19th September 2024
Effect of screening mammography on the risk of breast cancer deaths and of all-cause deaths: a systematic review with meta-analysis of cohort studies
Auteurs : Philippe Autiera, Karsten Juhl Jørgensenb ,Michel Smansa ,Henrik Støvringc
a International Prevention Research Institute (iPRI), Lyon, France
b Cochrane Denmark and Centre for Evidence Based Medicine, Odense, Denmark
c Steno Diabetes Center, Aarhus, Denmark
https://www.jclinepi.com/article/S0895-4356(24)00181-1/fulltext#fig1
A study that challenges the assertion that screening reduces breast cancer mortality
This publication, which appeared this summer in the Journal of Clinical Epidemiology, is likely to raise quite a few questions.
The authors argue that women who undergo breast cancer screening are different from those who do not, and that this finding undermines the validity of certain studies validating the effectiveness of mammographic screening in reducing breast cancer mortality.
The differences in the results of studies examining and seeking to quantify the reduction in the risk of death thanks to mammographic screening are linked to different behaviour patterns among women who undergo screening. These differences in results have nothing to do with the value of screening.
Without breast cancer screening, the results on the risk of death from breast cancer would be equivalent. Let’s look at this in detail.
For almost three decades, breast cancer screening has been less and less effective, and above all the benefit-risk balance is far from positive. Two major studies published in 2023 suggest that the risk of death from breast cancer is falling (which is excellent news) for both screened and unscreened women. The other demonstrates that no screening (apart from colon cancer screening by sigmoidoscopy) can extend life expectancy.
However, the French health authorities and the INCa (French national Institute of cancer)repeatedly claim that screening can reduce mortality by around 15% (previously 20%, but this figure has now been revised downwards), based on various studies of varying quality and reliability.
We have already explained what this percentage means in real life, which is why women must demand that the data be presented not as flattering percentages but as raw values.
So let’s assume a 20% reduction in mortality: this means that out of 2,000 women not screened over 10 years, you will see 5 deaths from breast cancer, and out of 2,000 women screened over 10 years, you will see 4 deaths from breast cancer, so the relative reduction in the risk of dying from breast cancer is 1 in 5, or 20%. But in absolute terms, that’s just one less woman dying.
It is clear that a very large number of women (2,000) need to be monitored over a very long period (10 years) to see just one fewer death in the group screened, which is a very small benefit, and one that is still uncertain…
Read here: https://cancer-rose.fr/en/2024/08/23/why-do-we-value-screening-so-much/
What does this study show?
The authors of this review of cohort studies show that women who take part in screening at baseline already have a much lower risk of death from all causes than other women. ‘Subjects who do not participate in screening are generally less aware of their health, are more economically deprived, have more co-morbidities or disabilities and have a shorter life expectancy.’
In other words, the mortality results would be the same without any screening, and the wealthiest women would still have a better life expectancy than the others, screening or no screening.
The authors argue that the methodology used in certain observational studies aimed at proving a reduction in the risk of dying from breast cancer thanks to screening are biased by the very sample of women chosen for these studies, and that they present a false and misleading methodology.
Sociological studies and demographic surveys have already shown that socio-economic level is the primary factor influencing life expectancy[1].
The study in simplified language
We quote the authors:
“The ability of screening mammography programmes to reduce the risk of dying from breast cancer (BC) is often evaluated using studies that compare the risk of BC deaths among women who attend screening with the risk among women who do not attend screening. Such studies find that this risk is lower for women who attend screening. However, women who attend screening are generally more resourceful and behave in ways linked to better health outcomes. Also, screening mammography finds occult cancers that would not have progressed into potentially lethal disease (overdiagnosis). These two factors mean that we would expect such studies to find lower risk in women who attend screening, even if screening has no benefit. Unfortunately, how much these two factors affect results is not known with sufficient certainty to reliably correct for them.
Because screening mammography cannot lower the risk of death from causes other than BC, we assessed if the risk of death from all causes combined was also lower for women who attend screening.
We found that women who attended screening had a much lower risk of death from all causes and that this lower risk was similar to their lower risk of death from BC. This means that results from studies that compare risk of breast cancer deaths between screening attendees and nonattendees are entirely due to shortcomings in their design and should therefore not be used to evaluate the benefit of screening mammography programmes.”
The healthy user or self-selection bias
Healthy user bias (HUB[2] [3]) is already well known.
The level of compliance with screening, explain the authors of the study, should be considered as a marker of risk, and not as a causal factor for advanced breast cancer and a higher risk of dying from breast cancer, or even from a cause other than breast cancer.
In fact, women whose compliance with screening is low correspond to a group of women who are generally less compliant and less concerned about their health, less attentive to their body and its changes; this group has more advanced-stage cancers and a lower overall survival rate than the regular screening group.
The difference in the risk of death from breast cancer is therefore explained by a difference in behaviour, and without screening the situation would be identical.
Apart from the fact that women are largely uninformed about these new data, breast cancer screening is proving to be ineffective in reducing the risk of death, as it has been promoted for decades, and is associated with a large number of undesirable effects that are now well known (false alarms, over-diagnosis, irradiation, over-treatment).
Where is the problem?
The main problem is, as the authors explain, that the benefits of breast cancer screening on mortality are completely biased in observational studies:
“The evaluation of the effect of cancer screening programmes on the risk of cancer-specific death is often assessed through observational studies (ie, case control and cohort studies) despite the longstanding knowledge that these studies are vulnerable to many biases, mainly the self-selection bias (or “healthy user bias”, or “healthy screening bias”) [1,2]. Self-selection originates from differences in personal characteristics and lifestyle between subjects choosing to attend and not attend screening. Subjects not attending screening are generally less aware of health, are more deprived, have more comorbidities or disabilities, and have a shorter life expectancy, all characteristics associated with increased risk of cancer mortality, irrespective of the effect of screening on this risk [4–5].”
The observational studies mentioned by the authors involve making observations in a population and collecting information, in this case health data, on groups of people.
These studies do not require any particular action to be taken; they simply involve observing subjects’ responses to the intervention being studied (in this case, screening) as a function of certain parameters (in this case, mortality).
But not all scientific studies are created equal, and there is a ranking.
Cohort studies, which are either observational or interventional, retrospective or prospective, do not have the same standard of evidence as a randomised clinical trial, for example. (Read here)
Researchers conducting observational studies on the effect of screening on mortality have made various attempts to limit this self-selection bias. However, these attempts have varied from one researcher to another, and have not been backed up by precise studies.
These attempts at rectification have never been unanimous or satisfactory, and in the end the observational studies in no way prove that screening is effective in reducing mortality; rather, the opposite is true. They strongly suggest that screening does not have the desired effect on mortality;
The reductions in the risk of death from breast cancer reported by cohort studies of the effectiveness of mammography screening are entirely due to self-selection bias…’ The corrective methods used to date by observational studies “cannot fully correct for the true effect of self-selection on mortality risk estimates”.
The corrective methods used to date by observational studies ‘cannot fully correct for the true effect of self-selection on mortality risk estimates.’
The benefits of this study
To quote the authors:
‘-This is the first study to show that observational studies of the effectiveness of cancer screening are subject to biases that cannot be controlled.
-Participation in mammography screening is an indicator of characteristics associated with a lower risk of death from any cause, including breast cancer, which observational studies have falsely interpreted as an effect of screening.’
In conclusion, ‘given that screening mammography has no effect on causes of death other than breast cancer, and that screening mammography is supposed to reduce the risk of death from breast cancer, cohort studies comparing women who participate in screening with those who do not participate after being invited should find significantly greater relative reductions in the risk of death from breast cancer than from all causes combined.
This has not been the case in cohort studies of the effectiveness of screening mammography.’
The changes requested by the authors, practical application
The fundamental question is: ‘What are the implications and what should change now?
This question is fundamental because the whole promotion of screening has been based on these studies, which falsely demonstrate, due to methodological flaws, that breast cancer screening is beneficial when it is not.
The answer is therefore: “observational studies based on individual screening history, no matter how well-designed and conducted, should not be regarded as providing evidence for an effect of screening”
This has led to a demand for practical application. Indeed, in 2002, an international group of experts within the IARC, the International Agency for Research on Cancer, reviewed the evidence on breast cancer screening, following the publication of the Nordic Cochrane Centre’s systematic review in The Lancet[6] [7].
