The Journal of the Royal Society of Medicine published an article this week levelling the following criticisms at the NHS breast cancer screening review, 2008: the paper cited statistics that were hard, sometimes impossible, to source; it was not peer reviewed; failed to mention crucial negative elements (overdiagnosis and false positives) of the 20-year programme; and sold the concept of screening in a paternalistic way.
When I spoke to the co-author Karsten Jorgensen, he went further: "Those responsible for the breast-screening programme are also those responsible for evaluating the programme … I would not say for financial gain, but maybe for professional gain. This is something you've worked on for a long time, if the programme fails you fail as a professional. It's often taken very personally."
At this level, the row has the makings of a bar-room brawl. Julietta Patnick, the director of NHS cancer screening, responded that the "annual review doesn't pretend to be a scientific publication" – that's why it isn't peer reviewed. Indeed, all the criticisms fundamentally misunderstand its nature, which is as a 20-year hurrah for the programme, not an independent study.
Professor Stephen Duffy conducted many of the randomised trials that make the NHS such a clear, rich source of screening data. He queries one of the controversial article's main assertions – that the age group that wasn't screened saw the same drop in mortality as the age group that was.
And its data points are misleading, according to Duffy, taking as its mortality highpoint 1988 (before screening started) and its low point 20 years later. It is more illuminating to take averages over periods of years: the first pre-screening; the second as screening begins; and the third as the programme matures. "It's a shame my paper isn't published till March," Duffy said, mildly. "But when we did that, taking the average, in those screen-affected ages, the mortality went down in the screened ages 28% more than all other ages combined." His paper – whose subject he summarises as, estimating the absolute numbers of breast cancer deaths prevented in the UK breast screening programme – appears in the March issue of the Journal of Medical Screening, having been peer reviewed: I'm not trying to wade in and arbitrate, like a medieval layman.
In the meantime, there is one element of this programme – indeed, of screening as a concept – on which everyone broadly agrees, though they differ on its prevalence and impact, and that's overdiagnosis. Jorgensen and his co-author Peter Gotzsche estimate that one in four diagnoses are overdiagnoses: cancers that, left untreated, would not develop to be life-threatening. Duffy says the figure is between one in 15 and one in 20, but concedes: "You diagnose a cancer, you can't know what would have happened if you hadn't diagnosed it."
Jorgensen expands on this, pointing out that you never meet overdiagnosed patients, only people delighted to have overcome cancer. Those who would be fine without treatment only show up in statistics. "It's a very severe harm to get an unnecessary diagnosis, to have to live the rest of your life fearing that this disease might come back."
How to digest this risk, as a patient, is perplexing for philosophical more than mathematical reasons. It is much easier, first of all, to conceive of oneself as an individual rather than part of a cohort, so the premise – that some will be obviously saved, while others will be invisibly harmed – nudges one towards screening. There is much more drama and clarity in being identified for a life-saving treatment, than in potentially having a cloud of uncertainty that you didn't need to have. And many of us, furthermore, would probably think we had the emotional resilience to cope with a threat, and would take that over being struck down without warning. But that's easy to say when you haven't been diagnosed with breast cancer.
In fairness to the Danish medics, the NHS attitude is rather paternalistic here. Duffy likens screening to vaccination in infants: "When you're a child, you don't want an injection to protect you from this or that disease. That bit of discomfort can sometimes be worthwhile." Patnick, meanwhile, comments: "I know it's not for everybody. We all make decisions that would run contrary to good advice all the time, that's part of the richness of human life." They're unwilling, in other words, to brook the unnecessary anxiety of overdiagnosis as a real concern: whereas, for some patients – perhaps for many – anxiety is a pressing inhibition and its avoidance something to consider very seriously. That said, who's to say the relief from anxiety after being screened and given an all-clear isn't a comparable boon?
The debate offers a fascinating insight into the grey areas of screening. We think of it as a no-brainer, but maybe there is more to it; maybe we undervalue the importance of our own emotions when it comes to accepting best practice, and yet have expectations of the NHS that are highly emotional.
It's surprising, too, to see this terrain so swarming with people whose intent is good, yet who are so much at odds with each other. Out of respect for their collected expertise, I'm going to have one breast screened, and the other left alone. It'll be my very own randomised control trial.
• This article was amended on 22 January 2009. The original said that Stephen Duffy's paper on breast cancer screen was awaiting peer review in due course. This has been corrected. The paper covers the period 1974-2005.
