The Canadian National Breast Screening Studies are compromised and their results are unreliable. They should not factor into decisions about breast cancer screening

Abstract

The Canadian National Breast Screening Studies were compromised by an unblinded allocation process and poor quality mammography. Contrary to the requirement that allocation in a randomized controlled trial (RCT) be blinded to avoid any possible intentional or unintentional subversion of a random allocation, all women in the CNBSS trials underwent a clinical breast examination prior to assignment to the study arm or the usual care arm. Women with abnormal clinical breast examinations were identified, and this information was available to the coordinators who then assigned the women on open lists. It was, therefore, possible to assign women to whichever arm the coordinator chose. Although subversion was likely unintended, a significant number of women with four or more positive axillary lymph nodes were assigned to the screening arm of CNBSS1. This explains why there were more breast cancer deaths among the screened women in the first ten years of the trial and why the 5 year survival of the control women was better than 90% when the background survival in Canada at the time was only 75%. The trials were further compromised by the poor quality of the mammography which was confirmed by a review conducted by the trials' organizers. These fundamental problems compromise the CNBSS and make their results, which are major outliers in the RCT's of breast cancer screening, unreliable. Consequently, they should not be used to establish guidelines for breast cancer screening.

Keywords: Breast; Breast cancer; Mammography; Randomized trials; Screening.