Letter to the Editor Flawed methods explain the effect of mammography screening

van Schoor et al (2011) apply a case–control design to evaluate the Nijmegen breast screening programme, which began in 1975. They estimate a 65 % reduction in breast cancer mortality during the period 1992–2008 by calculating the odds ratio for the breast cancer death rate for screening participants vs non-participants. It is not the first time that this study design has been used to claim a large effect of screening mammography in Nijmegen (Verbeek et al, 1984). In 1984, some of the same authors used a similar design to claim a 52 % reduction in breast cancer mortality by 1981. This large effect was surprising, as the same authors reported that there was no actual reduction in the breast cancer mortality rate in Nijmegen at that time, when participants and non-participants were combined (Verbeek et al, 1984). Interestingly, although the new study covered a longer observation period (1975–1991), it identified an effect of only 28 % in Nijmegen by 1991 (van Schoo