When the 18th-century naval surgeon James Lind compared citrus fruits with other remedies for scurvy among sailors, he could hardly have guessed that what he was doing would eventually become routine in medical research. Still less would he have predicted that the business of making simple comparisons could ever generate hostility of the kind experienced earlier this year by the paediatric department of North Staffordshire Hospital trust.
During the past 50 years, the clinical trial has moved centre stage in medicine. And until everything worth knowing about the treatment of illness has been discovered, there it will stay. But as trials have grown bigger, more elaborate, more costly, more regulated and more hedged about with ethical dilemmas, so too have the headaches they pose for their organisers.
The principle on which all trials are founded is deceptively simple. To find out if a treatment works you give it to one group of patients, withhold it from another, and then see how the two lots make out. A similar procedure can be used to compare one treatment with another. The devil, of course, is in the detail, and trial organisers now go to great lengths to outwit their chief enemy: bias.
When assessing different treatments, might a clinician be tempted to, say, allocate the sicker patients to one rather than the other?
Randomisation can take care of that: determining the choice by chance eliminates the doctor's preferences.
But that is not enough - the patients' expectations of benefit from any prescribed treatment also have to be discounted. So those allocated to a 'no-therapy' group have to share the illusion of being treated. Hence the use of placebos.
The patients, clearly, must not be party to this knowledge; they must be kept 'blind' to it. The same rules apply when a new drug is compared with an old one; the very fact that the drug is newer will enhance its placebo effect.
And then there are the doctors' attitudes. If they know who is receiving the active or the newer or what they may suspect to be the better alternative, might they convey their enthusiasm or lack of it to patients?
Hence the final elaboration of this already ornate edifice: the double blind controlled trial, in which neither patient nor doctor knows, until a code is broken, who is getting what. There are other and subtler variants, but this core model is the standard.
The hurdles which now confront anyone planning a trial are the result of a range of social and technical developments. The effect of the first antibiotics on life-threatening infections was so dramatic that trials were scarcely necessary - although they were, in fact, done. Most new treatments offer benefits which are more modest. So a trial with the statistical power to reach a definite conclusion may require not hundreds but thousands or even tens of thousands of subjects.
The arch exponent of big trials, Professor Richard Peto of Oxford University, points out: 'It might be reasonable to hope that a new treatment for acute stroke or acute myocardial infarction could reduce recurrent stroke or death in hospital from 10 per cent to 9 or 8 percent. . .but not to hope that it could halve in hospital mortality.'
This, though, is worth having: even aspirin prevents 10,000-20,000 deaths from heart disease per million people treated. But to demonstrate improvements which - though proportionately small are potentially beneficial to large numbers of people - requires big multi-centre trials.
Organising these takes more than amateur enthusiasm and back-of-the-envelope calculation.
The Medical Research Council - which pioneered trials, and is still the main source of state funds - runs about 150 at any one time, half of them on screening and other nondrug forms of intervention. Most are multi-centre studies. You can glean an insight into the logistics of running a modern trial by glancing through the MRC's own guidelines.
Once the principal investigators have persuaded the council to accept their research protocol and fund the project (itself no mean feat) they have to recruit a trial management group and devise elaborate systems of safety reporting, data handling, and recordkeeping. They have to organise a procedure for dealing with complaints from dissatisfied subjects.
They have to make arrangements for independent supervision of the trial, which means setting up a trial steering committee, and a data and ethics monitoring committee . They need to take account of quality assurance, quality control and audit.
They must design patient information leaflets, consent forms and case report forms. And so on.
In Britain, most patients who enter trials are recruited by the doctors involved, or referred by sympathetic colleagues. Advertising plays a limited role, though this could change. In America, lists of clinical trials are already posted on the web.
Some organisers make use of media contacts. But recruitment isn't always successful, and trials can fail before they have begun.
The advantage to patients of joining a trial lies in their chance to benefit from a new drug before it is generally available, and the possibility of receiving care which may not be better but is likely to be more systematic. The disadvantage, of course, is equally self-evident: the new may be less efficacious than the old, or even damaging.
Here, as elsewhere in medicine, a clutch of ethical issues has emerged.
