Friday, October 28, 2005

Puffed up arguments

The UK government has announced a new health bill to introduce a ban on smoking in most public places in England. Only bars that do not serve food and private members clubs will be excluded. The bill also gives the Welsh Assembly the power to impose a total ban in the principality. Meanwhile, plans are in place for a complete ban on public smoking in Scotland and Northern Ireland. Health secretary Patricia Hewitt said: 'This package is a huge step forward for public health and will help reduce deaths from cancer, heart disease and other smoking related diseases.' According to the Scientific Committee on Tobacco and Health (SCOTH), exposure to secondhand smoke raised the risk of lung cancer by 24 per cent and heart disease by about 25 per cent in non-smokers.

While a link between consistent exposure to other people's smoke and disease does not seem implausible, the evidence for it is remarkably weak - far weaker than is ever acknowledged in discussion of this new bill.

The most recent SCOTH report notes that in 'most studies considered individually the observed odds ratios failed to reach statistical significance'. In other words, any relationship between secondhand smoke and disease was so weak that it was impossible to tell whether the result was simply down to chance. The SCOTH figures were produced by tacking together all these studies, with all their potential confounding differences of subjects and methods. This seems to be a rather dubious approach, perhaps designed to produce the desired answer rather than the true one.

Figures for death tolls from secondhand smoking are not arrived at from examining the cause of death in individual patients. Instead, they simply take these questionable figures and extrapolate them to the whole population. Is the population of bar staff, who often work part-time or move on to other careers after a year or two, really comparable with that of the non-smoking wives in these studies who are exposed to cigarette smoke in confined and badly ventilated spaces every day for decades?

A 24 per cent increase in lung cancer rates for non-smokers sounds like a lot - but this increase in relative risk has to be set against the rarity of the disease in this group. The absolute risk of lung cancer for non-smokers exposed to secondhand smoke remains tiny. As for the risk from heart disease, SCOTH does not say at what age this increased risk applies. If it is at a young age, the absolute risk is again small. If it is in old age, when heart disease is commonplace, it seems implausible.

There are good reasons why transport companies and work places prohibit smoking that usually have nothing to do with fears about health. Few complain about being asked not to smoke on public transport because non-smokers have no choice but to use buses, trains and planes. But pubs are quite different, and for the government to intervene like this, it needs to demonstrate that substantial harm can be avoided. It has failed to do so.

New health bill will deliver commitment to ban smoking in majority of workplaces, UK Department of Health, 27 October 2005

Scientific Committee on Smoking and Health (SCOTH)