Memorandum submitted by Dr Fiona Mathews,
University of Oxford (J12)
I would like to submit the enclosed paper for
consideration by the Agriculture Select Committee. This raised
serious concerns about the statistical power of the badger cull.
I was alerted to these issues some time ago, when I discussed
the design of my own project (an investigation of the Risk to
Cattle from TB in wildlife other than badgers (SE3009), MAFF,
£960,051) with members of the Independent Scientific Group
(ISG). My figures show that it is probable that the trial, as
it currently stands, has statistical power of only 50 to 60 per
cent to detect a reduction in TB incidence of 20 per cent within
five years, not the 90 per cent power assumed.
It is appropriate to draw to the committee's
attention the fact that the enclosed paper has not yet been published
in the scientific press. It was recently submitted to a leading
academic journal. Although five of the six referees agreed that
the points raised were fundamentally correct, the editor concluded
that I would find myself "considerably outgunned" in
what might be very public fora. In addition, several referees
raised concerns about the use of simulated rather than real data
on breakdowns in the trial areas to illustrate the points made.
The summarised data published to date are unsuitable for sample
I therefore respectfully request that the committee
1. What is the minimum reduction in TB
incidence that the ISG consider important to detect, in both proactive
and reactive areas, and within what timescale? (The Krebs
Report cites a 20 per cent reduction within five years).
2. What reduction in TB incidence do
the ISG expect to detect, given the data available to them, in
both reactive and proactive areas?
3. What do the ISG consider is the power
of the trial to detect such reductions?
4. Why are no trial areas located in
Wales? The trial zones are meant to be a representative sample
of all areas with high rates of repeat and contiguous breakdowns
in order that the results can be generalised. Two areas of Wales
fall into this category but are not being sampled. The Policy
Unit of MAFF have denied anecdotal reports that the Countryside
Commission for Wales declined to approve the trial.
5. Will MAFF now allow historical data
on the incidence of TB in the trial areas (prior to the commencement
of the current trial) to be made publicly available? It is
important that such an anonymised database permits the identification
of repeat and contiguous breakdowns.
6. Will MAFF commission the collection
and publication, by independent observers, of data on the survival
of badgers within trial areas (including sites where permission
for culling was denied)? Considerable badger survival is widely
reported by local Badger Groups. This contrasts with information
released by MAFF. Neither source of data has been independently
confirmed. Badger survival clearly has important implications
for the extent to which TB incidence in cattle can be expected
7. Will the Veterinary Laboratories Agency
release historical data on the proportion of badgers, culled in
previous badger removals, found to have open lesions at post mortem?
These data, indicating the prevalence of infectious individuals,
are important to build better models of the epidemiology of TB
in badgers and cattle, and assist in understanding the likely
impact of the cull.
8. What, explicitly for both proactive
and reactive areas, is the mechanism being tested in the badger
If the aim is to reduce TB in cattle by removing
all (or most) infected badgers in particular areas, then the completeness
of the cull is an important issue. Further, it is difficult to
see how a "reactive" cull could be effective, unless
the trial is primarily concerned with repeated breakdowns at the
same or adjacent site. It is important in this case to note, as
discussed in the attached manuscript, that repeated and contiguous
breakdowns are not independent of each other and pose particular
difficulties for sample size calculations and analysis.
Conversely, if the aim is to reduce badger density
generally in trial areas to below a threshold level, such that
transmission within the badger population is reduced and TB maintained
at a low prevalence, then there are important policy implications.
Badger density would need to be maintained at low levels for long
periods. This scenario contrasts with the information I understand
was presented to the Bern Convention by the Central Science Laboratory,
which suggested that the culls would not depress badger density
in the longer term. Finally, unless there are a great many culls
within the reactive area, the potential for lowering badger density
is, in any case, limited.
27 October 2000