That argument is likely a distortion of the medical argument which goes something like, “People who are overweight by the medical definition of BMI between 25 and 30 are not necessarily unhealthy. There are some circumstances such as being an athlete or genetics that are associated with denser body compositions. BMI is a crude tool that is useful for some things but should not be used on its own to indicate health status.”
BMI isn’t science. Its an almost 200 year old equation that was arbitrarily fitted to data. There is no go reason square ones height except that it fit the data. The cutoffs are arbitrary and, at least in the US, shifted in 1997.
And there is a growing body of evidence showing it’s not accurate for many cases in addition to the one you provided.
We have better indirect measure and far better direct measures for assessing disease progression and likelihood of disease development. Getting rid of BMI won’t stop fat shaming, but I hope it gives people pause.
I completely agree, we need to move on from BMI. But it’s a bit silly to say it’s BMI isn’t science when it’s been used for the entirety of modern health sciences. People would be shocked by how many crude, yet useful enough measures that health sciences use even today. And it’s notoriously slow/stubborn in adopting the best tools for many methods. Still, humanity has continued to make scientific progress with them.
Show me the science in the particulars and I’m happy to change my mind. Its widespread use in the modern medical system doesn’t make it scientific. We continue to use generally true ideas such as drink water and then wrench them into prescriptive positions like drink 8 cups of water per day. Literally no science to support that claim.
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 Whether it makes for good science or not, it’s use makes it part of science and replacing outdated tools is part of the broader scientific process–that doesn’t make the tools “not science”.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 appreciate the systematic review and meta-analysis. It’s a good starting statement and if I worked with children, I would look at the paper more closely. As a whole, these studies don’t address the most at risk groups with a high level of evidence. Perhaps that last paper will be part of a meta-analysis that gives clearer evidence of BMI indicating CVD in children. This would be great.
I focus on medical practices because it’s my area of expertise and where I do my work. So I see the negative effects of people’s conceptions around weight, BMI, obesity, and how difficult it is to change even with the best applied efforts. I wrote my initial response when I saw an avalanche of self-righteous, care trolling with vague allusion to science and medicine with a level of certainty that isn’t warranted. At best, I was being confrontationally polemical, at worst, I lack nuance or sensitivity to work in the field.
The ease at which people fat shame and delude themselves that they are helping is astounding. I was a little surprised to see it on Lemmy.
Admittedly, my statistical training isn’t the best, but I appreciate the role it plays in making sense of large datasets. Still, I appreciate the reminder to dive deeper into how statistics are used in observational studies. For me, at least, I wish that much of this was done before the wide deployment of BMI in the populous. I’m not saying that fat-shaming wouldn’t continue, but there doesn’t need to be poorly applied scientific ammunition either.
What science would change your mind? There’s never going to be a magical cutoff number for cholesterol or height or weight that separates healthy and not healthy.
Heuristics are useful tools and sometimes that’s the best you get. You need water to live, clogged arteries cause heart attacks, insulin resistance leads to diabetes. Exactly how much of any given thing causes bad outcomes is going to vary case by case, but doesn’t negate trends.
I say all this as a former wannabe body builder who hasn’t had a BMI under 25 in about 20 years, but I still know a BMI of 60 or 80 is no good.
I though I was clear about this, but I’ll reiterate.
That the heuristic is accurate.
That the heuristic is more accurate than other easily applied heuristics.
That when the heuristic makes categories, the categories are backed by studies. These studies would show a statistical increase for specific health outcomes above this cutoff. That line would be tested relative to other proximal lines.
These heuristics would include different recommendations for different populations such as race, biological sex, and age.
A better alternative, as I had previously linked to, would be abdominal fat as measured at the waist. Easy heuristic and closely correlated to CVD.
All of what you say is true, but you’re not address my particular issues.
Thx sorry I didn’t read all your comments in the post, I was using that question as a proxy to whether or not your discussion was in good faith. It seems like the answer is yes.
