It’s easy to assume the earlier women are screened for breast cancer, the better. And a recent US study, which found screening women with mammography from the age of 40 saved the most lives, generated headlines around the world.
We need to be cautious, however, when interpreting studies like this and the media reports they create. That’s because with screening, its benefits – less risk of dying of cancer – are clear, and are easily exaggerated. But the potential harms of screening are harder to recognise and readily overlooked.
Read more: When talking about cancer screening, survival rates mislead
The recent US study compared an intensive screening strategy (strategy a) of annual mammograms for women starting at age 40 all the way to 84 years of age, with two less-intensive screening strategies.
Strategy b) offered annual mammograms from 45-54 years, then every two years until age 79. Strategy c) offered mammograms every two years from 50-74, the same screening policy we have in Australia.
Strategy a) has become known in the media as “screening from age 40” but it is really screening more often, and until an older age (when breast cancer is more common), as well as starting earlier. It prevented the most deaths, according to the modelling. But at what “price”?
By screening longer and more often, the more intensive strategy a) required women to have three times as many mammograms. It caused three times as many false positives or false alarms (when women didn’t end up having a breast cancer despite an abnormal mammogram), as the least intensive strategy c). In a major omission, the authors did not address potential harms of overdiagnosis and overtreatment (more below).
So what questions do you need to consider when reading reports about studies like this?
1. Who’s giving the advice?
Three specialist radiologists and a medical physicist, all employed by departments of imaging or radiology, authored this study. Screening mammography in the US is big business. The total annual cost of screening mammography there was estimated to be US$7.8 billion in 2010.
So, why should we be concerned? Because previous research has found financial conflicts of interest increase the risk of bias, and lead committees towards recommendations that are more favourable towards mammography screening.
Non-financial conflicts of interest can also affect recommendations. Guideline panels with radiologist members are more likely to recommend screening for women from age 40 years than recommendations issued by panels without radiologist members.
2. What aren’t they telling me?
A mammogram may seem harmless, but it can cause long-term problems that many people would never think of. An important one is finding harmless, idle or dormant cancers, a major factor in overdiagnosis.
Overdiagnosis is common not just in breast cancer, but in screening for prostate, thyroid and lung cancer. How common? When a UK panel carried out an independent assessment of the benefit and harms of screening mammography, it found the chance of a woman being overdiagnosed by screening was three times greater than the chance screening would save her from dying of breast cancer.
Even the chief medical officer of the American Cancer Society urges accepting overdiagnosis and overtreatment as harms of breast cancer screening.
Yet the authors of this latest US study didn’t consider overdiagnosis and overtreatment when concluding annual screening from age 40-84 years is best.
Read more: Five commonly over-diagnosed conditions and what we can do about them
Overdiagnosis is important because it isn’t a good idea to have cancer treatments (surgery, radiotherapy and antihormone pills) for a harmless cancer (overtreatment).
Each of these treatments comes with risks of side-effects, as UK woman Elizabeth Dawson describes in her blog. Two and a half years after starting treatment she was still wondering whether the cancer that was found by screening was overdiagnosed or not, and whether she needed all, or even any, of the treatments she’d had. She hates that the drugs she’s still taking to prevent a recurrence make her bones frailer. She’s been told not to go out when it’s icy because she might fall and fracture, but she hates the idea of being housebound at 56 when she feels so well and active.
The US study did include false positives in its calculations, but may not have recognised fully the impact. Being recalled for an abnormal mammography is scary. But what is less well known is that even three years after being declared free of suspected cancer, women with false positives consistently report worse psychosocial outcomes; they report feeling more dejected and more anxious, and report worse sleep and negative impacts on sexuality than women with normal mammograms.
Mammography uses radiation, so there’s a small chance the screening process itself may induce cancers over time. But starting screening from 50 and screening every two years is estimated to reduce the number of induced cancers five-fold compared to annual screening from age 40.
3. What’s the health-care context?
The US has a very different health-care context to Australia. In the US, mammography screening costs are paid by many different organisations. So debates over recommendations may have implications for whether health plan organisations cover services or not.
In contrast, as part of our national cancer screening programs, BreastScreen Australia provides mammograms in a national, publicly funded program that offers high-quality screening to eligible women, for free.
The health-care context is also relevant when we consider an individual woman’s risk of breast cancer. This debate (about when to start screening and how often) is relevant to women at average risk of breast cancer. For women with a strong family history, or who know they carry a breast cancer genetic mutation, screening more intensively offers greater benefits.
So which strategy really is best?
The ultimate aim of screening is to reduce deaths from breast cancer. Yet, whichever screening strategy we use, screening is not 100% effective.
It probably reduces the risk of dying from breast cancer by about 20%, at most by 40%, and perhaps as little as only a few percent.
So we must balance this limited benefit with a clearer picture of harms like overdiagnosis and overtreatment to avoid tipping over into net harm.