DANISH professor Peter Gotzsche has been in Australia in recent weeks, arguing that mammographic screening does more harm than good. The controversy is not new to researchers or cancer specialists.
Mammograms can detect cancers that are curable and they save lives in individual cases. But they also miss cancers that are curable, cancers that grow and even spread between one mammogram and the next.
To minimise the risk of missing anything, we have to recall a woman for extra tests whenever we see something on a mammogram that is remotely suspicious. This creates anxiety, even though most women recalled are ultimately reassured.
Mammography also detects pre-cancerous changes and slow-growing cancers, some of which would not trouble a woman in her lifetime. We can’t predict which cancers (or pre-cancers) are going to be a problem, so we treat them all. This is what has been called “overdiagnosis” and “overtreatment”.
The benefits and the problems are acknowledged on both sides of the screening debate. The controversy is about the balance of good and harm and the argument is not easily resolved.
Survival from breast cancer has improved, but there are many confounding factors that make it difficult to know how much screening has contributed to that.
The amount of harm resulting from overtreatment can’t be measured directly, because we don’t know who has been overtreated. Furthermore, the psychological effects of screening can be difficult to assess. Some women feel reassured by being screened, but screening also creates anxiety through false alarms.
Despite these difficulties, we can usefully estimate the maximum potential benefit of screening. About 13per cent of women will get breast cancer, and only 3per cent of women will die from it. Ignoring the possibility that women are killed by overtreatment, and making some other assumptions favourable to screening, we would still have to say that mammographic screening prevents less than a third of all breast cancer deaths, and thus prevents no more than 1per cent of all deaths in women.
Or, to put it another way, we can’t save more than one life for every 100 women on a screening program. It is important to remember that such an overall benefit has not been proven – this is just an optimistic estimate of the possible benefit from our current screening techniques.
Further, we need to understand that the benefit doesn’t come with one mammogram – 100 women would need to attend diligently for annual or biannual screening for 20 or 30 years in order to (potentially) save one life.
Of course, a life saved will not be saved forever. The target population varies, but the minimum age is 40, and the average age of people being saved would be over 60. The life expectancy of those people would be less than 90. Thus the real “best case scenario” for mammographic breast screening would be 30 years of life saved for every 100 women on a screening program, or an average improvement in life-expectancy of less than four months per person. (Being a non-smoker will let you live, “on average”, more than 10 years longer than a smoker.)
Of course, luck and life aren’t shared evenly like this. For most women, mammography won’t save their lives. For some, it will cause them unnecessary harm. But for a few, it will allow them to see their grandchildren grow up.
Mammographic screening, from the consumer’s point of view, really is a bit like life insurance. You make a lot of small investments over a long period, and you hope to look back and say it was a waste of your money.