blood would be a sensible and practical measure to take against the hypothetical risk because if infectivity were to be present in blood, it would most likely be in these cells. It introduced this measure ? leucodepletion ? in July 1998. At this time the number of cases of vCJD was rising sharply ? from three in 1995 to 10 in 1996 and 1997 and 18 in 1998. It seemed a very real probability that a big epidemic was starting. But even if subsequent events suggest otherwise ? it currently appears that case numbers peaked in 2000 (28 cases, so far in 2003 there have been 16 deaths) this measure remains uncontroversial, even if it costs money that could be spent on other things. It is practical ? filtering out the white cells is a routine procedure ? and has other benefits, because it reduces the number of febrile reactions to transfusions, reduces the risk of transmitting cytomegalovirus (which can cause major problems in patients with organ transplants) and reduces unwanted immune responses.

Leukodepletion could be seen as a good example of the precautionary principle at work. But it was not a hard test for politicians to pass. There was a consensus view from their scientific advisors ? whose opinions they hardly ever went against ? that it would be a good thing. It was technically achievable. It brought other benefits to transfusion recipients. And it demonstrated to the public, still concerned about vCJD and BSE, that vigorous action was being taken. A far harder test of the principle is policy making about GM food, which epitomises a lack of scientific certainty. Bearing in mind Sam Goldwyn's aphorism "making predictions is difficult, particularly about the future", I am glad I am a scientist who responds to uncertainty by saying "more research needs to be done", rather than a politician, who has to take decisions in the here and now!