The randomization question

From the outset, the Canadian National Breast Screening Study has been a source of controversy. Data from the first arm (CNBSS-1), released in 1993, questioned the effectiveness of mammography screening for women aged 40 to 49.

Critics at the time argued that the study’s methodology was flawed because the randomization procedure was not blinded. Women were examined first and then assigned to a screened group or control group on open lists. As a result, too many women with advanced breast cancer may have been allocated to the screening arm.

The latest information as published in CNBSS-2 brings the study's methodology under fire once again. In an official statement, the ACR said "the study design, involving volunteers, cannot be used to draw conclusion about the general population. The authors correctly point out that the death rate in their control group is far lower than in the average Canadian population."

The same held true for CNBSS-1, which showed a greater than 90% survival rate among women ages 40 to 49, strongly suggesting a "randomization imbalance," the ACR stated.

But in its paper, the University of Toronto team defended its methods, pointing to a meta-analysis done by Danish researchers that concluded that the study was sound. The authors, from the Nordic Cochrane Center in Copenhagen, reviewed eight breast imaging trials for bias and found that there were too many inconsistencies in six of the eight studies conducted to render them valid. However, they praised the 1993 CNBSS study methods.

Another major criticism of the study has been that during the early 1980s, the healthcare providers involved were not properly trained in breast imaging techniques.

"Mammography technique was poor even at the time the study was done. The radiologists interpreting these mammograms had no special training, unlike those involved with clinical breast examination (CBE)," said Dr. Carl D’Orsi, professor, vice-chairman and director of diagnostic radiology at the University of Massachusetts Medical School in Worcester.

The ACR added that healthcare professionals taking part in the CNBSS were "highly trained and skilled" in CBE, but that the same could not be said for the radiologists.

Next page: Mammography versus clinical exam?

1, 2, 3

Click here to discuss this story.

Page 1 of 569
Next Page