The scale of the slaughter in 2001 UK foot and mouth epidemic was unprecedented, the true scale of the epidemic is still unknown.
4.2 million adult animals were slaughter on 10509 premises, a further 2 million animals were slaughter for welfare reasons
less than 15% of the animals culled were on premises where foot and mouth virus or antibodies were isolated at lab. test. The Institute of Animal Health-Pirbright- maintain that lab. tests are at least 90% accurate.
80% of the infected premises with positive test results are in Scotland or in the north of England
in the first 5 weeks of the epidemic only 11% of infected premises were slaughtered within 24hrs. 33% were slaughtered within 24 hrs overall.
Only 4.6% of contiguous farms were slaughtered within 48hrs, 29% were slaughtered within 72 hrs.
In 8 counties all or most of the contiguous culling occurred around infected premises with negative test results or no test results.
A confirmed negative laboratory test result is usually available within 48 hrs, on occasions it may take 96 hrs. 50% of contiguous farms were not slaughtered within 96 hrs.
DEFRA will not reveal how 88% of infected premises in this epidemic became infected because if they did it would enable us to understand the epidemic.
If we understood the epidemic we could determine1. The true scale of the epidemic2. Whether the adopted culling policy was appropriate3. How many premises were culled illegally andunnecessarilyAnd more importantly we would ensure, that the rightlessons are learned for the future.The aims of this report are:-- to explain why data regarding disease spread is not available- to present published epidemic data and anecdotal evidence- to demonstrate that all the published evidence indicates that the culling policy used in the epidemic was flawed.
If this culling policy remains part of the UK contingency plan all epidemic data must be made available to allow independent researchers to evaluate this policy.
Understanding the 2001 UK foot and mouth epidemic
To be able to understand the UK 2001 foot and mouth epidemic it is essential to establish how the disease spread from farm to farm: that is disease transmission.
None of the Inquiries have investigated disease transmission they have all accepted DEFRA's explanation that: 78% of infected premises became infected as a result of local spread; 10% are under investigation and the remainder were infected by animals, vehicles, people, windborne or other (source: DEFRA's epidemiology report of 21 Oct 2001).
DEFRA will not reveal the precise method or methods of spread for 88% of premises in this epidemic because if they did it would show:-
1. The true scale of the epidemic
2. Whether the adopted culling policy was appropriate
3. How many premises were culled illegally and unnecessarily
1. The true scale of the epidemic
Of the 2026 infected premises, 1324 have positive laboratory tests, 401 have negative laboratory tests and 301 were untested (JCC data analysis DEFRA).
Given that laboratory testing is at least 90% accurate (Pirbright) we can say for certain that:in this epidemic there were animals on 1324 farms which had foot and mouth disease.
If we knew how these 1324 premises became infected we could almost certainly establish:-
a) whether the 401 negative premises and the 301 untested premises were infected.
b) the true risk of exposure faced by the 8226 pre-emptively culled premises: that is, contiguous and dangerous contact premises.
2.Whether the adopted culling policy was appropriate
If it can be shown that the disease was spreading by the movement of animals, vehicles and people then traditional methods of control: rapid slaughter of infected premises; accurate tracing and rapid slaughter of dangerous contacts; movement restrictions and strict biosecurity; properly implemented would have brought the epidemic under control.
The only justifications for embarking on complete removal of all premises within a certain radius of an infected premises are:
a) traditional methods, implemented properly, are not working
b) it is not known how the disease is spreading and any investigations are inconclusive
c) there are no restrictions on physical or financial resources
3.How many premises were illegally culled
a) Powers of Slaughter
Under domestic law it is illegal to cull animals which cannot be shown to have been in contact with a source of infection or been in any way exposed to foot and mouth disease.
