Cancer Research UK on Google+ Cancer Research UK on Facebook Cancer Research UK on Twitter
 

Breast screening benefits and harms

Breast screening outcome statistics for mortality reduction and overdiagnosis for the UK are presented here. Breast screening literature varies in its definition of breast cancer using different combinations of ICD codes. Here we have included ductal carcinoma in situ (DCIS) tumours as well as invasive cancer and so the ICD codes for statistics on this page are mostly ICD-10 C50 and D05.1, which is different to the other breast statistics on this site.

 

About breast screening

Breast cancer (invasive) is the most common cancer in the UK (2010), with 49,564 cases in 2010 accounting for 31% of all female cases. There were a further 5,765 cases of in situ breast carcinoma (all in situ ICD-10 D05) in women in the UK in 2010, of which 83% were DCIS (ICD-10 D05.1). In 1986, the Forrest Report recommended the introduction of a national breast screening programme on the basis of trials undertaken in the United States and Sweden.1

Across the UK women aged 50 to 70 are invited for breast screening with mammography every three years by the NHS Breast Screening Programme (NHSBSP). Women over 70 are eligible for breast screening but are not automatically invited. In England a trial is taking place to look at the possible benefits of extending breast screening so that women aged 47 to 50 and 70 to 73 are also invited.2

The principle benefit of breast screening is earlier detection, which facilitates less aggressive treatment, improving prognosis and ultimately, it is hoped, saving lives. However, there are some well-documented harms related to screening including radiation exposure, pain and the physiological consequences of false positive, false negative and true positive results. Many observers consider overdiagnosis, the detection of cancers which would have never caused harm in the woman’s lifetime, to be by far the greatest harm.

Screening works by detecting cancer early and this means that a screened population is, at least initially, likely to have a higher incidence of breast cancer. An increase in breast cancer cases can be clearly seen in the incidence trends increasing in the early 1990s for the 50 to 64 age group and later for the 65 to 69 age group. As the breast cancer has been detected earlier via screening we then expect to see a lower incidence in later age groups. Some screen-detected cancers, however, may never progress to become symptomatic. These are known as overdiagnosed cases.

section reviewed 30/10/12
section updated 30/10/12

 

Benefit: mortality reduction

A number of randomised trials of women aged 50 and over have also suggested that mortality from breast cancer is reduced in those offered screening compared with unscreened controls (for those trials where there is a reduction, it is not always statistically significant) (Table 2.1).

Table 2.1: Relative Risk of Breast Cancer Death in Screening With Mammography Versus No Screening in Randomised Control Trials with 13 years follow up

Randomised Control Trial (Year) Relative Risk for Women All Ages (Confidence Interval) Relative Risk for Women Aged 50 and Over (Confidence Interval)
New York (1963) 0.83 (0.70-1.00) 0.78 (0.60-1.01)
Malmo (1976) 0.81 (0.61-1.07) 0.86 (0.64-1.16)
Kopparberg (1977) 0.58 (0.45-0.76) 0.55 (0.42-0.73)
Ostergotland (1978) 0.76 (0.61-0.95) 0.71 (0.56-0.91)
Canada I (1980) 0.94 (0.74-1.27) N/A
Canada II (1980) 1.02 (0.78-1.33) 1.02 (0.78-1.33)
Stockholm (1981) 0.73 (0.50-1.06) 0.64 (0.41-1.01)
Gothenburg (1982) 0.75 (0.58-0.97) 0.83 (0.60-1.15)
UK Age (1991) 0.83 (0.66-1.04) N/A

CR Xls Icon Download this table (33.5KB)

Adapted from the Cochrane Review Analysis 1.2 and 1.6.3

Controversy, however, continues over the effectiveness of breast screening programmes.3,4

For example, the randomised control trials have been criticised for the poor quality of randomisation, for differential exclusions between study and control groups, and some have suggested that the findings from the trials are no longer relevant to the current screening programme because drug treatment has improved so markedly.

Reduction in the relative risk

A number of interested parties have carried out meta-analyses of the randomised control trials to try to improve the estimate for the reduction in breast cancer mortality due to screening. Although there were differences in which trials were included, how the trials were run, and the follow up period used in the analyses, the results from the meta-analyses are reasonably consistent (Table 2.2).5

Table 2.2: Relative Risk of Breast Cancer Death in Screening With Mammography Versus No Screening in a Number of Meta-Analyses

