At 15:12 20/01/04 +0000, Tony Greenfield wrote:

>During the seventies there was a  multi-centre study into cot deaths in
>England and Scotland run from Sheffield.  The director was Professor John
>Knoweldon and the principal paediatric pathologist was Professor John
>Emery.  They have both died since.  I was the medical statistician
>responsible for analysing the data and worked closely with Knoweldon and
>Emery for two years:  1978 to 1980.
>I no longer have the data or reports but guess I could obtain them from
>the archives of Sheffield University's department of community medicine.
>One of the main findings was that the risk of sudden death of a child was
>increased if a previous sibling had died.  This is not necessarily
>genetic: more likely environment and management.  In other words, the
>probabilities of two deaths in the same family are not independent.

Indeed - but one obviously has to be careful when one moves away from
constitutional (i.e. genetic) factors contributing to SIDS to
'environmental' ones as reasons for the lack of independence - since
someone (and we have a few candidates available!) would be bound to claim
that a major reason for the lack of independence was that parents who had
killed one of their infants would be more likely to kill another!  As
valuable ammunition for the defence in such cases, one would need evidence
of the lack of independence in cases which were 'definitely not' cases of
parental murder - really impossible to establish, since we know so little
about most of these deaths that we cannot rule out parental
murder.  However, if one could establish the existence of constitutional
(or non-homicidal environmental/management) reasons for non-independence,
that would help a lot.

>The evidence by Professor Meadows,  in which he presented a very small
>probability (one in 76 million, I think) of two cot deaths in the same
>family, is totally absurd. It struck me as such when I first heard of it
>and I, unforgivably, assumed that the lawyers and their advisers would
>understand that.

Indeed - it's absurd and statistically naive - I presume probably as
simplistic as the square of the probability of a single cot death.  As you
say, it is extraordinary that this suggestion was not effectively (or at
all?) challenged by the defence, who presumable will have had statistical
advisors - so their seems to be legal incompetence at work here as well.

Kind Regards,


John

----------------------------------------------------------------
Dr John Whittington,       Voice:    +44 (0) 1296 730225
Mediscience Services       Fax:      +44 (0) 1296 738893
Twyford Manor, Twyford,    E-mail:   [log in to unmask]
Buckingham  MK18 4EL, UK             [log in to unmask]
----------------------------------------------------------------

******************************************************
Please note that if you press the 'Reply' button your
message will go only to the sender of this message.
If you want to reply to the whole list, use your mailer's
'Reply-to-All' button to send your message automatically
to [log in to unmask]
*******************************************************