On Tue, 14 May 2013, Brian Wood wrote:

> Mike Miller:
>
>> On Tue, 15 May 2012, Brian Wood wrote:
>>
>>> http://www.telegraph.co.uk/health/healthnews/9260091/Forget-BMI-just-measure-your-waist-and-height-say-scientists.html
>>
>>
>> Thanks, Brian!  I did a little searching around for info about 
>> waist-to-height ratio (WHR) and how it compares with body mass index 
>> (BMI) as a predictor of various disease outcomes and mortality for 
>> different age groups, etc.  It looks WHR does not work better for 
>> children.  But it may be better for predicting cardiovascular disease 
>> in adults.  The correlation of WHR with BMI is about .9, though, so 
>> they aren't very different.
>>
>
> This seems to indicate there is a difference:
>
> http://www.telegraph.co.uk/health/healthnews/10054519/Waist-to-height-ratio-more-accurate-than-BMI.html


I wasn't saying that there was no difference, just that it is a small 
difference, and that the literature I reviewed at the time found the small 
difference for adults and not for children.  A woman named Ashwell, cited 
in the newspaper article you shared, has links to her papers here:

http://www.ashwell.uk.com/publications.htm

The most recent one, Obes Rev 2012;13:275-286, claims to show in 
meta-analysis that "Compared with BMI, WC improved discrimination of 
adverse outcomes by 3% (P < 0.05)" which isn't exactly overwhelming in a 
total sample of 300,000 people.  Like I said, they correlate about .9, so 
they are almost the same thing.  The advantage of WtHR is probably that it 
doesn't mislead when people vary a lot in muscle mass.  A disadvantage is 
that the waist isn't very easy to measure.

WtHR will not replace BMI because it requires that the waist be measured, 
which isn't even as easy as it sounds, and waist measurement is not 
available in many data sets where height and weight are available (e.g., 
drivers license records).  For research where data is being collected from 
individuals (not from publica databases) and the individual has to go to a 
nurse or doctor for assessment, there are even better ways to measure 
adiposity than either WtHR or BMI.

Mike