Hacker News Reader

Comment by throwaway2037 | Hacker News Reader

throwaway2037

To be clear, it looks like the health stat term "overweight" means anything greater than "normal", which includes obese.

Quick Google search:

    > what percent of australians are overweight?

First hit:

    > Over the last decade, the proportion of adults who were overweight or obese has increased from 62.8% in 2011–12 to 65.8% in 2022.

Source: https://www.abs.gov.au/statistics/health/health-conditions-a....

US NIH says: 73.1% are overweight (includes obese). Ref: https://www.niddk.nih.gov/health-information/health-statisti...

Sure, 73.1% > 65.8%, but Australia is still plenty overweight. Both are appalling.

View on HN

> Both are appalling.

Both are also largely meaningless because they're based on BMI, which is literally just mass/height^2. No measure of percent body fat, no measure of any other aspect of health, just mass by the square of height.

If you're comparing the BMI of two countries with very similar gene pools it's not a bad point of comparison (though the raw number still doesn't tell you much without more context about build types), but when you're comparing Australia to the US the gene pools of the non-European minority groups are sufficiently different to make BMI pretty worthless as a point of comparison for public health.

How do they even pretend that works when some South Asians are shorter and the Danes are so large? Are the Danish just 100% obese?

This says it's <20% there[1][2]

1: https://en.wikipedia.org/wiki/Average_human_height_by_countr...

2: https://academic.oup.com/eurpub/article/33/3/463/7058153

Height is factored in, albeit in a very simple way

Yeah, it's not height itself that's the problem, it's body shape. As just one example: some people have long torsos and short legs, others long legs and short torsos. Those two groups will have wildly different BMI curves even at similar health levels.

BMI is a very crude measure. It is only useful in the extreme, and not particularly relevant for an individual. If you have a BMI of 40 you should think about losing weight, but exactly nobody need a composite number to know they are that fat. Similarly, you might have a normal BMI but an LDL of 300 or a strange lump on your thyroid. These things are vastly more important than your BMI

To be fair BMI is based on Belgians from the 1850s with all the implications it has. Modern people are much taller (+15 cm for males) and eat a lot more protein.

So the line between normal and overweight is somewhat blurry. e.g. someone who is 6'3" and weighs 200 pounds is overweight. Which might or might not be the case (but you certainly don’t need to a body builder or invest a lot of time to maintain at least reasonably healthy 18-20% body fat ratio).

>To be fair BMI is based on Belgians from the 1850s with all the implications it has.

It's a pretty straightforward formula, and even though the cutoffs might be arbitrary, there's undoubtedly a u shaped mortality curve centered somewhere around 20-25. At a population level I don't think any of that is a relevant factor. No one thinks that your risk of death hinges on crossing an arbitrary line, but being fatter definitely isn't good for your health.

But the question here is whether you can compare the BMI curves of two countries and come to any conclusions, and the answer is pretty solidly "no".

Bone density, muscle mass, torso height relative to leg height—none of these things are factored into BMI and all of these vary wildly depending on genes.

> vary wildly depending on genes

Being born with worse attributes doesn't mean those attributes are out of the equation. A lot of people die much younger than they should just by poor luck, with little or no control on their end.

If you happen to have higher bone density and higher muscle mass just because you have higher testosterone, you're still better off. Sometimes people just get the short end of the stick. That doesn't necessarily mean BMI doesn't work.

I've even met smokers, smokers! Who have better cardiovascular health than me and are skinnier. Which really fucking sucks for me.

Veins have a 2D cross section where body’s are 3D. That’s one reason why using height ^2 not height ^3 may better correlate with health.

Similarly, cancer risks may scale with the amount of tissue or the surface area for things like skin / colon cancer but it’s hard to see how being taller is beneficial.

Reply on news.ycombinator.com