creating and using facts in science

http://www.techcentralstation.com/042605E.html

They Don't Embarrass Easily

By John Luik Published 04/26/2005

A few years ago The New York Times ran a cartoon that showed two
Washington DC policy experts having a conversation. "In Washington the
search for truth is a creative process. First, you create a premise.
Next you create a statistic to back it up. Then you create an audience
by repeating it over and over again, until the media pick it up. That's
when you know that you've done it."

"Done what?"

"Created a fact!"


Just add Atlanta -- the home of the Centers for Disease Control -- to
Washington and you have a pretty good idea of how obesity science and
policy are made these days. Despite the fact that the CDC has been
caught out creating "statistics" to back up the "fact" of an obesity
epidemic, it appears to be neither embarrassed nor remorseful. As the
Associated Press reported:



"CDC Director Dr. Julie Gerberding said because of the
uncertainty in calculating the health effects of being overweight, the
CDC is not going to use the brand-new figure of 25,814 in its public
awareness campaigns and is not going to scale back its fight against
obesity."



So let's get this straight. When the CDC's numbers said that obesity
was overtaking tobacco as the leading cause of death, allegedly killing
400,000 Americans a year, it was all right to use that 400,000 figure
non-stop to scare us into losing weight. But when the real number
turns out to be just short of 26,000 then the CDC is so worried about
"uncertainty" that they aren't going to use the figure. Obviously, Dr.
Gerberding did not get her doctorate in logic.



The reason, of course, that Dr. Gerberding does not want to use the
supposed 25,814 deaths (remember, we're not even certain there are this
many) from obesity as the basis for a new campaign in the war on fat is
that, well, as things go, it just isn't terribly useful. Gerberding
knows that it is pretty hard to talk about an epidemic if you've only
got 25,000 victims, and without talk of an epidemic it's pretty
difficult to get the media's attention, Washington's money and, most
importantly, push through a host of coercive policy measures that tell
Americans what they can and what they cannot eat.



But notice, the good doctor isn't going to let a few hundred thousand
fewer deaths stop a good thing. Despite the lack of credible
statistics, we just "know" that obesity is still killing millions of
Americans -- even if we can't find where they live -- and there is no
reason to "scale back" the fight against obesity. Now that's what makes
the CDC and much of official Washington such scary places. Typically a
good public policy process first finds clear evidence of a substantial
problem and then goes about finding a policy solution. Evidence drives
and shapes policy. But here we have policy "creating" evidence. Even
though the supposed justification for intervening into the lives of
millions of Americans is now defunct, the claim that fat causes early
death is still made and the campaign against obesity still goes
forward.



Of course, what Dr. Gerberding was too modest to tell the Associated
Press is that the 25,814 figure is not destined to survive for long as
it is just too inconvenient. Already someone, somewhere, has started
the tedious but very necessary work of crunching some more acceptable
number that will bolster the "fact" of obesity=premature death back
into headline-grabbing territory. But just to make that enterprise as
difficult as possible, here are a few non-Washington statistics from
the last 50 years about the alleged dangers of being overweight and
obese.



First, take the study that really started the "obesity epidemic", the
1999 JAMA published research "Annual Deaths Attributable to Obesity in
the United States" which suggested that being overweight was
responsible for about 300,000 deaths a year in the US. If you look at
this data you find that individuals with BMI's of 25 -- overweight --
have a lower risk of dying prematurely than those with BMI's of 20 or
normal weight. Again, those with BMI's of 20 have the same risk of
premature death as those with a BMI of 30 -- obese.



Second, take another widely cited study, "Overweight, Obesity and
Mortality from Cancer" (New England Journal of Medicine, 2003) which
like the JAMA article generated huge headlines with its claim to
provide a definitive account of the obesity-cancer relationship. The
authors claimed that up to 90,000 cancer deaths a year could be avoided
if every adult kept his or her BMI below 25. Yet the data actually
shows something quite different. People with BMI's of 18.5-24.9 -
normal -- had a cancer mortality rate of 4.5 deaths per 1000 study
subjects while individuals with BMI's of 25-29.0 -- the overweight --
had a cancer mortality rate of 4.4 deaths per thousand -- actually
lower. In other words, for the 100 millions plus Americans who are
classified as overweight there was a negative correlation between being
overweight and dying from cancer.



Nor are these findings flukes. Ancel Keys, the proponent of the
cholesterol-heart disease theory, who over 25 years looked at fat and
mortality rates in Japan, Greece, Italy, Yugoslavia, the Netherlands,
Finland and the United States came to similar conclusions. Published as
the Seven Country Study in 1980, Keys found the risk of premature
mortality due to excessive weight increased only at the extremes of
over and under weight. A recent study looking at a group of Europeans
from his original data found men with BMI's of less than 18.5 -- that
is thin men -- had almost twice the mortality rate of either normal or
overweight men, even controlling for smoking. The study further found
that being overweight -- a BMI of 25-29.9 -- had no effect on mortality
and even those men who were obese still had lower death rates than the
thin men.



A similar pattern was found in a 1996 US study that re-analyzed data
from previous studies involving more than 600,000 subjects. The study
found white males with BMI's in the normal range of 19-21 had the same
mortality rate as those with BMI's of 29-31 -- overweight and obese.
Again, for non-smoking white males, the lowest mortality rate was found
with BMI's of 23-29, a range that includes the overweight.


Perhaps as we wait for the next set of statistics that purports to show
us dying by the thousands from our fat we can console ourselves with
the one genuine fact in this sorry mess which is that we have more to
be worried about in junk science than in junk food.

--

TC

.

Relevant Pages

  • Obesity Danger May Have Been Overstated
    ... calculation from the CDC. ... estimate 14 times higher: 365,000 deaths. ... The new analysis found that obesity - being extremely overweight - ...
    (sci.med.cardiology)
  • Re: Britains obesity death rate
    ... More people are dying in Britain due to being overweight or obese than ... Around one in every 11 deaths in the UK is now linked to carrying ... Experts also warned that the number of fatalities due to obesity may ...
    (misc.survivalism)
  • is fat the next tobacco?
    ... "Is Fat the Next Tobacco?" ... reported that about 300,000 deaths per year were associated with ... overweight and obesity, and warned that those health conditions might ...
    (misc.consumers)
  • Study Was Very Contentious, Figures Loosey-Goosey
    ... More Errors in CDC Obesity Deaths Study Says Center for Consumer ... CDC's Own Researchers Said Internal Debate Over Obesity Study Was ... Consultant to Ortho-McNeil ...
    (sci.med.nutrition)
  • Re: English immortality imminent, says top boffin, says Channel 4
    ... that they could scare you into buying a policy by forcing you to ... I'm sure Y & Y causes more deaths than secondhand smoke. ... very specialised and directed trolling. ...
    (alt.usage.english)