Relative and Absolute Risk

A headline says a habit “doubles your risk.” Doubles from what? If the risk was one in ten thousand, doubling makes it two in ten thousand. The relative change sounds huge. The absolute change, one extra case in ten thousand people, may matter very little.

Ask for both numbers. Relative risk compares two rates and hides how large either is. Absolute risk tells you how many people in a group are affected. A big relative change on a rare outcome is a small absolute change; a small relative change on a common outcome can affect many people.

The same trick works in reverse. “Cuts your risk by half” can mean a trivial benefit. Turn each claim into a plain count: out of a thousand people like me, how many are affected with and without the thing?

The scary headline

A news story says a common medicine “raises the risk of a rare condition by 100 percent.” Readers stop taking it. The condition affects about one person in ten thousand.

A doctor explains that a 100 percent rise means two in ten thousand instead of one. For most patients the medicine’s benefit is far larger than the added risk. The headline was true, and it told them almost nothing about what to do.

Try this today When a headline gives a percentage change, look for the plain count: out of a thousand people, how many?