What is the difference between a false alarm and cancer overdiagnosis?

The confusion between the two concepts is very often made. We must admit that the difference is not intuitive, and the concept of overdiagnosis is relatively recent.

So let’s go back to the definitions.

FALSE ALERT (or false positive)

It is the suspicion of cancer on a mammographic image, which will not be confirmed after additional examinations.
A false positive is, therefore, a non-cancer that we know after performing examinations other than the mammogram.
These additional examinations that need to be done to invalidate the suspicion can be heavy and sometimes even result in biopsies, the number of which has greatly increased since screening was introduced, this situation being made more likely by the double reading performed in organized screening (a second radiologist examines the images taken in the first radiology center).
Knowing about a false alarm is often very stressful, as the woman must sometimes wait several days or even weeks for confirmation of the absence of disease, especially for biopsy results, where depending on the region, the delay is between a week and a month.

For every 1000 women over 50 participating in screening for 20 years, there would be about 1000 false alarms in France, leading to 150 to 200 unnecessary biopsies.
(Revue Prescrire, February 2015/Tome 35 N°376)

However, other assessments exist, giving higher figures. For example, here is an Australian evaluation 1. It provides a result over 25 years, corresponding to the overall screening duration in a woman’s life if she undergoes screening from age 50 to 74.

Excerpt from the book “Mammo ou pas mammo?” ed.Thierry Souccar


Overdiagnosis is a real diagnosis of cancer. On a microscopic examination, there are cancerous cells, but the diagnosis is pointless for the woman because this cancer would never have impacted her life or health. It is not a misdiagnosis but a lesion that is unnecessary to detect.

How is it recognized? On an individual level, seeing a patient, we cannot know. For the attending physician, the pathologist, the radiologist, the surgeon, there is only one “diagnosis.” That is why everything detected for a particular woman will be treated.
But on a population level, which is true for all screenings, epidemiologists note that there are always more cancer diagnoses in screened populations than in those not screened 2. And that the greater the screening intensity in a population, the greater the number of diagnoses 3.

In the study cited in reference 2, the Oslo researchers compared two groups of women of the same age; one was screened every two years, and the other was not screened and examined after 6 years. They found 22% more cancer diagnoses in the screened group (without counting carcinoma in situ, see the question on this subject). Moreover, with an identical mortality rate in both groups…
What can we conclude from this? If all cancer cells and all cancerous lesions were meant to become a “cancer disease,” we would have the same rates of women with cancer in both screened and unscreened groups since both groups were composed of women with identical profiles. And there should have been a significant difference in mortality.
However, this is not what we observe. Therefore, when we do not seek them, some cancers do not progress or progress very slowly without impacting the person’s life, and the person’s immunity can control in such a way that there is no progression to cancerous disease. Having cancer is not synonymous with being ill with cancer.

Autopsy studies are also available. In autopsies of people who have died from causes other than cancer, we see that many people have cancerous lesions in their bodies that are unexpressed and dormant 4.

Yet what is the problem with overdiagnosis in breast cancer or other cancers? You might say that it is good to detect as many lesions as possible are detected…

The first problem is that over-diagnosis, i.e., a useless diagnosis made in a healthy person who is not complaining about anything, will inevitably be followed by over-treatment, which is useless by definition. The onset of a “disease” is declared as soon as a microscopic examination confirms the presence of cancerous cells.
Yet, cancer treatments are not harmless. They all carry risks, some of them fatal 5. The somewhat cynical argument that the woman will receive “lighter” treatment because a small lesion has been found is not acceptable. It is not a question of giving a woman a light treatment (the “lightness” of which, for those who have been through it, is relative) but of doing nothing if the woman did not need it.
The reduction in surgery remains to be proven. Moreover, wherever screening is carried out, mastectomies only increase, and in no case can we claim that women will have less surgery thanks to screening 6.

Our mini-video on overdiagnosis here: https://www.youtube.com/watch?v=Mr995i_Hetg&feature=emb_imp_woyt

The second problem is that if one accepts an overdiagnosis of a medical procedure with a potentially harmful effect, there must be a compensating beneficial effect for the population, for example, a drastic reduction in mortality, an impact on the overall mortality of the population, a drastic reduction also of the most serious and deadly cancers, and as we have seen above a very significant decrease in heavy surgical and radiotherapy procedures.
None of these expected beneficial effects of screening have been achieved… Serious cancers escape screening because they are often immediately disseminated as soon as they are detected, have intrinsically pejorative biological characteristics 7, and are rapid and voluminous at the time of diagnosis because they have a high growth velocity 8.

We invite you to refer here for a more detailed explanation.

The figures for overdiagnosis vary according to estimates and studies.
The best known and so far uncontested estimate is that of the independent Cochrane Collaboration, which you will find presented on our poster here.

This is the result of a meta-analysis, i.e., an analysis of the results of several studies (trials) carried out at a time when cohorts of unscreened women were available to compare them with cohorts of screened women (in the 1970s and 1980s).
In a population of 2000 women screened over 10 years, we can expect a prolonged life thanks to screening, but at the same time, 10 women are overdiagnosed.

