Forty years ago, an American biochemist discovered a new "cure" for cancer. Laetrile, synthesised from an apricot kernel, was reputed to wipe out the disease that had mystified scientists for decades.

Americans, desperate for a cure, sold their homes to move to Mexico where the treatment was freely available. The US government set up controlled clinical trials to see whether the drug was as revolutionary as claimed.

It was found that there was no evidence that it made any impact on cancer and, indeed, that it was highly toxic when used with other treatments or taken in large doses. Cancer patients on the trial taking Laetrile died sooner than those who were having conventional therapy. Yet, despite all the evidence, the Internet still has 825 Web sites for Laetrile, also known as "amygdalin".

Professor Alan Horwich, director of the clinical research at the Royal Marsden Hospital, understands why people want to believe in anecdotal evidence which appears to hold out hope when nothing else does.

"People in desperate circumstances are impatient with the slow process of scientific research," he says. "They feel abandoned when they are told that conventional medicine cannot offer them a cure. They'll say `I'm not content with that and I'll try anything, however extreme'."

However, cancer specialists agree that the controlled clinical trial is the only way that new treatments and therapies can be brought safely to the patient's bedside.

The journey from laboratory to patient is long and costly and a potential treatment will have been studied in a laboratory for 15 or 20 years before it gets to the clinical trial stage.

A typical trial can involve thousands of patients divided into two groups at random, one of which takes the new treatment while the other continues with conventional treatment. Trials can be "single" blind, where the doctors know which patient is receiving which treatment, but the patients do not, or "double" blind where neither patient nor doctor knows who is receiving what.

The double-blind controlled clinical trial is the "gold" standard, eliminating any bias on the part of the trial doctors. "In a double-blind trial you have removed all the possibilities where the doctor subconsciously selects a good quality patient to give the new treatment," says Dr Trevor Hince, scientific director of the Cancer Research Campaign.

It is a myth that new treatments given in a clinical study are automatically better than existing ones. Only 1 per cent of new drugs ever reach the market, often because what worked so spectacularly on a mouse fails to have the same impact on a human being. In any case, trials are not only designed to investigate the possibility of a cure.

"A new drug may not improve the cure rate, but it could be better because it has fewer side-effects. Conversely, it might improve the survival rate a little bit, but the side-effects might be so awful that patients might prefer to have a shorter life of better quality," explains Dr John Toy, director of clinical programmes at the Campaign.

Nevertheless cancer patients who are on clinical trials - even those not receiving experimental treatments - do remarkably better than those who are not on a trial, probably because they are being closely monitored and getting more medical attention.

Ironically, despite the thousands who die of the disease every year, there is still a shortage of suitable cancer trial patients. Less than 5 per cent of people with cancer are recruited on to clinical trials. This is not usually the fault of the patients, most of whom are eager to take part.

"People like to feel they're contributing to treating an illness from which they are suffering," says Professor Horwich. "They feel this is a terrible illness and if they can contribute something to fighting it they'd like to."

The biggest hindrance to clinical trials comes not from a short supply of patients - the most minor piece of publicity on a hopeful study results in sacks full of letters from willing patients - but doctors who, in the pared-down NHS find little enough time for their clinic patients, let alone trial patients. "Doctors simply cannot find that spare few minutes in each day to get patients into clinical trials," says Dr Toy. "The patient must understand fully why he is being asked to go into a clinical trial and will have lots of questions which must be answered. "

In addition, clinical trials need additional nurses: if you need one nurse for every 10 patients in a cancer ward you need two nurses for every 10 patients in a clinical trial - and additional pharmacologists to monitor the dose of drugs to be used.

Results from cancer studies world-wide are fed into the computer at the Cochrane Unit in Oxford, which is building up a data base of "evidence- based medicine" so that data from studies is applied in a systematic way across the country. The heart-breaking anomalies in cancer treatment in the NHS, which have meant that a woman with breast cancer in one part of the country may get lower-quality or outdated treatment compared to a woman in another part of the country, are becoming a thing of the past.

There is also a sense that the work of decades is finally bearing fruit. "We've been on a bit of a plateau for the past 10-20 years, seemingly not making much progress, but we are now suddenly at a time when we're at a corner in the street and we are about to go round it," says Dr Toy. "There are new and exciting treatments just around that corner."