Friday, March 26, 2010

Breast cancer, diet, and exercise...I smell a rat

"Up to a third of breast cancer cases in Western countries could be avoided if women ate less and exercised more..." -- Associated Press

When stories like this show up on the front page of the paper, they always get my attention. I'm naturally a skeptic and am always suspicious when technical subjects get sensationalized. Normally, when I read something like this, I suspect statistical douchery. For example, suppose you search a database and find that the breast cancer mortality rate for women who don't exercise is 27 deaths per 100,000 and the mortality rate for women who exercise is 18 deaths per 100,000. In the world of cancer research where the competition for funding is fierce, you may rush a paper into publication that states that exercising can reduce breast cancer deaths by a third. However, suppose you do a little additional searching of the data base and sort the exercising and non-exercising groups by household income where you find that the death rate for breast cancer in both groups is much higher for low income households than high income households. You also find that, since high income households have more disposable income and more leisure time than lower incomes, a much greater percentage of exercisers belong to high income households. Consequently, your original exercise correlation was merely telling you that poor women have a higher breast cancer death rate than women who can afford more frequent doctor visits, regular screening, etc.

Purely statistical correlations such as the one I've hypothetically described above are entirely useless when it comes to reducing breast cancer rates. However, when they are used to identify mechanisms by which cancer cell growth is triggered, they can be a valuable tool. To continue the above example, suppose you used the exercising correlation to look into the hormone levels of exercisers and non-exercisers and found that the exercisers had lower levels of certain hormones that have been associated with cancer occurrences. This is useful information.

So...with this in mind, I went on the Internet to track down the source of the "one-third reduction" number. The report ended up being based on a talk given by Carlo La Vecchia at a conference in Barcelona. The source of La Vecchia's figures was cited to be a study by the International Agency for Research on Cancer. If you go to the IARC website, you'll find that the figures come from a report (Volume 6) that was written in 2002 and is now out-of-print. Hmmm...data from an 8 year old report suddenly makes front page news. What the hell? First, understand that in any technical field, 8 years ago was ancient history. I am supposed to believe that a supposedly critical connection between diet and exercise and breast cancer was announced 8 years ago and that there have been no follow-up studies that have made more recent or updated data available? I think not. So what is going on here? Here are my best guesses:

1. Carlo La Vecchia wanted a boondoggle to the Barcelona conference, so he submitted an abstract for essentially a throw away talk using old data. Consequently, he gets the University of Milan to pay for his trip. Meanwhile, Maria Cheng, the AP reporter, is at the conference looking for something, anything, she can write a story on. La Vecchia mentions his 25 to 30 percent reduction in breast cancer number, Cheng grabs that, rounds it up to "one-third", makes a few calls to some breast cancer workers she knows who give her some broad general comments, and she's got an easy byline.

2. The story has an interesting quote by La Vecchia, "What can be achieved with screening has been achieved. We can't do much more. It's time to move on to other things." By "other things", the professor is obviously talking about preventing breast cancer through diet and exercise, which will, of course, require tens of millions of dollars in new funding to study...and as he has just established himself as the lead expert in this subject -- if only by dredging up some work probably done 10 years ago -- La Vecchia looks to be leader in the clubhouse as the recipient of a good chunk of that money.

3. A potentially darker interpretation of the timing of this story, since it is coming out of the Associated Press, who can't issue any story without injecting their political agenda into it, is that it is intimately connected to the recent health care debate. Anyone with a brain recognizes that the health care bill that just passed is going to result in higher taxes and reduced services -- if you don't believe this, please email and tell me about the rest of your life on Planet Pollyanna...I'll bet it's nice there. Taxes will go up, but tax increases are never very palatable to the American public, even during prosperous times. Any congress person that votes for a significant tax increase these days might as well go home and campaign for their successor. Reducing services is much easier to slip by the public, since they won't know they're missing them until they need them. Herein lies a possible subtext in the AP's story...if you're a fat woman, who drinks, doesn't exercise, and is diagnosed with breast cancer, then "I'm sorry, ma'am...you've brought this on yourself. We need to save our very expensive treatment for people who are more deserving. Why don't you go take a walk and quit eating Big Mac's?" The AP article mentions drinking (bad), exercise (good), and obesity (doubly bad)...I'm surprised they didn't go for the trifecta and throw in smoking.

1 comment: