The Los Angeles Times' headline was "Tomatoes fight off cancer." The New York Times' read "Study says tomatoes cut risk of cancer." The Washington Post touted "Study says tomatoes fight off cancer; Finding confirms scores of reports."
The only thing the "finding" confirmed is that the media won't question anything claimed by the public health establishment.
Harvard researcher Edward Giovanucci claims in the Journal of the National Cancer Institute (Feb. 16) that 57 of 72 studies report consumption of tomatoes and tomato-based products reduces the risk of cancer.
Giovannucci further claims that 35 of the studies reported statistically significant results -- that is if you define statistical significance the way Giovannucci has.
According to Giovannucci, a result was statistically significant if the relative risk had a p-value of 0.05 or less.
But it is elemetary that statistical significance requires an appropriate p-value and a confidence interval that does not include the null result (a relative risk of 1.0).
If we look at the 72 studies with the correct definition of statistical significance, the results are much less newsworthy.
For example:
- For lung cancer, only 2 of 16 studies reported statistically significant results.
- For stomach cancer, only 1 of 14 studies reported statistically significant results.
- For digestive tract cancer (not including stomach cancer), only 4 of 15 studies reported statistically significant results.
- For genitourinary tract cancer, only 4 of 14 studies reported statistically significant results.
- For female reproductive tract cancer, only 3 of 14 studies clearly reported statistically significant results.
At best, Giovannucci can only claim that tomatoes and tomato-based products don't appear to increase cancer risk. But that doesn't make for very exciting news.
Is the media too gullible or too eager for headlines?
Comments on this posting?
Click here to post a public comment on the Trash Talk Bulletin Board.
Click here to send a private comment to the Junkman.