Chalmers and Altman didn't; but Eysenck did!
In his comments on the paper by Richard Doll and his colleagues, Dr
Calinas-Correia suggests that Chalmers and Altman analysed the work of
Janerich et al about passive smoking and lung cancer. We didn't, but
Eysenck (1) did in a book that Doug Altman and I co-edited.
Hans Eysenck, like Ronald Fisher before him, promoted the notion that
the association between smoking tobacco and lung cancer is not causal, but
reflects a predisposition to both in people with certain genetic or
psychological characteristics. As Doll and Peto note in their response
(2) to Calinas-Correia, however, selective citation of evidence (whether
inadvertent or calculated) can be misleading.
The main message of the book that Calinas-Correia cites incorrectly
is that people reviewing research evidence should take steps to reduce
bias, and when possible, the effects of the play of chance. I hope that
the forthcoming 2nd edition (3) will succeed in communicating this message
more effectively than the 1st edition appears to have done!
1. Eysenck HJ. Problems with meta-analysis. In: Chalmers I, Altman
DG, eds. Systematic reviews. London: BMJ Books, 1995.
2. Doll R, Peto R. Smoking and dementia in male British doctors. BMJ
3. Egger M, Davey Smith G, Altman DG (eds). Systematic reviews in
health care: meta-analysis in context. 2nd edition, BMJ Books, 2000.
Competing interests: No competing interests