SINCE federal government support of the national mammography program (Breastscreen) began in 1991, mortality from breast cancer in Australia has declined by 28%, but how much of this improvement can be attributed to screening and how much to improved adjuvant treatments?
Disentangling the contributions of screening and improved treatment to reduced mortality is challenging. However, the incidence of breast cancers (BCs) that are advanced at the time of diagnosis is one measure that could be useful. For screening to have an independent impact on mortality it must be able to reduce the incidence of advanced tumours, but recent evidence does not indicate that this is the case.
A recent study, which included 14 countries and regions with at least 7 years of mammography screening, has failed to find a reduction in the incidence of advanced cancer in countries with sustained use of screening.
It is now also clear that the improvements in BC mortality are not restricted to populations with screening. A paper in the BMJ, reported similar mortality reductions for breast cancer in countries with and without screening programs for the same period: Northern Ireland (screening) versus Ireland; the Netherlands (screening) versus Belgium; and Sweden (screening) versus Norway.
This follows reports in 2009 and 2010 of analyses of trends in breast cancer mortality that concluded that improvements in BC mortality could not be attributed to screening in Denmark, Norway and the UK.
In Australia, the 5-year survival rate improved from 71% to 83% between 1986 and 1995, before Breastscreen could have had an impact, and improvements in adjuvant chemotherapy and endocrine therapy received by Australian women during this period could explain most, if not all, of the observed decline in BC mortality.
The introduction of screening has been associated with a sustained increase in the incidence of invasive BC, most evident in Australia for women aged 60-69 years who have had the highest Breastscreen participation (almost 60%), but lowest mortality decline (19%). This raises the question of whether screening is responsible for overdiagnosis of BC.
Overdiagnosis refers to BC that would not have manifested clinically in a woman’s lifetime and therefore would not have been diagnosed in the absence of screening. In Australia, overdiagnosis from screening has been estimated at 30%-40%. Overdiagnosis would be even greater if ductal carcinoma in situ detected by screening was also taken into account.
The issue then becomes: what is the balance between lives saved and harm from overdiagnosis and unnecessary treatment? Estimates derived from randomised screening trials have produced a range of ratios between harms and benefits, with the possibility that the balance could be as unfavourable as 10 women overdiagnosed for every life saved.
Why are we saying this now when the message in the past has been to encourage women to be screened?
The original trials demonstrating the benefits of screening were carried out between 1960s and 1980s, when 5-year survival rates for women with BC was much poorer than it is today and adjuvant therapy was received by only a small minority of women. The remaining potential benefit of earlier diagnosis is now much less and, considering the issue of overdiagnosis, a simple message of unqualified benefit from screening is no longer appropriate.
So what should women be advised about screening mammography?
The harm of screening is not currently being explained in the UK screening program, leading to a call for a better balance in the information provided to women.
We consider that Australian women should also be given more information about the balance of benefits and harms of screening so that they are better equipped to make an informed choice.
Associate Professor Robin Bell is the deputy director of the Women’s Health Program, School of Public Health and Preventive Medicine, Monash University. Professor Robert Burton is from the School of Public Health and Preventive Medicine, Monash University.
Full reference details are available on request from firstname.lastname@example.org
Posted 26 September 2011