This video explains the concept simply and clearly. I have a nitpick issue with the “raw milk” example in the video. In that example, the “risk” of drinking raw milk, on an individual basis is very small. However, if tens or hundreds of millions were to drink raw milk, then the % who get sick each year would be a large number.
This is why it’s important to understand statistical concepts and to think about what they mean in real world scenarios.
The pharmaceutical industry makes much use of the term “relative risk” to draw attention to potential benefits of drugs. But when it comes to side effects, they use “absolute” percentages.
For example, during the short run clinical trials, they find that 4% of users experience joint pain – but that 3% of non-users (given a placebo) suffer joint pain. This will be reported as 1% of users experienced joint pain.
Meanwhile, during the drug test, 13% of those who took the drug experienced, say, a heart problem but 15% of those who did not take the drug experienced a heart problem. This will be reported by saying that “there are 13% fewer heart problems among those who took the drug”.
See the problem?
The absolute risk went from 15% to 13% – but they do not report that as 2% benefited – instead, they report that there was a 13% reduction in heart events (2% divided by 15%).
The consumer hears: 13% reduction in heart events and just 1% had side effects! Sounds good!
The first comparison reports a relative risk reduction, and the latter reports an absolute percentage. That sounds much better than saying there was a 2-percentile point reduction in heart events, doesn’t it?
The use of absolute and relative percentages is used to make benefits sound greater than potential side effects or harms.
When side effects do occur, there is rarely further investigation to explain why the side effects are occurring. The primary purpose of drug trials is to (1) show that it works better than some alternative therapy, and (2) does not kill people.
That’s it!