I frequently wonder how many better metrics are available that just aren’t as easy to capture as stepping on the scale, grabbing blood oxygen, and taking blood pressure. I’m sure that part of the balance is value of vitals versus time or effort to collect them.
That argument is likely a distortion of the medical argument which goes something like, “People who are overweight by the medical definition of BMI between 25 and 30 are not necessarily unhealthy. There are some circumstances such as being an athlete or genetics that are associated with denser body compositions. BMI is a crude tool that is useful for some things but should not be used on its own to indicate health status.”
BMI isn’t science. Its an almost 200 year old equation that was arbitrarily fitted to data. There is no go reason square ones height except that it fit the data. The cutoffs are arbitrary and, at least in the US, shifted in 1997.
And there is a growing body of evidence showing it’s not accurate for many cases in addition to the one you provided.
We have better indirect measure and far better direct measures for assessing disease progression and likelihood of disease development. Getting rid of BMI won’t stop fat shaming, but I hope it gives people pause.
I completely agree, we need to move on from BMI. But it’s a bit silly to say it’s BMI isn’t science when it’s been used for the entirety of modern health sciences. People would be shocked by how many crude, yet useful enough measures that health sciences use even today. And it’s notoriously slow/stubborn in adopting the best tools for many methods. Still, humanity has continued to make scientific progress with them.
Show me the science in the particulars and I’m happy to change my mind. Its widespread use in the modern medical system doesn’t make it scientific. We continue to use generally true ideas such as drink water and then wrench them into prescriptive positions like drink 8 cups of water per day. Literally no science to support that claim.
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/
I appreciate the systematic review and meta-analysis. It’s a good starting statement and if I worked with children, I would look at the paper more closely. As a whole, these studies don’t address the most at risk groups with a high level of evidence. Perhaps that last paper will be part of a meta-analysis that gives clearer evidence of BMI indicating CVD in children. This would be great.
I focus on medical practices because it’s my area of expertise and where I do my work. So I see the negative effects of people’s conceptions around weight, BMI, obesity, and how difficult it is to change even with the best applied efforts. I wrote my initial response when I saw an avalanche of self-righteous, care trolling with vague allusion to science and medicine with a level of certainty that isn’t warranted. At best, I was being confrontationally polemical, at worst, I lack nuance or sensitivity to work in the field.
The ease at which people fat shame and delude themselves that they are helping is astounding. I was a little surprised to see it on Lemmy.
Admittedly, my statistical training isn’t the best, but I appreciate the role it plays in making sense of large datasets. Still, I appreciate the reminder to dive deeper into how statistics are used in observational studies. For me, at least, I wish that much of this was done before the wide deployment of BMI in the populous. I’m not saying that fat-shaming wouldn’t continue, but there doesn’t need to be poorly applied scientific ammunition either.
PS. You might like this study that examined some of the boundaries for BMI.
What science would change your mind? There’s never going to be a magical cutoff number for cholesterol or height or weight that separates healthy and not healthy.
Heuristics are useful tools and sometimes that’s the best you get. You need water to live, clogged arteries cause heart attacks, insulin resistance leads to diabetes. Exactly how much of any given thing causes bad outcomes is going to vary case by case, but doesn’t negate trends.
I say all this as a former wannabe body builder who hasn’t had a BMI under 25 in about 20 years, but I still know a BMI of 60 or 80 is no good.
I though I was clear about this, but I’ll reiterate.
A better alternative, as I had previously linked to, would be abdominal fat as measured at the waist. Easy heuristic and closely correlated to CVD.
All of what you say is true, but you’re not address my particular issues.
Thx sorry I didn’t read all your comments in the post, I was using that question as a proxy to whether or not your discussion was in good faith. It seems like the answer is yes.
I frequently wonder how many better metrics are available that just aren’t as easy to capture as stepping on the scale, grabbing blood oxygen, and taking blood pressure. I’m sure that part of the balance is value of vitals versus time or effort to collect them.