The following originally appeared on The Upshot (copyright 2015, The New York Times Company).
In 1985, about three-quarters of women age 40 to 79 had never been screened for breast cancer with mammography. By 2010, things had changed enormously. More than 70 percent of women of similar age had been screened in the previous two years. Many people attribute the large improvements we’ve seen in breast cancer mortality over that period to early detection through ambitious screening programs.
In recent years, though, many researchers have concluded that annual screening mammograms for a wide range of women may be doing more harm than good. Because of this, the American Cancer Society has updated its recommendations on how often women should be screened. But after years of public awareness campaigns, getting physicians, and the public, to accept the new guidelines will probably be very hard.
The cancer society now recommends that most women not begin annual screening mammography until age 45. When they turn 55, it recommends that they switch to every-other-year screening. That should continue until a woman is expected to have less than 10 years of life remaining. The society now recommends that no clinical breast exams — where physicians or nurses feel for lumps — be done as a screening procedure at any age.
This is a big change. For years, many organizations, including the cancer society, recommended annual mammograms beginning at age 40 and clinical breast exams every few years starting as early as age 20. After 2003, the society changed its guideline development procedures to be more transparent and evidence-based; these new recommendations are a result.
For instance, a new study examined whether annual or biennial screening is superior for women of different ages. The researchers used data from the Breast Cancer Surveillance Consortium to look at more than 15,400 women 40 to 85 who were found to have breast cancer within a year of an annual screen or within two years of a biennial screen. They wanted to determine if the screening interval affected prognosis.
This is important because many believe that stretching out the time between mammograms means that when tumors are diagnosed, they will be more advanced, and therefore more dangerous. The researchers found that in premenopausal women, women who had screens every other year were more likely to have a higher proportion of cancers diagnosed at stage IIB or higher, or to have a tumor greater than 15 mm. After menopause, however, these differences were no longer found. This is why annual screens are recommended from age 45 to 55, and biennially after that.
But it’s important to talk about the benefits of screening not just in relative risk reductions. The overall impact, or absolute rate reduction, of screening is often small. Systematic reviews of randomized controlled trials estimate that almost 2,500 women age 40 to 49 need to be screened to prevent one death from breast cancer in 15 years. More than 1,600 women over age 50 need to be screened to achieve the same result. Cohort studies and case-control studies are more supportive of screening, but still find that many hundreds of women need to be screened to prevent one death from breast cancer over 15 years.
Harms also result from screening mammography. More than half of women screened annually for 10 years will have at least one false positive finding. These can result in as little as a second exam and as much as abiopsy, which also carries small but real risks. Fewer false positives result from biennial screens. A modeling study showed that using biennial screens instead of annual screens in women 50 to 69 might result in 4,000 fewer false-positive biopsies for every 100,000 women over 10 years (and 57,000 fewer false positives over all).
The evidence looking at clinical breast exams is surprisingly thin. One cohort study from the United States in 2005 found that the addition of the clinical breast exam to mammography detected 0.4 extra cancers per 1,000 women, but also an extra 20.7 false-positives. This means that more than 50 false positives were obtained for every real positive. Cohort studies fromCanada and Japan found similar results, and no apparent benefit from clinical breast exams alone.
The American Cancer Society recommendations still involve more screening than the United States Preventive Services Task Force, a panel of experts appointed by the federal government, suggests. That body recommends no universal screening until age 50, and that it occur every two years after that until age 75.
Both of these sets of recommendations, though, argue that old recommendations were too aggressive. They maintain that our default should be screening later and less often than we once thought. Neither recommends clinical breast exams anymore.
These changes will be very hard to put in place. Many studies have shown that once physicians become conditioned to be more invasive, it’s hard for them to reverse course. This year a study highlighted the results of a randomized trial to reduce the rate of C-sections, which have become too common in the past decade. Conducted over one and a half years in 32 hospitals in Quebec, the positive results mentioned included a reduction from 22.5 percent of cases to 21.8 percent. That’s what success often looks like, after spending lots of money and many man-hours in interventions. It’s unlikely that will occur generally for this recommendation.
A more relevant study was published even more recently in JAMA. It discussed the effects of changing guidelines on the practice of glycemic control in intensive care units. In 2001, a study was published that seemed to say that tight glycemic control was preferred in elderly patients. Over the next seven years, more and more admissions used this algorithm. But in 2008, a bigger and better study was published that showed this practice was actually harmful. Since that time, the adoption of tight glycemic control has slowed, but it has still increased, not decreased.
It’s a problem with treating breast cancer as well. Even though recommendations argue that prophylactic mastectomy be considered only in women with a mutation in a cancer-related gene like BRCA, more than two-thirds of women who have the procedure don’t have such a mutation. This is in spite of the fact that a mastectomy doesn’t improve survival in those women. Even though breast conservation therapy has become a “standard of excellence,” studies show that people aren’t listening.
It is very hard to get doctors to do less. I have discussed this before in the context of the placebo effect in operations. Once physicians believe they are doing good, it is hard to get them to change their minds. Others think that doing more protects them from lawsuits, although that’s most likely untrue. Some have seen screening pick up diseases, leading to treatment and positive outcomes, and have assumed that it was the screening that made the difference (although evidence refutes this). And, yes, some are probably influenced, overtly or subtly, by financial incentives that still lean toward everyone being more invasive.
But in this case, it will most likely be even harder to change the public’s expectations. For decades, we’ve been told, over and over, that more screening is better, that early detection is the key to a cure. That’s true, up to a point. We seem to have passed that point, though, and more and more experts are trying to reverse course.