Breast cancer screening: conflicting evidence? what are the important questions for health?

.
Several new, and sometimes contradictory, studies about breast cancer and screening mammography have recently appeared in the medical literature. In a recent study from Norway, published in the New England Journal of Medicine, Kalager et. al[1] looked at the impact of mammography screening of women 50-69 years old in that country by comparing groups of women who were screened and those who were not screened between 1996 and 2005 and comparing them both with comparable groups for the 1986-1995 period. They discovered a significant decrease in the rate of breast cancer detected in women who were screened – but also in women who were not screened. The decrease was greater in the women who were screened, but this only accounted for 1/3 of the reduction in breast cancer mortality in that group; while the reductions in each group were statistically significant, the difference between them was not (quite). Overall the reduction in the death rate in screened women was 7.2 per 100,000 woman-years, and in the non-screened group 4.8 per 100,000 woman-years.

Meanwhile, Cancer published an article from Sweden looking at mammography screening for women who were 40-49 years of age. Because not all Swedish counties screened women that age, they were able to do a retrospective comparison between those counties who began screening at 40 compared to those counties who began at 50. Hellquist, et. al.,[2] found a significant risk reduction (a relative risk of 0.74 for women invited to screening and of 0.71 for those actually screened). Over 16 years of the study “…there were 803 breast cancer deaths in the study group (7.3 million person-years) and 1238 breast cancer deaths in the control group (8.8 million person-years)”, or converting to the same units as the other study, a reduction of 3 deaths per 100,000 woman years, from 14 to 11. The authors note that 1252 women this age would need to be screened to prevent one death.

The Swedish study provides data to support screening women at 40 (although the reduction was greatest in the 45-49 group, which is not surprising as the incidence of breast cancer increases with age). The Norwegian study shows that the rate of breast cancer has decreased anyway. The Swedish study is not a randomized controlled trial, but rather solely a historical comparison, and thus did not control for differences between the women who got screening and those who did not that might have affected breast cancer risk.

So which are we to believe? Both, to the extent that we can understand the methods used. The question is what are we to do? “We”, here, being both women who might choose to undergo mammography screening, and doctors who have to decide whether to recommend it, especially to those under the age of 50. For me, it will involve discussions with each woman, dependent on her individual risks and preferences. The population health benefit is small but significant; while for any woman the risk reduction might be of value regardless of the cost, that woman (and everyone else) has to also bear the cost of screening all those women – 1252 to save the life of one woman 40-49 in the Swedish study. Clearly, women who have a higher risk (family history) will benefit more, since they are more likely to have breast cancer than those without such a family history.

The cost is not only all those mammograms, but the cost of further tests (additional views or ultrasound) in women with questionable results (more common in younger women who on average have denser breasts), in biopsies – not without risk, though low – and pathology reports, discomfort etc. Not to mention what that money might otherwise be spent on. Feeding the homeless? Providing prenatal care? Fighting HIV/AIDS? Building more bombs? Bailing out big financiers? I would certainly agree that there are social programs that could more effectively use the money currently being spent to screen low-risk younger women for breast cancer with far more social benefit; I would also acknowledge the very low likelihood of any saving in mammography screening being used for such purposes rather than other high-tech medical care or further health system profit.

A more important question, perhaps, is why, at any age level, are certain parts of the population – poorer and minority women – less likely to be screened? To the extent that screening does detect cancer in earlier stages, when intervention can be done, and thus make a difference in mortality, why is it not being done as effectively with minority women? For example, African American women are less likely to get breast cancer, but more likely to die from it.

But people do want “answers”, definitives. Newspapers like to publish “breakthrough” research that seems to provide them. When the “answers” seem contradictory, it leads to frustration for many, and a decision to do what they had planned to do anyway, to do what seems to make “sense” (and what they can afford). After all, if sometimes we hear that Vitamin C or E or beta-carotene is a wonder drug to prevent cancer, or Alzheimer’s, and later studies show that it isn’t, and sometimes may even create greater risk, what are we to do?

One thing we could choose to do is to listen to those who are sure that they have the answers, even when these are based upon no data or carefully-selected data. If someone advises people based upon a single study, regardless of the scientific quality of that study and how often its results might have been refuted by later studies, they can be more confident and sure of themselves. This can inspire confidence on those they advise. (I am reminded of the story told by a colleague, fresh out of residency, who was practicing with farmworkers in the Imperial Valley of California. These folks, he said, are used to going to a doctor who listens to their problem, nods, tells them what they have, and gives them a shot to fix it. “I,” he noted, “listen, then tell them to take off all their clothes, examine them for a long time, and then finally tell them I’m not sure what it is, but it probably doesn’t need a shot – and that the shot might even be bad for them – and likely it will go away with time. Who are they going to trust?”). Showing confidence in your opinion is very persuasive. Thus the term “confidence men” (now shorted to “con”).

