First of all, it's not fair to compare the numbers 67, 12, and 58, because they were based on different denominators - you have to start by converting them to rates (per 1,000 women, say): 67/20200 = 3.3 per 1,000, 12/10800 = 1.1 per 1,000, and 58/31000 = 1.9 per 1,000. So the rate of breast cancer detection in the examined group was 3.3/1.1 = 3 times higher than in the refused group and 3.3/1.9 = 1.8 times higher than in the control group. This is certainly enough to establish an association between screening and an increase in breast cancer detection, because only the examined group got screened, but of course it doesn't mean that screening caused an increase in breast cancer. All that is going on is that screening is better at *detecting* breast cancer, which after all was the point of the screening.