Sunday, April 02, 2006

A Comedy of Errors

When the British Cattle Movement Service (BCMS) was set up in 1996 to record the births, marriages and deaths of all UK cattle in response to EU tracing requirements, little did those of us who have to operate within it realise just how skewed its information would become when let loose on the great and the good of the scientific world.

We have already told you of the '14 million' movements made by GB cattle - which are in fact movement of data relating to a single bovine movement in our post:
http://bovinetb.blogspot.com/2006/02/mistake-mischievious-or-misleading.html. But not content with this much publicised gaff, the 'ologists of Oxford with a little help from our very own VLA appear to have made another. (only one Matt? !!)

Matthew 1, has been trawling. No, he's not taken up fishing instead of farming - same minister, little point - he was searching the net (fishy terms again) for information on something totally different and came across a piece from 'Nature'. published in 2005.

Cattle movements and bovine tuberculosis in Great Britain
M. Gilbert
1, A. Mitchell2, D. Bourn3, J. Mawdsley2, R. Clifton-Hadley2 and W. Wint3

For 20 years, bovine tuberculosis (BTB) has been spreading in Great Britain (England, Wales and Scotland) and is now endemic in the southwest and parts of central England and in southwest Wales, and occurs sporadically elsewhere. Although its transmission pathways remain poorly understood, (
no they aren't - ed) the disease's distribution was previously modelled statistically by using environmental variables and measures of their seasonality1. Movements of infected animals (that would be all types of 'animals', we hope?) have long been considered a critical factor in the spread of livestock diseases, as reflected in strict import/export regulations, the extensive movement restrictions imposed during the 2001 foot-and-mouth disease outbreak2, 3, the tracing procedures after a new case of BTB has been confirmed and the Government's recently published strategic framework for the sustainable control on BTB4.

Since January 2001 it has been mandatory for stock-keepers in Great Britain to notify the British Cattle Movement Service of all cattle births, movements and deaths5.

Biological Control and Spatial Ecology CP160/12, Université Libre de Bruxelles, avenue F.D. Roosevelt 50, B-1050 Brussels, Belgium
Veterinary Laboratories Agency, Weybridge, New Haw, Addlestone, Surrey KT15 3NB, UK
Environmental Research Group Oxford Limited, PO Box 346, Oxford OX1 3QE, UK
Correspondence to: W. Wint3 Correspondence and requests for materials should be addressed to W.W. (Email: william.wint@zoo.ox.ac.uk).
Received 20 January 2005; Accepted 18 March 2005


Despite the extraordinary pedigrees of those involved with this 'research' - using those 14 million postcards - we would point out, (with the greatest respect of course) that ;

1. BCMS was set up in 1996, (not January 2001.)

2. Since 01/07/1996 (July 1996) all cattle have had to have passports which were issued by BCMS and which accompanied them on any movement, tearoff postcards from which are sent back to BCMS and sometimes generate 4 movements of data, even to an abattoir. And to generate a passport, first one has to register a birth...

3. Cattle identification at this time consisted of a herd number which was both letters and numbers, a 5 digit individual number plus a UK stamp.

4. In January 1998, double tags were made compulsory. A distance readable one (which falls out for a hobby) plus another in the opposite ear, which could be either plastic or metal.

5. On July 1st 2000, all numerical tags were introduced. A 6 figure herd number (no letters) plus UK, and a 6 figure idividual number, or 5 plus a revolving 'check' digit.

And so it has been ever since. No more changes. And animals recorded in that 1996 database, holding passports can now enter the food chain.

It is difficult reading the 'Nature' report to understand that, as the authors state quite authoratively that BCMS database started in 'January 2001'. But by this time Matt 5 had registered 419 cattle births on its non existent database and about the same 'Off' movements. He had no 'On' movements of bought in cattle on the nonexistent database - but was about to celebrate the New Year by going under restriction with bTb, just like Mr. Jones in our post below.

2 comments:

JM O'Donnell said...

Some of your objections seem to be based on a misunderstanding of the paper. Firstly, they are aware that the movement data has been taken from the BCMS from 1996, but they specifically use the data from 2000-2003 because it is the most reliable from what I understand reading the paper and the supplementary paper.

I'll also quote a portion from the supplementary materials published with the paper that might address some of the other issues you raise:

[b]For all events, except births and deaths, the BCMS receives two records: one for the location
the animal has moved off and the other record for the location the animal has moved onto. For any
meaningful analysis of movement patterns, “off” and “on” records have to be “paired.” This process
resulted in unresolved movement histories leading to the rejection of some animal’s movement
records, in particular in the earlier years of the CTS (this procedure is detailed at length in Wint et al10,
and in Mitchell et al.11). However, this significantly improved in the recent years with 79%, 70%, 75%
and 88% of animals born in the years 2000, 2001, 2002 and 2003 respectively having logical
movement histories.
The paired movement database from the period 2000-2003 was then queried to estimate the
following variables in each square: total number of inward movements, number of movements from
infected areas, and the proportion of movements from infected areas, and these layers were added to
the series of predictors obtained for each 1 km cell (see Supplementary Table 1).[/b]

Gilbert M., A. Mitchell, D. Bourn, J. Mawdsley, R. Clifton-Hadley and W. Wint. Supplementary material, Cattle movements and bovine tuberculosis in Great Britain (2005). Nature, 435.

The paper itself is rather unobjectionable as the computer model based on cattle movements does, sufficiently accurately, predict where BTB outbreaks in outlying areas of Great Britain should occur.

Incidentally, they also do use badgers as a key 'predictor' in their model for the presence of TB. They also hypothesise that some regions where imported cattle go from core 'infected' areas to uninfected areas without significant numbers of breakdowns are the result of a lack of wildlife reservoirs or rapid slaughter.

Matthew said...

Thankyou for taking the time to post that comment Mr. O'Donnell.

Perhaps we can be forgiven for misinterpreting:
"Since January 2001 it has been mandatory for stock-keepers in Great Britain to notify the British Cattle Movement Service of all cattle births, movements and deaths", ..when of course what the scientists who wrote the paper meant to say was "We have used data from 2000 - 2003 on BCMS' database which started in 1996."

That the scientists who wrote the paper factored in a 'paired' movement is encouraging. It is our experience that the leader of the ISG, Professor John Bourne - who oversaw the RBCT Krebs trial - has yet to make that quantum leap.

When the computer modelling was undertaken using these paired movements, there is no information as to whether extra 'movements' via markets and dead end hosts such as abattoirs were factored out. But we are encouraged that badgers were part of the predictor, and would be even more encouraged if Defra would release the number of 'closed' herds, (that is those with no 'On' movements of bought in cattle) who are under bTb restriction.

And you are absolutely correct with the observation in your final passage, that the movement of tb via an infected cow into areas with either no bTb, or a different strain, rarely lasts longer than the identification and slaughter of any cattle involved. In fact a spoligotype map of GB shows that in hundreds of cases, the strain of bTb in both badgers and cattle is unique to their indigenous area, with only very isolated cases of a 'foreign' spoligotype appearing out of its own patch. If cattle/cattle spread were significant, then one would expect the map to be very muddled indeed. (More on that in a later post).