Friday, May 16, 2003

Passive smoking: report rejects cancer link

Researchers have reanalysed data produced by the American Cancer Society and concluded that no risk from environmental tobacco smoke (ETS) can be demonstrated from the figures.

What the paper illustrates is the difficulty of drawing firm conclusions from such a longitudinal study. Although the numbers of people involved are very large, there are so many different factors to be taken into account that any result may just be 'noise'. For example, it would be impossible to measure the effect of ETS on smokers, because the effect would be drowned out by the known relation between smoking and cancer. So, the focus is people who have never smoked who live with a partner that does smoke. The problem is that there is no accurate measure of how much smoke someone consumes, there can only be an estimate. How do you compare these estimates both with each other and with reality? Then there is the problem of whether the 'never smokers' really do abstain completely. Is it really possible to find people who have no exposure to smoke? Thus, it is difficult to know just how much exposure there really is to ETS.

Then, there is the problem of working out a rate of cancer that takes into account class, diet, age, sex etc. Many of these factors are themselves subject to discussion about the level of risk associated with them. So, it is difficult to produce a definitive comparison.

Therefore, a longitudinal study can only be of much use if the result indicates a strong association. Hence, reported risk factors of less than 2.0 are usually ignored. Yet, the ETS debate has been driven by studies showing much weaker relationships, made statistically significant only by combining different studies together, a technique itself riddled with problems.

The only safe thing to say about ETS and ill-health is that a relationship has not been established, and if it is, it is likely to be weak. Which, given that non-smokers in smoky environments inhale only a fraction of the smoke that the smokers themselves inhale, seems to make sense.

Row over passive smoking effect, BBC News, 16 May 2003


Post a Comment

<< Home