Smoking causes what now?

Well we're used to the Smoking Kills, apparently we're now going to get Smoking causes fatal lung cancer. Now I've kind of let the first one go, after all everyone who smokes dies, but then again everyone who drinks dies, everyone who lives dies - Life Kills.

So why the bee in my bonnet? Because I have a problem with "causes". To demonstrate here's another little statistic - the bald men have more testosterone in their system then others. So high levels of testosterone causes baldness? Well possibly, first you have to look at the opposite - baldness causes high testosterone levels? Shave someone's head and measure their levels - no increase so nope.

Maybe it's not the high levels maybe it's simply linked to something else? Again testable, percentage of bald men with high levels vs percentage of bald men without high levels, then percentage of bald men with high levels vs percentage of hirsute men with high levels.

You can see it's complicated.

So let's go back to the "Smoking causes fatal lung cancer" while looking at the tests we did for baldness.

This is difficult, you can look at the ASH website or the Cancer Research UK site for hard statistics and find nothing except links to other publications which are normally only abstracts (if online at all). There are some figures though so let's have a look.

For 2002 there were 37,700 new cases of lung cancer. Only 1 in 100 are people under 40, the majority are over 60. That's interesting if smoking does cause fatal lung cancer it seems to be a delayed response possibly cumulative. So if our hypothesis is correct we'd expect the number of lung cancer sufferers to be smokers. Well Cancer Research states

Around 90% of lung cancer cases in the UK are caused by tobacco smoking

But that's cart before horse and will you look at that I can't find a single statistic on their site that explicitly shows the percentage of smokers with lung cancer.

So here's another hypothesis - the risk of lung cancer simply increases with age. If those 1 in 100 under 40's are all smokers (or the majority) then we might have a causal link, otherwise we need the numbers for the 90 in 100 broken down for smokers and non-smokers; nope not going to get that. Okay let's assume that the majority are smokers, does that give us a link? Well no, now we have to do a comparison between overall numbers of smokers and non-smokers of that age. Well the chart links to a 404 so I'm having to estimate.

Looking at 40 year olds and assuming they started at age 18 then around 1980 around 40% of the adult population was smoking. Looking at 60 year olds around 1970 we get about 50%. So if smoking causes lung cancer and assuming that the same percentage of 60 year olds to 40 year olds started at age 18 and continued smoking up to 2002 then we'd expect about a 10 % difference between the numbers of lung cancer sufferers of those ages. (yes lots of assumptions I'm afraid)

Per 10,000 for 40 year olds around 6.6; 60 year olds - about 138. Ouch that's a big difference.

But what about that 1950's report where they compared 1,300 smokers vs 1,300 non-smokers, the smokers had an increase in lung cancer rates that's pretty conclusive isn't it? Well sure, but now ask why were these people smoking. Were they under stress, in situations that might cause them to want to smoke? You could say that people who smoked also drank alcohol, maybe it's the combination that causes problems?

Okay I'm not saying there is no link between smoking and lung cancer what I'm trying to show is that it's more complicated then the simple slogan on the side of a packet. Smoking causes fatal lung cancer? Smoking may increase the chances of fatal lung cancer in later life; not as punchy though.