What’s the evidence for Evidence-Based Medicine?

•As a medical student, I attended a lecture by breast surgeon Michael Baum. At that time, how breast cancer was treated depended largely on where the patient lived and on how hard she argued for the treatment which she believed to be right. Surgeons often rejected trials because they thought they knew best; so did the patients. There was no place for evidence-based medicine. But I have never forgotten what Baum said: “Every woman who is not in a controlled trial is in an uncontrolled trial.” Finally his trials established what treatments work better than others, and mutilation and death rates have decreased.

As an SHO I contributed to a DRCOG revision book. I tried to track the origins of the words of wisdom from Ten Teachers that were quoted to all students and repeated in all the other textbooks. I could find no evidence to support many of the ex cathedra statements which determined how young doctors were going to treat their patients.

Early converts to evidence-based medicine sometimes preached their message with an air of religiosity which irritated their colleagues. Now, when new treatments are presented we expect to see the research that supports them. So why is evidence-based medicine still so often regarded with suspicion?

Firstly, is the evidence really ALL the evidence? Dr Ben Goldacre has shown that we rarely have all the data because trials are withheld. So available data is skewed – generally in favour of the drug under consideration. AllTrials may make a difference, but meanwhile if you read a trial in which doing nothing, or providing a cheap treatment, wins over something new and expensive, believe it. Regard the rest with a raised eyebrow.

If you aren’t a statistician, who do you believe?

Secondly, even if all the evidence is available, how trustworthy are the conclusions? Even papers which appear to be game-changers may be re-interpreted by other statisticians, leading to very different conclusions about the appropriate clinical strategies. If you aren’t a statistician, who do you believe?

Even where analyses are based on all the available evidence and are generally supported, there are problems. If I can’t understand how to interpret the statistics, few of my patients are going to. David Spiegelhalter, Professor of Public Understanding of Risk in Cambridge, knows how to get a point across. I like his nice graphics showing NNT (Number Needed to Treat) and NNH (Number Needed to Harm). Take a look at his website Understanding Uncertainty and play around with the animations. Find out what works for you, and so what may work for your patients.

Spiegelhalter’s graphics may help you, and the patient across from you, to make decisions about treatment. But in real life how many patients are the white middle-aged male with a single problem who is recruited for trials? Mrs Hussein is 84, has at least two other chronic diseases and is already taking six different medications. Will the extra pills add to her expectation of life or to her problems? If your patient is not quite like those envisaged in the guidelines you are likely to be directed off the pathway and into a swamp of uncertainty.
Another problem with EBM is the assumption that one size fits all. David Sackett, who introduced EBM, observed that EBM is not cookbook medicine. But managers and governments like recipes. EBM provides them with yet more tools for measuring what is easy to measure about doctors’ performance, not what is meaningful. EBM is about populations while GPs are treating Mrs Jones. But we are under pressure to tick boxes. So patients, whose views are supposed to be paramount, can find themselves under pressure to conform. And locums too. Do your employers ask you whether the patient was satisfied or whether you ticked the QOF boxes?

What price clinical acumen these days? Isn’t that what makes an experienced doctor different from a computer? Isn’t that what makes triage safer in the hands of doctors than non-doctors with algorithms? But how do you tell good judgment from dogmatic intransigence? A cynic once said that a surgeon’s idea of a trial is “In my experience …” (one patient); “in my series … “ (two patients); “time after time after time …” (three patients). Alternative practitioners call it ‘custom and practice’. When is doing something for which there is no evidence inspired and when is it like prescribing snake oil?

EBM looks backwards. Each recommendation needs to be reviewed as new evidence become available. Progress is made by challenging authority and dogma. Anecdotal medicine, good observation and making links between two hitherto apparently unconnected pieces of knowledge have to be fostered. So we need more than a register of clinical trials register. We need to share and develop experience, n-of-1 trials, good ideas, speculations. We need evidence-informed medicine, in recognition that EBM, like all tools, needs to be used with skill and judgment.

We need to change from EBM to EIM (Evidence-Informed Medicine) in recognition that evidence, like all tools, needs to be used with skill and judgment. •

Judith Harvey

Judith Harvey was a research scientist, ran the VSO programme in Papua New Guinea and taught in a Liverpool comprehensive school before going to medical school. She has been a partner, a salaried GP and a locum, an LMC chair and a long-time supporter NASGP. Her charity, Cuba Medical Link, enables medical students to go to Cuba for their electives.

No Comments Yet.

Leave your comments