!! spellx done
180  Statistics And Law
     The Washington Post  November 29, 1969

     Part of the confusion about the "safety" of food additives and
other environmental risks is the gap between statistics and law. The
1958 (Delaney) revision of the Food, Drug, and Cosmetic Act required
that a food additive be tested and found "safe" for human consumption
before it could be certified for sale.  However, "safe" has never been
legally defined, and its interpretation is left to the discretion of
the Secretary of Health, Education and Welfare.  The act does make no
restriction, that nothing can be called safe if it causes cancer on
feeding to any animal.

Theoretical objections can be raised against this principle, but it 
has had the practical effect of forcing administrative actions in
doubtful circumstances where the public was bearing the risks of the
uncertainty.  For example, there is no certain proof at this time 
that cyclamate causes cancer in man.  However, industrial chemicals
like naphthylamine that are well known to cause human bladder cancer,
have a latent period averaging fifteen years before they show their
effect. Were we to wait another decade before pushing back on cyclamate
(which produces cyclohexylamine in the body), we might be committing
a million Americans to bladder cancer while this massive 
"clinical trial" was going on.

                Against Essentials

     It is also true that the Delaney clause might be invoked against
essential dietary constituents that could cause cancer in animals in
very high doses.  In fact, these may have something to do with an
irreducible burden of cancer that arises as a byproduct of normal
metabolism.  However, cyclamate is quite dispensable, and it is folly
to take such risks with it.

     The Delaney clause could indeed be modified for greater scientific
precision.  New agents should be tested for mutations as well as
cancer.  As we develop better theories of and experimental models for
other chronic diseases like hypertension and atherosclerosis, these
should also be covered.  Cancer has taken first place in this list 
mainly because we already have well-established animal tests for it.

     On the other hand, if a quantitative meaning can be attached to
ideas like "zero tolerance" and "safety" we might make the law more
flexible.  Cancer in an animal should be taken as a grave presumption
of hazard for man but not a final proof of it.  It is also that every
mouthful of food that any Americans eats for the next century will 
contain at least one molecule of DDT, so that a "zero tolerance" taken
literally would be absurd.

                Firm Standards

     The law must set firm standards of expected safety as a logical
basis of the tests that attempt to show whether the standard is met.
An additive surely should not harm as many as one per million of its
innocent  users.  This standard would be regarded as criminally lax if
we knew that 200 Americans a year were being killed.  On the other
hand, it is very difficult to get advance assurance about any additive
with such precision.  We can do experiments in animals with high doses
to look for possible trouble.  We can investigate the biochemical
mechanism of action.  We can try to find the formula that governs the
relationship of effect to dose.  If there is a presumption of hazard,
we might then still reply to it, and set tolerances for human use--but
only if there is a convincing argument that they meet the safety 
standard.

     Secretary Finch has recently announced the relaxation of his
categorical ban on cyclamate.  The current state of research put
cyclamate, as commonly used, far above the one-per-million standard;
but much more work is needed to understand the exact dimensions of the
hazard.  If another standard of safety underlies these administrative
judgments, this is a number that ought to be advertised on the labels.
------------------------------------------------------------------------