EvidenceHealthScreeningCancer
Mammography Screening for Breast Cancer: Mortality and Overdiagnosis
- Study type:
- Cochrane-style systematic review of randomised controlled trials
- Topic:
- Randomised mammography screening trials re-analysis
- Added:
- 2026-02-25
- Status:
- published
Strength of evidence
🟡 Moderate – large randomised trials with hard endpoints, but heterogeneous quality; several influential trials show serious risks of bias.
Source
Gøtzsche PC, Jørgensen KJ. Screening for breast cancer with mammography. Cochrane Database of Systematic Reviews. Cochrane-style review re-evaluating randomized mammography screening trials.
Open original studyRandomised mammography screening trials in average‑risk women have been re‑analysed in a Cochrane‑style review. The data show: - at best a small reduction in breast cancer‑specific mortality, - no clear reduction in all‑cause mortality, - and substantial overdiagnosis and overtreatment (extra cancers found and treated that would never have caused symptoms). Several major trials suffer from serious methodological problems (flawed randomisation, exclusions, contamination of control groups), and modern breast‑cancer treatments have improved independently of screening.
Key findings
- When all trials are combined, invitation to screening is associated with a relative reduction in breast cancer‑specific mortality on the order of roughly 15–20 % in some analyses.
- The effect on all‑cause mortality is small and not clearly statistically significant.
- When analysis is restricted to trials judged as methodologically stronger, the apparent benefit on breast cancer deaths becomes smaller and may disappear.
- Overdiagnosis emerges as a major concern – screened groups consistently show a higher cumulative incidence of breast cancer that does not disappear with extended follow‑up.
- Some analyses suggest that 20–50 % of screen‑detected cancers may be overdiagnosed.
- Harms include false‑positive results leading to repeated imaging and invasive biopsies, psychological distress, and overtreatment.
- Improvements in breast‑cancer treatment over the decades likely reduce the incremental benefit of early detection compared with the original trials.
Background
- Mammography screening has been promoted for decades as a way to detect breast cancer earlier and thereby reduce mortality. Beginning in the 1960s–1980s, several large randomised trials in Europe and North America compared women invited to regular mammography with women not invited. Later, a Cochrane‑style review re‑evaluated these trials and raised serious concerns about their methodology and the true size of any benefit, particularly in the context of modern treatment options and the problem of overdiagnosis.
Methods
- The core evidence consists of eight major randomised trials, often cluster‑randomised by geographic region, in women mostly aged 40–74 years. Intervention groups received invitations to periodic mammography (typically every 1–3 years), while control groups received no systematic screening invitations (but could still receive diagnostic mammograms if clinically indicated). Outcomes included breast cancer incidence, disease stage, breast cancer‑specific mortality and all‑cause mortality, with follow‑up periods of around 10–25+ years. The Cochrane review by Gøtzsche and Jørgensen assessed trial quality, distinguishing adequately randomised trials from those at high risk of bias (e.g. exclusions of early deaths, uneven randomisation, contamination of the control group). It pooled results separately and combined, examined absolute as well as relative risk differences, and estimated overdiagnosis (persistent excess incidence in screened groups after long follow‑up).
Results
- When all trials are combined, invitation to screening is associated with a relative reduction in breast cancer‑specific mortality on the order of roughly 15–20 % in some analyses. However, the effect on all‑cause mortality is small and not clearly statistically significant. When analysis is restricted to trials judged as methodologically stronger, the apparent benefit on breast cancer deaths becomes smaller and may disappear. Overdiagnosis emerges as a major concern. Screened groups consistently show a higher cumulative incidence of breast cancer that does not disappear with extended follow‑up, indicating that many screen‑detected cancers represent lesions that would not have surfaced clinically during the woman's lifetime. Depending on methods and assumptions, estimates of overdiagnosis vary widely, but some analyses suggest that 20–50 % of screen‑detected cancers may be overdiagnosed. Harms of screening include false‑positive results leading to repeated imaging and invasive biopsies, psychological distress, and overtreatment (surgery, radiotherapy, endocrine therapy) for cancers that might never have caused symptoms. Improvements in breast‑cancer treatment over the decades (more effective surgery, radiotherapy and systemic therapies) likely reduce the incremental benefit of early detection achieved through screening compared with the situation in the original trials.
Interpretation
- From conventional public‑health messaging, mammography screening is presented as a proven way to "save lives". From the WorthDay perspective, the actual evidence is more nuanced and considerably weaker. Randomised trials show, at best, a small reduction in breast cancer‑specific mortality, no clear effect on all‑cause mortality, and substantial overdiagnosis and overtreatment. Methodological flaws in several key trials, plus changes in treatment standards and background use of opportunistic mammography, further limit confidence in pooled estimates. The overall balance is that many women will undergo false alarms, biopsies and potentially unnecessary treatments so that a small number may avoid a breast‑cancer death – and this does not translate into a detectable improvement in total survival. For WorthDay, S0040 therefore represents moderate‑quality evidence that population mammography screening is far from the life‑saving intervention it is often advertised to be. Any decision to participate should be based on a clear understanding of absolute benefits and harms, not slogans.
Limitations
- Trial quality is heterogeneous; several influential trials show serious risks of bias (flawed randomisation, exclusions, contamination), which can exaggerate benefits.
- Overdiagnosis estimates rely on modelling and different definitions; published numbers range widely and are not directly comparable.
- Treatment and diagnostic practices have evolved since the original trials, so current benefit–harm profiles may differ from historical data.
- Most data come from high‑income countries; applicability to other settings with different baseline risks and health systems is uncertain.
- Trials rarely report comprehensive quality‑of‑life outcomes, psychological impact or long‑term consequences of overtreatment.
- Women's individual preferences and value judgements regarding small potential mortality reductions versus the risk of overdiagnosis are not captured in aggregate statistics.
WorthDay perspective
- S0040 shows that the common claim "screening saves lives" is, for mammography, heavily overstated. The best randomised evidence suggests only a small possible reduction in breast‑cancer deaths, no clear effect on overall mortality and a substantial burden of overdiagnosis and overtreatment.
- For WorthDay, this is moderate‑quality evidence against blanket encouragement of population screening. Users should be given absolute numbers (how many deaths prevented and how many overdiagnosed per 1 000 women screened) and supported in making informed, value‑based decisions rather than being pressured by fear‑based campaigns.
- In the Health category, mammography screening is a prime example of why WorthDay insists on clear distinction between disease‑specific and overall outcomes, and on considering harm as well as benefit. Elements related to screening should emphasise preparing questions for doctors, understanding absolute risks and considering personal values, not simply "get every available test".
Permalink: https://www.worth-day.com/science-compass/s0040-mammography-screening-for-breast-cancer-mortality-and
Related studies
- Low-dose CT lung cancer screening in heavy smokersS0045
- PSA-Based Prostate Cancer Screening: Small Disease-Specific Benefit, No Clear Survival GainS0041
- General Health Checks in Adults: No Clear Benefit on Morbidity or MortalityS0042
- Adjuvant Chemotherapy After Surgery for Colon Cancer (Stage III)S0037
Science meets practice