11) A large randomized trial (n = ~30,000 in each of two groups) is published evaluating routine mammography versus no r
Posted: Sun Feb 13, 2022 8:55 am
11) A large randomized trial (n = ~30,000 in each of two groups) is published evaluating routine mammography versus no routine mammography for reducing the mortality from breast cancer. After randomization, the authors decided to drop patients who had the onset of breast cancer before enrollment in the trial. This resulted in 800 patients being dropped from the study arm and 300 from the control arm. Among the study group, 153 patients died of the disease versus 196 in the control group (p<0.05). The authors of this widely quoted study concluded that the screening test was effective in reducing mortality from the disease. What principle is violated in conducting this analysis? How does violation of this principle affect the estimate of treatment effect (increase estimate of effect, no change on effect, or decrease estimate of effect) and why?