An overview of the Swedish randomised mammography trials: total mortality pattern and the representivity of the study cohorts

Abstract

Objectives: To analyse the cause of death pattern in the cohorts of women included in the Swedish randomised mammography screening trials by comparing the groups of invited and control women both with each other and with the general population of Swedish women.

Setting: Since 1977 four randomised trials of mammography screening have been performed in Sweden: Malmö, Kopparberg and Ostergötland (the two county trial), Stockholm, and Gothenburg.

Design: Overview of four randomised mammography screening trials.

Results: The total numbers of deaths in the invited and control groups respectively were 15 695 and 11 887 corresponding to a relative risk (RR) of 1.00. There were no significant differences between the invited and control groups for cause-specific mortality, except for breast cancer. When the total mortality in the invited and the control groups was compared with that for Swedish women in general the standardised mortality ratio was close to 100.

Conclusions: The cause of death pattern in the invited group was, except for breast cancer, very similar to that in the control group, showing that the groups were comparable. Similarly, the total mortality including breast cancer mortality in the control group was almost identical to that in Swedish women in general. The same was true, with the exception of breast cancer, for the invited group. These observations confirm that the trial cohorts are representative of Swedish women and indicate that the quantitative results from these trials may safely be generalised to the Swedish populations.