« Rivers North | Main | Value »
June 19, 2007
mammograms
Please remember that I am not giving advice. Rather, I am recording information and trying to think through these issues.
Mammograms are one of the givens of modern medicine. A recent review in American Family Physician states correctly, "Breast cancer is one of the most significant health concerns in the United States. It is the most commonly diagnosed cancer in women and the second leading cause of cancer death in women." (So it is madness for me to challenge this sacred cow. But mad I am and challenge I must.)
Over the past few decades, several methods of screening have been advocated. Screening makes prima facie sense. Find the cancer early and you can prevent more significant problems. With this reasoning, any screening method is better than none. So we have the options of Breast self-examination (BSE), clinical breast examination (CBE), mammography, ultrasonography, MRI, scintimammography, PET and ductal lavage.
Compared to other conditions for which one might screen, at least breast cancer is relatively common. A woman has a one in 11 risk of developing breast cancer, so the saying goes.
But a woman's risk is 3-4 times higher if she has a first degree relative with breast cancer. This means that one's risk is much lower if one does not have such a relative.
No screening test is perfect. The sensitivity and specificity of these various tests vary, but most are about 60-80% (except ductal lavage, which has a sensitivity around 20%).
There are many studies on the effectiveness of these various methods of screening. There are 9 randomized controlled trials since the 1960's. Both USPSTF and Cochrane have reviewed these, coming to slightly different conclusions. But both agree that BSE is not helpful. cochrane concluded that SBE increased the number of unnecessary biopsies and was thus harmful.
CBE evidence was inconclusive.
Mammography gave mixed results. Cochrane concluded that a few of the trials were flawed. Excluding these, they found no reduction in mortality with mammography. USPSTF included those trial and found some benefit. Computer-aided detection (CAD) should in theory improve the sensitivity, but Fenton's NEJM study found otherwise. False-positives increase and more unnecessary biopsies are done.
At best, in women 40-49, screening 1,792 women with mammography (with or without CBE) will save one woman from breast cancer death in 14 years. For older women, 50-69, the number needed to screen (NNS) is 838, much more reasonable. I am not sure of the NNH (number needed to harm) here but it is significant. A single false positive can induce worry enough to erase any benefit. Anyone who dismisses this with the possibility that the next test will prove no cause for worry has lost touch with reality. No amount of retesting will erase the worry for some women.
But not to be side tracked, let me reiterate. We must screen almost 1000 - 2000 women to prevent one cause-specific death in 14 years. To do that we risk the time, radiation exposure, worry, unnecessary biopsies, unnecessary mastectomies, and hubris. Hubris more than anything else. We think that by preventing one death from breast cancer, since breast cancer is such a great killer, that we are fighting back death on a grand scale.
Russell Harris, MD, MPH, has a good editorial in the same issue, reasonably summarizing this data. I agree with his one sure conclusion: teaching SBE is a waste of time. And some of the genetically at risk women might benefit from careful screening.
What are we trying to accomplish focussing on such small risk-reduction? We think that if we can conquer even a fraction of such a devastating disease, we are further on the road to immortality. Or so it would seem.
Social iatrogenesis | By Robert Maddox | 10:32 PM