Epidemiological Certainty

The following post is drafted by Gasdoc from an idea supplied by a new member of the blogging team called "PearlyQueen".

This is the way to design a study so that you can tell if anyone died due to smoking:

To start you will need to clone a man and woman. You will only need two pairs of subjects, but it is essential that they have identical genes and identical lives. From the test tube, these two couples must live together and everything they do must be identical. They must eat and drink the same stuff and follow the same paths of education and vocation. They must live in the same house and go absolutely everywhere together. There will always be a flaw in the work as they will not be able to sleep in the same bed. It is probably better that they do not smoke but if one was to do so then they must all smoke exactly the same amount and make of tobacco. They should of course all do this at exactly the same time and place.

The next stage of the study will occur after both couples have fallen in love, married, coincidently conceived and given birth to identical same sex infants. It would be best to plan for these deliveries to be by caesarian section at the same age of gestation and by non-smoking, equally qualified obstetricians with the same training or at least adhering to the same written protocol, which must detail the anaesthetic and surgical technique.

At this stage it will be necessary to decide whether you wish to examine the effects of smoking or passive smoking.

So, if the keen epidemiologist would like to contact me, I will further elaborate on the details of the next stage of the study.