Day 034 - 11 Oct 94 - Page 10


     
     1        means (f), because it just comes before (g) in the next
     2        study and (f) is fried potatoes.
     3        A.  I would presume that is correct since (f) was not
     4        referred to in the chart, yes.
     5
     6   MR. JUSTICE BELL:  But if it is fried potatoes and it is in the
     7        vegetable fat column, is it fair to assume it means fried
     8        in vegetable fat?
     9        A.  That would be my conclusion, yes.
    10
    11   MR. MORRIS:  So we have one up on the Surgeon General in that
    12        we have spotted a mistake!
    13        A.  In the final of these four categories, the cohort
    14        studies, in which individuals are identified -----
    15
    16   Q.   Sorry, just before we move on from case-control, I count
    17        seven studies out of 11 that showed a positive
    18        association.
    19        A.  In the case-control studies?
    20
    21   Q.   Of fats of some kind, mostly animal fats; is that correct?
    22        A.  I find nine studies.
    23
    24   Q.   Well, some find two; the same study finds two
    25        correlations?
    26        A.  I see.  Yes, I am sorry.  I am referring to nine
    27        specific comparisons; you are quite correct.
    28
    29   Q.   The rest found basically no association but not negative?
    30        A.  Yes.  It is important again to understand what "no
    31        association" means.  That does not disprove the
    32        hypothesis.  It is, in the best of times, difficult to
    33        establish links between dietary factors and a disease such
    34        as cancer, which takes a long time to form.  It is also
    35        difficult to assess diets accurately.
    36
    37        When the studies show a statistical relationship, that is
    38        quite important.  When they fail to do so, it quite often
    39        is an error in method rather than meaning that there is
    40        simply no relationship between those two.  To draw on my
    41        analogy yesterday, it is rather like if 20 search parties
    42        go off in the woods and 15 of them are unable to find any
    43        evidence of a culprit or other individual there.  That
    44        does not mean there was no individual in the woods.  But
    45        if five do find evidence of a person's passage, the 15
    46        search parties that found nothing do not outweigh those
    47        five that did find clear evidence of it.
    48
    49   Q.   Perhaps in this situation it is a bit reverse because it
    50        is 15 that are finding it and five are not, in your 
    51        analogy? 
    52        A.  Yes.  When one looks at the evidence that we have 
    53        looked at so far, the overwhelming weight of the evidence
    54        is that there is a clear association of a direct nature --
    55        by "direct" I mean the higher the fat intake, the higher
    56        the risk of cancer.  It is found not in one research
    57        methodology but in three.
    58
    59   MR. JUSTICE BELL:  Just pause a moment.  If what you have said
    60        about "no association" as against where there is a

Prev Next Index