"Joel Rosart" wrote:
>
> Hi Keith,
>
> Rob talks about epidemiological studies on page 54 of NHE. As he points out,
> numerous studies have correlated fat intake with cardiovascular disease. But
> if any of us thought that these studies were true, then we wouldn't be
> following NHE! The problem of course is that all of these populations also
> have high sugar intakes, so its impossible to find the real culprit based on
> these studies.
>
> So an epidemiological study is pretty useless on its own, especially when a
> researcher's bias can determine what conclusions are made.

Joel, sorry for my tardy reply. As I read it, you are not criticizing
epidemiological studies as
such, just the poor methodology employed by some epidemiological researchers.

I agree with you that it's easy to link food intake of cholesterol or saturated
fats with heart
risk if you leave out all other aspects of diet and lifestyle. Another problem
is that lifestyle
diseases take many years to manifest themselves and so brief experiments are
bound to
miss important variables. Self-reporting is another problem (most people lie
about their
food intake, either deliberately, through forgetfulness or ignorance).

But to me this means that the epidemiological studies have been sloppy and that
better
research design, within the epidemiological paradigm, would be the best
solution. In fact, I
can't think of any better research design. Can you?

Keith