Tuesday, April 20, 2010

Winn’s Legacy: The Fifty Fabrication


In a previous post I explained how Deborah Winn, an epidemiologist at the National Cancer Institute, misinformed the American public and America’s doctors about the oral cancer risks of smokeless tobacco use.
Winn is also responsible for the widespread statement that smokeless tobacco users are 50 times more likely to develop oral cancer. That statement is a complete fabrication.

Winn’s 1981 study in the New England Journal of Medicine is the source for this completely erroneous information. As I discussed previously, Winn reported that the relative risk for oral cancer among powdered dry snuff users was 4. The puzzle is how a relative risk of 4 has been transformed into 50. Here are the keys to the puzzle’s solution.

First, oral cancer is a very rare disease, but smokers and heavy drinkers have considerably elevated risks. The most common sites for oral cancer are the sides of the tongue and underneath the tongue; other surfaces of the mouth are hardly ever affected. There are rare exceptions: oral cancer infrequently occurs among dry snuff users in the location where they place the product -- the gum and cheek. Winn’s 50 number applies only to the gum and cheek, not to the entire oral cavity (for which the number is still only 4). Even a few cases of cancer in the gum and cheek, which are virtually never otherwise involved, can raise the risk calculation.

The second key doesn’t involve Winn’s main findings, but a secondary analysis she performed on the duration of dry snuff use. Winn found that the highest risk for cancer of the gum and cheek was among women who had used dry snuff for over 50 years. That risk number was 47.5, and it was based on only 15 cases (the small number of cases determined that the 95% confidence interval – the range within which the estimate lies – was very wide, from 9 to 250!). Winn rounded up her estimate to “50-fold” in the abstract.

It is important to point out that the 50 and the 4 refer to the same number of oral cancer cases. Here is an example of how these numbers relate to each other. Let’s say that you have $100 in a savings account, and the average savings account in that bank has $25. This means that you have 4 times as much money as the average customer at that bank. But you decide to move your account to another bank, where the average savings account only has $2. Now you can correctly boast that you have 50 times more than the average account holder, even though you still only have $100.

So, Winn’s “50” number (actually 48) is appropriately used only in the following specific description: Women who used powdered dry snuff for over 50 years were 48 times more likely than nonusers to develop a cancer of the gum or cheek where the snuff was placed.

That is not how anti-tobacco extremists use the “50” number. They are eager to portray smokeless tobacco use as life-threatening, so they say that users of any type of smokeless tobacco for any duration have a fifty-fold elevated risk of oral cancer. This is a total canard.

The “50” fabrication continues to be peddled by organizations and government agencies that know better. Here are some examples:

The American Cancer Society (here)

The Centers for Disease Control and Prevention (here)

The University of Iowa (here) and Brown University (here)

State agencies in Kentucky (here), Rhode Island (here) and Wyoming (here).

It appears that Winn’s “50” legacy will persist until Americans demand that organizations and government agencies provide accurate information about the risks of tobacco.

No comments: