Bias in Obesity Research

If you are overweight, that means that your probably of dying early is greatly increased, right? I’ll bet your doctor has told you exactly that (my VA doc has — he told me that all of the “obesity research” said so).

Researchers led by CDC senior epidemiologist Katherine Flegal conducted a meta-study of several major federal health studies done between 1976 and 2000, with a very interesting measurement criterion for the endpoint — namely, death. The intent was to determine how much influence excess weight has on all-cause mortality, so they took some care in isolating weight from other confounders such as smoking, alcohol consumption, and race.

When the researchers discovered that being overweight actually reduced all-cause mortality, the entire study was thrown out and started over with all new data from different studies from roughly the same time period. After all, everybody knows that overweight is unhealthy. They just assumed that something in the study was giving the “wrong” answer, and they just needed to find the “right” meta-data to use to reach the “correct” answer.

The second study showed the same results.

Two long-term, large, important, and arguably well-constructed studies which both reached the same conclusion, namely, overweight people live longer than people of “normal” weight.  Having a few extra pounds of fat on your body appears to be not only protection against several types of disease, but aids in the recovery from various types of injury.

How many of you have heard of either of these studies? When I told one of the “doctors” at the VA Medical Center in Dallas that overweight people live longer than skinny people, he flatly “informed” me that he knew better, and that the thinner you are, the longer you will live. I think one of the things that they teach in medical school is that a doctor must never admit to having made an error. Another item medical schools must teach is that you can get all of the medical knowledge you really need from pharmaceutical brochures and television commercials.

One reason you might not have heard of these studies is because they are really hard to find. Most of the links I have found to these studies have been orphaned, and the CDC site itself no longer lists the studies anywhere I could find. For some time after that, the CDC was still claiming that obesity is “responsible for 400,000 excess deaths per year,” a “fact” which has no basis in reality, and which has been repeated all over the internet.

However, I did find the 2nd study in JAMA April 20, 2005, Vol 293, No. 15. Here is an excerpt:

Relative to the normal weight category (BMI 18.5 to <25), obesity (BMI ≥30) was associated with 111 909 excess deaths (95% confidence interval [CI], 53 754-170 064) and underweight with 33 746 excess deaths (95% CI, 15 726-51 766). Overweight was not associated with excess mortality (−86 094 deaths; 95% CI, −161 223 to −10 966).

Digging a little deeper into that study, I found the following graphs which plot all-cause mortality versus BMI (click on graph for larger, more legible version):


Looking closely, you can see that all-cause mortality is lowest in the group with BMI between 25 and 30 in all three age groups. And the group that fared worst overall was the group with BMI < 18.5 — especially in the age group that I currently occupy.

Note the bias, even here! Can you see it? The “normal” group (BMI 18.5 to 25) is arbitrarily baselined as 1.0 relative risk. If you instead baseline the group from 25 to 30 as 1.0, the relative risk of “normal” weight would be a bit more obvious, and you would see more clearly that being obese (BMI 30-35) is only very slightly worse than being “normal.”

I recently learned from a friend about a very interesting blog (now inactive) called Junk Food Science which has a lengthy series of articles on this “obesity paradox.” The author of that blog has been the target of some fairly vicious attacks, so you might want to go read her articles before they get taken down by some spurious complaint.

About Mr. N=1

I am co-owner of, along with my wife, Georgene (Mrs N=1). Read more about me on Howard's about page.
This entry was posted in Diet, Health, Nutritional Research. Bookmark the permalink.

Leave a Reply