Map of life expectancy at birth from Global Education Project.

Wednesday, December 20, 2006

An Epochal Medical Breakthrough!

Our company has developed a genuine medical miracle. A large scale, 12 year randomized controlled trial has shown that Alivea™ reduces your risk of dying from Himmingburden's disease by two-thirds! That's right, according to a report in the New England Journal of the American Medical Association Archives of Internal British Medicine, people who took Alivea™ daily had a relative risk of .333 (95% CI .23 to .48) of dying from Himmingburden's disease at the conclusion of the trial compared to subjects on placebo.

Side effects of Alivea™ include bad breath, compulsive ass scratching, and inability to see the humor in Marx Brothers movies and other comic stories based on objectively implausible premises, but this is a small price to pay to save your life. Ask your doctor about Alivea™ today and rest assured, Alivea™ can save your life.

Okay, that sounds pretty good. This stuff really can save my life. Now what if I told you that the risk of dying from Himmingburden's disease over 12 years is 1 in 10,000, which means that if 30,000 people take this stuff, at the end of 12 years only one of them will have died from Himmingburden's, compared with three who would have died otherwise? I might also remind you of the sad news that if Himmingburden's doesn't get you, something else will. After we do the calculations, we discover that the average increase in life expectancy from taking Alivea™ is 12 minutes and 37 seconds. And oh yeah, taking it for 12 years costs $17,000.

Not so great after all. The problem is that we have been informed about the relative risk, but not the absolute risk. When you tell people that doing something reduces a risk by half, or two thirds, or some such impressive sounding ratio, they perceive that as a huge advantage. But if you tell them the absolute risk instead, and it is very small, the perception is completely different. It turns out that Himmingburden's disease is not something worth worrying about in the first place.

Lisa Schwartz and colleagues, in the new British Medical Journal, find that the majority of medical journal articles report relative risks in the abstract, but not the absolute risk. About a third of articles don't report the absolute risk anywhere. This has the psychological effect of making interventions seem more beneficial, and risks more grave, than they really are.

I suppose this happens because researchers want their findings to seem really, really important; and because the drug companies that pay for much of medical research and still maintain editorial control over much of the resulting literature want their drugs to seem really, really beneficial. But a much better question to ask is, how many people do we need to treat to get a specified benefit? And, having asked that, how important is that benefit in the context of the other risks and the expectancy for quality life of the individual? Once we ask those questions, the case for intervention usually looks much less compelling.

Arm yourself with skepticism.

No comments: