Flawed Canadian study shouldn't deter women from mammograms

March 19, 2014

Breast cancer experts in New Zealand and around the world have warned that a "deeply flawed" Canadian breast screening study should not deter women from having mammograms.

The study has attracted widespread attention since it was published in the British Medical Journal in February, but has yet to be endorsed by breast cancer specialists and the wider medical community. Objections to the study of 90,000 women have been raised on three major grounds.

  • The "primitive" technology employed - the study followed women who had mammograms between 1980 and 1984, and used mammogram equipment that was substandard even for the time. "Having been one of the experts called on in 1990 to review the quality of their mammograms I can personally attest to the fact that the quality was poor," says Professor Daniel Kopans from Harvard Medical School. Only 30% of the breast cancers found in the Canadian study were detectable by mammography. Modern mammograms can detect 70-80% of cancers.
  • Suspect "randomisation" - while the study is described as assigning women randomly to mammography or no mammography, women were in fact given clinical breast examinations by a doctor or nurse. If a lump was detected, the woman was put into the mammography arm of the study. While that's understandable from a patient care perspective, it does mean that women with more advanced cancers were put into the mammography arm, which increased the chances of dying in the mammography arm.

  • Several larger and better controlled studies contradict the Canadian study, showing a reduction in breast cancer mortality of between 20-40% for women who have mammograms.One of the best known is the Swedish Two-County study, which has followed 130,000 women for 29 years and has found significant survival benefits for women undergoing screening mammography.

"That [20-40% mortality reduction] is certainly what we observe clinically," says Dr Patrick Borgen, chairman of surgery at Maimonides Medical Center in New York. He describes the Canadian study as "very flawed", and says a number of the doctors who were involved with the study resigned their positions in protest.

Here in New Zealand, NZ Breast Cancer Foundation medical advisers Dr Ron Kay and Dr Belinda Scott agreed that the Canadian study's outdated technology and poor randomisation make it unreliable. Belinda Scott says the Swedish Two-County study is a far more robust study, demonstrating "a substantial absolute reduction in mortality from breast cancer" as well as a decrease in diagnoses of advanced cancers.

Read Professor Daniel Kopans' discussion of the Canadian study's technical flaws, published in the British Medical Journal


-