Why that ‘cancer’ on a mammogram might not be anything to worry about

Too often, both doctors and patients fail to appreciate how much uncertainty there is about the benefits of tests and treatments – and this can lead to unnecessary harm. That’s the suggestion in a new book by Steven Hatch, an assistant professor of medicine at the University of Massachusetts in the U.S., who says we need to properly weigh up the risks and benefits before we make decisions about our healthcare.

Picture a game where you have to identify snowballs thrown from 100 ft away in the midst of a raging blizzard.

You don’t know how many will be thrown, how often or how fast. You must simply look into the whiteness and decide whether you see randomness or something worthy of attention.

Too often, both doctors and patients fail to appreciate how much uncertainty there is about the benefits of tests and treatments – and this can lead to unnecessary harm, says professor Steven Hatch

This concept has a special resonance in medicine, for it is how radiologists often describe the challenge of detecting a breast tumour on a mammogram.

Mammography does not give black-and-white answers, but is quite literally a grey technology – images produced by an X-ray of the human body show up as various shades of grey.

Separating cancer from non-cancer is not as simple as is commonly thought, and the interpretation can depend upon whose eyes are doing the reading.

• Patients who pay £1,000 for life-saving drugs online that…
  

  Simple test can prevent thousands like Terry bleeding to…
  

  ME AND MY OPERATION: Blast of heat reduces the damage of…
  

  Can suncream RAISE your risk of skin cancer? Think…
  

DOCTORS AREN’T ROCKET SCIENTISTS

It’s not just mammograms (a subject I will revisit shortly) – looking for a snowball in a blizzard is a useful metaphor for uncertainty, which is pervasive in medicine.

It may surprise you to learn doctors do not often ‘know’ what they are doing with the same mathematical precision we associate with rocket scientists or chemical engineers. A diagnosis is often a conjecture; a prognosis typically less certain than that.

See it as a spectrum of certainty – on the left is the ideal, where we have a high level of confidence that we really know something and that this something indicates clear benefits. For example, there is overwhelming evidence that vaccination saves lives and is associated with almost no harm.

Doctors do not often ‘know’ what they are doing with the same mathematical precision we associate with rocket scientists or chemical engineers. A diagnosis is often a conjecture [file photo]

On the centre-left of the spectrum might lie the treatments that we are reasonably confident have real benefits. For instance, some diabetes drugs save lives, but several have pretty serious side-effects, so we can’t assure every patient they will be beneficial.

Approaching the middle is the realm of pure speculation, where evidence is contradictory or lacking altogether. For example, much current research is devoted to the impact of the gut microbiome – the billions of bacteria in our intestines – on human behaviour, though exactly what is happening is anyone’s guess.

Towards the right side is where we become more certain again, this time about the harms of a drug or diagnostic approach.

Finally, the far right side is where we’re quite confident that a practice is harmful – such as taking antibiotics long-term with no evidence of bacterial infection.

I would argue that uncertainty is the great unspoken secret of medicine, and by ignoring it we do real harm to ourselves.

‘RISKY’ CLOTS THAT MAY NOT HURT YOU

Possibly the most important real-world example of this misguided faith in certainty is when we find a disease that isn’t a disease – or overdiagnosis.

Take a blood clot on the lungs, a pulmonary embolism (PE). Potentially life-threatening, they are difficult to detect because their very nonspecific symptoms – such as fever or rapid heart rate – can be confused with other conditions.

And until the Nineties, there was no reliable test for them. Then came CT pulmonary angiography, which provides a detailed picture of the lungs and their blood supply, and in just eight years the incidence of PEs nearly doubled.

However, the death rate hardly budged during this time. The CT enabled us to find many PEs, but those we have found largely haven’t helped prevent death.

Moreover, any treatment comes with risks of harm – blood-thinning medication for PE, for example, carries potential side-effects such as uncontrolled bleeding.

We’ve created technologies that are ever more sensitive at detecting abnormalities we call cancer

THE CASE AGAINST CANCER SCREENING

The same problem exists with conditions such as prostate and breast cancer.

An analysis of U.S. cancer statistics from the Seventies to early this century shows that several cancers have remarkably similar patterns.

The number of diagnoses year after year tends to rise, while the rate of death remains basically the same. Why? It’s almost certainly because we’ve created technologies that are ever more sensitive at detecting abnormalities we call cancer.

At cell level they are cancers, but not the kind of cancers that threaten lives – basically ‘cancer’ that isn’t cancer in the way doctors and patients understand the term.

I would argue that the practice of using mammograms to screen otherwise healthy women under the age of 50 (ie, those with no symptoms such as lumps) may be on the centre-right spectrum of certainty.

In other words, there is a minimal to moderate amount of evidence that mammography in this group is harmful – in addition to the stress of a false-positive mammogram, there are more sinister complications, such as unnecessary mastectomies, chemotherapy and radiation.

One of the more sobering assessments came in 2004, when Scandinavian researchers estimated that as many as one in every three women told they had invasive breast cancer in fact had no cancer at all.

Overdiagnosis also affects cancer survival rates. If you are alive and cancer-free five years from your initial diagnosis, you are considered cured.

Yet if twice as many people are diagnosed with a given cancer, but the death rate from that cancer remains unchanged, the five-year survival rate will appear to double, making the screen look like an even better bargain.

THE TRUTH ABOUT DEPRESSION PILLS

The issue of uncertainty can be seen in common treatments, too.

Take SSRIs (selective serotonin reuptake inhibitors) for depression. Their effectiveness is often measured using a ‘rating scale’ – where a patient’s responses to questions suggestive of depression are given points. The worse the symptoms, the more points.

It’s much harder to distinguish mild from moderate, and moderate from severe, during tests for depression

A depression scale can be fuzzy and subjective, but one can still separate those who are not depressed from those who are severely depressed with great confidence.

Unfortunately, it’s much harder to distinguish mild from moderate, and moderate from severe. And that fact becomes very important when we consider the value of SSRIs.

Taken as a group, most studies have shown they produce a mild clinical benefit, while those with high scores (ie, those more depressed) tend to reap the biggest benefits.

Yet this hasn’t stopped SSRIs from being marketed as something of a panacea, with approximately one in 15 Britons taking an antidepressant. And once patients start an antidepressant, they tend to stay on it.

Whether there are any discernible benefits for patients is less clear – long-term use of antidepressants is very much in the middle of the spectrum in terms of benefits.

Drugs are neither miracles nor curses – they are both. Their value can be properly assessed when the size of the benefit is weighed against the risks of their use.

Patients should therefore ask their doctor lots of questions, not only about the benefits but the risks of treatment, and the consequences of declining it.

Another useful question is: how many people like me need to take this drug before one life is saved?

The more you read about statistics and probability, the better you will become at grasping what’s really at stake with any medical recommendation.

Extracted from Snowball In A Blizzard: The Tricky Problem Of Uncertainty In Medicine by Steven Hatch. Copyright Steven Hatch 2016, published by Atlantic Books on June 2 at £14.99. To order a copy for £11.99 (20 per cent discount), call 0844 571 0640 or visit www.mailbookshop.co.uk before June 7, 2016. PP free on orders over £15.