Meta Analysis Analysis Details Relative Risk (95% Confidence Interval)
Independent Breast Screening review 13-year follow-up in trials reported in the Cochrane Review, Random-effects meta-analysis 0.80 (0.73-0.89)
Cochrane Review 13-year follow-up in ‘adequately’ and ‘sub-optimally’ randomised trials, Fixed-effect meta-analysis 0.81 (0.74–0.87)
Cochrane Review 13-year follow-up in ‘adequately’ and ‘sub-optimally’ randomised trials, in women 50 and over, Fixed-effect meta-analysis 0.77 (0.69–0.86)
Cochrane Review 13-year follow-up in ‘adequately’ randomised trials, Fixed-effect meta-analysis 0.90 (0.79–1.02)
US Task Force Trials selected from a systematic search, women aged 50–59 years, Random-effects meta-analysis 0.86 (0.75–0.99)
US Task Force Trials selected from a systematic search, women aged 60–69 years, Random-effects meta-analysis 0.68 (0.54–0.87)
Canadian Task Force Trials selected from a systematic search, Random-effects meta-analysis 0.79 (0.68–0.90)
Duffy et al Review of all trials and age groups 0.79 (0.73–0.86)

CR Xls Icon Download this table (38KB)

Adapted from the Independent Panel on Breast Screening.5

Absolute risk

There has been much greater disagreement about the absolute risk reduction due to screening in comparison to the relative risk, as shown by the varying estimates of the numbers of women who would need to be screened to prevent one death (Table 2.3).6

This stems in part from the variation in several factors, including the age at which regular screening starts, the period over which it continues, and the duration of follow-up after screening.7

Table 2.3: Estimated Numbers of Women Needed To Screen to Prevent One Breast Cancer Death

Study Analysis Details Number Needed to Screen / Invited
Independent Breast Screening review Based on an RR reduction of 20% for women aged 55–79 years in the UK. 235 women invited
Cochrane Review Absolute risk reduction based on the 13-year follow-up in the trials considered ‘adequately randomised’. 2000 women invited
US Task Force Based on 7 years of screening and 13 years of follow-up. 1339 women invited aged 50–59 years
377 invited aged 60–69 years
Canadian Task Force Women aged 50–69 years screened every 2–3 years for about 11 years. 720 women screened
Duffy et al Based on 22-year follow-up of women aged 50–69 years in the Swedish Two-County trial, assuming that the absolute risk reduction for the 7 years of
screening can be multiplied up to reflect 20 years in the UK screening programmes.
113 women screened
Beral et al Women aged 50–70 years regularly screened for 10 years, based on summary of published evidence. 400 women screened

CR Xls Icon Download this table (38KB)

Adapted from the Independent Panel on Breast Screening.5

section reviewed 30/10/12
section updated 30/10/12

 

Harm: overdiagnosis

There is an ongoing debate about how many women are overdiagnosed because they have been to breast screening. Indeed, there is no agreement even on how overdiagnosis should be presented, in terms of which population to report it from.

It has been shown there are at least eight different ways of presenting overdiagnosis estimates,8 and this variety has contributed to the confusion about the level of overdiagnosis.

Two methods of presentation, which make particular intuitive sense, are:

  • the proportion of cancers diagnosed over the lifetime because of breast screening – this represents the additional burden caused by breast screening and is perhaps the most useful figure from a public health perspective
  • the proportion of cancers diagnosed during the screening period – this represents the probability of a cancer diagnosed being an overdiagnosis and is perhaps most useful from an individual woman’s perspective.

Overdiagnosis can be calculated by comparing a screened population with an unscreened population. This can be achieved by evaluating:

  • two areas of the same country, one with and one without screening,
  • different age groups, some which have experienced screening or not,
  • the same area before and after screening.

However, the risk of breast cancer varies between populations and none of these methods can guarantee that the risk of breast cancer in the comparison population is the same as the screened population, and adjusting for differences in risk is difficult.9

Estimates of overdiagnosis for invasive breast cancer range from −4% to 7% of invasive breast cancer incidence in the unscreened population for women aged 40 to 49 years, 2% to 54% for women aged 50 to 59 years, and 7% to 21% for women aged 60 to 69 years.9

The variation in both the presentation and the methodology used to calculate the estimates of overdiagnosis arises partly because the randomised control trials were not set up to measure overdiagnosis as an outcome. Also, in many of the trials, the control group also received screening after a number of years.

Of the trials given in Table 2.1, only three didn’t screen the control group at the end of the intervention period. It is possible to use these trials in order to estimate overdiagnosis in a screened versus non-screened setting. Table 2.4 shows how the proportion of overdiagnosed cancers changes under different definitions. In each case the numerator – the additional cancer cases in the screened group over the control group after long term follow-up – remains the same and the population used for the denominator varies.

