09/20/2007 09:53 pm ET Updated Nov 17, 2011

Give Medical Epidemiologists a Break

Whenever we start arguing over a healthy diet - and whether
science has anything credible to say on the subject - an image of the
witch in Disney's "Snow White" pops into my head.

She's holding out an apple. It's red, it's delicious, it's
reputedly good for you, except for maybe that eternal sleep thing. But
there's a very low probability of falling into a coma. In fact, science
would tell you zero probability since research has so far failed to
confirm that black magic can turn an apple into a potent sedative. So
go ahead, dearie, take a bite.

The whole question of what science can tell us about what to
swallow is, of course, more than fantasy. It's been getting an
intellectual workout this week, following a story in the most recent
New York Times Sunday Magazine, titled "Do We Really Know What's

To greatly simplify the case made by author Gary Taubes: health
studies in large populations, partly observational studies, are
complicated and produce uncertain results. Such uncertainties in
epidemiology often lead to erroneous conclusions. Taubes provides
exhaustive - and sometimes damning - detail to support those

As we know, the result is a medically induced mental whiplash -
learning that alcohol, coffee, salt, fat, chocolate and their ilk are
good for us, no bad, good, no partly bad, no partly good, no wait, no,
maybe good. Taubes does a thorough, even exhaustive job, of exploring
the ways that science has led us into this state of confusion. The most
consistent complaint about the story, at least from those studying
science communication, is that it is overly negative and, even worse,

Myself, I don't think science reporting means science boosting.
The negativity doesn't bother me. I just wish he'd been negative in a
more interesting - or maybe a more comprehensive - way. I'm not
denying that epidemiology is an uncertain science. But what isn't? It's
not just that we don't understand a healthy lifestyle or say a healthy
diet. We don't even fully understand the individual health components
of that diet.

To return to my original example of the apple - beyond the fact
that witch-poisoned fruits are probably bad - we're not even sure how
to evaluate this single fruit. As it turns out, I'm a journalist-style
expert on apples because I just researched them for the magazine,
Science & Spirit (upcoming November/December issue). So I can tell you that
apples are loaded with chemical compounds that appear to protect
against cell damage and against diseases linked to such damage,
especially cancer.

But, as it turns out, all apples are not created equal. Research
by food scientists at Cornell University shows that apples growing in
the sunny crown of the tree contain much more of the good chemistry
than those hanging in the shade. Other studies show variation between
species. Golden Delicious apples are less potent than Red Delicious
(which I personally regret because I'm not a Red Delicious lover).
Fuji apples look promising on the chemical scale, McIntosh less so. But
those apple comparison studies have barely started - meaning that no
scientist can tell a consumer which apple in the fruit bowl contains
the best chemistry or whether it would be better to replace it entirely
with, say, some cranberries.

Further, scientists are also unsure which, if any, of the healthy
chemicals consumers actually absorb. This question of bioavailability
is another important unanswered question in the realm of food and
health. How much of the good stuff goes into my body? How much does
gets absorbed by Gary Taubes? Unknown. Do the chemicals in apples boost
the healthy chemistry of onions? Does it matter when you eat the apple,
early in the day, late at night? You guessed it, more unknowns. "Most
of research papers related to health benefits of apples are based on
laboratory [testtube] or animal model studies, therefore, we are not
completely sure about the REAL health benefit of apples on humans,"
says Chang Y. "Cy" Lee, chairman of the department of food science and
technology at Cornell.

And that's just using apples as an example - one could make the
same case for any fruit, vegetable, or other component of a healthy
diet. Adds Lee, "We need a long term human clinical study with those
special foods and bioactive compounds. We need to know the up- and
down-expression of genes due to special foods and compounds. Until that
time, scientists as well as industry that related to human health must
be cautious on claiming any specific foods and compounds on health

So let's give the medical epidemiologists a break. But only a
short one. We need to encourage medical researchers to spend more time
exchanging information with food scientists, like Lee, and experts from
other down-in-the-field specialties. It's from that kind of
cross-disciplinary exchange of information - which occurs too rarely in
science - that the important questions are often answered. And I
suspect we'll need that and more if we really want to keep the witch
from the door.