Wednesday, July 11, 2007

Parkinsons & Smoking

passive smoking, smoking bansThe Times yesterday demonstrated a fairly high standard of journalism yesterday when reporting on the latest puplication into the association between Parkinson's Disease and smoking. Using words like "may reduce risk" and "research suggests", Joanna Carpenter managed to present the facts rather than the hysteria.


However, the fact that it was reported is the issue. As usual the publication was an epidemiological study but not only that it was a study that simply pooled the available past data. By doing this the authors managed to satisfy the God of statistics and produced their Nirvana, statistical significance. Of course it still has no real significance and certainly doesn't mean that smoking reduces your chances of developing Parkinson's Disease.

The statistically significant risk ratio produced was 0.13 to 0.32. This was of course referred to as 13% to 32% as these are bigger numbers and no-one understands decimals. Hang on, does anyone underststand percentages. However, the God of epidemiology is not satisfied as such small relative risks are not supposedly reliable. Nevermind, so long as the God, Statistics is happy!

This report and the academic publication is an example of the level of science reporting we are exposed to today. It has an exact parallel with "Passive Smoking" reports and puplications as they too glorify the God of Statistics and fail to prove anything at all at exactly the same spurious or irrelevant relative risk ratios.
Put another way...
If the anti-smoking lobby is right about the passive smoking risk studies then I am definitely 32% less likely to develop Parkinson's disease! Hurray!