[The scientific basis for population screening for breast cancer in the Netherlands]

A recent Cochrane review stated that there was a lack of evidence for a decrease in mortality as a result of population breast-cancer screening. The principal data were drawn from five Swedish randomized controlled trials and one Canadian trial. However, the studies cannot be so easily combined because there were important differences in the attendance rate, detection rate, technical quality, referral rate, clinical baseline situation and screening interval. For example, in one of the studies the women from the control arm underwent an annual clinical palpation carried out by a trained nurse or physician, which could have led to an underestimation of the screening effect. Further breast-cancer mortality might not be a good outcome measure because this was not reliably determined; only total mortality was to be observed. This is clinically and methodologically incorrect because breast-cancer mortality was meticulously studied, documented and validated. In the Cochrane review it is suggested that the randomisation was inadequate, but evidence for this was not supplied. The discussion about age differences as a marker for incorrect randomisation is out of date and has been revealed to be unjust. It seems likely that an important part of the decreasing trend in breast-cancer mortality in several countries (including the Netherlands) is due to screening programmes. However, the evaluation of breast-cancer mortality over the next five years is crucial, if greater certainty is to be gained about this.