Nine RCTs (371,618 women) were included in the review. Number of participants ranged from 14,613 to 160,840 women. Mean follow-up ranged from 10.7 to 17.5 years. Quality scores ranged from 0.595 to 0.919. The two trials with scores of 0.595 were considered low quality and were excluded from the primary analysis.
The pooled relative risk from the seven higher quality RCTs was 0.83 (95% CI 0.72 to 0.97) which indicated that 17% (95% CI 3 to 28%) of breast cancer deaths would have been avoided with universal regular screening of women aged 39 to 49. When all nine RCTs were included, the relative risk was 0.85 (95% CI 0.75 to 0.95). When three trials which randomised women before 1980 were excluded, the pooled relative risk did not show a statistically significant effect of screening (RR 0.87, 95% CI 0.56 to 1.13).