A general point: It is generally helpful to express risks as absolute risks (or number needed to treat) whenever possible, as odds ratios or percentages changes are hard to interpret, and often mislead if the ‘base rate risk’ is minute: often people’s brains mistake ’25% increased risk of Y if you do X’, as risk of ‘25% of Y if you do X’.
E.g. pretend mobile phone use really caused an exotic brain cancer, and the OR is 100. But the base-rate of this brain cancer is something like 1/ten million. So although mobile phone users are 100 times more likely to get exotic brain cancer, their absolute risk goes up from 1/ten million to 1/million, so the absolute risk reduction of avoiding mobile phones would be a 0.0009% mortality risk benefit—just over a million people would have to stop using mobiles to avoid a single case of exotic brain cancer.
A general point: It is generally helpful to express risks as absolute risks (or number needed to treat) whenever possible, as odds ratios or percentages changes are hard to interpret, and often mislead if the ‘base rate risk’ is minute: often people’s brains mistake ’25% increased risk of Y if you do X’, as risk of ‘25% of Y if you do X’.
E.g. pretend mobile phone use really caused an exotic brain cancer, and the OR is 100. But the base-rate of this brain cancer is something like 1/ten million. So although mobile phone users are 100 times more likely to get exotic brain cancer, their absolute risk goes up from 1/ten million to 1/million, so the absolute risk reduction of avoiding mobile phones would be a 0.0009% mortality risk benefit—just over a million people would have to stop using mobiles to avoid a single case of exotic brain cancer.