Most of us presume that doctor knows best, so when they smile benevolently and say, "Take three a day. It's the newest treatment so it must be wonderful", or, "It's the most time-honoured treatment, it must be superb", we swallow the pill. We rarely ask doctors how they can be so sure that what they're offering will help - not harm - us.
Much of Sir Iain Chalmers' professional life has been devoted to just this question, and to ensuring that we are likely to get treatments that will do us good. As he says, "Right now, the NHS is determining what treatments we will be allowed, but do we know what criteria they are using to decide? Testing treatment must be fair, but some trials end up concluding that useless or even harmful things are good, or that useful ones are useless."
Chalmers is currently running the James Lind Library, a website where we can find out how decisions are made about what treatments work. It explains the idea of "fair tests" and illustrates how these have evolved.
Named after the 18th-century naval doctor who proved that citrus fruit cured scurvy, Chalmers explains: "In 1747 Lind was at sea with a bunch of sick sailors. He wanted to test various treatments and took 12 ill men, gathered them in the same bit of the ship and fed them the same diet. The only major difference was what he treated them with."
Lind compared the effects of cider with tonics of sulphuric acid vinegar, seawater, nutmeg or two oranges and a lemon. Within six days, one of the sailors given fruit returned to normal duty, while the other was so well he got the job of nursing the rest. Lind concluded that the fruit had worked, and the admiralty eventually insisted that lemon juice be given to all sailors.
Chalmers says: "Lind's experiment progressed many areas of medicine, from understanding more about vitamins to public health and the organisation of trials - hence my wish to celebrate him in the name of this multinational venture."
The website contains a host of beautiful reproductions from original works, including Lind's 1753 Treatise of the Scurvy, and facsimiles of the 1969 report suggesting that thalidomide caused limb deformities and the 1974 study which showed that aspirin averted heart attacks.
There is also information about how trials should be conducted. The many stages of "fair tests" include deciding what questions need to be asked and minimising effects of chance and bias. Chance can largely be avoided by testing treatments on enough people: a new drug that helps 500 out of 700 people is more convincing than if it helps five out of seven. But abolishing bias is trickier.
Chalmers says: "Bias takes many forms. Researchers might report only what shows the drug they're testing in a good light, or publish positive results in an English journal and bury negative findings in a foreign language paper. Then there are studies where new treatments are tested on relatively well people, the old ones on an iller group. Sounds incredible, but it happens."
Chalmers also points out that registration of trials is not legally required. "Results can be ignored. Some trials are publicly registered when they begin, but most aren't. The bad stuff can just go in the bin. Patients who agree to trials assume they're contributing to a knowledge base, but this behaviour leads to biased, imprecise estimates of treatment effects, can harm us all and is scientific as well as ethical misconduct."
Strong stuff from a man who was knighted for services to medicine, but Chalmers has his reasons for being so forthright. "My motivation comes from making some really bad mistakes as a doctor, or doing what I was told because someone thought it worked, and harming patients in the process." Chalmers gave up clinical medicine after "seven pretty uneasy years" mainly in obstetrics and paediatrics and he is passionate about helping patients to make informed choice based on the evidence of what works.
He recalls his early days as a paediatrician. "I bought Dr Spock's book in 1965, which said put babies to sleep on their fronts. I promulgated its message. But Spock had got it shatteringly wrong, basing his advice on theory not practice, and babies died as a result."
Jan Vandenbroucke, professor of clinical epidemiology from Leiden University in the Netherlands, is one of Chalmers' three co-editors of the library. In Chalmers' Oxford office, their loving, almost parental glances switch between the computer screen where the library winks away, and a pile of 18th-century books including Lind's Treatise in English and French. Vandenbroucke fears that people may not realise how much these books can bring to understanding 21st century medical care.
"It's really important because it shows how we got to the present organisation of healthcare. It shows that current methods to decide whether treatments work are part of a centuries-old continuum. We hope people will learn humility from it, openness of thought.
The library owes its existence to the Royal College of Physicians in Edinburgh, which has one of the world's best collections of historical documents of this kind. He says: "We want the library to contain information which people wouldn't see, and show them how testing has evolved."
The library's future thus lies in the past, as such material is gradually incorporated. But Chalmers says there is one area where he hopes breakthroughs will force them to widen its scope. "Sometimes a new treatment is hyped before it's tested, so people have really high expectations. This must affect trial outcomes, but how do you measure it? When someone cracks this, we'll cover it as part of trial evolution."
As the library develops (comments are invited) it will not shirk from the dark side of experimentation. Chalmers says: "A section will look at atrocities in the middle ages where prisoners were poisoned to see if the antidote worked, and the appalling abuses during the second world war in Nazi-occupied Europe and Japanese-occupied China. One way to honour those who suffered and died is to ensure that their data are not lost, or kept secret as has been done by American authorities."
The library is dedicated to patients and professionals who have helped to generate reliable evidence. Chalmers believes that making such work better known will empower us to ask the right questions about treatment that we are offered, rather than trotting off like lambs to the chemist with a prescription for whatever it says on the coffee mug that the drug company gave the doctor.
