See also the interim submission made to the Lessons Learned Inquiry by the authors Nicola and Andrew Morris

 

 

 

 

EXTRACT

 

The interim conclusions which can be drawn from this report are:

1.The mathematical models developed for the 2001 UK FMD epidemic may not accurately

reflect the reality of the epidemic; the magnitude of this inaccuracy could be quantified if

more actual epidemic data was released by DEFRA.

 

2.There are inaccuracies in epidemic data, currently in the public domain, particularly

regarding disease diagnosis

 

3.The actual situation regarding farm infectivity is crucial in determining disease control

strategies. The assumption that farm infectivity is constant is from day after infection to

slaughter as used in the mathematical models, is contrary to present scientific

knowledge, previous epidemics and some evidence from this epidemic ( namely the

observation that cattle were 15 times more susceptible to infection than sheep).

 

It is essential that, at the very least, epidemiological case studies are carried out to

establish whether farm infectivity for all types of infected premises did in fact remain

constant, from infection to slaughter, throughout the epidemic. The impact of any

identified increased infectivity should be assessed particularly as increases in infectivity

might explain the uncontrolled spread of disease in the 'hot spots'.

 

4.Given that there is evidence to suggest that many of the 2.5 million animals culled as

dangerous contact or contiguous contacts were not incubating the disease at the time of

slaughter, any evidence to the contrary should be put in to the public domain.

 

 

 

Over reliance on mathematical models, which do not accurately reflect the reality of the 2001 UK epidemic, to explain the evolution of the epidemic will almost certainly mean that any assessment of the success of disease control policies is flawed .

 

 

 

 

 

 

Submission to the Royal Society Inquiry into infectious diseases in livestock

 

From Nicola and Andrew Morris

Eatons Farm

Church Lane

Tibberton

Droitwich

Worcs WR9 7NW

(01905) 345 248

 

 

 

Our submission relates entirely to foot and mouth disease and only covers certain sections of your inquiry:-

diagnosis of disease- in particular misdiagnosis

modelling of disease - accuracy of models

-evidence required to establish accuracy

of models

control of an outbreak- effectiveness of culling

 

 

 

Our submission is only intended to be an interim report.

 

If DEFRA provide us with the information we have requested a further report will be

produced which will hopefully conclusively show that :-

a substantial number of animals have been unnecessarily slaughtered

control of the epidemic could have been achieved through rapid slaughter

of infected premises

most disease transmission occurred as a consequence of delays in slaughter

 

 

We became involved in the FMD epidemic in April 2001, when we were told our cows were to be culled in the contiguous cull. We could see no logical nor scientific reason for the culling of our animals so we opposed the cull.

Our experience prompted us into researching FMD particularly with regard to mathematical modelling. The enclosed paper is the culmination of that research.

 

 

 

Thank-you

Late addition: DEFRA have now confirmed that to 29/03/01 there were 749 laboratory confirmed cases, over 900 clinically confirmed cases awaiting laboratory tests results.

Most of those 900 are now laboratory confirmed (see table 1)

 

2001 UK foot and mouth disease epidemic: An interim report to establish the potential effects of relying mainly on mathematical models rather than actual epidemic data to assess the success of culling policies

 

 

Summary

Mathematical models can reliably be used, with limited actual epidemic data, to explain the evolution of an epidemic, providing the models accurately reflect the reality of the epidemic. There is evidence to suggest that the mathematical models used in 2001 UK FMD epidemic do not accurately reflect the reality of the epidemic. Until more actual epidemic data is released into the public domain, the mathematical models used in the 2001 epidemic can not reliably be used to assess the success of disease control policies .

 

 

Introduction

 

The 2001 UK foot and mouth (FMD) epidemic has resulted in the slaughter of around six million livestock on UK farms. In April 2001 mathematical models were formulated to predict the likely spread of disease, and the disease control policies implemented were determined with regard to those initial predictions. Subsequently, mathematical models have been used to assess the actual effect of the policies implemented, since in implementing control measures the epidemic has evolved differently than if there had been no intervention.

 

To make a valid assessment of the success of disease controls, it is important that a model accurately reflects the reality of the epidemic. Once a model is formulated model output can be compared to actual data both through time and cumulatively, a close match would obviously indicate that the model is accurate.

 

In the 2001 UK FMD epidemic, mathematical models have featured more prominently than actual data in any assessments of the success of disease control policies. If the model accurately reflects the reality of the epidemic then detailed actual data is unnecessary and gross actual figures will suffice. However, if data is limited or inaccurate, errors in the data

may impact on model formulation such that the model is not accurate. In addition, if model output is then compared with actual data which contains inaccuracies, a close match, in itself, will not necessarily support any conclusions regarding the success of any disease control policies.

 

In 2001 UK FMD epidemic, some actual data is limited, in particular identification of source of infection; some data has been minimised or excluded from analyses of the epidemic, namely disease misdiagnosis; and it has been assumed that farm infectivity does not increase over time. The potential effects of non-identification of source of infection and disease misdiagnosis will be examined. Actual epidemic data will be adjusted to include disease misdiagnosis and the significance of the adjusted data will be discussed. Evidence will be presented to show that farm infectivity does increase over time, in some situations, the implications of this evidence will be discussed with regard to current slaughter policies

 

 

 

Source of infection

 

In a FMD epidemic indentification of source of infection for each infected premises may require extensive investigations and such investigations might not ultimately identify infection source. In 2001 UK FMD epidemic actual source of infection for most farms is uncertain (Ferguson et al 2001), for two main reasons:

a) prior to both discovery of FMD on 19 February 2001 and the imposition of movement

restrictions on 23 February initial spread of disease was extensive.

b) the disease has mainly occurred in sheep, symptoms in sheep may be mild and transient

(Donaldson et al 2001) and as a result FMD may pass through a flock unnoticed.

For the purpose of model formulation precise identification of source of infection is not necessary since probability equations can be used to determine both the likelihood of farm 'A' infecting farm 'B', and also the likelihood of disease transmission over a particular distance at a particular time.

 

Disease transmission (spread of disease)

 

For 2001 UK epidemic, where possible, contact tracings were undertaken by DEFRA (formerly MAFF) to determine the source of infection for farms infected in the early stages of the epidemic (DEFRA 2001). These tracings showed that once movement restrictions were imposed the likelihood of disease spreading over a distance of 9 km fell from 38% to 12%, and farms closest to index cases of FMD were at greatest risk of becoming infected (Ferguson et al 2001). By determining the source of infection, the risk of disease transmission (spread) as a function of distance from an infected premises can be estimated. A number of tracings will enable a spatial transmission kernal to be developed.

A spatial transmission kernal is essentially a predictor of the likely transmission of disease as a function of distance from an infected premises. From the contact tracings undertaken by DEFRA , the spatial transmission kernal developed by Ferguson et al (2001) showed that farms 0.5,1 and 1.5 km from a single farm affected by FMD would have probabilities of 0.26,0.06 and 0.02, respectively of becoming infected .

 

Subsequently, in light of the way the 2001 epidemic has evolved, Ferguson et al ( 2001) have established an alternative spatial kernal which predicts considerably more long distance transmission events. The median distance of the alternative spatial transmission kernal, is about 4km: that is, most disease spread occurred from the infected premises up to about 4km. The results from the alternative kernal imply that the original contact tracings process,undertaken by DEFRA, was biased such that closer contacts were more easily identified (Ferguson et al. 2001).

 

It may be that neither the original nor the alternative spatial kernal are reliable predictors of spatial transmission. The reliability of the spatial transmission kernal depends on accurate assumptions being made regarding the infected premises and transmission of disease. Indentification of incorrect assumptions, or errors in actual data may help to establish the reliability of the spatial transmission kernal.

 

 

 

The most likely incorrect assumptions, or errors in actual epidemic data which would impact on the spatial transmission kernal are :-

 

a) index premises have been wrongly classified as secondary infections

b) disease misdiagnosis is more extensive than gross figures suggest and may vary through

time and from region to region

c) farm infectivity may not be constant from the day after the farm is infected until the day

animals are culled

 

 

Classification of index premises

 

Index farms (primary infections) in an epidemic can usually be identified on date of confirmation of infection, such an approach is appropriate for cattle, since clinical infection is obvious but this approach is not entirely appropriate for sheep. In 2001 UK epidemic, 223 infected premises were identified on laboratory diagnosis not clinical diagnosis (DEFRA 2001), in these 223 cases, clinical diagnosis was not possible or was inconclusive because infection date was such that FMD lesions were healed when the infected premises was identified. In such cases, it can be difficult to accurately pinpoint date of infection.

If index premises have been solely identified on the basis that all infections occurring after 3 March (an infection occurring on 23 February would be reported on average 9 days later) are essentially secondary infections then some index premises will be incorrectly classified.

An incorrect classification of an index premises implies that disease transmission has occurred when there has been no transmission.

 

Another situation where transmission is implied but may not have occurred is when the source of infection is the farm itself. Early on in the epidemic it was not unusual, on mixed farms with infected sheep and uninfected housed cattle, that only the sheep were slaughtered. Under such circumstances, when cattle are turned out onto pasture previously grazed by infected sheep, cattle may become infected (Keeling et al 2001 ). For data collation purposes, infected premises with more than one slaughter date the latest date only is used (Ferguson et al 2001(b)), such an approach may effectively mean that an index infection is reclassified and in addition within farm transmission is mistaken for farm to farm transmission.

 

For the epidemic as a whole, incorrect classification of index premises and unidentified within farm transmission is likely to be of minor importance, but it may be of local importance, and may explain situations where there has been resurgence of disease in areas which have been disease free for several weeks. Both incorrect classification of index premises and within farm transmission will have impacted on the transmission kernal such that transmission rates may be overestimated, the magnitude of this overestimation could be calculated by reanalysing the data from those farms infected in the early stages of the epidemic. Such an analysis would also identify those farms where FMD has been misdiagnosed.

 

 

Clinical misdiagnosis of infected premises

 

Clinical misdiagnosis of FMD is a feature of 2001 UK epidemic. Of 2030 infected premises, 1318 have positive laboratory test results, 346 have negative test results and 366 have no test result (DEFRA 2001). False positive test results are unlikely but false negatives can occur. In 1967,of a sample of farms clinical diagnosed with FMD, 98.3% returned positive laboratory test results (Anon 1968).

 

Donaldson and Kitching (2001) have stated that, for the 2001 UK epidemic, they would expect at least a 90% correlation (agreement)between clinical diagnosis and laboratory diagnosis: the correlation is in fact 73.7% if all untested farms are assumed to be positive and 64.9 % if all untested farms are assumed to be negative. The status of untested premises could probably be ascertained from epidemiological investigations; such investigations would enable a more accurate correlation between clinical and laboratory diagnosis to be determined for the 2001 UK epidemic.

 

The effect of clinical misdiagnosis, on both model formulation and any subsequent analysis of the effectiveness of disease control measures, might be negligible if the degree of misdiagnosis did not vary over time and from region to region.

 

Clinical misdiagnosis over time

 

In the first 37 days of the epidemic (up to 28 March), confirmation of an infected premises, on the whole, required a positive laboratory test ; thus, it is unlikely that many farms clinically diagnosed as being infected with FMD which were in fact disease free (and therefore tested negative at laboratory test) would be confirmed as infected premises up until 28 March. Thus, clinical misdiagnosis is unlikely to be a significant feature of the initial contact tracings undertaken by DEFRA .

 

From 29 March, infected premises were confirmed on clinical grounds alone; if these farms subsequently returned negative test results their status as infected premises was retained (DEFRA 2001). Keeling et al (2001) noted that over-reporting of the true number of cases (misdiagnosis) was largest (up to 25%) shortly after the peak of the epidemic ( on 26 March 2001).

 

Regional variations in clinical misdiagnosis

 

There are regional differences in the degree of misdiagnosis: for example in Southwest England 25% of infected premises had negative laboratory test results and 21% were not tested (DEFRA 2001) ; in Hereford and Worcester the results were 45% and 35 % respectively (Morley 2001).

In 2001 UK epidemic the amount of clinical misdiagnosis has varied over time and from region to region. The impact of clinical misdiagnosis in this epidemic is twofold: it implies that disease transmission is greater than it actually is, and it is likely that model outputs match confirmed FMD cases more closely than would be the case if misdiagnosed premises were excluded from actual epidemic data.

