I believe that what you are stating that the insurance companies believe is WRONG. I have no independent reason to believe or to disbelieve that they believe it and don’t particularly care.
The current scientific data is that in Whites, Asians, and Hispanics, there is no significant difference in mortality by BMI between genders. (Multiple large prospective studies available at this point.) There may be a gender difference for Blacks in the “overweight” group with “overweight” Black males at greater risk than “overweight” females.
If the insurance tables tell you something else then they are not correct.
Sorry if I read your “so what?” to my comment about central obesity and where the fat is stored as mattering more than the BMI itself as illustrating your lack of grasp of its importance. Again, BMI is a crude but clinically very useful screening tool, equally useful and equally limited for both males and females. I believe it is a mistake to use it, for males or females, as the definition for overweight or obesity, let alone for health risk prognostication above and beyond fitness and nutrition habits. It is however “where the light is good” so it is used in many studies.
Another fact check. Also wrong. Using the BMI definition, it is more males than females in the United States who are overweight and obese, 69.9 to 56.6.
Sorry for such multi-posting, but more detail that does document some kernel of truth in your claim. Currently, 20 to 74 years old, significantly more men are overweight, about the same percent male and female obese, and more women extremely obese. Female obesity varies greatly by race and ethnicity with non-Hispanic Black over 20 at 58.5% obese compared to non-Hispanic White at 22.9%. Male variation by race and ethnicity OTOH is minimal, all in the 36 to 39% range.
Well that’s a bit strange given that your own chart supports my position.
Let’s assume you are an actuary and you are trying to figure out the BMI cutoff for preferred risk categories. Just looking at the chart, the curve for healthy non-smoking men passes the 6% extra risk ratio at a BMI of about 26.0. That’s as good a place as any to draw the line for the preferred risk category for men.
Ok, what about for women? At BMI 26.0, the extra risk is already 9% or more. If 6% extra risk is the standard, you will have to draw the line at a bit less than 25.0.
I think the mistake you are making is assuming that “very similar” means “same.”
Ok, but there is a lot of impact when the government publishes an official definition of “overweight” and “obese.” And has a web site where you can put in your height and weight and tells you if you are “normal” “overweight” or “obese.”
I’m not sure, I would have to check back on the reading I did. I note that it’s a bit circular to use the BMI-based definition of obesity on this issue.
From the table (9 a and b) in appendix the 95% confidence interval for healthy never smokers at BMI of 25.0 to 27.4 is 1.01 to 1.12 for males and 1.05 to 1.14 for females. From the scientific and statistical POV that is not a significant difference. In both the significant jump occurs when one moves over a BMI of 30 (obese) and both male and females have the exact same 44% mortality risk increase associated with that. And note, if the insurance companies believe what you think they believe (and a link could be helpful), and are going by actual current best data, then they would be less concerned about never smoker females with BMIs of 25 to 27.4 (marginally overweight) than with low normal BMI (18.5 to 19.9) never smoker females, who have a 14% increase in mortality rate compared to those with BMIs of 20.0 to 24.9. Higher than the risk they’d label marginally overweight.
In both cases the significance, especially for those in the overweight labelled group, is marginal. Actually knowing where the fat is, how much muscle mass there is, and what fitness level people have and what their nutrition choices were would tell you much much more. You may be interested in some of the discussion from this past thread.
Yes, BMI (while still of some utility) is overplayed, especially for the “overweight” category. For both genders.
It depends on what you mean by “broadly” and “nearly.”
Anyway, your position is a bit silly. You cite a study which you claim shows that the insurance companies are wrong. When I point out to you it shows that the insurance companies are right, you then claim that the study is not statistically significant. Essentially it’s just a coincidence, according to you, that your own study backs up my position.
Besides which, if your current (unsupported) claim is true, then your study certainly doesn’t show that the insurance companies are wrong. At best your study is inconclusive.
No need to teach me anything, since I have already taken both basic and advanced stats. Just show me the calculation you did. You did actually do a calculation, didn’t you? Or did you just assert your position without doing any calculation at all?
Life insurance companies might do better to factor in height*, as far as any differences between men and women go:
Of course, being overweight may reduce lifespan by a similar mechanism (postulated as being due to a bigger body having more cells and requiring more upkeep and cell divisions, which are limited).
*Forget that I suggested this, or I’d have to pay more than most people.
Honestly Brazil, at the risk of a mod-slap, I have to say that I have a hard time believing you. No one who has taken any stats would ask that question. Were you awake? Did you pass?
95% confidence that the true number is between 1.01 and 1.12 vs 95% confidence that the true number is between 1.05 and 1.14? You are really claiming that shows a significant difference?
No there is no need to calculate that. The study is significant and shows no significant difference. If you are claiming that the authors are incorrect to conclude that the two groups showed no siginificant differences (which is why they pooled them for the rest of the analyses) please feel free to do the math that proves them wrong.
I, btw, am not claiming the insurance companies are wrong since I have no idea what they think. I only know what *you think *they think, based on your interpretation of an unsourced, uncited table of unknown vintage, while an article by acturaries posits the opposite. And I don’t care what they think as I have no reason to defer to acturaries.