Informed consent, for example - research ethics committees insist on it, but a handful of clinicians claim that too much information may upset patients and even subvert the purposes of the trial. And because ethics committees don't operate to an agreed set of guidelines on this or anything else, a multi-centre trial approved by one institution may be damned by another.
Compensation for untoward effects of the treatment can also be contentious. In cases of negligence, the trial volunteer has the same rights as any other patient who has come to grief. But should there also be compensation for 'non-negligent' harm?
The MRC says it gives 'sympathetic consideration to claims', but other organisations may take a tougher stance. The deluge of publicity which descended on North Staffordshire Hospital trust during February this year illustrates some of the pitfalls.
The trial at issue was comparing the standard method of ventilating newborn babies with an alternative system. The study left some babies dead or brain-damaged.
The parents said they didn't know they were taking part in a clinical trial; the researchers countered by showing signed consent forms. The parents pointed out that signing a form when you are distressed about your baby is hardly the time to expect informed consent - and that the alternative ventilation was not as good as they had been led to believe.
The researchers conceded that although some babies in the experimental group came to grief, the numbers who did so were not significantly different from those receiving the conventional treatment.
And so it went on.
The sadly self-evident moral of this tale is that trial organisers have to make great efforts to be sure that patients, or parents, really do understand what is being proposed.
One remedy may be to enlist patients in helping to design trials. This, like several other radical developments in methodology, grew out of AIDS activism in the US. After a series of noisy protests, the US National Institutes of Health began to hold open meetings of some its AIDS research committees, and eventually conceded the principle that patients should have a voice. The trend is less advanced in Britain, but there is now a body called the Consumers' Advisory Group for Clinical Trials (see box, right).
A newer ethical twist concerns the fate of patient volunteers once a trial is over. In general, patients stand to gain if a trial demonstrates that a particular treatment is effective. But now, with biotechnology offering us ever more expensive drugs, their funding is not automatically assured (see 'Rollercoaster ride' box, top of previous page).
A great many trials rely on the drug industry. Companies which have funded a trial of one of their products, only to see it failing in some way, are understandably reluctant to publicise the findings. Periodic protests by aggrieved researchers illustrate that this does happen - with consequences which range from inappropriate prescribing to bias in the medical literature.
The chair of GlaxoWellcome, prompted perhaps by criticisms of his own company, recently contributed an editorial in the British Medical Journal outlining a more open policy on clinical trials.
'By law, ' he wrote, 'we are required to include all trials involving a product in the regulatory submission for that product. The problem for decision makers and prescribers is that much of this information is not in the public domain.'
GlaxoWellcome has decided to organise a register of all its future trials. 'The objective of this policy is to help those undertaking systematic reviews of clinical data and to help reduce the impact of publication bias. . . We have also committed to publishing all clinical trials, as far as this is possible.'
As yet it is still too early to judge whether Glaxo's policy will become the industry standard.
Funding bodies - along with whatever happens to tickle the intellectual fancy of individual academics - also have a powerful influence on what research gets commissioned in the first place.
Jenny Simpson, chief executive of the British Association of Medical Managers, would like to see a closer correspondence between the need for research, and what actually gets done.
But drug companies, medical charities and most other nongovernmental agencies have agendas set by the interests of shareholders, trustees and donors. Only the state may be left to pick up what she refers to as the essential but 'non-sexy' stuff.
Many doctors who organise and take part in clinical trials are full-time academics; research is part of their job. But for most NHS staff, contributing to a trial makes extra demands on time and resources.
'This is a concern of both medical and non-medical managers, ' says Jenny Simpson. Indeed, some are less than enthusiastic about the prospect of their hospitals taking part in trials.
When all the practical, ethical and financial hurdles have finally been overcome, and the results of the trial are published, there is still no guarantee that anyone will read them or, having read them, act on them. The evidence that 'clot-busters' - drugs designed to dissolve blood clots following a heart attack - increased survival was available years before their use became routine.
One response to these hold-ups is the Cochrane Collaboration, an attempt to extract useful findings from the world literature and present them in a form that clinicians and managers can use to choose treatments and formulate policies.
Besides the sense in which it is used here, the word 'trial' can also mean 'a painful test of one's endurance, patience, or faith'. There must be many triallists for whom this second meaning is easily the most apt.