Margaret Beckett 12 Nov 2001 : Column 577 Animal Health Bill
The current powers have proved insufficient in a number of cases. In stamping out the disease, it has for many years been the accepted policy approach to kill infected animals. However, it has also long been regarded as necessary to cull animals on contiguous premises as well as on infected premises, in order to protect against the spread of disease. Although that has been established policy, there has always been an issue of when animals could be culled. As required by existing legislation, only animals which have been "exposed" to FMD can be culled. Inevitably, when that is the criterion, there are liable to be, and have indeed been, disputes about whether animals have been in contact with or exposed to the disease.
b) Confirmation of disease and Laboratory testing
Domestic law does not require laboratory testing to confirm disease ( ie a clinical diagnosis will suffice) nor does it require the testing of pre-emptively culled premises.
EU law (directive 85/511/EEC- the minimum community control measures) requires the testing of animals clinical infected, suspected of being infected or exposed to disease. A negative test rules out the presence of disease; a positive test confirms disease (Articles 4 (1) and 5(1))
The only exemption to testing is for clinical infected animals found on a holding, where that holding can be epidemiologically linked to another holding where disease has already been confirmed at laboratory test (Article 5(3)).
Therefore, under :-
Domestic law - where a veterinary surgeon confirms that animals are clinically infected with
FMD he will declare that the premises on which they reside is an infected
place, a pre-emptive cull around this infected place can occur if, there are
reasonable grounds to suggest that, contact or exposure has occurred and
this contact or exposure occurred at a time when the infected animals were
likely to have been infectious. Such a cull is legal and necessary.
- any farms which were pre-emptively culled due to their contact or exposure
to farms where FMD had been ruled out, by virtue of a negative laboratory
test result, were culled illegally and unnecessarily ( 1700-3000 farns).
- any farms which were pre-emptively culled due to their contact or exposure
to farms which had FMD but where contact or exposure did not occur
were culled illegally and unnecessarily( up to 5000 farms).
EU law - all 8226 pre-emptively culled premises should have been tested to see if
contact or exposure had occurred. The timing of this test, relative to the time
of slaughter, is not stipulated and could occur at slaughter: obviously if taken
at slaughter the main benefit is to show whether the pre-emptive culling
policy is being accurately targeted.
Published Evidence regarding Disease Transmission in 2001 UK FMD epidemic
1. Epidemic Models and Expert Advice
a) Mathematical Model
The mathematical model produced by Imperial College recommended:-
- all infected premises should be slaughtered within 24 hrs
- all contiguous farms should be slaughtered within 48 hrs
- for every 1 infected premises it was predicted that up to 6.7 contiguous premises would need to be pre-emptively culled to bring the epidemic under control
b) Biological Model
The InterSpread model was developed in the early 1990's and has progressively been refined The model has been used for FMD policy evaluation in Europe. The model was used by MAFF/VLA to analyse control strategies being considered March 20, 2001, and when a more comprehensive national farms database became available in early April the control strategies were then re-simulated using case reports to April 10, 2001. The only results which appear to be in the public domain are those for the second set of model runs
(Veterinary Record, August 4 2001).
The InterSpread model used by MAFF/VLA recommended:-
- slaughter out of infected premises within 24 hours, a predicted slaughter of about 1.3 of the surrounding farms ( high risk farms) per infected premises within a further 48 hours, and
minimal interfarm movements of susceptible animals to bring the epidemic under control
c) Opinion from Pirbright
Drs Donaldson and Kitching both expressed grave doubts about the mathematical models: that is, the Imperial College and Cambridge models. Their main concerns were:-
1. The species being model were hypothetical yet initial the disease was confined to sheep
2. The quality of the input data was poor 'the quality of the input data - it relates primarily to the role of sheep. It is very hard in certain circumstances to know what the lead in time has been on certain infected premises. There has been a trend for sheep to be infected silently and then to infect cattle, and cattle are the indicator species. But putting a time frame to the period from the recognition of the disease in cattle back to the introduction of the infection of the sheep is, I would suggest, very difficult, and could have a wide amount of variation. ' (EFRA select committee 25/04/01 p67 413)
3. Dr Donaldson could not see how this epidemic was any different than any other and therefore could/should be controlled in the traditional way.
