Breast cancer screening: one life saved at a cost of three overdiagnosesBMJ 2012; 345 doi: http://dx.doi.org/10.1136/bmj.e7460 (Published 05 November 2012) Cite this as: BMJ 2012;345:e7460
An independent expert panel estimated the risks and benefits of screening for breast cancer in the UK, where women aged 50-70 years are invited for mammography every three years. It seems that for every 235 women who are invited for screening, or every 180 women who are screened, one life is saved, but three women are diagnosed and treated for a cancer that would not otherwise have become apparent in their lifetime (overdiagnosis)⇑.
Nearly a fifth of women who receive a diagnosis at screening are overdiagnosed, but it is impossible to know which ones. However, a woman who attends all invitations to mammography screening has only a 1% chance of being diagnosed with a cancer that would never have caused problems if she had not been screened.
These are best estimates from inadequate data, the panel emphasised. They reviewed the evidence available from the literature and heard testimonies from leading experts. Trial evidence was considered most valid, although only three of 10 available trials had been randomised properly. Together, the trials looked at nearly 700 000 women and were done between 1963 and 1991. Have improvements in treatment over the years made these findings irrelevant? The panel found no data to support this and thought that relative risk reductions achieved in the trials (20%, 95% CI 11% to 27%) should still hold today.
The best evidence on overdiagnosis came from three trials in which women randomised to the control group weren’t offered screening at the end of the study period. Overdiagnosis was assessed from the population perspective, as the proportion of all cancers ever diagnosed in women invited to screening who are overdiagnosed. It was also assessed from the perspective of a woman invited for screening, as the probability that a cancer diagnosed during the screening period represents overdiagnosis.
Many observational studies were also considered, but results varied greatly and contributed little to the panel’s final estimates.
Cite this as: BMJ 2012;345:e7460