By the way, I don’t think this is such a good choice. That studies may seem contradictory is part of the nature of science, of how the studies were conducted, on whom, with what controls, using what methods. A future blog will discuss the work of Dr. John Ioannidis, whose work has raised questions about how much of the work published as research is true, made recently au courant by David H. Freedman’s article in the November Atlantic “Lies, damned lies, and medical science,” and some of the important lessons to be learned – and not learned – from his work. Nonetheless, this confusion does not justify complete nihilism – do whatever you want and it’s all the same – or much less undertaking treatments that have never really been studied. (See Drugs, Tobacco, Doctors and the Health of the Public, September 10, 2010.)

In the meantime, remember that there are probably not too many magic shortcuts to better health. A healthful diet and more exercise will make you fitter and stronger and able to do more things; reading and thinking and doing puzzles are interesting, stimulating and fun, whether they prevent Alzheimer’s disease or not.

[1] Mette Kalager, M.D., Marvin Zelen, Ph.D., Frøydis Langmark, M.D., and Hans-Olov Adami, M.D., Ph.D, Effect of Screening Mammography on Breast-Cancer Mortality in Norway, N Engl J Med 2010; 363:1203-1210.

[2] Hellquist BN, et. al., Effectiveness of population-based service screening with mammography for women ages 40 to 49 years, Cancer published online Sept 29 ,2010
.

Breast Cancer Screening and Evidence-based Medicine

.

In case you’ve been in a coma for a while, the US Preventive Services Task Force (USPSTF), a federally-funded-but-independent group of scientists who evaluates the evidence regarding preventive care, has announced new guidelines on screening for breast cancer. It has generated an amazing amount of comment, from physicians, patients, advocacy groups, politicians, and journalists. Every possible position on the issue, from thoughtful and balanced to alarmist and opportunist, has been taken and published by someone. For a quick review of articles just in the New York Times in the last few days we have:

November 17:
“Panel Urges Mammograms at 50, Not 40”
November 18:
“New Mammogram Advice Finds a Skeptical Audience”,
“Many Doctors to Stay Course on Breast Exams for Now”
November 19:
“Screening Policy Won’t Change, U.S. Officials Say“
Columnist Gail Collins: “Breast Brouhaha”
November 20:
Kevin Sack, News Analysis: “Medical Science and Practice in Conflict”
“Mammogram Debate Took Group by Surprise”

Therefore, the wise course might be for me to stop here, let you read everyone else, and not get involved. Of course, I won’t. Let me start by discussing the discovery, use, and application of evidence in medicine, and in particular with the USPSTF.

Disclaimer: I don’t work for, or have any relationship, financial or otherwise, with USPSTF, but I do believe that the responsible practice of medicine requires keeping up with the evidence and changing practices as new information becomes known; it should not be a “faith-based” effort.

Medical evidence for anything, including appropriate preventive services, gathers slowly. Studies are first done on high-risk populations, then later on average or low risk. Depending on the variables looked at, and the population studied, different information can emerge. Rarely (but sometimes) is the data from one good study on the same population directly opposite that of previous studies; more likely it will be similar, but might be of a greater or lesser degree of magnitude. Or just different enough to tip the risk/benefit balance. Because virtually never is anything – a treatment, a diagnostic test, a preventive activity – all good or all bad. There are benefits, real or potential, and risks, of varying degree. As new evidence accumulates, it tends to move the scales, or the seesaw, more down or up on one side or the other. Usually not enough to drop one end to the ground, but sometimes enough to tip the balance. And new studies are being done all the time, and it is not only hard to keep up, it is hard to assess the changes in risk. But it must be assessed, because it would be wrong to just keep doing what you were doing when the evidence changes.

That is the incredibly valuable service that the USPSTF has been providing since it was first convened in 1984. Evaluating, the existing studies and making recommendations – to clinicians – on what they should discuss with their patients. They are not really for patients, although this seems to be where much of the confusion is. Much of the coverage is about individual women trying to decide what to do – or reacting “against” the recommendations. The Times article cited above, “Many Doctors to Stay Course on Breast Exams for Now”, includes this: “Patients are already trying to figure out what the recommendations mean.” It means they should discuss them with their health care provider.

