Reader Comments

Post a new comment on this article

Referring to Bradford Hills criteria of causality

Posted by Faggiano on 30 Aug 2011 at 15:31 GMT

I agree with the authors that the crucial point is the risk of confounding of the relationship between smoking scenes and risk of smoking onset. But there are at least two considerations lacking in their reasoning.

The first is about causality. Any “natural” association must be studied trough observational studies, and thus it is always not possible to exclude the role of known or unknown confounders. Bradford Hill suggested, many years ago, several criteria to be used to sustain, or exclude, the causality of a risk factor (see I believe that the authors should admit that these criteria cannot exclude the causality of the smoking scenes in the onset of tobacco smoking.

The second is about opportunity. As for drug use, violence, and sex scenes, compared to the suspicion of a role of smoking scenes in the onset of smoking, I don’t see major reasons to not restrict the use of smoking scenes to adult rated movies.

No competing interests declared.