Katherine Flegal Discusses the Prevalence of Obesity in the US
Interview From the Special Topic of Obesity, September 2010
Over the years, your most controversial article has been a 2005 JAMA paper—"Excess deaths associated with underweight, overweight, and obesity," (Flegal KM, et al., 293[15]: 1861-7, 20 April 2005). That's not among your most-cited papers, but it's certainly garnered a lot of attention. Can you tell us what you found and why it's controversial?
It doesn't surprise me that it's not as highly cited as these prevalence of obesity articles, because the prevalence articles are always really highly cited. Almost every article on obesity starts out with something like, "the prevalence of obesity is very high...", and they cite one of our articles. We just published two more of these articles in JAMA; they'll probably get cited, too. The high citation rates are for practical reasons.
That 2005 article was a project we undertook to look at the mortality associated with different weight categories. We used all the NHANES data that we had at that point—NHANES I, NHANES II, and NHANES III—and we had mortality follow-up, and we got some top-notch statisticians involved in this at the National Cancer Institute.
We defined our process and went through our process and got these estimates, and they were much lower than the previous estimates for the overweight category. We actually didn't find any excess mortality—if anything, we found slightly fewer deaths than expected.
So mortality in the BMI 25 to 29 group was less than, say, 20 to 25?
We estimated that there were fewer deaths in the overweight category (BMI 25-29) than there would have been if those same individuals had been normal weight. And overall our estimates for the number of excess deaths for obese and overweight people, relative to normal weight, were quite a bit lower than previous estimates.
We got a lot of publicity for this. A lot of people thought it was excellent and some people didn't like it, and the people who didn't like it were pretty vocal.
Why didn't they like it? Did they have data that disagreed with what you published?
You can't really say that they had data that disagreed. You'd have to ask them exactly why they didn't like it. It was just kind of a strong reaction. Previously there had been two other articles—by David Allison in 1989 and by Ali Mokdad in 2004—both of which had much higher estimates.
People just seemed to like the higher estimates. People seem to like high estimates of everything. They see it as something you can use to motivate people. Our estimates dropped the numbers really by a lot. That took people aback.
Have you followed up on that initial paper to see if these numbers have changed any?
We had another article in 2007 with a little bit more data, and with cause-specific mortality (Flegal KM, et al., "Cause-specific excess deaths associated with underweight, overweight, and obesity," JAMA 298[17]: 2028-37, 7 November 2007). It really matched the earlier results.
We found that the excess mortality associated with obesity is primarily from cardiovascular disease, which is no surprise. Not from cancer and not from the "everything else" category—non-cancer, non-cardiovascular. And, in fact, the reason why the overweight category has lower mortality is that it has lower mortality from all the non-cancer, non-cardiovascular causes—infections, injuries, and the whole group of miscellaneous conditions, of which there are hundreds.
"...we published an article in 1994 in JAMA showing a fairly large increase in the prevalence of obesity. The increase was eight percentage points so it was not huge, but it was noticeable and surprising."
You can see that there are differences by cause in the effects of weight on mortality. I think that's part of what the whole picture is. All-cause mortality is a very blunt instrument.
So according to your data, the price for excess fat, at least in terms of mortality, is only paid when you're actually obese?
Well, there's also what's called the obesity paradox, which is not a very good term. Let's say you have acute lung injury and you're in the ICU, research seems to show that the heavier you are, the better your survival is, up to a certain point.
There's quite a bit of literature on this so-called obesity paradox. You can have different weight levels and they can increase your chances of getting sick but improve your survival. There can be a balance between those two effects. There can be a tipping point where survival starts to increase because any disadvantages of getting more disease are counterbalanced by advantages.
What do you consider the most difficult or challenging part of doing the kind of statistical research you've been doing on obesity and overweight?
Well, if we're talking only about the prevalence of obesity, it's not too difficult. It's pretty straightforward. Oddly, though, one of the more difficult things about this is not to go beyond the data.
Journalists often try to elicit some sort of comment about how this is so important, or this is what I predict, or why this is happening. There are a lot of questions that come up all the time and I really don't know the answer to those questions. It's certainly a temptation to say, "Well, it's probably caused by TV watching" or something like that. But we really don't know. I don't.
So one difficult thing is not going beyond the data, trying to not over-interpret what we have. This is just a factual report of data on the US population. This is what it is. It's not saying anything else.
Katherine M. Flegal, Ph.D.
National Center for Health Statistics
Centers for Disease Control and Prevention
Hyattsville Maryland, USA
KATHERINE FLEGAL'S MOST CURRENT MOST-CITED PAPER IN ESSENTIAL SCIENCE INDICATORS:
Flegal KM, et al., "Prevalance and trends in obesity among US adults, 1999-2000," JAMA-J. Am. Med. Assn. 288(14): 1723-7, 9 October 2002 with 2,878 cites. Source: Essential Science Indicators from Thomson Reuters .
KEYWORDS: OBESITY, WAIST-HEIGHT INDICES, ANTHROPOMETRY, PREVALANCE, EPIDEMIC, BMI, DEFINITION, MORTALITY, WEIGHT CATEGORIES, HANES, CAUSE-SPECIFIC MORTALITY, OBESITY PARADOX.