I was instructed to self-promote, so here is a post I wrote a few days ago that overlaps with some of Dynomight's favorite topics. I was frustrated by the lack of readily available information about what BMI healthy people should aim for, so I wrote a post to try to fix that.
It looks like more like a 7% reduction (see Aute et al. (2016) Table E, which I show in OP). 7% is still not huge—a BMI of 30 has a hazard ratio of 1.85, and 45 has a hazard ratio of 4.18—but it's big enough that I care about it.
Edit: Ok I see now that you are directly quoting me...I don't know why I said that because I'm pretty sure it's wrong. I will update it to be correct.
On the height thing, I don't really have a sense, I don't know why the denominator uses height squared instead of height cubed. Presumably bodily width and depth vary sub-linearly with height, but why is it squared instead of height^2.5 or height^1.7 or something?
FWIW I do know that height is associated with increased mortality in general. Part of the mechanism is that taller people have more growth hormones which leads to more cancer. Could also be that height puts more strain on the heart.
I found this quote extremely interesting: "The 2015–2020 Dietary Guidelines for Americans sets healthy weights as those corresponding to BMIs between 18.5 and 25. […] Panel members agreed that the risk of heart disease, diabetes, and high blood pressure begins to climb at a BMI of 22 or so. But they didn’t feel justified choosing such a low number as the cutoff between healthy and unhealthy weights, because doing so would have labeled a large majority of the U.S. population as overweight. […] [B]ut many people with a BMI of 23 to 25 are not at their healthiest weight."
There's some discussion of this in the footnotes, American BMIs are so high and the effects so small that it's kind of justified. Certainly when you add in various other effects (like exercise and muscle mass)
I basically agree with the panel that at a policy level, getting people from 30 to 25 BMI is like 100x more important than getting people from 25 to 22. I personally have no trouble losing weight so I wanted to know what BMI I should aim for. And my BMI is a bit on the high end (~24) so I wanted to know if I should be concerned about that. I've started trying to lose weight since writing this post.
One question: I haven't thought this through very well, but suppose that caffeine hurt performance by way of harming sleep. Could that change any of the conclusions? I've considered similar experiments, but it seemed like the only way to really test it would be to do a bunch of multi-week periods of using caffeine, abstaining, using caffeine, etc.
I didn't read the literature on this but this is my impression from reading articles/watching YouTube videos by people who did read the literature:
Yes, caffeine can hurt next-day performance by interfering with your sleep. If you take caffeine in the afternoon every day, that's probably bad for your performance. I don't know if the poor sleep is enough to cancel out the benefits of caffeine, my guess is that it is if you accumulate bad sleep for enough days in a row.
But if you take caffeine early in the morning (which I do), it appears to have little effect on sleep. I recently watched this video which cites a meta-analysis saying that it's ok to have small doses of caffeine >9h before sleep, and large doses >13h before sleep.
This is great! I must admit that I had to stare a long time at this figure to understand it, but once I did, I felt like it was good evidence for your conclusion. I also thought the discussion of exercise was very interesting. (It's frustrating that it's basically impossible to have an RCT for BMI, and really really hard to have one even for exercise.)
Add comment