Views & Reviews Personal View

Harms from breast cancer screening outweigh benefits if death caused by treatment is included

BMJ 2013; 346 doi: (Published 23 January 2013) Cite this as: BMJ 2013;346:f385

Re: Harms from breast cancer screening outweigh benefits if death caused by treatment is included

Sidelining overdiagnosis

The UK independent review panel refutes professor Baum’s calculations on deaths from treatments, remaining on one side of the binary opposition between scientific positivists where no real communication or meaning is taking place.1 This is a form of empty speech. However, even within this opposition, the panel isn’t making sense. Its non-sense points to an ideology that fantasises about the power to control cancer, an ideology that demands people to freely believe that too much information is bad for them because it might prevent them adhering to public policy. The policy is being aggressively marketed, not in terms of ‘need’ the distributive factor that might be best thought to apply to the market sphere of health care, but in terms of ‘for healthy people, the potential to prevent a death from cancer – a life to be saved’. A turn to fantasy may well be at work here, fantasy that provides the subject with an imaginary reality that prevents us from accepting our being-unto-death, an ensures the resilience of the product, breast cancer screening, in the market.2 Professor Baum’s response may be a polite call for radical reform but remains locked into a binary opposition, calling for a ‘truly informed choice’ that ignores the ideology at work and that will not ‘unwind’ overdiagnosis?

Marmot et al state3:
‘The Panel agrees that overdiagnosis is important but it is impossible to say, for an individual woman with cancer detected at screening, if her cancer represents overdiagnosis. It is therefore difficult to estimate the magnitude of overdiagnosis.’

Even from a purely positivist perspective this doesn't make sense. I thought the magnitude of overdiagnosis was calculated by comparing the numbers of cancers presenting clinically in a non screened group with a screened group. If there was no overdiagnosis then the numbers of cancers presenting clinically in the screened group would be reduced by the number of screened diagnoses, all other things being equal. With overdiagnosis the number of cancers presenting clinically in the screened group would only be reduced by a proportion of those presenting clinically in the non screened group. So, what would this calculation have to do with knowing whether or not a particular woman's screens diagnosed cancer represents overdiagnosis or not, in a binary sense? Although it could be argued the estimate of the amount of overdiagnosis would provide some information about the likelihood or probability of any one particular screening diagnosed cancer being an overdiagnosis, which might influence decision making, this seems to be contradicted in Hersch's paper: they say that, in relation to screening diagnoses4:
' ......once cancer is detected, evidence based treatment is virtually always indicated because potentially threatening cases cannot be differentiated from those representing overdiagnosis.'

This seems to preclude the possibility of a woman making any decision at that stage - she will just have to have the indicated treatment. But doesn't this contradict the whole notion of providing woman with information about overdiagnosis prior to screening because this will enable them to make a choice. This is one of the points Miriam Pryke makes in her rapid response in this series.

So, why is this important? What might this be signifying?

By conflating the difficulty of estimating the magnitude of overdiagnosis and the impossibility of knowing whether a particular cancer would ever present clinically, the panel, and Hersch et al, may be suggesting that the level of overdiagnosis becomes somehow irrelevant once a woman has been screened. In turn this may suggest that overdiagnosis doesn't really need to be played into the policy decisions about whether women should continue to be invited to screening or should continue to be offered screening. So, why might this be happening?

This would be consistent with other aspects of their letter being part of a Master Discourse delivering an imperative, a ‘judgement’ from on high. Language and context are important. The panel is eminent, and apparently highly respected, it is British, they make a 'judgement' that is portrayed as superior to the Nordic Cochrane review. The concept of ‘clarity’ is invoked as the rationale for the judgment, a concept used several times by pro-screening commentators following the review.5 But ‘clarity’ as a signifier skates over the underlying concepts of reassurance and accuracy which may, in this case, be mutually exclusive, as the Citizen’s Jury implied. If there is much uncertainty then why invoke clarity which implies certainty? Kirwan in the editorial even suggests that the programme should continue in the light of the uncertainty!6 This seems to suggest, incoherently, that screening programmes should be started even when evidence of net benefit is uncertain, as in this case.

The medical profession must be wary of allowing itself to be the arbiter of what information it provides to the public. Withholding information on harms to ensure attendance, as disturbingly suggested in Bekker's editorial in 2010, would seem to give a green light for any politically motivated programme to be marketed at a vulnerable population via a 'consider an offer' approach.7;8 Even whilst ideologically people feel 'cared for' and reassured and free to in effect work for the system, and maintain and reproduce the means of production. This might explain the resilience of the programme

The Marmot panel disagrees with Cochrane. Some experts are strongly in favour of screening. Why would anybody be strongly in favour of screening? Is this evidence of a fantasy of divine power on the part of the policy makers, coupled with fantasies of 'lives being saved and futures guaranteed ' on the part of an anxious Cartesian subject invoked as always uncertain, always needing faith.

Reference List

(1) Baum M. Harms from breast cancer screening outweigh benefits if death caused by treatment is included. BMJ [ 2013 Available from: URL:BMJ2013;345:f385
(2) Glynos J. The place of fantasy in a critical political economy: the case of market boundaries. Cardozo Law Review 2013; 33(6):2373-2411.
(3) Marmot M. re: harms from breast cancer treatment outweigh benefits if deaths caused by treatment is included. BMJ [ 2013 Available from: URL:
(4) Hersch J, Hansen J, Barratt A. Womens' views on overdiagnosis in breast cancer screening: a qualitative study. BMJ [ 2013 Available from: URL:BMJ 2013;346:f158
(5) Times. Keep having scans despite the downsides, women told. The Times 2012 Oct 30.
(6) Kirwan C. Breast cancer screening: what does the future hold? BMJ [ 2013 Available from: URL:BMJ2013;346:f87
(7) Bekker H. Decision aids and uptake of screening. BMJ [ 2010 Available from: URL:BMJ2010;341:c5407
(8) Entwistle VA, Salkeld G, McCaffery K, Irwig L, Flitcroft K, Trevena L et al. Communicating about screening. BMJ [ 2008 337:[789-791] Available from: URL:BMJ 2008;337:a1591

Competing interests: GP and PhD student studying overdiagnosis, EBM - positivisms ideology and the subject - a Lacanian discourse analysis

07 February 2013
Owen P Dempsey
Manchester Metropolitan University