USPSTF assigns both a grade and a level of certainty to its recommendations. The grades are A, B, C, D, and I, and the levels of certainty are high, moderate, and low, and are described at the USPSTF website, at http://www.ahrq.gov/clinic/uspstf/grades.htm. When the grade is A or B, the procedure is recommended to clinicians in practice; when the grade is C it is generally not recommended, but there may be individuals or situations in which the benefit would exceed the risk; when it is D, it is not recommended. A grade of I means there is insufficient evidence to assess whether there is net benefit.

USPSTF is not the only group that makes recommendations. Many medical professional organizations and advocacy groups (such as the American Cancer Society and the American Heart Association) also make recommendations. However, the USPSTF is independent and has no “dog in the fight”, no financial or emotional attachment to an outcome. For example, after these recent recommendations were announced, suggesting most women start receiving mammograms later, and have them less often, the American College of Radiology (ACR) announced its disagreement with them. I’m sorry, but the fact that radiologists have an obvious financial stake in doing more mammograms has to make their opinion more suspect.

Certainly doing fewer mammograms will save money for insurers (including the government, for Medicare patients). In the current climate of our debate on health reform, some have seen these recommendations as an effort by these insurers to save money, and others have noted that, because of this, the timing of the announcement was “unfortunate”. However, unlike the ACR, the members of the USPSTF have no financial stake in their truly independent recommendations. I, for one, absolutely consider them to be the most valid source of independent analysis and advice.

What about these specific recommendations? They make sense to me, and are supported by the evidence. The recommendation that mammography begin at 40, rather than 50, only dates back to 2002, and was controversial at that time.
“In 2002…,” reports the Times in Panel Urges Mammograms at 50, Not 40, “When the group recommended mammograms for women in their 40s, some charged the report was politically motivated. But Dr. Alfred Berg of the University of Washington, who was the task force chairman at the time, said ‘there was absolutely zero political influence on what the task force did.’ It was still a tough call to make, Dr. Berg said, adding that ‘we pointed out that the benefit will be quite small.’ In fact, he added, even though mammograms are of greater benefit to older women, they still prevent only a small fraction of breast cancer deaths.” In the last 7 years more studies have come out, which have weakened the relative benefit to risk ratio for women between 40 and 50 who are at average risk for breast cancer.

In addition, the body of evidence does not suggest that there is significant additional benefit to screening every year rather than every two years. It also recommends against teaching self-breast examination (not against women doing it) – a “D” recommendation -- because there is good evidence from large population studies that it offers no advantage death or morbidity from breast cancer. Continuing screening of women over the age of 74, and doing clinical breast examination (by a physician or other clinician) in addition to mammography get “I” – insufficient evidence recommendations.

In the news analysis cited above, “Medical Science and Practice in Conflict”
“Mammogram Debate Took Group by Surprise”, Kevin Sack notes that “The backers of science-driven medicine, with its dual focus on risks and benefits, have cheered the elevation of data in the setting of standards. But many patients — and organizations of doctors and disease specialists — find themselves unready to accept the counterintuitive notion that more testing can be bad for your health.” But it certainly can be.

More, it should be obvious, is not always better, not only with fatty foods and salt, but with screening. The risks of unnecessary (read: too often for the risk level of the woman under consideration) mammograms is more than anxiety for false positives. It is also biopsy for false positives. And sometimes complications. And radiation exposure, which is not insignificant, and is, as we know, linked to causing cancer. Think: men get breast cancer also, but not at the rate that women do (about 1%). We do not screen men, because the risk/benefit ratio is way over to risk. Let me make clear that all of this discussion is about screening; by definition, someone who has NO symptoms. No lump, no discharge, no skin dimpling. They are not about diagnostic mammograms – examining someone with symptoms or physical findings, or a previous abnormal mammogram, and certainly do not apply to follow-up of people who have had breast cancer.

In “Many Doctors to Stay Course on Breast Exams for Now”, Dr. Annekathryn Goodman, director of the fellowship program in gynecological oncology at Massachusetts General Hospital, says “It’s kind of hard to suggest that we should stop examining our patients and screening them….I would be cautious about changing a practice that seems to work.” I hope that is not what she meant. What “seems to work”, while intuitively attractive, is not always correct. That is why we have independent bodies such as USPSTF continually examining the evidence.
.