Donald A. Berry, a member of the Physicians Data Query panel of cancer experts, which advises the National Cancer Institute about cancer screening tests, is chairman of the department of biostatistics at M.D. Anderson Cancer Center in Houston.
Question: Were you surprised by the controversy (over the necessity of mammograms)?
Answer: In a way, I was. Our deliberations had been low-key. We proposed a modest change in the mammogram guidelines because we found that the scientific evidence that mammography is beneficial is pretty weak. Women have been oversold on screenings because there's so much anxiety out there about breast cancer. Consequently, anything that might help is viewed with great joy.
Q: What prompted your committee to review their guidelines?
A: Our conclusions about the benefits of mammograms were based on the outcomes of seven studies that have been conducted around the world. We knew these studies weren't perfect but there was an article a few months ago in a scientific journal (the Lancet) which detailed some serious flaws in this research. Our subsequent review agreed with some of their critiques, and there was certainly a smoking gun or two, in terms of potential biases in these studies.
Q: So your committee concluded there is no benefit to having mammograms?
A: We went through lengthy discussions, and the issue is quite complicated. But the key question is: Does having regular mammograms reduce mortality? And we decided that was uncertain -- and that's our bottom line. No definitive studies have been done that really prove mammograms are beneficial.
Q: If there is no benefit, then why do doctors continue to enthusiastically recommend that women get them?
A: Some (people) don't appreciate the importance of biases in making observations in screening programs. What they observe is that for women whose breast cancer is identified from mammograms, the prognosis is extraordinarily better than for those whose cancer is detected from a physical breast exam (by their doctors, a procedure called palpation). But it doesn't logically follow that a mammogram is what made the difference.
Q: What are these biases?
A: The first is the lead time bias. What that means is that you can detect an aggressive breast cancer earlier with a mammogram than you would by mere palpation. So instead of dying three years after being diagnosed, patients might live seven more years. But all you're really doing is moving the goal posts, and these women didn't survive longer than they would have without the mammogram. They were just diagnosed earlier.
The other key bias is length bias, which is more subtle. Mammograms tend to detect the smaller, slower-growing cancers that take longer to kill women. Or the tumor might have grown so slowly that they died of something else, and the cancer might not have been discovered otherwise.
Palpation, in contrast, tends to identify the faster growing, more aggressive cancers. These tumors often don't show up on routine mammograms because they may develop between screenings. As a consequence, since physical exams tend to find the more dangerous tumors, mammograms -- which pick up the less risky cancers -- look good in comparison.
Q: How do you account for the difference in the findings of your committee and those of the U.S. Preventive Services Task Force, which based their recommendations on the same research?
A: Two groups looking at the same data can come to different conclusions.
Brainerd Dispatch ©2013. All Rights Reserved.