New study suggests mammography benefit lower than previously believed
Readers may remember the uproar last year when the U.S. Preventive Services Task Force (USPSTF) recomended against routine mammography of women between 40 and 50 and that low-risk women over 50 go from having mammography yearly to every other year. The Task Force was not very politically savvy in how it presented its recommendations, failing to adequately explain to the public the basis for its position.
The recommendations were brutally attacked as potentially leading to “countless deaths,” often by those with vested interests in the status quo, such as the American College of Radiology, whose members stand to lose a lot of business should the recommendations be implemented. The recommendations were even compared to not screening airport luggage for bombs, thus using fear of terrorism to trivialize the report.
I found the quality of the debate to be deeply disturbing. Proponents of the current, or even increased testing ridiculed the idea that testing has risks. “What’s a little anxiety against a saved life,” they argued. This argument has several weaknesses. One is that “anxiety” caused by false positives, of which there are hundreds per correct diagnosis, can be severe and debilitating in some cases. Also, the fear of cancer may remain with some women for a long time and not be allayed by repeat negative findings, thus potentially reducing quality of life.
Another major concern is that the result of false positive cancer screenings is not simply “anxiety,” but is also unneeded biopsies in about half of those receiving false positive screenings, with greatly increased “anxiety,” as well as all the risks of having an unneeded medical procedure. Furthermore, many cancers that are detected would otherwise have turned out to be asymptomatic, producing no problems and never being detected before the woman died of other causes. Since medicine currently has no way of determining whether a detected cancer is one of these ultimately benign asymptomatic ones or a dangerous, potential deadly, disease, most of these women with asymptomatic cancers will wind up receiving surgery and/or chemotherapy that ultimately was unneeded, with all the risks, side effects, and, yes, “anxieties,” that accompany breast cancer treatments. Plus these women will live the rest of their lives fearing the recurrence of a disease that, unknown to them and their physicians, was never a threat.
To avoid being misunderstood, this is in no way intended to slight the potential value of mammography for many women. I am simply arguing for recognition that screening tests like mammography, like most medical interventions, are not without risk and that the risks should be realistically evaluated by physicians, communicated to patients, and considered alongside the possible benefits.
All of this is by way of introduction to a new study in the New England Journal of Medicine that adds additional complications to the proper evaluation of the value of mammography. [There is an excellent detailed analysis of this study by David Gorski in Science-Based Medicine on which I base much of my description of the study. The New York Times also has a briefer, and less detailed, but still useful article.]This study developed a methodology to substitute for a randomized controlled research design in evaluating mammography benefit.
The only true randomized controlled trials of mammography were conducted decades ago; these studies found a relatively large benefit in saved lives. Due to ethical concerns, these studies can likely never be repeated as some women would have to be denied mammography after it was proven beneficial.
The new study’s authors set out to examine mammography’s benefit today, when treatments for breast cancer are much more effective, due to the development of targeted interventions, than when the original randomized trials were conducted. Due to improvements in radiological techniques over time, the researchers expected to find even greater value of mammography today than in prior decades.
Since the researchers could not use a randomized procedure to assign large numbers of women to receive or not receive mammography, they sought an alternative way to control for the likelihood that women who receive mammography are likely different from those who don’t receive it in ways that may affect their chances of dying. Thus, for example, women who go for mammography may be more likely to follow medical recommendations regarding healthy lifestyles or may be more likely to follow through on medical care. Thus, they might be less likely to die from breast cancer than a group not receiving mammography for reasons not connected to the actual value of the test.
The researchers, conducting their research in Norway, developed a clever methodology using a historical accident in Norway to approximate a random design. In the late 1990′s, Norway gradually rolled in a program to offer mammography to most women aged between 50 and 69. The researchers compared deaths from breast cancer over time in geographic areas where mammography was not offered over the time period examined in the study o estimate how much of the reduced death rate was due to historical improvements in treatment approaches. They then compared the death rates before and after mammography was offered in other communities, to determine the total benefit, including both the historical improvements due to treatment as well as benefits due to mammography. For those who’d like to hear this design described in the researchers’ own language:
First, we compared women in the nonscreening group with their historical counterparts to determine the temporal change in mortality that was not attributable to the introduction of the screening program and that was likely to reflect improved treatment and earlier clinical diagnosis. Then, we compared women in the screening group with their historical counterparts to determine the change in mortality after implementation of the screening program. In this second comparison, the difference in the rate of death between the two groups can be attributed both to the screening program and to temporal trends in mortality that were unrelated to the screening program. Thus, the reduction in mortality that was related to the screening program was the difference between the rate ratio for death among women in the screening group as compared with their historical counterparts and the rate ratio for death among women in the nonscreening group as compared with their historical counterparts.
Through this design they concluded, to their surprise, that approximately two-thirds (18% of 28%) of the reduced death rate over time — which simpler research designs would ascribe to the apparent benefit of mammography — was really due to improved treatment or other factors than the screening procedure. This study suggests that the benefits of mammography screening, while real, are smaller than previously believed. The New York Times described the magnitude of the effect this way:
In the new study, mammograms, combined with modern treatment, reduced the death rate by 10 percent, but the study data indicated that the effect of mammograms alone could be as low as 2 percent or even zero. A 10 percent reduction would mean that if 1,000 50-year-old women were screened over a decade, 996 women rather than 995.6 would not die from the cancer.
To the Times‘ credit, they appropriately point out the costs of mammography:
And screening has a cost, Dr. Welch [author of an editorial accompanying publication of the study] said. Screening 2,500 50-year-olds for a decade would identify 1,000 women with at least one suspicious mammogram resulting in follow-up tests. Five hundred would have biopsies. And 5 to 15 of those women would be treated for cancers that, if left alone, would have grown so slowly they would never have been noticed.
The reception of this study is considerably better than that of the Prevention Task Force report last year. The methodology of the new study is being praised by many of those who believe that mammography is still valuable. The Times points out that the study’s authors’ are not in complete agreement as to the appropriate recomendations for mammography given the evidence that its benefit is likely less than previously believed:
Marvin Zelen, a statistician at the Harvard School of Public Health and the Dana-Farber Cancer Institute, who was a member of the research team said even though the mammography benefit is small, if he were a woman he would get screened.
“It all depends on how you approach risk,” Dr. Zelen said. His approach, he says, is “minimax” — he wants to minimize the maximum risk — which, in this case, is dying of a cancer.
Dr. Kalager [study lead author] came to the opposite conclusion. She worries about the small chance of benefit in light of the larger chance of finding and treating a cancer that did not need to be treated.
“Since I’m a breast cancer surgeon, I know what being treated is like,” she says. The decision to be screened, she says, “is a matter of personal preference. Is it worth it to risk becoming a patient without it being necessary?”
Many women may still want mammograms, she says, and that is fine.
“I think we have to respect what women want to do.”
The study makes life more difficult for many physicians, especially primary care doctors. As the reported benefits of mammography decline, one hopes that these physicians will become better at describing the risks and benefits of the test and at helping women come to an informed decision whether they wish to be tested. If this study holds up, it will be less clear that mammography should be the default for all women in any age group.
In any case, the reception of the new study suggests that progress is being made in media portrayals of the complexities of real-world medical research. Further, it suggests that the costs of receiving medical tests may start to get the attention it deserves. When it comes to medical tests, it is time to stop simply assuming that more is always better.
September 27th, 2010