November 24, 2004
Data on Deaths From Obesity Is Inflated, U.S. Agency Says
By GINA KOLATA

he Centers for Disease Control and Prevention says that its widely
publicized estimate that 400,000 Americans die each year from being too fat
is wrong and that it will submit a new, lower figure to the medical journal
that published its original estimate last March.
It will be a simple correction of an inadvertent calculation error, the
centers said.

But that figure of 400,000 deaths has taken on a life of its own. Those
concerned about obesity cite it to show that being fat is almost as bad as
smoking. And tobacco opponents attack the number, saying it is grossly,
possibly purposely, inflated.

Yesterday, The Wall Street Journal reported that the centers planned to
revise the estimate and had undertaken an internal review of the study,
published in The Journal of the American Medical Association.

The estimate originated with an effort by the C.D.C. to determine the
effects of the nation's growing number of overweight and obese people.

The result was a paper by Dr. Ali H. Mokdad and colleagues published in the
medical journal that was controversial from the start, with some of the
most vigorous attacks coming from antitobacco groups.

But beneath the dispute lies a truth about science and its uses in an age
of limited resources.

If obesity is a leading cause of death, more money should be spent to try
to prevent it, and to treat it. Insurers should pay for diet programs or
weight-loss surgery. If it is not so deadly, some fear, the impetus to pay
for such things might vanish.

On the other hand, the more money that goes to fighting obesity, the less
will be available for other programs, like antitobacco ones.

"The tobacco people are afraid that it's a zero sum game," said Dr. Eric
Oliver, a political scientist at the University of Chicago who is writing a
book in the politics of obesity. "If obesity gets declared Public Enemy No.
1, it's going to come at their expense."

Now, with disputes and reviews under way, it is by no means certain what
the final, official consensus on obesity deaths will turn out to be. The
correction to the original article may be just the beginning as the agency
has now asked for outside advice on how to do the calculations.

In response, said Dr. Dixie Snider, the chief scientist at the centers, the
Institute of Medicine will have a meeting next month to discuss the
appropriate statistical methodology to calculate obesity deaths. "One of
the things we're trying to do is not to come up with any one person's
favorite method," Dr. Snider said.

But Dr. Snider insisted that the paper in question was actually of only
minor importance in the nation's fight against fat.

"We regret that we published a number that in retrospect was incorrect,"
Dr. Snider said. "And we regret the confusion. But the underlying message
does not change. The paper makes clear and the erratum makes clear that
tobacco and obesity are the two biggest killers."

Others disagreed, often forcefully.

"The kind of policies one would develop for something that is killing about
as many people as tobacco or a quarter as many people as tobacco are very
different," said Dr. Stanton A. Glantz, professor of medicine and director
of the Center for Tobacco Control Research and Education at the University
of California, San Francisco.

Dr. Glantz estimates that the number of deaths from obesity to be more like
100,000 than 400,000. And the inflated numbers of obesity deaths, he added,
represent " a very, very fundamental mistake that was made in the paper,
which they have done nothing to address."

"This is not some esoteric little detail over which there is huge
uncertainty," he said.

Others, who are not part of the antitobacco movement, agreed with Dr.
Glantz that the 400,000 figure made little sense.

Dr. Oliver, for example, said obesity, like tobacco, had little effect on
mortality in people over 65. So with two million deaths a year in the
United States, 70 percent of which are among people over 65, virtually
every younger person who dies would have to die from obesity. "The numbers
simply don't add up," he said.

That is the same argument made by two statisticians at the disease control
agency, Dr. David Williamson and Dr. Katherine Flegal, who published papers
this year reporting that the statistics used to calculate the obesity
deaths were wrong.

They noted that the way to estimate deaths from obesity was to look at each
age group and ask how many deaths might be due to obesity and then add the
numbers to get the total deaths. That is the way the agency calculated
tobacco deaths, coming up with a figure of 435,000. But for obesity, the
agency looked at the death risk in younger people and extrapolated it to
the entire population.

In response to the obesity analysis, the Office on Smoking and Tobacco at
the centers circulated its own analysis of tobacco deaths. Using the same
methodology as had been used for obesity, it came up with about 640,000
deaths from tobacco.

Dr. Snider says he knows about that tobacco office analysis and is aware of
that criticism of the mortality estimates. But, he said, he does not think
that is the entire answer to getting the statistics right, and he is
counting on the Institute of Medicine Review to provide guidance.

But that institute is not going to satisfy someone like Dr. Glantz.

Antitobacco advocates are furious, he said, and he hears from them
regularly. But others, too, are outraged, Dr. Glantz said. When he
mentioned his criticisms of the centers' statistics at a National Cancer
Institute meeting on tobacco control, he said, "it was like turning on a
spigot."

The Institute of Medicine meeting, he added, "is more politics."

Copyright 2004 The New York Times Company | Home | Privacy Policy | Search
| Corrections | RSS | Help | Back to Top

-------------------
Problems/Questions? Send it to Listserv owner: [log in to unmask]


To unsubscribe, send the following message in the text section -- NOT the subject header --  to [log in to unmask]
SIGNOFF SDOH

DO NOT SEND IT BY HITTING THE REPLY BUTTON. THIS SENDS THE MESSAGE TO THE ENTIRE LISTSERV AND STILL DOES NOT REMOVE YOU.

To subscribe to the SDOH list, send the following message to [log in to unmask] in the text section, NOT in the subject header.
SUBSCRIBE SDOH yourfirstname yourlastname

To post a message to all 1000+ subscribers, send it to [log in to unmask]
Include in the Subject, its content, and location and date, if relevant.

For a list of SDOH members, send a request to [log in to unmask]

To receive messages only once a day, send the following message to [log in to unmask]
SET SDOH DIGEST

To view the SDOH archives, go to: http://listserv.yorku.ca/archives/sdoh.html