Table 2.4: The Proportion of Overdiagnosed Cancers in Three Randomised Control Trials Using Three Different Denominators

Definition Used Malmo I (55-69) Canada I Canada II
Excess cancers as a proportion of cancers diagnosed over whole follow up period in women invited for screening. 10.5% (82/780) 12.4% (82/663) 9.70%
(67/693)
Excess cancers as a proportion of cancers diagnosed during screening period in women invited for screening. 18.7% (82/438) 22.7% (82/361) 16.0% (67/420)
Excess cancers as a proportion of cancers detected at screening in women invited for screening. 29.1% (82/282) 29.4% (82/279) 19.8% (67/338)

CR Xls Icon Download this table (37KB)

Adapted from the Independent Panel on Breast Screening.5

Overdiagnosis as a proportion of cancers diagnosed over the lifetime because of breast screening varies between 10 to 12% and between 19 and 23% as a proportion of all cancers diagnosed during the screening period.

section reviewed 30/10/12
section updated 30/10/12

 

Balancing the benefits and harms

The wide variation in the estimates for both the benefits and harms have led to an inevitably wide range of estimates of the ratio of the number of women overdiagnosed to the number of lives saved. These range from 10 overdiagnosed cases for every death prevented to one overdiagnosed case for every two deaths prevented, reflecting a twenty-fold difference.3,10

In October 2011, Professor Sir Mike Richards, National Clinical Director for Cancer, and Dr Harpal Kumar, Chief Executive Officer of Cancer Research UK, asked Professor Sir Michael Marmot (Professor of Epidemiology and Public Health and Director of the Institute of Health Equity at University College, London) to convene and chair an Independent Panel to review the evidence on benefits and harms of breast screening in the context of the UK breast screening programmes.5,11

After reviewing the evidence, the Panel took the assumption that UK women aged 55 to 79 invited to breast screening had a 20% relative reduction in the risk of dying from breast cancer per year, compared with unscreened women. Because breast screening programmes in the UK had been established for more than twenty years they assumed that the benefit of breast screening has already accrued and so instead of reducing the current mortality rate by 20%, they calculated what the mortality rates would have been had there been no screening programme. They did this by increasing the current mortality rate by 25% (equivalent to a fall of 20% from the higher level). This provides an absolute reduced risk of 43 deaths prevented per 10,000 women invited to screening.

The panel were, however, clear that considerable uncertainty surrounds this estimate. The main types of uncertainty are:

  • statistical - the 95% confidence interval around the relative risk reduction of 20% is 11% to 27%,
  • bias - there are a number of potential sources of distortion in the trials which could have affected the estimates of the mortality benefits, and
  • relevance of the findings of these old trials to the current screening programmes.

To calculate the magnitude of overdiagnosis in women, UK screening programmes require information on women who have been in a screening programme from age 50-70, then followed for the rest of their lives, which is not available. Any estimate of this will, therefore, include uncertainty. Having reviewed the evidence, the Panel assumed an absolute risk of overdiagnosis of 19% (based on a meta-analysis of the three studies given in Table 2.5) and then applied this to the risk of being diagnosed with breast cancer per year in UK women aged 50 to 69 invited to breast screening. This provides an absolute risk of 129 women being overdiagnosed per 10,000 women invited to screen.

The Independent Panel, therefore, concluded that the Breast Screening Programmes in the UK extend lives but at a cost. The Review estimated that while breast screening prevents around 1,300 breast cancer deaths in the UK per year, but it can also lead to around 4,000 women each year aged 50 to 70 in the UK having treatment for a condition that would never have caused them harm. Table 2.5 shows the various ways which the figures around the deaths prevented and overdiagnosis can be presented.
 

Table 2.5: Comparison of the Benefits and Harms of Breast Screening, in numbers

Measure Screening Benefits Screening Harms
Relative Risks Women invited for breast screening have a 20% reduced chance of dying per year in UK women aged 55 to 79 from breast cancer compared with what it would be without a screening programme. 19% of breast cancers diagnosed in women aged 50 to 70 invited for screening would not have caused any problem if left undiagnosed and untreated.
Number of deaths prevented and women overdiagnosed per year in the UK 1,307 3,971
Number of deaths prevented and women overdiagnosed per 10,000 invited to screening 43 (0.43%) deaths prevented 129 (1.29%) women overdiagnosed
Number of deaths prevented and women overdiagnosed per 10,000 who actually attended screening 56 deaths prevented 168 women overdiagnosed
Ratio of overdiagnosis to deaths prevented 1 death prevented 3 women overdiagnosed

CR Xls Icon Download this table (38KB)

section reviewed 30/10/12
section updated 30/10/12

 

Results for women that attend screening

The figures above refer to women invited to screening as the analysis is based on intention to treat. Analysis based on intention to treat minimises the impact on study results if drop-out rates differ between the groups being compared. Intention to treat analysis is useful for assessing the effect of a policy, rather than the effect of a treatment. Because women receive no benefit from receiving an invite to screening alone, both the benefit and harms associated with screening only accrue to women who actually attend screening. The risk reduction and the rate of women overdiagnosed will therefore be higher in women attending screening than in women invited to screen.

