What were they thinking?
Last week the US Preventive Services Task Force - a group of independent screening experts - released new recommendations for breast cancer screening. But instead of the standard suggestion of doing more - they suggested doing less. Screening mammograms should be routinely begun at age 50 (not 40), should be done every two years (not every year) and self-breast exams should be discouraged (not encouraged).
Perhaps the question should be rephrased: What were they smoking?
Breast cancer screening has a reputation of being the "3rd rail" of screening policy. Arguably the best studies have been ignored or impugned, scientific consensus panels have been overruled by politicians, and histrionic (and personal) attacks have been leveled. Traditionally, there has been only one safe policy proposal: propose to do more. Otherwise, nobody wanted to touch it.
Some of the reaction to last week's recommendation lived up to this reputation. But there is a changing climate of medical opinion about screening. It is also changing among breast cancer advocacy groups. You might reasonably wonder why.
First, understand exactly what screening is: it is the systematic search for cancer in people who have no signs or symptoms of it. All the debate about mammography has nothing to due with the question of how to evaluate a new breast lump that has become apparent to a woman (or her partner). She should go to the doctor; the doctor should order a diagnostic mammogram. That's not screening. There's no debate there.
The changing climate around screening reflects a fact that has become more broadly recognized: the effort to detect cancer early is a double-edged sword. It has benefits; it has harms.
The benefit of breast cancer screening is that some breast cancer deaths can be avoided. Unfortunately, it doesn't happen very often: most women destined to die from breast cancer, will still do so - even if they are regularly screened. To get a sense of magnitude consider this: among 1000 women age 50 about five will die from breast cancer in the next 10 years. Screening mammography over the 10 years will take that number down to around four. In other words, out of 1000 women -- one benefits (the number might be as high as two for 70 year olds or as low as 0.2 for 40 year olds).
But that means 999 women were screened for 10 years and gained nothing. There is a growing appreciation that what happens to them is also relevant. They can be harmed.
All have to be engaged in the screening process. While many may find it not at all bothersome, others may find it to be a hassle-filled, scary and/or painful experience. And many -- somewhere between 200-500 in this country -- will have an abnormal mammogram at some point over the 10 years. About half of these will be biopsied. Some will find these false alarms terrifying. Some will be made to worry that, although they don't have cancer, their breasts are somehow abnormal.
Then there's the problem of overdiagnosis: women who are diagnosed with a cellular abnormality that meets the pathologic definition of cancer, but turns out to be a form of cancer that was not destined to cause symptoms or death. Because we doctors don't know which cancers are which, we treat everybody. Therein lies the problem: among those 1000 women, it looks like somewhere between two and 10 will get treated for a cancer that was never going to bother them. Since there was nothing to fix, the treatment can only cause harm.
Some may ask: Shouldn't doctors do everything possible to avoid breast cancer deaths? But consider the implications were that the only goal: we would start screening women at age 20 (yes, women can get cancer in their 20s), we would screen every month (yes some cancers grow so fast this would be the only way to catch them early), and we would never stop. In fact, if we wanted to really do everything possible, we would suggest prophylactic mastectomies in all women.
But you know this is crazy. Screening programs require a delicate balancing act. It's hard to make healthy people better, so it's doubly important that we don't make them worse.
One of my colleagues likens designing screening programs to operating a control console with multiple dials. There's a dial to turn for how young to start. There's a dial for how old to end. There's a dial for how frequently to screen. There's a dial for how intensive a test to use (e.g. physical exam, mammography, MRI). And there's a dial for how hard to look for abnormalities in the examination (the threshold at which an exam is labeled "abnormal").
There is one thing doctors are sure of: the right position is not to turn the dials all the way up. That's a recipe for lots of unnecessary anxiety, false alarms, biopsies, surgeries and overdiagnosis. Dialing up screening programs is definitely not the path to a healthy population.
The Task Force recognized this. They found that most of the benefit of mammography could be preserved -- and the harms substantially reduced - if the dials were turned down a bit. So did the American College of Obstetricians and Gynecologists last week, when they dialed down their recommendation for how frequently cervical cancer screening should be performed.
These are example of how all of us - doctors, reporters, policymakers and the general public - need to think about screening tests in general. It's simply not healthy to look too hard for things to be wrong in the well. No one wants to make people sick in the pursuit of health.