Are mammograms safe

The short version: mammography trades a small radiation risk and a large false-positive burden for a modest reduction in breast-cancer deaths, with no effect on overall mortality.

The evidence for mammography is built on a foundation that looks solid until you check what it actually measures. The randomized trials show a 15 to 25 percent relative reduction in breast-cancer mortality for women aged 50 to 69 [3, 8]. But none of those trials found any reduction in all-cause mortality [3]. With 660,000 women enrolled, even a 2 percent drop in total deaths would have been detectable. It was not there. That means the breast-cancer deaths avoided are being replaced by other causes of death in the screened group, or the trials are picking up a real but narrow effect that does not extend life overall.

The harms are concrete and large. Over 20 years of biennial screening starting at age 50, for every 1,000 women invited, about 200 will get a false-positive mammogram, 30 will have a biopsy for a false alarm, and 15 will be overdiagnosed with a cancer that would never have caused symptoms [3]. Overdiagnosis means you get treated for something that was never going to hurt you. That is not a side effect. It is a direct harm: surgery, radiation, chemotherapy, anxiety, all for a lesion that would have stayed quiet.

The radiation risk is small but real. A 2014 estimate put the lifetime risk of radiation-induced breast cancer from biennial screening at about 10 per 100,000 women, with about 1 death per 100,000 [16]. That is tiny compared to the false-positive burden, but it is not zero, and it is higher for women with BRCA mutations or those who start screening before age 30 [15]. The Dubai study from 2025 shows organ doses vary by breast thickness and age, with younger women and denser breasts receiving higher doses [14].

The benefit is also smaller than the headlines suggest. The Norwegian cohort study that found a 43 percent mortality reduction in screened women compared to non-screened women is an observational comparison [4]. Screened women are healthier, richer, and more health-conscious than non-screened women. The same study that validated the Danish model found that total mortality in screening regions was close to expected without screening, and the small excess they found was not specific to screening areas [13]. When you adjust for selection bias properly, the benefit shrinks. The 2018 German cohort study showed that even after extensive adjustment for comorbidity, the hazard ratio for all-cause death in screened versus non-screened women was 0.54, meaning screened women were half as likely to die from anything [1]. That is not a mammography effect. That is a healthy-volunteer effect.

My call: for a woman at average risk, the evidence does not show that mammography extends life. It shows a trade: a modest reduction in breast-cancer deaths exchanged for a high rate of false alarms, unnecessary biopsies, and overdiagnosis. The burden of proof is on the people recommending the test, and they have not shown a net benefit in the outcome that matters most. Confidence: moderate.