Hacker News new | ask | show | jobs
by Zababa 744 days ago
I think part of it is regression to the mean. Like, your body heals itself naturally from many things. Say you get a cold. For most people at some point it'll heal. If you have a group of people with the cold, and try to determine how effective vitamin C or zinc or COLDKILLER777 is, you can't just give it to them and say "look, they're healed", because they would heal naturally. You have to prove that they feel less symptoms or heal faster than people in the same circumstances that don't receive the same molecule.

I also remember some similar stuff for back pain and surgeries. In that context people were seeking treatment when their back issues peaked, and the question was, when you take the cohort of people that had back surgery and the cohort of people that didn't, did the back surgery make a difference? Because some people healed naturally.

I don't know if this is true in that specific context, but to take a more pedestrian one, I've had lots of small cuts, burns and things like that during my life, and they all healed.

1 comments

The placebo effect controls for all of these factors, and is entirely real. It's the reason placebo control groups exist in the first place. If the placebo effect was just statistical noise, then you could greatly simplify trials by just taking a random group of people as the control, and do away entirely with trying to hide from both patient and technician who is taking the drug - because it wouldn't matter at all.

You can see the opposite effect with the less well known nocebos [1]. People can experience objectively measurable side effects (such as bloating) that are in no way associated with a treatment, but that a patient believes to be a side effect. It can even be fatal. The article references aboriginals who will 'curse' one another resulting in the victim rapidly dying, because he believes so strongly that he is going to die! A similar thing in contemporary medicine has been observed with those who receive a fatal prognosis of cancer with them ending up dying long before there is any way the cancer could have killed them.

[1] - https://en.wikipedia.org/wiki/Nocebo