Every pharmaceutical company hopes to make a blockbuster drug - one that will net them more than a million dollars and boost their chances of making it into the highest bracket of one of the top earnings leagues in the world. The money to be made is astronomical. Pfizer, the biggest drug company on earth, made a gross profit of $42.7bn in the year to April 2005.
But, the companies will tell you, they spend a fortune on failed drugs. Thousands of compounds are tested in the labs, but few show promise. Of those that get as far as trials, six out of seven fall by the wayside. That's why the drug industry claims that it costs it $800m to research and develop each medicine. The price tag includes the failures as well as the marketing of the successes. Yet if it's a lottery, the companies show no signs of wanting to tear up their tickets.
The long journey from lab to medicine cupboard used to begin with a gleam in the eye of a researcher, usually at a university - an inkling that a certain molecule could perhaps play a part in treating a disease. Now it's more likely to start with the computerised mass screening of potential compounds. Something is identified that looks promising. Work in the lab confirms that it has an effect on, say, tissue samples or a virus. The scientists get excited. Once they have done all they can to establish what the product does and how, and whether it is helpful or harmful, the animal tests begin.
The point of the animal tests - generally on rats or mice - is to prevent what happened at Northwick Park. If the mice die or exhibit disturbed behaviour, it's time to think again. As of 2004, the Medicines and Healthcare Products Regulatory Authority, which licenses drugs in the UK, has to see the results before it will allow any human trials to begin. Investigators will now be looking very closely at the dossier submitted by TeGenero, the German company whose drug was given to trusting young volunteers in Northwick Park.
The pharmaceuticals used to run their own clinical trials. Increasingly, they now hand over the work to professional companies such as the US-based Parexel, which recruits volunteers, gets ethical approval and has its own medical staff to apply the experimental medicine. On their UK website they advertise for healthy male volunteers (women are not recruited for most trials because of potential risks to any foetus if they get pregnant). "What's in it for you?" the website asks. The answer is the satisfaction of having helped in the hunt for cures for diseases, free food with no washing-up, a medical checkup, plenty of time to study, browse the internet, play pool or sleep - and cash. Healthy volunteers are overwhelmingly students - whether from the UK, or passing through on a gap year from Australia or South Africa - in need of money.
Those first trials, involving a small number of healthy volunteers, are known as phase one. For phase two, patients with the disease that the new drug is intended to treat are recruited. This time there may be a couple of hundred, and very likely the trials will take place in several countries.
If the drug has a beneficial effect on their disease, phase-three trials are launched. This is the big time. Several thousand patients are recruited worldwide, either through doctors who specialise in the disease or through GPs. These trials have to be double-blind, randomised and controlled. The subjects are split randomly into two groups, with half given the trial drug and the other half a placebo. Nobody is supposed to know who gets which. When the code is finally broken, after a year or two, scientists will finally know whether they have a wonder drug or an expensive failure.
The company now applies for a licence, armed with boxes of trial data. In fact, the hurdle they have to jump with the MHRA is not high - they must prove the drug is safe and "effective". That means that it has an effect. It could be a small effect - it is not the business of the regulator to worry about how good the drug is. That falls to doctors and, in recent years, the National Institute of Clinical Excellence, which is supposed to compare drugs and recommend what works best for the NHS.
This long trial process, with the ultimate stamp of approval of a licence, is supposed to safeguard patients. The tight regulatory framework we have was devised in the post-Thalidomide days, in the horror of babies born with truncated limbs after their mothers took a drug for morning sickness. Yet now we have Vioxx. Thousands of patients are now suing manufacturers Merck after the company was forced to withdraw the painkiller, widely used by arthritis sufferers, in 2004. Data from its trials had emerged showing that it increased the risks of heart attack. David Graham of the Food and Drug Administration, the US drug regulator, said it could have caused up to 140,000 cases of heart disease in the US since it was licensed in 1999.
How did this happen? Merck said it tested Vioxx not against placebo - since it would be unethical to take people off their painkillers completely - but against another drug that had a protective effect on the heart. But some doctors have accused Merck of trying to cover up the problem for years.
The fact is that drug trials demonstrate only what they are designed to demonstrate. Seroxat and other antidepressants of the SSRI (selective serotonin reuptake inhibitors) class have had to be forbidden to children after a number of awful cases. A close analysis of the trial data would have shown children becoming more suicidal than they did on placebo - but no one was looking for that at the time.
It is only in the years after the trials, when not thousands but hundreds of thousands of patients are taking the pills, that we truly find out about them. All medicines have side-effects. If they affect one in 100,000, we won't know about them until the doctors are handing out prescriptions in vast numbers.
Most doctors will say that the European regulatory system is as safe as it can be. But the drug companies are heading out of Europe and the US in droves. They have found that they can cut their costs and speed up their trials if they test their new medicines in developing countries instead. This is the story at the heart of John le Carre's book The Constant Gardener, which has just been made into a film.
Could we see a real Constant Gardener scandal in Asia or Africa, as the pharmaceutical industry tests drugs on people who - unlike UK students - may not fully understand what could happen to them, or may say yes under the influence of a father or a chief?
The companies insist not. The Declaration of Helsinki, drawn up in the aftermath of Nazi experiments in the second world war, and other international ethical codes should prevent it. Even so, questionable things happen. In 2001, a scandal broke in Kano, Nigeria. Pfizer was accused of failing to get ethical approval for a trial it carried out on a meningitis drug. Its doctors had flown into the heart of a meningitis epidemic in 1996 and begun treating children randomly with either the new drug Trovan or with the "gold standard" cephtriaxone. Five died on Trovan and six on the older drug. The letter of approval from a Nigerian ethics committee was dated a year later, it emerged. But what upset Medecins sans Frontières, who were treating children nearby, was that Pfizer's doctors treated only 200 children in an epidemic that claimed the lives of 15,000.
The bald truth is that drug trials play out for high stakes, and it's hard to be sure, with so much at stake financially, that what should be at the centre of all this - improving human health - doesn't end up being sidelined.
