Day 254 - 22 May 96 - Page 31
1 instance, when they allow for body mass index they may do
2 some statistical calculation which irons out what they
3 think is the element of fat intake in relation to body mass
4 index or not.
5 A. I think what they have tried to do in this table is to
6 relate the total energy intake from fat as a proportion of
7 the overall total energy intake, so they are trying to look
8 at only fat. Therefore, they have taken total energy
9 intake as a factor which must be taken into account when
10 trying to compare two groups of people. For example, it is
11 not only -- I think what they have tried to do is to
12 exclude fat from the total energy intake when looking at
13 this particular variable. Basically, this is a statistical
14 method which is, with all its criticisms, trying to look at
15 a specific factor in relationship to the development of, or
16 the risk of, patients developing breast cancer.
17
18 Q. It is one of a number of approaches?
19 A. Yes.
20
21 MR. RAMPTON: My Lord, can I intervene? I did notice something,
22 it is about this question of quintiles. What they have
23 actually done, but not how they have done it, is explained
24 on page 358 under the heading "Studies Specific and Pooled
25 Results". Then one sees that there is an asterisk besides
26 the word 'table 2' in parenthesis and then it says at the
27 bottom of the page, "See NAPS document number...", whatever
28 it is "five pages of supplementary material", which plainly
29 -- and then you have to send off to Grand Central Station,
30 "has not been included in the table".
31
32 MR. JUSTICE BELL: No. There we are. I mean, we can only work
33 for better or worse on what is in a particular paper, and
34 the witness's views of the particular paper among many
35 others.
36
37 MS. STEEL: OK. If we go to page 360 in the table we were
38 looking at there, the numbers of women with a less than 20
39 per cent of energy from fat is quite a small number is it
40 not?
41 A. That is correct, yes.
42
43 Q. It would not be very fair to draw any great conclusions
44 from the number that were in the study?
45 A. I think that is fair comment. What you actually see on
46 this are the confidence intervals, you know, the sort of
47 range and, as you can see, they are very wide. What they,
48 in fairness to them, have said in the article is that there
49 is no specific trend in any of these figures according to
50 the percentage of energy intake from fat. Even those women
51 having large intakes, there is no statistical trend to
52 suggest that they are at greater risk than women who have a
53 low intake. The black square is the, sort of, mean value.
54
55 If you look at women who have 25 to 30 per cent of their
56 energy from fat, their dock comes slightly higher up the
57 risk chart than women who have greater than 45 per cent of
58 energy from fat. Those numbers are reasonably large, there
59 is nearly 500 in one group and over 800 in another group,
60 and all they are saying is they cannot see a trend in any
