Correction to This Article
A May 10 Health section story about a study exploring aspirin use and breast cancer prevention incorrectly labeled hormone receptor positive cancers the most dangerous kind. That description applies to hormone receptor negative breast cancers.
Overstating Aspirin's Role In Breast Cancer Prevention
How Medical Research Was Misinterpreted to Suggest Scientists Know More Than They Do

By Lisa M. Schwartz, Steven Woloshin and H. Gilbert Welch
Special to The Washington Post
Tuesday, May 10, 2005

Medical research often becomes news. But sometimes the news is made to appear more definitive and dramatic than the research warrants. This series dissects health news to highlight some common study interpretation problems we see as physician researchers and show how the research community, medical journals and the media can do better.

Preventing breast cancer is arguably one of the most important priorities for women's health. So when the Journal of the American Medical Association published research a year ago suggesting that aspirin might lower breast cancer risk, it was understandably big news. The story received extensive coverage in top U.S. newspapers, including The Washington Post, the Wall Street Journal, the New York Times and USA Today, and the major television networks. The headlines were compelling: "Aspirin May Avert Breast Cancer" (The Post), "Aspirin Is Seen as Preventing Breast Tumors" (the Times).

In each story, the media highlighted the change in risk associated with aspirin -- noting prominently something to the effect that aspirin users had a "20 percent lower risk" compared with nonusers. The implied message in many of the stories was that women should consider taking aspirin to avoid breast cancer.

But the media message probably misled readers about both the size and certainty of the benefit of aspirin in preventing breast cancer. That's because the reporting left key questions unanswered:

· Just how big is the potential benefit of aspirin?

· Is it big enough to outweigh the known harms?

· Does aspirin really prevent breast cancer, or is there some other difference between women who take aspirin regularly and those who don't that could account for the difference in cancer rates?

This article offers a look at how the message got distorted, what the findings really signify--and some broader lessons about interpreting medical research.

How Big a Benefit?

Just how big is the potential benefit of aspirin?

The 20 percent reduction in risk certainly sounds impressive. But to really understand what this statistic means, you need to ask, "20 percent lower than what?" In other words, you need to know the chance of breast cancer for people who do not use aspirin. Unfortunately, this information did not appear in any of the media reports. While it might be tempting to fault journalists for sloppy, incomplete reporting, it is hard to blame them when the information was missing from the journal article itself.

In the study, Columbia University researchers asked approximately 3,000 women with and without breast cancer about their use of aspirin in the past. The typical woman in this study was between the ages of 55 and 64. According to the National Cancer Institute, about 20 out of 1,000 women in this age group will develop breast cancer in the next five years. Therefore, the "20 percent lower chance" would translate into a change in risk from 20 per 1,000 women to 16 per 1,000 -- or four fewer breast cancers per 1,000 women over five years.

For people who prefer to look at percentages, this translates as meaning that 2 percent develop breast cancer without aspirin, while 1.6 percent develop it with aspirin, for an absolute risk reduction of 0.4 percent over five years.

Another way to present these results would be to say that a woman's chance of being free from breast cancer over the next five years was 98.4 percent if she used aspirin and 98 percent if she did not. Seeing the actual risks leaves a very different impression than a statement like "aspirin lowers breast cancer risk by 20 percent." (See "Research Basics: How Big Is the Difference?")

Against What Size Harms?

Is the potential benefit of aspirin big enough to outweigh its known harms?

Unfortunately, aspirin, like most drugs, can have side effects. These, according to the U.S. Preventive Services Task Force, include a small risk of serious (and possibly fatal) bleeding in the stomach or intestine, or strokes from bleeding in the brain -- harms briefly noted but not quantified in the original study or in most media reports. To decide whether aspirin is worth taking, women need to know how the potential size of aspirin's benefit in reducing breast cancer compares with the drug's potential harms.

Sound medical practice dictates doing the same kind of calculation -- of potential benefits against potential harms -- anytime you consider taking a drug.

We provide the relevant information in the "Aspirin Study Facts," below. The first column shows the health outcome being considered (e.g., getting breast cancer, having a major bleeding event). The second column shows the chance of the outcome over five years for women not taking aspirin. The third column shows the corresponding chance for women taking aspirin. And the fourth column shows the difference -- the possible effect of aspirin.

As the table shows, the size of the known risk for stomach bleeding to a woman taking aspirin daily nearly matches the size of the still-hypothetical benefit in terms of breast cancer protection. That kind of comparison might lead some women to conclude that the tradeoff doesn't warrant the risk.

While it may take you some time to become familiar with this table, we think this sort of presentation would be helpful in many situations; for example, whenever people are deciding about taking a new medication or undergoing elective surgery.

Is It Really Aspirin?

Does aspirin really prevent breast cancer, or is there some other difference between women in the study that could account for the difference in cancer rates?

Can we be sure that aspirin was responsible for the "20 percent fewer" breast cancers that the Columbia researchers found among aspirin users compared with nonusers?

To understand why not, it is necessary to know some of the details about how the study was conducted.

The researchers collected information from all of the women in New York's Nassau and Suffolk counties on Long Island, who were diagnosed with breast cancer in 1996 and 1997. For comparison, they matched these women with others who did not have breast cancer, but who were about the same age and from the same counties. The researchers asked all the women about their use of aspirin.

They found that aspirin use was more common among the women without breast cancer. While the researchers were careful to report that the use of aspirin was "associated" with reduced risk of breast cancer, the media used stronger language, suggesting aspirin played a role in preventing breast tumors.

Unfortunately, this kind of study -- an observational study -- cannot prove that it was the aspirin that lowered breast cancer risk. Strictly speaking, the researchers demonstrated only that there is an association between aspirin and breast cancer.

Consider how an association between aspirin and breast cancer could exist even if aspirin has no effect on breast cancer.

It could be that women who use aspirin regularly are already at a lower risk of breast cancer. Imagine, for example, there was a gene that protected against breast cancer but also made people more susceptible to pain. Women who carried this gene would be more apt to use aspirin for pain relief. The lower breast cancer risk in aspirin users might simply reflect the fact that they had this gene. In other words, aspirin might have nothing to do with the findings. To really know if aspirin lowers breast cancer risk would require a different kind of study -- a randomized trial. (See "Research Basics: Cause or Association?")

Nonetheless, observational studies are important (and often crucial) in building the case for doing a randomized trial. In this instance, the researchers had a theory for how aspirin might prevent breast cancers. They predicted that it would only be true for certain kinds of cancers (so-called hormone receptor positive cancers, the most dangerous kind, which account for about 60 percent of all breast cancers). And that is just what they observed: The association between aspirin and breast cancer was not seen in hormone receptor negative cancers. That the researchers' prediction was correct supports (but does not prove) the idea that aspirin reduces risk. The next logical step would be a randomized trial.

The difference between "cause" and "association" may seem subtle, but it is actually profound. Even so, people -- like the headline writers in this case -- often go beyond the evidence at hand and assume that an association is causal. Readers should know that many associations do not reflect cause and effect.

The Bottom Line

In a large observational study, researchers found slightly fewer breast cancers among women who took aspirin regularly compared with women who did not. Because aspirin's benefit in reducing breast cancer (assuming it can be proven) was small, it may not outweigh the drug's known harms. While it is possible that aspirin itself reduces the risk of breast cancer, we cannot be sure from this study. It would take a randomized trial to be certain. Fortunately, one has just been completed by researchers at Harvard Medical School, and the results are expected in the very near future. Until then, it is too soon to recommend taking aspirin to prevent breast cancer. ·

Lisa Schwartz, Steven Woloshin and Gilbert Welch are physician researchers in the VA Outcomes Group in White River Junction, Vt., and faculty members at the Dartmouth Medical School. They conduct regular seminars on how to interpret medical studies. (See The views expressed do not necessarily represent the views of the Department of Veterans Affairs or the United States Government.

© 2005 The Washington Post Company