Randomized trials of mammography have demonstrated the efficacy of mammographic screening for breast cancer in terms of preventing deaths, but various issues of particular interest remain, including: quantification of overdiagnosis; evaluation of service screening outside the research setting; absolute benefit in terms of number needed to screen per life saved; which types of tumours benefit most from early detection; use of screening data to investigate tumour biology and natural history. This paper describes examples of approaches to the above issues, along with some important results.