Day 022 - 12 Sep 94 - Page 46


     
     1        that kind of very long history, if you are going to try to
     2        identify when they started and why they started?
     3        A.  I think almost certainly most cancers have a long, as
     4        we say, gestation period.
     5
     6   Q.   It is not like a virus that you either get in a week or
     7        you do not?
     8        A.  No.  You have introduced viruses.  I mean, virus
     9        infections may be a responsible factor in the development
    10        of some cancers.
    11
    12   Q.   That was not quite the point I was making.  All I mean is
    13        cancer is always going to be some kind of long process?
    14        A.  Usually, I think that is right, yes.
    15
    16   Q.   So really for people doing cohort studies, there is a
    17        great responsibility on them to do it over a substantial
    18        period of time for it to have any significant evidential
    19        result?
    20        A.  Of course, this is one of the criticisms of case
    21        controlled studies is that, you know, when you start
    22        looking at the controls, you take a healthy group of
    23        people who do not have the illness, you are comparing
    24        people with the illness, right?  You have actually
    25        identified a population who have, say, breast cancer; you
    26        then select out a group of people who do not have breast
    27        cancer and see what happens to them, OK?
    28
    29   Q.   An equivalent population?
    30        A.  Yes, and because they are healthy, you know, again it
    31        may be some time, you have this health delay factor (as it
    32        is called) where people, you know, you might see a lower
    33        incidence of breast cancer in the group that you have
    34        selected because they are actually healthy at the time you
    35        start to study them.  So, this can give misconceptions
    36        about cancer incidence in this new group that you have
    37        started to look at compared with the group you have taken
    38        who have got the disease and, therefore, are a group from
    39        the past.
    40
    41   Q.   Yes.  Is another area of problems with cohort studies, and
    42        probably case controlled studies as well, the accuracy of
    43        actually recording people's diets or people recording
    44        their own diet even probably more of a problem?
    45        A.  I think with all of these epidemiological studies, be
    46        they prospective or retrospective, one of the major
    47        difficulties is identifying diet.  This is, when we talk
    48        about methodology, most of the authors in all of the
    49        publications state that this has been a real problem for
    50        them is to identify it because, you know, they talk about 
    51        average portion size.  Now, what does an "average portion" 
    52        mean, and things like that.  There has been real 
    53        difficulty in identifying the actual diet that people
    54        eat.  People's recall of diet is very poor.  I mean, I am
    55        not sure that I, without a great deal of effort,
    56        can remember what I ate a month ago, for example.
    57
    58   Q.   Yes.  I cannot remember if I ate anything at all!
    59
    60   MR. JUSTICE BELL:  I was going to say, if you were asked what

Prev Next Index