“There are two potentially effective screening tools for prostate cancer, one which
will reduce their chances of dying from the disease by between 20 and 30 percent,
while the other will save one life after 10,000 person-years of screening [...] the
two programmes are the same, they were just packaged differently.” – Professor Michael
Baum in Doctoring Data by Malcolm Kendrick
Let’s say I give medication A to 100 people and nothing to 100 people. At the end
of the year, 1 person dies from people taking medication, and 2 people die from the
• Absolute difference is 1 in 100, or 1%.
• Relative difference is 1/2, or 50%. I can say medication A reduces the risk of
dying by 50%.
I give medication B to 1000 people and nothing to 1000 people. At the end of the
year, 1 person dies from people on medication B, and 2 people die from the other
• Absolute difference is 1 in 1000, or 0.1%.
• Relative difference is still 1/2, 50%. I can say medication B reduces the risk
of dying by 50%.
Benefits mentioned in most articles we read are relative benefits, and absolute benefits
are often buried somewhere. Are the researchers lying? Absolutely not. But do you
see the problem with just looking at relative benefits/risks? I hope so.
I think it is important to know both relative and absolute benefits/risks, so you
can make an informed decision. If you don’t know, you may be in for a bad surprise.
I highly recommend Doctoring Data by Malcolm Kendrick..