– What is “hazard ratio” in medical research, and how is the number interpreted?
In: Other
The hazard ratio is
(how often people in group A die of something) ÷ (how often people in group B die of something)
…and we use it to compare treatments. Generally groups A and B will be people who have the same disease/problem, but we give them different treatments and then see who does better.
Let’s say I have 200 people who are having suicidal thoughts. I split them into two groups of 100 each, A and B. I give Group A antidepressants and have them meet with a specialist twice a week. I tell Group B to suck it up and quit complaining. (Because the world sucks, I will now state explicitly that this is poor medical practice and you shouldn’t do this.)
After two years, 3 people in Group A and 8 people in Group B have committed suicide. The hazard ratio is 3/8 = 0.375, and that tells us the antidepressants+counselling thing works better than the manly approach.
If the hazard ratio comes out really close to 1 (0.98, 1.03, that kind of thing), it means both things work about equally well; the further away from 1 it is, the bigger the advantage one treatment offers. If the difference is big, it’s hard to justify *not* using the treatment with lower hazard. Basically it lets you say not just “A is better” but “A is *this amount* better”.
Latest Answers