Wednesday, February 28, 2007

Breast cancer drug trial results appear biased

In research and statistics classes, one idea is often repeated—be skeptical of statistics. It is not meant to degrade the value of statistical information, but more as a guide to pursue the purpose of the provided statistical information. The professors want students to evaluate the source of the research, usually by examining who provided the funding for the research. To be objective, research should be free of bias.

Unfortunately, bias seems to be the case being reflected in a world where advances and growing awareness are regularly reported—breast cancer research. In a Mcclatchy Newspapers article, it is noted that
Funding seems to affect breast cancer drug trial results.

According to the article:
“Some 84 percent of company-supported drug studies published in 10 major medical journals in 2003 reported positive results about the breast cancer drugs they investigated, according to an analysis by Dr. Jeffrey Peppercorn, a cancer physician and researcher at University of North Carolina-Chapel Hill's School of Medicine, and colleagues at Harvard's Dana-Farber Cancer Institute.”

“Non-industry supported studies were far less likely to be upbeat, publishing favorable results just 54 percent of the time.”

Compare and digest those numbers: 84% favorable when it is company funded research versus 54% favorable when studies were independent.

It doesn’t take a statistics major to figure that this equals a 30% differential when the drug companies are not funding the research.

If the statistical research being submitted is as skewed as this article reports, then This is truly sad. Breast cancer is a dangerous killer and honest research needs to be conducted to maintain a vigilant battle to defeat this beast.

We are taught that statistics are just numbers. We are also taught that just because the difference exists, this does not infer any causal relationship. However, recognizing this difference does make my eyebrows raise, oh, about 100% of the time.

No comments: