Last week, the British Medical Journal published two more studies on secondhand smoke.
In the first study, 37 published epidemiologic studies were pooled in a meta-analysis to come up with a reported increase in lung cancer incidence of 24 percent (RR=1.24, 95 percent confidence interval 1.13 to 1.36). [Hackshaw AK, Law MR, Wald NJ 1997; BMJ;315:973].
But just for a moment, we'll forget that the reported relative risk is very weak--i.e., the National Cancer Institute says increases in risk of less than 100 percent are too small to be relied on. We'll also forget that none of the studies had any reliable data on exposure to secondhand smoke--so who knows what any observed increase in lung cancer should really be attributed to.
Of the 37 studies, only 7--less than 19 percent--reported statistically significant increases in lung cancer incidence.
Exactly why should anyone believe a result supported by fewer than 1 in 5 studies?
In the second study, 19 published epidemiologic studies were pooled in a meta-analysis to come up with a reported increase in heart disease incidence of 30 percent (RR=1.30, 95 percent confidence interval 1.22 to 1.38). [Law MR, Morris JK, and Wald NJ 1997; BMJ;315:980].
Once again, forget that the reported relative risk is very weak and is not based on reliable exposure data.
Of the 19 studies, only 8--slightly more than 42 percent--reported statistically significant increases in heart disease incidence.
What is so compelling about a result supported by fewer than 1 in 2 studies?
The title of the accompanying BMJ editorial was very apt "Passive Smoking: History Repeats Itself." I have to admit this time they are right.
Meta-analysis of the secondhand smoke studies was a joke when EPA did it in 1993. And it remains a joke today.
Material presented on this home page constitutes opinion of the author.
Copyright © 1997 Steven J. Milloy. All rights reserved. Site developed and hosted by WestLake Solutions, Inc.