Wine’s link to breast cancer: Keep it in context
‘‘ Raising a risk that starts off small can result in a risk that’s higher but still very small in absolute terms.’’ Cornelia Dean, New York Times science writer
Many women obsess over their risk for breast cancer. That’s why news media flash blazing headlines whenever a study suggests a link between breast cancer and some “exposure.”
A new report cites evidence that drinking even one small glass of wine a day heightens a woman’s risk for breast cancer. Released by the renowned Fred Hutchinson Cancer Research Center in Seattle, it set off the usual alarms. It certainly got my attention while raising some questions.
The intention here is not to criticize the study’s validity. Rather, it is to interpret the findings in a way that’s useful to one who cares to protect her health but also loves a glass or two with dinner.
Helping in this process is a fascinating new book, “Making Sense of Science: Separating Substance from Spin,” by New York Times science writer Cornelia Dean. Its mission is to help non- scientists evaluate scientific claims, with much attention paid to studies related to health.
“People who want to scare us typically give us relative risk figures,” Dean writes. Raising a risk that starts off small can result in a risk that’s higher but still very small in absolute terms. This is best shown in an example. “Something that raises your risk of heart attack by 30 percent [ relative risk],” she explains, “could be no scarier than raising your odds from 6 in 1,000 to 8 in 1,000 [ absolute risk].”
When we hear that a small glass of wine a day may raise a premenopausal woman’s breast cancer risk by 5 percent and a postmenopausal woman’s by 9 percent, remember, those are relative risks. You want to know the absolute risk.
So what do we know about women and drinking? I asked Dean.
“We know drinking modestly raises the risk of breast cancer,” she said. But “it is widely believed that moderate drinking is associated with better heart health and longer life expectancy. And we know that almost half of American women die of heart/ vascular causes, whereas about 3 percent die of breast cancer.”
Another interesting point: “When an epidemiologist gets interested in an exposure- risk relationship, it’s typically when the exposure doubles the risk. Here we are talking about increases of about 8 percent or 9 percent. None of the wine studies gets near that.”
For contrast, consider the studies linking tobacco to cancer. Among smokers, the risk of lung cancer is about 900 percent higher.
A public that poorly understands risk also falls short on evaluating statistics. Some years ago, reports of a cluster of breast cancer cases on Long Island grabbed the headlines. The alarms were loud. They were also false.
When a cancer ( or autism, for that matter) is distributed randomly over a large population, there will inevitably be clusters of cases. This is a product of statistics, not evidence of something bad in the water.
“What Long Island has in relative abundance,” Dean concluded, “is not breast cancer but rather affluent women with access to good insurance coverage.”
So the problem with today’s unnecessarily scary reports is usually not the scientists suggesting elevated risks or the statisticians crunching the numbers. It’s how these risks and numbers are reported. It’s the job of journalists to interpret these findings for a general audience.
Non- scientists, meanwhile, would do well to arm themselves with some basic understanding of what studies and reports are really saying. As for me, I shall continue drinking wine — moderately, of course.