Dr Daniel Sokol
Many mistakes are all in the mind
Three years ago, Dr Jerome Groopman, a renowned Professor of Medicine at Harvard Medical School, was "on rounds" in a large hospital, teaching doctors and medical students about the patients on the wards and their various ailments.
At the end of the teaching session, he felt uneasy: "Something was profoundly wrong with the way they were learning to solve clinical puzzles and care for people".
This prompted him to ask a "simple" question: How should a doctor think?
His answer - published last month as a book, How doctors think - is a profoundly important contribution to medicine.
Don't expect tests to be infallible and don't be scared to ask for a second opinion
Its relevance extends beyond the medical profession. It also affects patients and potential patients alike. In other words, all of us.
Dr Groopman's insight is that doctors, especially when confronted with more complex cases, tend to think in ways that run the risk of error, but that the likelihood of error can be reduced if patients actively participate in the consultation.
Errors in thinking
When we think of medical errors, most of us imagine a GP prescribing the wrong dose of medication or a surgeon amputating the wrong leg.
Around 15% of all diagnoses are wrong
But Dr Groopman notes that most medical errors are not technical, but cognitive. They are, in other words, errors of thinking.
Diagnosis momentum, for example, refers to the tendency of a diagnosis to gain certainty as it is passed on from doctor to doctor.
Some way down the line, an originally tentative diagnosis is accepted as true - "Mrs Jones, a 36-year-old woman with multiple sclerosis" - and this may prevent the search for the real diagnosis.
Another is confirmation bias, which refers to selectively choosing evidence to support a presumed diagnosis and rejecting information which does not "fit".
To illustrate this, Dr Groopman recounts the story of Blanche Begaye who, struck by a flu-like illness, decided to drink plenty of orange juice and take aspirin.
Still, she got worse. Her doctor, who had recently seen several cases of pneumonia, made the diagnosis of subclinical viral pneumonia.
It was only later that she learnt that the patient did not have pneumonia, but aspirin poisoning.
She saw a range of symptoms typical of pneumonia (such as fever and rapid breathing) and dismissed the contradictory signs, such as the normal white cell count and the uncharacteristic chest X-ray.
Those signs, she believed, simply indicated an early stage of infection.
Many diagnoses wrong
These cognitive biases explain in part the alarming statistic that around 15% of all diagnoses are wrong (an optimist would say that a healthy 85% are correct).
Applying simplistic methods to complex conditions can lead to problems
In a recent US study, 100 radiologists were asked to comment on whether a series of 60 chest X-rays were normal.
They disagreed among themselves 20% of the time.
When re-reading the same X-rays later that day, they disagreed with their own previous conclusions between 5% and 10% of the time.
The lessons: don't expect tests to be infallible and don't be scared to ask for a second opinion.
Dr Groopman writes frankly about his own experience as a patient.
After repeatedly injuring his wrist, he visited six hand surgeons over three years and obtained four different opinions about the cause and treatment of the injury.
Different types of bias
Some of the surgeons displayed some sort of cognitive bias.
Consulation time is tight
One, for example, favoured action - surgical intervention - over inaction ("commission bias"), while another showed 'search satisficing', the tendency to stop thinking once a plausible explanation is found.
For Dr Groopman, he learnt that, contrary to popular belief, the 'surgeon's brain is more important than his hands'.
Why do mistakes occur? One reason is that doctors, usually out of necessity, often use shortcuts and 'rules of thumb'.
Reality, however, can be messy and applying simplistic methods to complex conditions can lead to problems.
Doctors, like everyone, are also subject to emotions and stresses that can affect their ability to think rationally.
They may be tired, depressed, or angry.
Not enough time
Finally, as Dr Groopman notes, "good thinking takes time" - and time is a scarce resource in the health service.
The average length of a GP consultation in the UK is roughly nine minutes and one study shows that doctors interrupt patients within 18 seconds of asking them to describe their symptoms.
Although medicine is fraught with uncertainty and some errors are unavoidable, patients can help reduce their frequency.
Asking simple questions such as "What else can this be?" to prompt the doctor to think of alternatives or, in A&E, even asking 'what is the worst thing this can be?' can prove useful.
Requesting a second opinion or questioning the value of an apparently ineffective treatment can also be worthwhile.
Such involvement from patients is not, for Dr Groopman, a confrontation.
On the contrary, through active participation the patient becomes a vital partner in the quest for health.
Dr Daniel Sokol is a medical ethicist at Keele University and Director of the Applied Clinical Ethics (ACE) programme at Imperial College, London.
Jerome Groopman's How Doctors Think (2007) is published by Houghton Mifflin Company.