I was talking about how widespread BMI is used in health sciences, I.e. everything from basic physiology to clinical trials to program evaluation to epidemiology. This is different from medical practice, e.g. family doctor taking your BMI. Whether it makes for good science or not, it’s use makes it part of science and replacing outdated tools is part of the broader scientific process–that doesn’t make the tools “not science”.
You’re asking about “accuracy” which is a good question, as well as “precision”. However in health sciences we usually evaluate such measures more thoroughly with similar concepts of validity (construct and discriminant) and reliability; you’ll also see sensitivity in the literature but it’s a kind of discriminant validity.
So if you do your own search using “BMI” and these terms on PubMed or even Google Scholar, you will find a range or scientific evidence. Most will say BMI is not good but not terrible, even good in some specific contexts. You will also find lots of evidence of how BMI is associated with other health indicators and health outcomes. I’m not going to spend an hour collating this for you. “Review” is also a useful search term. You seem smart enough to do it if you really want it. In any case, the argument is moot because we agree BMI should be replaced.
Edit: okay I was curious comparing BMI to WtHR and actually found a couple cherry-picked examples that might be interesting for you
I was talking about how widespread BMI is used in health sciences, I.e. everything from basic physiology to clinical trials to program evaluation to epidemiology. This is different from medical practice, e.g. family doctor taking your BMI. Whether it makes for good science or not, it’s use makes it part of science and replacing outdated tools is part of the broader scientific process–that doesn’t make the tools “not science”.
You’re asking about “accuracy” which is a good question, as well as “precision”. However in health sciences we usually evaluate such measures more thoroughly with similar concepts of validity (construct and discriminant) and reliability; you’ll also see sensitivity in the literature but it’s a kind of discriminant validity.
So if you do your own search using “BMI” and these terms on PubMed or even Google Scholar, you will find a range or scientific evidence. Most will say BMI is not good but not terrible, even good in some specific contexts. You will also find lots of evidence of how BMI is associated with other health indicators and health outcomes. I’m not going to spend an hour collating this for you. “Review” is also a useful search term. You seem smart enough to do it if you really want it. In any case, the argument is moot because we agree BMI should be replaced.
Edit: okay I was curious comparing BMI to WtHR and actually found a couple cherry-picked examples that might be interesting for you
https://www.mdpi.com/2072-6643/8/8/512
https://www.sciencedirect.com/science/article/pii/S2405457723021642
https://pubmed.ncbi.nlm.nih.gov/23775352/