4. Clinical diagnosis in sheep is very difficult and they were of the opinion that laboratory back up for an accurate clinical diagnosis in sheep was essential.(Vet. Record May 19 2001)
The published evidence suggests that:-
1.All the experts agreed that rapid slaughter of the infected premises within 24 hrs was essential
2.Neither Pirbright nor MAFF/VLA agreed with the pre-emptive cull policy as recommended by Imperial College team:
-Pirbright have publicly disagreed with the Imperial College pre-emptive cull policy
-the InterSpread (MAFF/VLA) model appears to recommend a pre-emptive policy 4 to 5 times smaller than the Imperial College model.
-Dr David Shannon, former chief scientist at DEFRA, was reported by Robert Uhlig (19/02/2002) as saying that the slaughter policy was based on flawed, biased and poorly thought through scientific advice.
2. Epidemic data
A The epidemic data used by the Mathematical Models
Assumptions made by the modellers( comments in brackets explain, why these assumptions were wrong )a) The first 70-80 infected premises all became infected pre-movement banb) Infected premises identified after about 3/03/01became infected after the movement ban was imposed 23/02/01 and were primarily due to local spread( we now know from the tracing project completed by the VLA that assumptions a) and b) did not reflect reality: there were many infections which occurred pre-movement ban which were not identified for weeks, possible months (that is after 3/03). Non identification of infection occurred because the disease can move silently through sheep)c) Confirmation of infected premises on clinical grounds only, did not start until late Marchd) The majority of infected premises identified up to 29/03/01 were infected with FMD ie misdiagnosis was minimal(assumptions c) and d) did not reflect reality. If the modellers had known that 90% of premises were being diagnosed on clinical grounds alone by 21/03/01, they would have known that the potential for misdiagnosis, particularly in sheep, was significant. Misdiagnosis causes 2 problems it gives the impression that the disease is spreading when it isn't and it is very difficult to explain how it is spreading since a misdiagnosed case, by its very nature, often has no identifiable source of infection.)e) delays in slaughter on the infected premises were down to 1to 1.5 days by end of March.( delays in slaughter to the end of March were still significant, 58% of IPs were not slaughtered within 48 hrs)
B The epidemic data presented to EFRA select committee 21/03/01
Mr Brown (8). All cases could be traced back to probable primary source ie Heddon to 20/03. There had been some animal to animal transmission but original source of infection can be traced back to Longtown (perhaps through 4 farms).J Sucdamore (20) in Devon and Cumbria initial seeding was from sheep, the disease is continuing within sheep but has also spread to cattle. The spread is farm to farm by people going on to farm or farm to farm where animals are grazing next door to each other. No wind borne .
(35) The shortest incubation time was 36 hours. Generally find it is between 4-7 days can be up to 14, depending on the dose and how animals infected. (38) 394 confirmed cases (IPs) 341 dangerous contacts (DCs) so far. (41) if disease is found on DCs they become IPs. They remain as DCs unless obvious disease at slaughter ( comment: implies weren't testing) (44) the interval from diagnosis to death is too long at the moment (65) We are confirming large numbers on clinical grounds. Some we are taking samples from just to make sure the clinical decisions are right. (126) Epidemiology- we are behind the game because we find disease and then we have to go back to see where it came from
C. Actual Published Epidemic Data
a) Links between FMD cases The tracing project recently completed by the VLA identifying links between infected premises has not been published. The EFRA select committee were presented with a simplified incomplete version of this data on 31 October 2001 ' links between FMD cases'. Over 300 cases have identifiable links with other infected premises, for many of these farms the farm they are linked to is not within 3 km: that is, infection was not necessarily due to local spread and often occurred before the movement ban was imposed.Confirmation on clinical grounds had been taking place in many cases from as early as 21 February 2001 (NAO 21/06/02 p.61 3.72). By 21/03 90% of infected premises were being confirmed on clinical diagnosis (EFRA select committee 21/03/01 (45)).
b) Clinical diagnosis
c) Laboratory testing Of 909 confirmed cases up to 1/04/2001, only 549 (60%) returned positive laboratory test results (PQ 6650 22/06/02).
d) Slaughter times on Infected Premises (PQ 5479)
% IPs Slaughtered within
24hrs 48hrs 72hrs >72hrs >96hrs
19/02 to 25/03 11 42 72 86 100
Epidemic as a whole 33 72 88 94 100
The assumptions made by the mathematical modellers were incorrect because the dataavailable at the time was incomplete and inaccurate.
