All doctors must have a bit of a God complex. It is not just the Harold Shipmans whose heads get turned by the power they have over their patients. Even a benign physician may stand next to us while we sweat and groan in our bed, and experience a frisson of omnipotence as he (it's still usually a he) ponders what cure to try out. It must feel marvellous, as Druin Burch, a hospital doctor from Oxford, admits: “I have injected people with morphine and seen their fear, their misery and terror dissolve.”

Burch argues that the God complex of doctors - their blithe conviction that their actions will surely produce benefits - is the most dangerous of the many hazards created by medicine over the ages. Until recently, doctors' belief that they did good was an almost total delusion. “Doctors, for most of human history, have killed patients far more often than they have cured them. Their drugs and their advice have been poisonous. They have been sincere, well-meaning and murderous.”

This thesis has already been put forward in David Wootton's ground-breaking history, Bad Medicine, published in 2006, a work to which Burch's owes a lot. Wootton showed that until 1865, when Joseph Lister proved the efficacy of antiseptic surgery, medicine had done more harm than good. The value of medicine is not relative. Either it works - by curing diseases, alleviating pain and deferring death - or it doesn't. Before 1865, Wootton demonstrated, very little medicine actually worked; and even after that date, medicine was far less important in saving lives than other factors such as better food and hygiene.

Burch's book lacks Wootton's sheer originality and tight historical focus. Where Wootton divided his story neatly into the Hippocratic age (bad medicine) and the modern age (good medicine), Burch veers forwards and backwards in time over 25 short thematic chapters. Each chapter, however, is a self-contained pleasure to read, like mini-fables on the perils of medicine. And there are ways in which Burch, with his robust and imaginative prose style, takes the story further than Wootton. As a doctor, he takes an even more pessimistic view about just how long bad medicine has continued. “Bleeding and mercury have gone out of fashion,” he writes, “untested certainties and over-confidence have not.”

One example he discusses is thalidomide, which doctors in the 1950s happily prescribed to pregnant women without any real information as to its effects. Thalidomide, for Burch, is a parable of misplaced confidence. The women who took it were confident that their doctors knew what they were doing, as were the doctors themselves, who accepted the reassurances of the drug company that it could be given “with complete safety to pregnant women and nursing mothers”. Only after thousands of children had been born with severe deformities were doctors shaken out of this complacency, and thalidomide banned. “The point of the thalidomide story is the fragility of knowledge.”

Doctors spoke then - and still speak - of “risks and benefits” when they should really speak, Burch argues, of “harms and benefits”, because the possible harm of any given drug is always every bit as real as its potential advantages.

Then again, even some genuinely useful drugs have encouraged doctors in dangerous habits. Burch does not deny that penicillin was a great discovery, “a miracle drug”. The trouble was that it encouraged doctors in their complacent belief that they could figure out what a drug did “simply by giving it and seeing what

happened” without a careful statistical weighing up of risks. Penicillin cemented the God complex, by making doctors trust even more in the power of their individual judgment.

Most therapies, however, unlike penicillin, cannot be judged meaningfully in this way. Take aspirin. It has been used (in the form of willow plants) since ancient times, yet its single greatest use - in preventing death from heart attack - was confirmed only in the late 1980s. This is because the percentages involved (it drops your absolute risk of dying after a heart attack by just 3%) were too tiny to be observed by individual doctors. It took many large-scale, randomised double-blind tests to confirm the truth. Yet, as Burch says, that 3% matters. “Aspirin, properly used, saves 100,000 lives worldwide” every year.

Burch thinks that statistical trials are “lovely” things, which have finally brought “comfort and healing” into hospitals. The rise of evidence-based medicine is the great remedy to the God complex of doctors. “The idea that even the most reasonable-sounding theories should be subjected to tests probably has more potential to make the world a better place than all the drugs that doctors possess... Humility, even more than pills, is the healthiest thing that doctors have to offer.”

It remains to be seen, though, whether we patients - sweating and groaning for someone to save our lives - are ready to stop looking for salvation from the men in white coats just yet.

Taking the Medicine by Druin Burch

Chatto £20 pp336

Times Recommends

• The 100 Best Books of the Decade
• The Times picks the 50 best paperbacks of 2009
• The conversation: James Ellroy