Monday, 24 January, 2005

Why are we fat?

A recent news story said that on average, American adults are 25 lbs lbs heavier now than the adults of 40 years ago.  That's a pretty sensationalist headline, evoking images of a nation in which every adult is 25 lbs overweight.  The media, as I've pointed out before, have a poor understanding of the term "average" in general, and they seem incapable of presenting a balanced story underneath their sensationalist headlines.  For example, some of that average weight increase is undoubtedly due to a reduction in malnutrition and also to an increase in average height.  If we all were an average of a foot taller, that 25 lb increase wouldn't be an issue.  The real issue is the percentage of people who are considered overweight or obese today compared to 1960.  But it's difficult or impossible for many reasons to make an accurate comparison of today's population with the population in 1960, so any comparisons you see are likely pulled out of thin air or somewhere that the sun doesn't shine.

More often, comparisons are made against 1990 for some reason.  The number tossed about these days varies depending on the source, but somewhere between 40 and 70 percent of American adults today are considered "overweight" or "obese".  The number who are considered "obese" (more than 20 lbs overweight) is reportedly between 10 and 30 percent.  That's roughly double the number from 1990 if you believe the reports, although again there are problems comparing today's population with 1990's population, the two biggest being changing criteria and detection bias.

Although today we have the Body Mass Index (BMI - more below), there was no generally recognized standard in 1990 for determining who was overweight or obese.  Different doctors used different criteria.  One could argue that the overs and shorts would balance--that is, that on average the doctors who had more stringent criteria were balanced by those who were more lenient.  It's a nice theory, but probably impossible to prove.  In any case, the lack of a generally accepted standard makes any comparison against 1990 somewhat suspect. 

The other problem with comparing today's population to 1990's population is detection bias:  we're seeing more fat people these days because we're looking for them.  If you go looking for something, you're almost guaranteed to find more of it than if you weren't looking at all.  Scientific studies in general, and medical studies in particular, often suffer from detection bias.  Another form of detection bias is trying to compare a population that you can see (today's Americans) with a population that you can't see (Americans from 1990).  It's possible to sample particular populations from historical records (military physicals, hospital admissions, etc.), but then you have to prove somehow that the sample is representative of the entire population.  That, too, is a difficult proposition.

The most widely accepted standard used today to determine one's "relative fatness" is the Body Mass Index, or BMI.  The BMI procedure calculates a height-to-weight ratio, and then uses that calculated number to put you into one of four categories:  "Underweight," "Normal weight," "Overweight," or "Obese."  Calculating the height-to-weight ratio is a useful measure of relative fatness, although it's not a definitive guide.  A body builder, for example, would be inordinately heavy for his height.  Still, the height-to-weight ratio is a good starting point.  In general, a higher BMI value is an indication of excess body fat.

My only real issue with the BMI standard is the grading scale, which I think is skewed towards an unrealistically thin ideal.  For example, my weight of 185 lbs and height of 5' 9" results in a BMI value of 27.5--halfway between overweight and obese.  Now I'll be the first to admit that I could stand to lose a pound or three, but I'd have to get down to 169 lbs to get out of the "Overweight" cagegory.  At 169 lbs, I'd be a lean, mean, biking machine with the physique of a professional bicycle rider.  That might have been realistic when I was 30 or 35, but certainly not today.  The BMI scale needs to be adjusted up a few points and probably adjusted for age as well.

That's my long winded way of saying that I don't put too much stock in the reported percentages of overweight or obese people, or in comparisons between today and 15 or 40 years ago.  It would be nice to know what those number really are, but I don't think that we can obtain them to any degree of certainty.  That said, it looks to me as though there are many more overweight and obese people today than there were 15 years ago, and I'm wondering why that is.  I'll be exploring that over the next few weeks.