But in 2015, a new group of international experts took stock of the knowledge accumulated after 2002[8]; however Autier et all criticise here a lack of transparency about the process of on-boarding working group members and the disproportionate inclusion of pro-screening experts.
The 2002 and 2015 groups both issued recommendations for the evaluation of breast cancer screening activities, but the notable difference is that in 2015 a large number of studies, of varying value, had been published in the meantime.
The 2015 group decided to adopt observational studies as the preferred study design for evaluating the effectiveness of screening in reducing mortality.
In view of this new study in the Journal of Clinical Epidemiology, showing the biases in observational studies comparing people who have taken part in screening with those who have not, it is requested that the recommendations issued by the IARC group in 2002 be reinstated, i.e. that not all observational studies be used.
What about women?
For women, it remains to draw all the consequences of the confirmed failure of screening mammography.
The observable consequences are essentially iatrogenic: over-diagnosis, with the corollary of over-treatment, which we can no longer tolerate and which is unethical, by inflicting non-harmless treatments on people who are a priori in good health, who complain of nothing, and thus seeing their lives destroyed, unnecessary amputations of organs, unnecessary ionising irradiation of a particularly radio-sensitive organ, an increase in the general level of anxiety for the individual and the population.
The adverse effects of over-detection also include an excessive and unfounded increase in the number of women said to be ‘at risk’, a trans-generational risk since it is needlessly transmitted to the offspring of women who are over-detected.
We should also mention the misinformation of women who have been mislead, avoidable medico-legal problems, useless medical activity with embolisation of radiology units and misdirection of care activity. Over-treatment and excess care for a population that is not in need is leading to under-treatment for another section of the population that needs it.
It is now imperative to provide women with information, and to refocus on fundamental research to understand the natural history of cancer, which is more complex than we imagined.
We need to address the socio-economic disparities which mean that women from poorer backgrounds have more problems accessing care, and accumulate more risk factors for cancer in general. And it’s not screening that will solve the problem; it has no positive role to play in the matter.
We need to accept the reality of the failure of screening and allocate resources to research and public information.
Persevering and encouraging women to take part without information is unethical and a waste of time, money and energy.
[1] https://www.insee.fr/fr/statistiques/3319895
[2] International Agency for Research on Cancer. Breast Cancer Screening. Lyon: IARC Press; 2002.
[3] Shrank WH, Patrick AR, Brookhart MA. Healthy user and related biases in observational studies of preventive interventions: a primer for physicians. J Gen Intern Med 2011;26:546e50.
[4] Mottram R, Knerr WL, Gallacher D, Fraser H, Al-Khudairy L, Ayorinde A, et al. Factors associated with attendance at screening for breast cancer: a systematic review and meta-analysis. BMJ Open 2021;11(11):e046660.
[5] McWilliams L, Groves S, Howell SJ, French DP. The impact of morbidity and disability on attendance at organized breast cancer- screening programs: a systematic review and meta-analysis. Cancer Epidemiol Biomarkers Prev 2022;31(7):1275e83.[6] Gøtzsche PC, Olsen O. Is screening for breast cancer with mammography justifiable? Lancet (London, England) 2000;355(9198):129-34. doi: 10.1016/s0140-6736(99)06065-1 [published Online First: 2000/02/16]
[7] Vainio H, Bianchini F, editors. IARC Handbooks of cancer prevention. Volume 7. Breast cancer screening. Lyon: IARC Press, 2002.
[8] Lauby-Secretan B, Loomis D, Straif K. Breast-Cancer Screening–Viewpoint of the IARC Working Group. The New England journal of medicine 2015;373(15):1479. doi: 10.1056/NEJMc1508733 [published Online First: 2015/10/09]
🛈 Nous sommes un collectif de professionnels de la santé, rassemblés en association. Nous agissons et fonctionnons sans publicité, sans conflit d’intérêt, sans subvention. Merci de soutenir notre action sur HelloAsso.
🛈 We are an French non-profit organization of health care professionals. We act our activity without advertising, conflict of interest, subsidies. Thank you to support our activity on HelloAsso.