Rollercoaster ride: multiple sclerosis patient in beta interferon trial
Anuradha Bahl, 35, from Chadwell Heath in Essex, has multiple sclerosis. In 1995 she enrolled in a phase III trial of beta interferon (which costs around£10,000 per year) financed by the drug company Schering.
She later discovered that she had been on the placebo during the blinded first phase of the trial. But for the past year she has been given the active drug in an unblinded extension of the trial, designed to give researchers a longer follow-up period.
'Since I've been on the active drug I haven't had a relapse for 11 months, ' she says. But the trial is due to finish in the autumn, along with drug company funding, leaving the future uncertain.
Will her local health authority, Redbridge and Waltham Forest, pick up the bill?
In a statement the HA has said it is reserving judgement on the use of beta interferon for secondary progressive MS until central guidance is produced by the National Institute for Clinical Excellence, which is expected to look at the treatment of MS with beta interferon as one of its priorities. The HA 'is not therefore making available any additional funding for treating this form of MS with beta interferon', it says.
But a spokesperson stressed that in the meantime, each case will be considered on its merits, after assessment in a specialist neurologist centre. The HA funds the drug if the patient's MS 'is similar to that of patients who were found to benefit from interferon therapy in the original trials', its statement says.
But with the deadline for completion of the trial approaching, Ms Bahl has still heard nothing. 'I'm starting to panic, ' she says. 'I know how helpful the drug's been. I feel it's giving me a new lease of life. If it's taken away in September. . .' She leaves the sentence incomplete.
The three types of clinical trial Clinical trials for new drugs normally fall into three categories.
Phase I studies are first attempts to investigate the safety of new drugs in humans. The researchers aim to find out if they have any toxic effects on a limited number of healthy, often paid, volunteers and, if so, the dose at which these appear. Phase I studies may also include laboratory investigations. Some agents, notably those for cancer, are tested at this stage on patients.
Phase II studies usually take longer, and are for investigating the potentially beneficial actions on patients. There will be more subjects than in phase I trials, and only half will receive the drug itself. The remainder, acting as a control or comparison group, will be given either a placebo (dummy pill) or - where there is one - the current treatment for the condition.
Phase III studies are similar to the above, but this time they involve from several hundred to several thousand patients in an attempt to assess the drug in something approaching normal use. Only when phase III trials are complete will companies apply to license it.
Putting consumers in the driving seat The Consumers' Advisory Group for Clinical Trials was started in 1994 by its chair, Mrs Hazel Thornton.
In 1991 Mrs Thornton was diagnosed with breast cancer, and invited to take part in a trial. On reading the information provided she decided that the outcomes chosen by the researchers were not right for her, and that the protocol left too much scope for over or under-treatment. She declined to take part, but began to think about trial methodology and informed consent in general.
An article in The Lancet led to an invitation to speak at a conference; a meeting with the surgeon Professor Michael Baum led to the setting up of CAG-CT 3.The group consists of patients and professionals. 'We see ourselves not as consumer advocates or representatives, ' says Mrs Thornton. 'We're advocates for high quality in research.'
But part of her definition of high quality is consideration for the needs and feelings of patients. If these are disregarded the research cannot, by definition, be of high quality.
So the group helps researchers plan trials which are acceptable and understandable to potential participants.
One such trial was the ATAC (Arimidex and tamoxifen alone or in combination), designed to compare the two drugs (Arimidex is one of the newer anti-oestrogens) in preventing the recurrence of tumours in post-menopausal women who have been treated for invasive primary breast cancer.
The reassurance that the protocol had been approved by CAG-CT may have contributed to the trial organisers' success in managing to recruit 9,000 patients in the first two years.
1 Peto R, Baigent C. Trials: the next 50 years. Br Med J 1998; 317 (7167): 1170.
2 MRC Guidelines for Good Clinical Practice in Clinical Trials. 1998.
3 Thornton H. Breast cancer trials: a patient's viewpoint. Lancet 1992; 339 (8784): 44-45.
4 Sykes R. Being a modern pharmaceutical company involves making information available on clinical trial programmes. Br Med J 1998; 317 (7167): 1172.
Geoff Watts is presenter of Radio 4's Leading Edge