Say you were worried about having a heart attack. Which drug would you rather take:
- Magic Drug A: reduces your risk of a heart attack by 50%
- Magic Drug B: reduces your risk of a heart attack by 1%
Presumably, you’d pick Magic Drug A.
But, what if you found out Magic Drug A and Magic Drug B were really the same drug?
Huh? How can that be? It has to do with something Graham Walker, an emergency department physician at Stanford, calls “statistical trickery.”Here’s how:
- There’s “absolute risk,” your risk of something bad happening, compared to everyone else. It’s a general population risk.
- There’s “relative risk.” This is your reduced risk of something bad happening in relation to something else, such as taking a drug. In our example, it’s our Magic Drug, to reduce risk of that heart attack. For people not working in numbers every day, this takes a moment to get your mind around, so bear with us for a walk-through.Let’s give our Magic Drug to 100 people. Time goes by, and the people who got the drug had 50% fewer heart attacks than 100 people who didn’t get the drug.Here’s where the statistical trickery comes in. How many people who were taking the drug actually had heart attacks? One.
And of the 100 people who did NOT get the drug? How many of them had heart attacks? Two.
So, yes, one heart attack is half as many heart attacks as two. So relatively speaking, it’s a 50% reduction. But, in absolute terms, the risk was reduced from 2% to 1%, a much less compelling 1% reduction.
Diana Stilwell of the Foundation for Informed Decision Making compares it to shopping. “It’s like going into a store that’s selling things at 20% off. The 20% is only meaningful to you if you know what you’re going to be buying.”
If you’re buying a pair of shoes, your 20% savings could be a lot of money. If you’re buying a loaf of bread, not so much.
- Which brings us to the next thing to know, “What was my risk in the first place?” If your risk of something bad happening is just two in a hundred, you may feel less excited about the prospect of reducing your risk by 50%. (You might also want to keep this in mind when you see drug ads, which tend to focus on relative risks).
- The next thing you should find out is “What are the side effects?” All drugs have them.
- Finally, our fifth thing to know. Statisticians (the ones not engaged in trickery) call it the Number Needed to Treat, or NNT. In our example, we gave the drug to 100 people and prevented one heart attack. The Number Needed to Treat was 100. In other words, for every heart attack prevented, 99 people were exposed to a drug and its potential side effects. Graham Walker believes those risks should be part of the equation when people are considering whether to take the drug. “You want to be that person who has their life saved, but the flip side is any time you expose someone to medicine, you’re going to give some harm to some of them.”
Walker is one of the founders of a terrific website, theNNT.com. Visitors can get easily digestible information about risks and benefits for a range of drugs and other medical treatments. Walker says the website is entirely supported by the handful of doctors who put it together. For fun, you can watch Walker’s video about the NNT.