An opinion poll commissioned for the government's 'Better Regulation' task force was curiously under-reported earlier this year. Not many people know, therefore, what a vivid picture it painted of the fallout from the BSE fiasco.

I have heard it said many times since the then agriculture minister, John Gummer, stuffed a beefburger down his poor daughter's throat that public confidence in the government's handling of risks to public health was at an all-time low. But here is resounding confirmation.

MORI asked its respondents who they would most trust to advise them on the risks posed by BSE. Top of the poll came independent scientists at 57 per cent, streets ahead of the next most trusted source, farmers (22 per cent). Government scientists scored a dismal 17 per cent. Right at the bottom came government ministers (4 per cent) and other politicians (2 per cent).

When the question was turned around and they were asked who they would least trust for advice, MORI's respondents put government ministers right out in front with 41 per cent, the next highest scores going to food manufacturers at 28 per cent and government scientists at 25 per cent.

It is unfortunate for Labour that it inherited this legacy of distrust. After all, it wasn't responsible for the botch-up. No wonder it decided to play safe and ban beef on the bone. But did that help? The same poll asked how strongly people supported various laws that had been - or could be - passed to help tackle general safety hazards. Eighty per cent strongly supported compulsory wearing of seat belts and 42 per cent said they would strongly support a complete ban on smoking in public places. But when asked about banning beef on the bone, only 5 per cent strongly backed this, with 32 per cent strongly opposed.

Opinion polls usually make good media stories, so why did this one receive so little attention? Were journalists too wounded to report it? When MORI asked people who they would generally trust to tell the truth (about anything), journalists came bottom with only 10 per cent. Government ministers scored an ignominious 17 per cent and government scientists 38 per cent. Doctors and teachers came top with 89 and 88 per cent.

Perhaps the poll didn't get reported because it wasn't very surprising. 'Public has no confidence in government's handling of health risks' might be considered old news, unworthy of even passing comment. But there was at least one stick of dynamite in the findings.

People were asked about a range of potential health risks, including too much alcohol, smoking, unhealthy diet, genetically modified food, MMR vaccines and unpasteurised milk. How well-informed did they feel about each of these? And did they think any of them posed a serious risk to themselves or their families?

Most people felt well-informed about smoking, alcohol and diet (90, 66 and 61 per cent respectively). Only 16 per cent felt well-informed about GM food, yet 31 per cent thought it posed a serious risk. In fact, more people thought GM food posed a serious risk than any other issue on the list, except smoking (43 per cent) and diet (32 per cent). People wanted new legislation on it more than on any other health risk. The poll was conducted in January, before the first scientific evidence came to light suggesting that GM potatoes damaged the immune system of rats.

Finally, 80 per cent of those polled agreed that, when the government is unsure of the facts, it should nevertheless publish what information it does have available. And 93 per cent agreed that the government should be more open about how it makes decisions.

These findings show that bungling one risk can seriously damage the government's chances of being trusted in the future. After a change of administration, the public is not inclined to trust the incoming government any more than the outgoing one.

For Labour, GM food is a risk just waiting to be bungled. People feel they are ill-informed about it and judge it to be a serious risk. The combination of Lord Sainsbury at the Department of Trade and Industry and Lord Haskins (of Northern Foods, major supplier to supermarkets around the country) as chair of the 'Better Regulation' task force, is worrying - regardless of their personal integrity.

The three-year moratorium on commercial production of GM crops provides a welcome breathing space. But what will happen when the three years

are up? It is most unlikely that

by then there will be incontrovertible evidence, backed by scientists and lobby groups on all sides, that GM foods are, or are not, safe to be eaten.

There will have to be another round of decisions - and that means another chance for government to restore or further demolish public confidence.

Three years ought to be long enough to overhaul the way government handles risks. People want to be protected against risks, but they know that accurate - if incomplete - information from a trusted source is a better form of protection than

half-truths and empty promises.

The government must stop thinking that its first duty is to reassure the public, because most of the time it can't. It must re-position itself as an honest broker between the experts and the people. That means owning up to its own lack of omniscience, setting out the facts as far as they are known, identifying gaps and making sure that, as far as possible, future research helps to improve our understanding of the risks that confront us.

It also means acknowledging that the quest for evidence is more often about opening up new uncertainties than about arriving at the truth and closing down the argument.