Randomized clinical trials (RCTs) have shown that initiating breast cancer (BC) screening between ages 50 to 69 and continuing it for 10 years decreases BC mortality. None of the existing RCTs included women over 74 and all included few women aged 70-74, such that the 95% confidence interval (CI) ranged from preventing 32.1 deaths to causing 17.2 deaths. However, there is no RCT data on when a woman should stop receiving BC screening, yet an estimated 52% of women over 75 receive screening mammograms in the US. A future randomized trial to study when to stop screening may be infeasible and potentially unethical. We used insurance claims data to emulate a (hypothetical) target trial of the effect of continuing screening on BC mortality among Medicare beneficiaries aged 70-74 or 75-84 years.
We emulated a hypothetical trial using a population-based cohort study with 20% random sample of Medicare (1999-2008). We selected 1,058,013 Medicare beneficiaries aged 70 to 84 with a life expectancy of at least 10 years and no previous diagnosis of BC, and who received a BC screening. We compared the following two strategies: continuing annual screening vs. stopping screening mammograms. The main outcome was 8-year standardized risk of BC mortality by age group (70-74, 75-84). Estimates are standardized by baseline and time-varying variables.
In the 70-74 age group, the estimated 8-year BC-specific death risk difference (95% CI) of continuing vs. stopping screening was -1.0 deaths per 1000 women (-2.3 to 0.1); hazard ratio 0.78 (0.63 to 0.95). In the 75-84 age group, the corresponding risk difference was 0.07 deaths per 1000 women (-0.93 to 1.3); hazard ratio 1.00 (0.83 to 1.19). The estimated 8-year risk of BC was 5.5% under the continue screening strategy (5.3% in the 70-74 age group, 5.8% in the 75-84 age group), and 3.9% under the stop screening strategy (3.9% in the 70-74 age group, 3.9% in the 75-84 age group).
Among women who have received at least one screening mammogram, we estimated that, compared with stopping BC screening, continuing screening past the age of 75 years results in no material difference in cancer-specific mortality over the following 8-year period. Our results in the age range of 70-74 match those of existing randomized clinical trials.
Clinical trial identification
Legal entity responsible for the study
US National Institutes of Health.
All authors have declared no conflicts of interest.