The Guardian – Thursday 9 February 2017
A new study shows that doctors are actually quite bad at estimating the benefit and harm associated with treatments they prescribe. It’s a wake-up call for doctors, but patients can also play a role in getting better treatment
An unimpressed nurse summons the oncology fellow to the chemotherapy chair. “I am not prepared to treat him with chemo. He can barely stay awake.”
“But his oncologist wants to push on,” the fellow responds.
“The patient doesn’t seem to understand how sick he is or how chemotherapy is doing harm. You’ll need to sort this out, I am afraid.”
The fellow sighs, caught on the horns of a dilemma.
Elsewhere, an elderly woman has taken warfarin, a blood thinner, for some time, and now presents with a massive cerebral bleed. She was going to the kitchen one moment and unconscious the next; she is expected to die shortly. As I console her stricken son, it emerges that she had sustained 50 falls that year leading up to the fatal one. There had been many doctor visits but no one had asked specifically about falls.
At the desk, as I solemnly write a note, I overhear the same exchange that’s going on in my head.
“Fifty falls!” one dismayed resident says. “Why would you put her on warfarin?”
“Because someone wanted to reduce stroke risk and someone else watched her heart disease but no one thought of the whole patient.”
“What were they thinking?”
If you listen to doctors and nurses, this is one of the most common questions you will find them grappling with and grumbling about. It reflects part genuine puzzlement and part exasperation that what one doctor has recommended seems ill-advised or even inappropriate to another.
The Grattan Institute estimates the cost of wasted healthcare dollars to be in the order of a billion dollars and the figure stings clinicians but as a disillusioned young doctor sighed, in the age of super-specialisation, it seems expedient to let every doctor manage “their own organ”. Except the practice harms patients who are after all, more than a collection of organs.
If highly trained doctors don’t understand their colleagues’ intentions it stands to reason that most patients feel even more hapless, caught in an endless tangle of tests and explanations but the knowledge and power asymmetry is such that it’s impossible to question the doctor, who must surely know better (if not best).
Unnecessary and expensive medicine is at an all-time high and the usual reasons given are patient expectations, financial incentives, therapeutic uncertainty, medico-legal fears and the sustenance of hope. Now a new study in JAMA Internal Medicine authored by two Australians points out that when it comes to unsound medicine, there is another element at play. It turns out that when prescribing a drug or ordering a procedure doctors are actually quite bad at estimating the benefit and harm associated with it.
In a systematic review of 48 studies performed in 17 countries and involving more than 13,000 clinicians, they found that doctors rarely had accurate expectations of benefits or harms. The inaccuracies were in both directions but more often, harm was underestimated and benefit overestimated.
No group of doctors fared well. As a result, children with acute ear infections may be overprescribed antibiotics and women with troublesome postmenopausal symptoms may be deprived of hormone replacement therapy. Obstetricians and neurologists underestimated the risk of birth defects from antiepileptic drugs and GPs overestimated the benefit of prostate cancer screening and underestimated the benefit of warfarin for atrial fibrillation, a common heart condition. Transplant surgeons were biased towards an inaccurately low estimate of graft failure and all types of doctors were unaware of the risk of radiation exposure from imaging.
Physicians overestimated the effect of some interventions on life expectancy by as much as 30% and for elective but by no means inconsequential surgery on the thyroid, lung, prostate and uterus, there were clinicians who believed that complications “never occurred or had a rate of zero”. Dermatologists couldn’t agree on psoriasis treatment and psychiatrists differed on the risk of harm from long-term antipsychotics. There was a reluctance to convey a numerical estimate of benefit and worryingly, clinicians “overwhelmingly recommend the interventions they provide”.
This study is a wake-up call for doctors because it speaks to our collective failure to appreciate that in prescribing more for our patients we don’t always help, and indeed, commonly inflict harm. The goal of good medicine is not only to avoid harm but also to provide actual benefit, a distinction that’s commonly blurred, including in oncology. Chemotherapy at the end of life improves neither quantity nor quality of life. It leads to more invasive procedures and greater likelihood of dying in an intensive care unit but patients continue to receive it.
In the reign of evidence-based medicine it is discomfiting news that doctors may not understand the data in the form of hundreds of thousands of studies poured upon us.
First, as any patient knows, the art of medicine matters as much as its science. Evidence applied without tact, consideration, empathy and an understanding of the patient’s perspective can be as harmful as evidence not applied at all. Doctors are increasingly exhorted to provide collaborative care and practice shared decision-making. The catch is that both art and science suffer when we don’t know the facts or struggle to convey them.
Part of the problem is the sheer volume of publications. Entwined in increasing bureaucratic demands many doctors lack the time and also the confidence to interpret academic research so we turn to (commonly paid) expert opinion, “peer influencers” and biased pharmaceutical advertising.
Medical schools run the obligatory statistics course but don’t ingrain in doctors that their interpretation of a journal article or more commonly, an “advertorial”, and their participation in marketing disguised as “literature” peddled by pharmaceutical representatives has a direct impact on patient experience, the cost of care and wasted healthcare dollars. Hospitals who should care even more about such education virtually ignore it and when it’s volume, not quality of care that’s rewarded, it all but extinguishes the desire to do better.
Meanwhile, what should patients do? The JAMA study suggests that doctors frequently don’t know and certainly, don’t know best. This is vexing but not all doom and gloom because doctors now have at their disposal an unprecedented number of sound guidelines, robust protocols and genuinely plain-language information for patients, not to mention easy web-based access to experts. When it comes to doctors seeking advice the world really is a global village. In a world of rapidly evolving information, patients should be prepared for a doctor to say, “I don’t know” provided this is followed by, “but I’ll find out.”
Here are three questions that every patient should ask of every new proposed drug or intervention:
If patients asked these questions more often and doctors took it upon themselves to answer faithfully, medicine might yet experience a new dawn.