The absolute reduction in risk of dying from breast cancer for women attending screening can be estimated as the absolute risk reduction in those invited to screening (0.43%), divided by the average coverage rate in the NHSBSP (77%): 0.43% / 0.77 = 0.56% or 56 breast cancer deaths prevented per 10,000 women attending screening.

A similar calculation can be applied to the percentage of women with an overdiagnosis, so 1.29% / 0.77 = 1.68% or 168 per 10,000 women attending screening.

The same calculation can not simply be repeated for the number of cancers diagnosed, because another factor – affluence of the population – interacts with the variables involved: the risk of breast cancer increases in more affluent populations, and more affluent women are more likely to attend breast screening.12,13 If on that basis, we assume the risk of breast cancer in women that attend breast screening is 10% higher than in women invited to screening then there will be 681 plus 10% (749) breast cancers diagnosed in 10,000 women attending screening.

Under this assumption then, after 20 years amongst 10,000 women who have attended screening:

  • 749 women will be diagnosed with breast cancer, and receive treatment.

Of these 749 women:

  • 157 will die from breast cancer, 56 fewer than in the group not attending screening.

Of the 592 who survive:

  • 56 will have their life saved by screening, see calculation above.
  • 168 will be diagnosed with a cancer and treated for a cancer that wouldn’t have caused problems in their lifetime (‘overdiagnosed’), see calculation above.
  • 369, the remainder, will be diagnosed with and treated for a cancer that would have been picked up without screening.

If the same 10,000 women were not able to attend screening, then after 20 years:

  • 582 will be diagnosed with breast cancer, and receive treatment, i.e. the 749 minus the number of overdiagnosed cases,
  • 168 women will have a breast cancer they never know about and that will not cause any harm during their lifetime.

Of the 582 women diagnosed:

  • 213 women will die from breast cancer, 56 more than in the group attending screening
  • 369, the remainder, will be treated for cancer and survive.

These figures have been illustrated in the infographic that explains the results of the breast screening independent review.

section reviewed 19/11/12
section updated 19/11/12

No Error

Rate this page:
Submit rating
Rate this page
Rate this page for no comments box
Please enter feedback to continue submitting
Send feedback
Question about cancer? Contact our information nurse team
 

References for breast screening benefits and harms

  1. Forrest, P Breast Cancer Screening: Report to the Health Minister of England, Wales, Scotland and Northern Ireland London, HMSO. 1986
  2. NHS Breast Cancer Screening Age Extension Full Randomised Control Trial (Accessed October 2012)
  3. Gøtzsche PC, Nielsen M Screening for breast cancer with mammography Cochrane Database System Reviewed 2011
  4. Autier P, Boniol M, Gavin A, Vatten LJ Breast cancer mortality in neighbouring European countries with different levels of screening but similar access to treatment: trend analysis of WHO mortality database British Medical Journal. 2011; 343(7818)
  5. Independent UK Panel on Breast Cancer Screening. The benefits and harms of breast cancer screening:an independent review. Lancet 2012; 380; 1778 - 1786.
  6. McPherson K Screening for breast cancer - balancing the debate BMJ 2010; 340:c3106
  7. Beral V, Alexander M, Duffy S, et al The number of women who would need to be screened regularly by mammography to prevent one death from breast cancer J Med Screen 2011; 18: 210–12
  8. de Gelder R, Heijnsdijk E.A, et al Interpreting overdiagnosis estimates in population-based mammography screening Epidemiol Reviews, 2011, 33(1):111-121
  9. Biesheuvel C, Barratt A, Howard K, et al Effects of study methods and biases on estimates of invasive breast cancer overdetection with mammography screening: a systematic review Lancet Oncol 2007;8:1129-38
  10. Duffy SW, Tabar L, Olsen AH, et al Absolute numbers of lives saved and overdiagnosis in breast cancer screening, from a randomized trial and from the Breast Screening Programme in England J Med Screen 2010; 17: 25–30
  11. Independent UK Panel on Breast Cancer Screening Independent Breast Screening Review. November 2012
  12. National Cancer Intelliegence Network Cancer incidence by deprivation England, 1995-2004. 2008
  13. Maheswaran R, Pearson T, Jordan H, Black D Socioeconomic deprivation, travel distance, location of service, and uptake of breast cancer screening in North Derbyshire, UK J Epidemiol Community Health. 2006: 60(3):208-12.
Updated: 30 October 2012