Memorandum submitted by Professor Denis
Mollison
COMMENTS ON
THE CHIEF
SCIENTIFIC ADVISER'S
REPORT, "BOVINE
TB IN CATTLE
AND BADGERS",
PUBLISHED ON
22 OCTOBER 2007
I write as the Independent Statistical Auditor
of Defra's Randomised Badger Culling Trial (RBCT).[3]
The Final Report of the Independent Scientific
Group (ISG) on the RBCT was painstaking, expert and balanced,
and I commended it to Ministers as an exemplar of how to bring
high quality science into public decision-making. The ISG's main
modelling and statistical analyses have been published in the
highest quality peer-reviewed journals, such as Nature and
the Proceedings of the National Academy of Science.
In stark contrast, the Chief Scientific Adviser's
comments published yesterday, as a report Bovine TB in cattle
and badgers, would not have passed my audit. It is unbalanced
and inexpert.
First, one has to question why he had to assemble
an "expert group" to comment on the ISG's Report at
all, when its analysis was of such quality and its conclusions
so clear. It undermines the concept of independent scientific
advice when it is filtered in this way.
Next, his report is mainly concerned with detailed
discussion of complex statistical modelling and data analysis
issues. Yet neither he nor any of his "expert" group
have expertise in this area comparable with Professors Cox and
Donnelly of the ISG.[4]
One consistent theme of Sir David King's report
is its attempt to play down or dismiss the ISG's findings of detrimental
effects near the edge of culled areas. These range from picking
apart statistical fine details in a textbook example of special
pleading (para. 43), to their wishful thinking that disruption
(under a continuing culling programme) may be transient (para.
37). They even suggest there may still be a role for the reactive
strategy (para 50), contrary to the conclusions of the ISG accepted
by Ministers in 2003.
Their epidemiological analysis is muddled in
its discussion of the basic concept of R0 (paras 9-11 and Annex
2); and on a substantive point, the results of Cox et al indicated
an R0 of about 1.1 not for cattle-to-cattle but for the whole
cattle-badger system, suggesting that cattle controls on their
own might well be sufficient to reduce this overall transmission
parameter R0 to a value less than 1, and thus bring the disease
into decline.
Nor do they discuss the economics: the ISG concluded
that the proactive strategy, if carried out over large enough
areas (of the order of 300 sq km), should have a beneficial effect,
but estimated this not to be cost-effective.
Finally, the language used is not always careful.
For example, "Strong action needs to be taken now" (para
4) sounds good, but if this refers to badger culling, the ISG
report makes clear that in many circumstances "strong action"
would be likely to make matters worse.
October 2007
3 My final audit report can be found at http://www.defra.gov.uk/animalh/tb/isg/pdf/statsaudit_rpt0607.pdf Back
4
Professor Sir David Cox is of particular international distinction,
including for his many books and papers on statistical modelling,
and for his integrity and the care and balance he brings to any
argument. He is a past President of the International Statistical
Institute (http://www.stats.ox.ac.uk/people/academic_staff/david_cox).
Professor Christl Donnelly is internationally recognised for her
experience and abilities in the detailed statistical analysis
of infectious disease data, and is also most careful and balanced
in presenting her conclusions (http://www1.imperial.ac.uk/medicine/people/c.donnelly/). Back
| 
