How Often Do Mammograms Find Cancer That Wouldn't Have Hurt You?
If you get screened between 50 and 74, roughly 10 to 20% of the cancers a mammogram finds are ones that would never have hurt you. That's worth knowing, but it isn't the '1 in 3' from older studies, and it isn't under 5% either. What moves the number most is your age. Through the core screening years, biennial mammography still comes out clearly in your favor.
The '1 in 3' figure came from a specific kind of study: count the extra cancers screening finds compared to unscreened populations, and call the surplus overdiagnosis. That math ignores lead time and self-selection. It counts a lot of cancers that would've surfaced later as if they were never going to. Fix those problems and the number falls.
Why the number ranges from 4% to 50%
Older ecological studies compared women in a screening program to women in earlier decades who weren't. In one influential 2009 review of organized programs, that math produced the '1 in 3' headline. A Dutch analysis of long-term screening came in closer to half. Neither method can tell whether an early cancer counted today would've been a later cancer counted tomorrow, so both inflate the excess.
Modern approaches fix that in different ways. A case-control study of England's national screening program found a raw excess of 9.5% among screen-detected cancers, which fell to 3.7% after adjusting for the fact that women who show up are already a healthier group. A Bayesian model of US women screened biennially from 50 to 74 landed at 15.4%. A 2023 re-analysis of the lower-bias randomized trials came in near 27 to 28% of screen-detected cancers, with wide uncertainty. The sub-5% figure is the floor when every correction gets applied at once. It isn't the consensus.
Age changes it more than anything else
The rate isn't spread evenly. In your 40s it's close to zero, because your life expectancy is long enough that even slow tumors will eventually surface clinically. Through your 50s and 60s it sits in the low double digits. After 70 it climbs steeply. The reason is arithmetic. Overdiagnosis has two sources: indolent lesions that would never have progressed, and progressive lesions found in someone who dies of something else first. The first source is small at every age. The second one balloons with competing mortality.
| Age group | Overdiagnosis rate | Deaths prevented per 10,000 over 10 years | What it means for you |
|---|---|---|---|
| 40-49 | Under 3% | About 3 | Real but small benefit; effect less certain in trials |
| 50-59 | About 11% | About 8 | Trade tips toward screening |
| 60-69 | About 15% | About 21 | Largest absolute benefit of any age band |
| 70-74 | About 31% | About 13 | Benefit narrows, still often favorable |
| 75 and older | 47% to 54% | Not established | Depends on your health and life expectancy, not age |
Two things stand out in that shape. Overdiagnosis climbs sharply past 70, and the number of breast cancer deaths a decade of screening prevents drops at the same time. The steepest turn is after 74, which is where the US Preventive Services Task Force stops making a general recommendation.
The benefit side of the ledger
The payoff isn't distributed evenly. Trials show a clear signal starting in the 50s and the biggest absolute gain in the 60s. In your 40s the relative benefit is smaller and less certain, but not zero. Both the Task Force and the American College of Obstetricians and Gynecologists updated their guidance in 2024 to recommend starting mammography at 40 rather than waiting for 50.
Where the trade actually turns
The one place the math stops favoring routine screening is after 74. Overdiagnosis climbs past 40%, mortality benefit falls, and other causes of death start doing more of the work. The Task Force concluded that current evidence isn't enough to recommend for or against screening past 74, and the American Cancer Society draws the line at 10 years of life expectancy rather than a hard age. Both are reasonable ways to say the same thing. Past a certain point, whether to keep going depends on your health, not the calendar.
The '1 in 3' figure is outdated methodology, and sub-5% is the floor when every possible bias is corrected out at once. If you're in the core screening years, the true rate is in the low double digits, and biennial mammography from 40 to 74 remains one of the few cancer screens where the arithmetic still clearly favors doing it. What would change that answer is a validated way to tell indolent from progressive lesions at biopsy, so the small share of true overdiagnosis, especially the slow, non-invasive ones, could be watched instead of cut.