Thus, many women must be screened over a long period to achieve a “life-saving” outcome, but at the same time, more cases of overdiagnosis are occurring. This means that women pay a heavy price for overtreatment, some of whom die. Some researchers also argue that the fatal effects of overtreatment accidents outweigh this meager benefit.9 According to Professor Baum, author of the referenced study, for every life saved, one life would be destroyed due to a fatal effect of overtreatment.

It should also be noted that this meta-analysis by independent Nordic authors is based on Canadian clinical trials, which were audited and confirmed to be free of bias.10

Thirdly, it is ethically unacceptable that, on the pretext that screening would save one woman in 2,000 over 10 years, another 200 to 400 women (false alarms) and another 10 women (overdiagnoses) should be harmed in return. This is ethically indefensible.
It doesn’t matter how large the overdiagnosis is. It exists. And the percentage “battles” you can read in the press from one source to another are of little interest when a woman is concerned because it will always be 100% for her.
At the very least, women should be fairly informed.

The benefit-risk balance of screening is, therefore not, when the adverse effects (overdiagnosis, false alarms, radiation) are added up in favor of a predominant benefit because the reduction in mortality is not significant and is not attributable to screening (a decrease in cancer mortality was observed as early as the 1990s, with the improvement in treatment, therefore before the introduction of screening, and this was also the case for other cancers that were not screened, according to the impact studies)

The term “overdiagnosis” has become a real public health issue with the multiplication of routine examinations performed on healthy populations. It has been officially included in the thesaurus of references in the biomedical field.

Illustration in video, animated presentation of these two concepts:

Related links

These concepts are included in our decision support tool, which you can consult.

See also, How does cancer develop?


[1] Jolyn Hersch. Aide à la décision pour le dépistage du cancer du sein pour les femmes à partir de 50 ansC’est votre choix. (Brochure australienne). Page 7. [En ligne : https://ses.library.usyd. edu.au/bitstream/handle/2123/16658/2017%20updated%20breast%20screening%20DA%20 (Hersch%20et%20al).pdf;jsessionid=F0396C69AD95F6431008EA16CB3B9195?sequence=1]. Consulté le 30 juin 2021.

[2] Zahl P-H, mæhlen J, Welch HG. The natural history of invasive breast cancers detected by screening mammography. Arch Intern med. 2008 Nov 24;168(21):2311–6.

[3] https://cancer-rose.fr/en/2020/12/29/analysis-of-harding-american-study-2015/
Harding C, Pompei F, Burmistrov D, Welch HG, Abebe R, Wilson R. Breast Cancer Screening, Incidence, and Mortality Across US Counties. JAMA Intern Med. 2015 Sep;175(9):1483–9.

[4] https://cancer-rose.fr/en/2020/12/29/frequency-of-incidental-breast-cancer-and-precancerous-lesions-in-autopsy-studies-a-systematic-review-and-meta-analysis/
Thomas, E.T., Del Mar, C., Glasziou, P. et al. Prevalence of incidental breast cancer and precursor lesions in autopsy studies: a systematic review and meta-analysis. BMC Cancer 17, 808 (2017). https://doi.org/10.1186/s12885-017-3808-1

[5] https://cancer-rose.fr/en/2021/03/02/heart-disease-is-the-biggest-killer-of-breast-cancer-survivors/

6] Our study about mastectomies in France https://cancer-rose.fr/en/2020/12/17/our-study-does-organized-screening-really-reduce-the-surgical-treatments-of-breast-cancers/

Our results are consistent with those found in other countries :

– In the United States, in a 2015 study of 16 million women, a 10% increase in screening activity was associated with a nearly 25% increase in lumpectomies and partial mastectomies (RR 1.24; CI 1.15-1.34), without a decrease in total mastectomies [8].

– In the United Kingdom, according to the 2013 Marmot report on breast cancer screening, the frequency of mastectomies is increased by about 20% in the screened population compared to the unscreened population [9].

– For all the randomized controlled trials worldwide that examined this issue in 2013, the Cochrane Collaboration estimates that the number of mastectomies is increased by 20% (RR 1.20; 95% CI 1.08-1.32) and the number of surgical procedures (mastectomies and lumpectomies) is increased by 30% (RR 1.31; 95% CI 1.22-1.42) [10].

[7] https://cancer-rose.fr/en/2020/12/17/are-small-breast-cancers-good-because-they-are-small-or-small-because-they-are-good/

8] https://www.youtube.com/watch?v=pbGZdyUCITc

[9] Baum M. Harms from breast cancer screening outweigh benefits if death caused by treatment is included. BMJ 2013 ; 346 : f385. doi : https://doi.org/10.1136/bmj.f385

[10] Bailar J. C, MacMahon B. Randomization in the Canadian National Breast Screening Study: a review for evidence of subversion. Canadian Medical Association Journal. Jan 15, 1997;156(2):193-199.

🛈 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.
Scroll to Top