- Fiona Godlee, editor, BMJ
When the Marmot report on breast cancer screening was published in November last year, many will have hoped for an end to this particular controversy. Prompted by calls in the BMJ for more honest information on the harms of screening (BMJ 2010;340:c3106, BMJ 2011;343:d6894), Marmot and his committee were charged with asking whether the screening programme should continue, and if so, what women should be told about the risks of overdiagnosis.
As nicely summarised by Nigel Hawkes at the time (BMJ 2012;345:e7330), the committee concluded that the programme should continue because it did prevent deaths—43 deaths from breast cancer prevented for every 10 000 women invited for screening. The downside was an estimated 19% rate of overdiagnosis: 129 of the 681 cancers detected in those 10 000 women would not have caused symptoms or death during the patient’s lifetime.
But despite this higher than previous estimate of overdiagnosis, the critics of breast screening have not been mollified. Michael Baum is one of them. In an article this week he takes the Marmot report to task (doi:10.1136/bmj.f385). Its estimate of harms was based on three old and shortish randomised trials, and the analysis takes no account of the improvements in treatment since these trials were done, which will reduce the benefits of screening. Nor does it make use of more recent observational data. With these data included, estimated rates of overdiagnosis as a result of screening increase to up to 50%.
This matters because it can change the decisions women make when invited for screening. Jolyn Hersch and colleagues explored attitudes to screening in a sample of 50 women in Australia (doi:10.1136/bmj.f158). The first point to note is that most of the women were surprised when they were told about overdiagnosis. As to whether they would attend screening, most said they would if overdiagnosis rates were 30% or lower, but a rate of 50% made most of them reconsider.
Attitudes might also harden if we were less coy about what we mean by overdiagnosis. In almost all cases of screen detected breast cancer, overdiagnosis means overtreatment. And according to Baum, this means that while deaths from breast cancer may be avoided, any benefit will be more than outweighed by deaths due to the cardiopulmonary and other adverse effects of treatment.
What of the future? In her editorial, Cliona Kirwan sees the benefits of breast cancer screening being eroded by more effective and less harmful treatments (doi:10.1136/bmj.f87). At what stage must we seriously consider whether this screening is a good use of £96m of the NHS budget? (www.cancerscreening.nhs.uk/breastscreen/cost.html) Since new and better trials are almost certainly out of the question, we must, says Kirwan, be meticulous in collecting and modelling population data to guide women and their medical advisers in this important decision.
Cite this as: BMJ 2013;346:f477