We now know that:
-many infected premises had identifiable links, links often occurred before the movement ban and some were not discovered for several weeks ( 171 farms were identified as infected premises on the basis of antibodies only)
-misdiagnosis on infected premises may have been as high as 40% in the first few weeks of the epidemic
-delays in slaughter on the infected premises were significant throughout the epidemic, 58% of infected premises were not slaughtered within 48 hours in the first 5 weeks of the epidemic.
The most likely explanations as to why the data was incomplete and inaccurate are:-
-insufficient resources were allocated to tracing the spread of disease. Tracing was hampered by the difficulty of diagnosing the disease in sheep; the precise start date of the epidemic was/is unknown and sheep movements were extensive, groups of sheep were split and infected lorries inadvertently further disseminated the disease.
- Information technology - there were significant problems entering data accurately on the databases ( eg. up until 18/12/01 there was still data for up to 500 infected premises missing from the disease control system database). Detailed slaughter data sent to the modellers was incomplete.
- there were problems with data transfer between Pirbright and Page Street throughout the epidemic.( Pirbright Submission to RSc Inquiry- Diagnosis 3.2.5)
3. Anecdotal Evidence
The published anecdotal evidence relates to examples where:-
a) mismanagement caused/ may have caused the disease to spread
b) pre-emptive culling policy
This evidence can be requested - e-mail firstname.lastname@example.org
Precise details of disease transmission in the 2001 UK epidemic have not been published. For 88% of the infected premises DEFRA have not published details of how those premises became infected or indeed which ones actually were infected.
If disease transmission was fully investigated not only would it reveal how well the epidemic was handled it would ensure that contingency plans for future epidemics are based on facts rather than assumptions. It is of some concern that present contingency plans include the 24/48 hour pre-emptive slaughter policy as recommended in March 2001. There is still no evidence in the public domain to suggest that the pre-emptive cull policy as implemented in the 2001 UK FMD epidemic was effective.
The published evidence shows that of the 10509 premises culled on1324 (12.6%) of those premises there were animals which had foot and mouth disease or had had foot and mouth disease sometime in the previous weeks (on 171 farms the only evidence of disease was antibodies to FMD).
To assume any of the other premises culled were infected, without first investigating disease transmission for the epidemic as a whole and publishing the evidence, is incorrect.
It is highly likely that, because the mathematical models were developed using incomplete and inaccurate data, the models significantly overestimated potential disease transmission once the movement ban was in place. In addition, the modellers were wrong to assume that the epidemic could not be brought under control using traditional methods.
Pirbright and, possibly DEFRA, disagreed with the pre-emptive slaughter policy recommended by the modellers.
The flaws in the models did not become apparent during the epidemic because:-
- of the 10509 premises culled only 2370 were tested (PQ 6649 19/06/02)
- there was no acknowledgement that the reason traditional methods were not working, in some areas, was because of mismanagement of resources and lack of resources
- resurgence of disease in Yorkshire and Northumberland, and the difficulties in bringing the
disease under control in Cumbria appeared to support the models. Further analysis of the disease transmission in these counties is required before any conclusions can be made.
- NFU and RCVS chose to ignore any members who criticised culling policy and refused to raise these concerns with any of the policy makers
All epidemic data and anecdotal evidence collated to date suggests that at best the pre-emptive cull policy resulted in the unnecessary slaughter of some 8226 farms and at worst it actually prolonged the epidemic by taking vital resources away from the truly infected premises and high risk dangerous contact premises.
To continue to assert that the pre-emptive slaughter policy saved lives is not enough; if the evidence exists to support this statement, that evidence must be published to allow independent researchers to evaluate the pre-emptive slaughter policy.