Comment »Posted on Tuesday 10 February 2009 at 10:49 pm by Jacob Aron
In Getting It Wrong, Health & Medicine

It’s not a very catchy headline, is it? It would have been much better to go with the Guardian’s Cannabis doubles testicular cancer risk, says US study – now that sounds scary. If you haven’t figured it out through my dripping sarcasm, we’re talking about two old friends: relative versus absolute risk.

We last saw the pair in the shocking news that drinking beer increases your risk of bowel cancer by a fifth (in other words, to about 0.073%). In case you’ve forgotten, let’s cover the basics again quickly.

Imagine there is a 1% chance of contracting a disease, meaning that in a given time frame (usually a year) we can expect 1 in 100 people to have this disease. Now, suppose that research finds that eating ice cream makes you more likely to get this disease – perhaps 3 in 100 ice-cream eaters have it, meaning there is a 3% chance you will contract the disease if you eat ice cream.

Relative risk is best thought of in terms of multiplication. By eating ice cream, your relative risk of the disease has trebled, because 1% becomes 3%.

Absolute risk is more like addition. An additional 2 people in every 100 will contract the disease if they eat ice cream.

Now, it’s clear why newspapers prefer relative risk to absolute risk. “Trebled” sounds much better than “plus 2″ in a headline. It’s a much less useful statistic however, especially when dealing with events with a low chance of occurring.

I must say though that in this case, if we look at the original paper, I think it’s actually the scientists who are to blame. In their words:

“We observed a 70% increased risk of TGCT associated with current marijuana use.”

That 70% is relative risk, not absolute. It’s also not double, which would be an increase of 100% in relative terms. That’s because the research showed a 70% relative increase for men in general, but the Guardian went with “double” which only applied to regular cannabis users, or those who started smoking before they were 18. To be fair, this is partially explained in the second paragraph, and then clarified towards the end.

So where did I get my un-catchy headline from? I once again turned to Cancer Research for some absolute risk statistics. In the UK, 7.1 men per 100,000 of the population contract testicular cancer. Dividing these two numbers gives us the absolute risk of 0.0071%, which I simply multiplied by 1.7 to calculate the 70% increase used in the headline.

I don’t know about you, but I think 0.01207% is pretty low. It’s because testicular cancer is actually one of the rarest forms of cancer that there is, so even a massive (relative) increase in risk results in a pretty small chance of catching it anyway. To me, that’s just not news – and yet pretty much anything involving cancer makes it in to the papers.

Remember: relative risk, and absolute risk. The difference is pretty important.


Sorry, comments for this entry are closed at this time.