In the early days of medicine, physicians might diagnose patients using bumps on their head, or dispense a couple of leeches to draw off "ill humours".
Doctors may be unaware of new practices
Yet a medieval doctor might give a more confident response than his modern equivalent if a patient asked for the evidence to support their treatment.
These days, it seems many of our "tried and tested" approaches to disease are nothing of the kind.
Researchers writing recently in the British Journal of Surgery concluded the practice of daubing patients with a disinfectant skin gel prior to operations made little or no difference to the rate of infections they suffered afterwards. Simple soap and water was just as effective.
However, despite this, it's more than likely that, in future, waking up after your operation in many British hospitals, you'll have that tell-tale orange stain around your wound. You'll have been given a treatment that doesn't work.
This isn't a single example. Many techniques in common use today don't have cast-iron evidence that they do any good. In some cases, firm evidence suggests the opposite is true.
Andrew Booth, from the School of Health and Related Research (ScHARR) in Sheffield, is assessing the proportion of modern treatments that are "evidence-based" - supported by "randomised controlled trials", which, if run correctly, give the best view on the value of a drug or device.
In the UK, researchers have assessed this in a variety of different parts of the health service, from busy GP surgeries to specialist hospital haematology units.
In many units, between 15% and 20% of the treatments offered did not have a shred of worthwhile evidence to support their use.
Eye-opener: But antibiotic drops don't work for many
Andrew Booth said the medical establishment was well aware of this.
"The public might be surprised at the low number of treatments which have evidence that they work - but doctors might be surprised that it is so high," he said.
He added that frequently, even when new research suggested clearly that doctors should stop using a particular treatment, nothing changed.
Michael Summers, chairman of The Patients Association, said patients would be "really surprised" to learn how little of what doctors did had been proven to work.
"We need to improve medical training, to make sure that doctors do know more about the effectiveness of the drugs they are prescribing," he said.
One of those doing this is Professor Paul Glasziou, director of the Centre for Evidence Based Medicine at Oxford University.
"I try to change the way individual doctors work," he says, "but really, the main thing we can hope to do is change the next generation of medical practitioners."
He can list dozens of examples where treatments are still widely used despite it being clear that all they give patients is side-effects.
"An example is PSA [prostate specific antigen] screening for prostate cancer. What the best studies tell us is that patients who have the test are equally likely to die from prostate cancer compared with those who don't," he says.
"This actually does harm, because patients who test positive may undergo unnecessary prostate surgery. But the test is still being carried out."
Even a simple antibiotic eye-drop prescription for a child's minor infection is likely to make no difference, and may help make the bacteria involved more resistant to treatment, he says.
Part of the problem for doctors is the sheer quantity of research emerging from hospitals, universities and laboratories across the globe.
"There are 90 new randomised controlled trials published every single day - this flood of information makes it very difficult for any doctor to stay up to date."
And when the evidence is disregarded, Prof Glasziou says, patients can be harmed.
When doctors measure blood pressure for the first time in a patient they should check both arms, as the readings may differ significantly.
But Prof Glasziou says this guideline isn't followed everywhere.
"I know of one case where a patient was being taken on and off his medication every couple of months simply because every time he visited the doctor, the reading was taken from a different arm.
"There are a lot of good things out there, but an awful lot of myths as well."