Abstract
Objective –
To examine the survival of patients with breast cancer diagnosed in different centres and by different methods in the United Kingdom trial of early detection of breast cancer, in order to investigate the contribution of different factors to the previously observed reductions in breast cancer mortality.
Setting –
A non-randomised trial of the early detection of breast cancer, in which women aged 45–64 in two districts were offered annual screening for seven years, women in a further two districts were offered education about breast self examination (BSE), and those in four districts formed a comparison group.
Methods –
Patients with breast cancer are classified according to the type of centre, method of detection, and attendance for BSE education. Univariate and multivariate survival analyses are carried out, including tumour size, dissemination status, and use of adjuvant treatment as additional variables.
Results –
In the univariate analysis, patients with breast cancer who are nonattenders for screening have a significantly worse prognosis than those in the comparison centres. Patients whose cancer is detected by mammography have the best survival rate. The inclusion of size and dissemination status in the multivariate analysis explains only about one third of the improved prognosis in these cases. There is a significant difference between prognosis in the two BSE centres.
Conclusions –
The use of prognostic factors as recorded in this trial to predict breast cancer mortality may be inadequate.