Recently, the British Medical Journal published the results of a study from Canada, the conclusion of which was that mammography does not save lives. This has generated a huge amount of press coverage (for a radiology topic), of course because it affects so many women. This study was performed in a large number of women in Canada, and the first author is an epidemiologist, Anthony Miller.
Since I do this intensely and read about it extensively, I do have some perspective on the issue that I can share.
The first thing to say is that, in all truth, no one knows with absolute certainty how many lives, if any, are saved by mammography. There are conflicting data. This is extremely common in medicine. Thus, the results of a single study are never definitive. One must get acquainted with the breadth of the information available from all sources, and then synthesize a sense of what is available into your own estimate of what is most likely to be true.
The New York Times led with a story about this publication that can be seen here:
(Note: the NYT’s firewall may or may not allow you to see it. On my second attempt to view it, I was prohibited)
When first read, this article seems convincing. If you have a little more knowledge about the situation, it is clear that the NYT has published something that they should be embarrassed about. Not that they misquoted, but the story that they wrote reads as if it was taken directly from the press release that the BMJ sent out, without any thought whatsoever. There is no mention that this is a very controversial topic. I have the distinct impression that this was a conscious decision on the part of the NYT to skew the story to a very negative viewpoint. Why do I say this? Because in the very same issue of the BMJ there was a comment written by Dr. Dan K0pans pointing out the very serious (and if true, fatal) flaws in the study. Kopans point was that this study was so badly done, that no valid conclusion can be made. (more on this later). This important viewpoint was in the very same issue as the article itself, and thus could not possibly be missed by a minimally conscientious reporter. Thus – in my view – it had to be a purposeful act to push an agenda of the NYT’s.
For a better report that includes the objections to the study, Diane Rehm on NPR did a good job –
Further – similar responses to Kopan’s (actually with many of the same sentences) were published immediately on the website of the Society for Breast Imaging, and the website of the American College of Radiology.
One thing that needs to be understood is that this is by no means a “new” study. The study was conceived in 1972, and actually started in 1980. They ended the patient enrollment in 1984, and began the phase of following up patients to see what their outcomes were.
Every few years, they publish results, which have been consistent since the first publication – that mammography did not change survival. I first heard the in depth analysis of the study at the 1992 Society of Breast Imaging. (22 years ago!) At that time the methods of the study were debated by the author, Miller, and by Kopans. I felt that Kopans discredited the study sufficiently that I couldn’t believe the results. (more on the specific problems later). Nevertheless, this study continues to be published every few years with a new eruption of debate. This is not new.