Also btw, I am not going to claim, and have not claimed, that males and females have the exact same mortality risk for each and every BMI point. My statement remains that the clinical utility of BMI, especially in the overweight range, is of the same limited clinical utility for both genders. There was as slight difference in the normal distributions when the data was first collected, and if statistical normal correlates with health outcomes then some small differences might be expected. This was and is a crude measure. Not precise stuff. A rough screen for obesity in its original intent and how it should be thought of today. The fact that that study shows the exact same 44% increased risk for the obese group in both genders is honestly a surprise to me.
No, I am skeptical of your claim. I do not have access to your paper and cannot see the shape of the data in question.
Please do not strawman me.
Please quote the part of the study where they conclude that there is no significant difference in excess mortality between (1) men of BMI between 25.0 and 27.4; and (2) women of BMI between 25.0 and 27.4.
So you are skeptical of the numbers from the article?
The article was already linked. The supplementary appendix is linked to there but to make it easier for you here. You made a claim from the data in the article that the curves are significantly different and support a contention that there is some different BMI point that should be used for declaring women overweight and at health risk than for men, like you claim insurance companies believe
The quote is simply one stating that they are similar enough to pool:
I still have no link to whatever your source is for what you have concluded about what insurance companies currently think, btw.
Look, I think we essentialy agree that BMI is a limited use tool. You want to try to find a way to finesse it and are trying to claim some big difference between males and females that the data does not support, one that would justify having different overweight and obesity cut-off numbers for males and females. My take is to not get too concerned about the BMI numbers in and of themselves, especially in the “overweight” group, other than as a screening tool. There are other means to assess that mean much more.
Perhaps, but I am mainly skeptical of the conclusion you claim that the article reaches.
No, I made a claim that the curves are different. You claimed that the difference was (statistically) insignificant. On top of that you claimed that the study affirmatively shows that the difference is insignificant. i.e. you are not claiming that the study is inconclusive as to a difference, you are also claiming that the study affirmatively shows that there is not a difference and therefore contradicts the actuarial tables I have referenced.
Please quote the study where it does so. Failing that, please admit that the study does no such thing and apologize for your misrepresentation.
I am happy to provide one if you first represent that you are seriously skeptical about my claim.
I have had odder conversations on these message boards but few.
Yes, I am “seriously skeptical” that insurance companies have a higher BMI cut off for men than for woman. In fact this cite claims the opposite for for both Prudential and TransAmerica - the “preferred rate” cut-offs translate to BMIs of 28.9 and 27.8 for males for the two carriers, and 29.8 and 28.3 for females. A higher BMI cut-off for women than men in fact. The opposite of what you claim. So let’s change that "seriously skeptical, to calling it out as an untrue statement. (I think what you are looking at, and pointedly NOT referencing, are the old and very outdated Met Life tables that date back to 1943 with an update in 1984. I do not believe your untruth is a lie; just very confused.) Allstate makes no gender differentiation. And I must question my first source, because Prudential’s actual pdf claims no difference based on gender with preferred for both up to BMI of 31 and up to 33 for non-smokers.
Yes, I think that Prudential and TransAmerica would be mistaken if they actually do give females more weight (that is, a higher BMI) before outgrowing the preferred rate, albeit they are consistent with the actuary article I linked to earlier. Given both a male and a female with BMIs of 29.1, Prudential would give the preferred rate to the female and not to the male (if the first ink was accurate). I know of no current scientific basis for that decision. TransAmerica would give the preferred rate to a female with a BMI 28.0 but not to a male with that BMI. Both sets of numbers from that cite imply that females take an additional 1 BMI point to be at risk from overweight. Again, I know of no basis for that. And again that is the opposite direction that you have claimed. They were likely wrong in 1943 when MetLife implied males were at significantly less risk for the same overweight BMI and they are likely wrong now if they imply the opposite.
You made some absurd claim about what the risk was for males and females at a BMI of 26. The data was actually collected for groups of 2.5 BMI units and when pointed out that the 95% confidence interval for the BMI containing 26 were almost identical for both genders, showing that the two groups are highly likely not significantly different you made some odd demand that I do the math, then doubted the data.
Your thesis is unfounded.
The sole support for your thesis, that insurance companies believe it, is untrue. Some may believe, marginally, the opposite, and *that *belief is without any scientific merit even though an actuarial article at least supports it.
So I’ll ask you: do you think the insurance underwriters who today conclude that males and females should get the same preferred rate for the same BMI, or possibly by some insurers’ reported assessments, that males must have a lower BMI cut off to get the same rate, have it wrong?
IF there really is any gender difference it is the opposite direction than what you have suggested. Such has not been consistently found however. To the degree it may exist it seems to vary with ethnicity.
The “overweight” BMI grouping is a very mixed bag and may be associated with different sorts of outcomes depending on many additional factors. Ethnicity and race, location of the fat, fitness levels, so on, all have impacts, some much more importantly than the BMI, especially for this group. In some studies of some groups slightly “overweight” was even associated with the least mortality.
BMI can be used to identify a population at greater risk to have unhealthy habits, and is useful for that, but even those obese by BMI (or in this study body fat percentage, which IS gender specific) are not at increased mortality risk if they are fit and consequently “metabolically healthy.”
Anyway, I have my own rules of debate and one rule is that when you claim someone said something, you have to back it up with a quote on request.
You have repeatedly claimed that the study you cited makes a particular claim but you refuse to back it up with a quote (or a calculation) even though I asked you multiple times.
I have no interest in taking part in such a discussion. This exchange is concluded.