Last week, the conversation about breast cancer diagnosis and treatment moved into new and uncomfortable territory. The New England Journal of Medicine published a large-scale study concluding that mammograms may not be as helpful as most people believe in treating breast cancer. Even worse, the study suggested that more than 1 million women may have been unnecessarily treated and diagnosed.
Everyone agrees that fewer women are dying from breast cancer. We—meaning the medical community and interested women—assumed that the increased survival rates were a result of early detection and mammography. “Early detection saves lives” is one of the main rallying cries of the Susan G. Komen Foundation-led pink-ribboned breast cancer community. But what if it doesn’t?
The study looked at breast cancer diagnoses over the last 30 years. As expected, screening coincided with an increase in detection of early stage breast cancers. But the number of metastatic cancers was unchanged. Even the number of late cancers— those that had spread to the nodes but not distant organs—did not decrease very much.
The idea of screening is to find disease early so it can be treated. People thought breast cancer was perfect for screening—that early breast cancer inevitably turns into late cancer. If this simple progression were true, then the number of increased early diagnoses weʼve seen during the last 30 years should be matched by an equivalent decrease in late diagnoses. But this did not turn out to be the case.
This study is flawed in the usual way that broad-based population studies are flawed. The claim is based only on observation and association. We just know that during the period of the study, mammogram use increased, the number of early breast cancer diagnoses increased, and late-stage cancer diagnoses barely decreased. We donʼt even know whether, for example, the women with metastatic cancer even got mammograms.
If a patient never had a mammogram and then came to her doctor with a huge palpable mass, we obviously would not expect early detection by mammogram. But that patient is still in the pile of late-stage cancers the authors use to demonstrate overdiagnosis. (A more meaningful study would only compare women whose cancer was diagnosed by mammogram.) Another reason the study can’t definitively show its touchiest suggestion—that mammograms were overdiagnosing cancers—is no one knows what is happening to these early stage cancers. The authors speculate that these cancers regress or just stay the same. But speculation is not the same as looking at the cells and watching them. That would require letting early stage cancer remain in someone’s body and sampling it over time without treating it. And at this point, when we are still aggressively treating early stage breast cancer, that study is nearly impossible to pull off.
Since I became a family-practice resident in 1998, I have been referring women for mammograms. Until 2009, it was pretty straightforward: Any patient over 40 got a mammogram referral every year, maybe every other year. When a federal task force recommended against screening mammograms in women between 40 and 50 in 2009, the backlash was fast and harsh. There were accusations of health rationing, of setting up women to die.
Though I found the 2009 recommendation credible, I have to admit I didn’t follow it. Mammograms were still the standard of care. The American College of Radiology opposed the recommendation. I felt too vulnerable. I worried I would be responsible if a cancer was missed. I, like most doctors, continued to refer all female patients over 40 for mammograms. But this new study may actually change my practice. I will probably still offer every woman over 40 a mammogram, but I won’t do it on autopilot. I will feel compelled to explain to my patients what the risks are, which is hard to do in the allotted 15 minutes. (It’s hard enough to explain it in an article.) Until we reach a tipping point, until the medical establishment as a whole decides that mammograms are not all that helpful, we doctors feel too vulnerable to be on the cutting edge.
Every patient has a mom or a friend or a cousin whose life was “saved” by treatment. Itʼs hard to tell someone that momʼs double mastectomy may not have saved her life. The breast cancer survivor is a modern heroine, clad in pink, surrounded by family, walking for the cure. It won’t be so easy to recast her as a stooge of bad medical information.
What makes this harder is that there are camps fiercely interested in one side or the other. H. Gilbert Welch, one of the lead authors of the new study and a known screening skeptic, published the book Overdiagnosed: Making People Sick in the Pursuit of Health in 2011. (He is the same researcher who worked to debunk the use of PSA screenings for men and prostate cancer.) “Overdiagnosis causes harm ranging from unnecessary worry to death in rare instances,” Welch once told Time. Welch’s skepticism about screening tests makes him a radical in the world of preventative medicine. He asks uncomfortable questions that threaten the equanimity of both patients and doctors.
And to be clear, my equanimity is sufficiently threatened by this study. I need to figure out how to integrate this new data into my practice. I suppose I will tell my patients:
We know a mammogram finds early cancer, but we have no way of telling if this early cancer would have killed you without treatment. Every mammogram may put you at risk for getting cancer treatment you don’t need. However, not getting a mammogram also has risks. We may not find a rare, early cancer that may become deadly.
In the past, when I would put that kind of decision to a patient, she would say, “Doc, what should I do?” In this case, I’ll have to say, “I don’t know.” But of course I won’t be able to leave it at that. People expect their doctors to have an opinion. People are looking for guidance. So I guess I’ll have to say, “It depends on what makes you more uncomfortable: the idea of unnecessary treatment or the idea of a missed cancer.”
But when she asks me if she’ll be OK if she skips the mammogram, I’ll be back to, “I don’t know.”