The
US Preventative Services Task Force (USPSTF) on Breast Cancer has been big news. Many people don't seem to understand how more diagnosis can be bad, and assume this is a nefarious plan to save money at the expense of women's health.
What the USPSTF recommendations say is that women who want a mammogram should get it, but unless the woman is eager for a mammogram or is at high risk, women under 50 should not get mammograms, and regardless, every two years is often enough for mammaography.
USPSTF reserved their harshest evaluation for manual breast exams, both self-exam and exams performed by a clinician. USPSTF found that such exams don't save lives, and lead to needless suffering.
I'm not an oncologist, and medical knowledge of cancer is quite incomplete, but I can explain why over-diagnosing breast cancer is harmful. From a treatment standpoint, there are three kinds of breast cancers:
- Slow growing cancers that are relatively easily treated, and may not cause a woman any medical problem were they to go undetected and untreated.
- Intermediate cancers that are likely to eventually become metastatic and deadly if untreated.
- Agressive cancers that grow very rapidly after appearance and are often quickly lethal.
Most of the women helped by mammograms are in the intermediate category. Their cancer will probably kill them if undetected, but if cancer is detected and treated in time, these women stand a high chance of survival. The problem is, there aren't very many women between the ages of 40 and 50 in this situation. What makes this state of affairs worse is that, both individually and in aggregate, early diagnosis of women with either slow-growing or aggressive cancer looks like progress, but doesn't necessarily make anyone live any longer.
- Women with slow-growing cancer are probably going to survive anyway, and some would never be diagnosed. When these women started being diagnosed with cancer, it caused the illusion of improved breast cancer survival. In the past, these women were surviving breast cancer for long periods without being diagnosed or treated. Also in this situation are women who are diagnosed with breast cancer but do not actually have cancer. That is, they are false positives. For women in either of these situations, diagnosis doesn't much affect or even reduces their life expectancy, because cancer treatment is somewhat dangerous. They may receive no benefit in return for this risk.
- Women with untreatable, aggressive cancer aren't much helped by diagnosis, because diagnosis doesn't lead to a treatment that benefits them very much. But as with slow-growing cancer, even if detection doesn't help these women live longer, it causes an illusion of improved survival, because these women tend to get their diagnosis earlier, even if they don't live longer.
Finally, there are risks to women who have negative mammograms, because mammograms require radiation. Each individual mammogram isn't very risky, but a woman following past recommendations will receive ten radiation doses to a radiation-sensitive tissue that they would otherwise avoid.
The USPSTF's analysis is complex, involving review of many complicated yet imperfect studies and meta-analysis with a half-dozen statistical models. I reviewed the USPSTF recommendations, but I obviously didn't try to duplicate their entire analysis. I lack both the time and the qualifications to repeat their work. But I find the USPSTF recommendations highly credible, because their conclusions are logical and well supported, and because other groups have been reaching similar concusions for many years.
Does that mean cancer diagnostics are a bad idea? No, of course not. Some of the improved survival from mammography has been an illusion, but mammography really does save lives. The problem with cancer screening is the problem with Dx in any population with a low incidence of disease: in populations where few people actually have the disease, a diagnostic has to be very, very good to have both low false positives and false negatives. If a significant number of people who get a postitive diagnosis don't benefit from treatment, the challenge is even worse. Over the next decade or two, diagnostics are going to revolutionize medicine, but many of the victories are going to be hard fought. USPSTF is telling us that women need to pick their battles with mammography. At some point, a diagnostic approach that works for women in their 40's will be developed, but we don't have it yet.
Labels: diagnostics
I reviewed this great new book by Atul Gawande here. For the P4 blog, I'll add a few comments realative to P4 Medicine. 