 

 

 

 

Adjusted actual epidemic data to allow for clinical misdiagnosis

 

Table 1 Confirmed cases of FMD from 20 February to 30 September 2001

Actual epidemic data Actual epidemic data adjusted to allow for misdiagnosed

assuming all confirmed and untested cases (DEFRA 3)

cases are positive

confirmed cases adjusted data

time period DEFRA DEFRA DEFRA 1 DEFRA 2

covered 1 2 total positive negative not tested total positive negative not tested

cases cases cases cases cases cases cases cases cases cases

20/02- 29/03 749 916 749 674 36 39 916 841 36 39

30/03 - 5/04 250 189 250 126 60 64 189 81 53 55

6/04 - 12/04 250 224 250 126 61 64 224 96 62 66

13/04 - 28/04 250 205 250 126 61 64 205 88 57 60

29/04 - 4/06 200 186 200 100 48 51 186 80 51 55

5/06 - 27/07 200 192 200 100 48 51 192 82 54 56

28/07 - 30/09 131 118 131 66 32 33 118 50 33 35

Total 2030 2030 2030 1318 346 366 2030 1318 346 366

 

Notes:

Data sources

DEFRA 1 - complete list of confirmed cases from DEFRA website, revised 31/07/01

DEFRA 2 - clinically confirmed cases extracted from the FMD daily situation report , version last updated 23/11/01.

Daily totals have been extracted from the graph and added to give period totals.

DEFRA 3 - correspondence from FMD communications branch- statistics given for IPs 1318 positive, 346 negative and

366 untested.

 

Explanations

Differences between DEFRA 1 data and DEFRA 2 data are to with changes in reporting policy. In the early stages of the epidemic most reported cases of FMD were not confirmed until positive laboratory tests were received (often 3 days), initially it was date of confirmation which was recorded. DEFRA 1 data has been altered to reflect this discrepancy and hence given rise to DEFRA 2 data. i.e. DEFRA 1 data is date of confirmation. DEFRA 2 data is date of clinical infection.

 

Assumptions

An assumption has been made that 92 % of the cases occurring up until 29/03 were positive, this assumption is based on the fact that confirmation of disease ,in most cases, required a positive laboratory test.

 

 

Analysis of the impact of misdiagnosis on atual epidemic data

 

The actual data for the number of confirmed cases with negative laboratory test result or no laboratory test result over time and by region has been requested from DEFRA. Without detailed actual data, it is not possible to assess whether misdiagnosis has been of significant magnitude to materially alter the close match between observed disease incidence and output from the mathematical models used in 2001 UK FMD epidemic.

The adjusted data, as presented in table 1, may well be a reasonable estimate of the actual data; however, the validity of any analysis using the data in table 1 could be questioned because assumptions have been made regarding the distribution over time of negative and untested premises. In addition, no allowance has been made for false negative tests and no valid assessment can be made regarding the actual disease status of untested farms.

 

It is, however, valid to examine case scenarios: -

 

Under the best case scenario (all negative and untested farms are uninfected) of infected premises clinically confirmed with FMD from 29 March to 30 September only 43% had FMD.

Under the worst case scenario (all untested farms are infected) of infected premises clinically confirmed with FMD from 29 March to 30 September 72 % had FMD.

 

 

Misdiagnosis has resulted in the unnecessary slaughter of not only the misdiagnosed infected premises but also many contiguous premises or those premises having a 'dangerous contact' with a misdiagnosed premises. It is highly likely that misdiagnosis, will only significantly impact on the match between model output and observed disease incidence in the decline phase of the epidemic. Although misdiagnosis may have given the impression that FMD was spiralling out of control outside the heavily infected areas, misdiagnosis does not explain disease transmission in the heavily infected areas.

 

 

 

FMD Transmission in heavily infected areas (' hot spots ')

 

As discussed in previous sections some transmission in the 'hot spot' areas was in fact almost certainly overestimated: that is, index premises wrongly classified as secondary infections, secondary infections occurring within the farm itself and some misdiagnosis of disease.

However, even taking into account some overestimation of transmission, there is no doubt that under the control measures in place, FMD was still spreading.

 

timetable of control measures

 

Date control measure

 

23 February Movement restrictions. Culling of IPs and DCs

IPs culled on laboratory confirmation

IP slaughter>24hrs after confirmation

15 March some culling of sheep within 3 km of IP

23 March some CP slaughter in hot spots

28 March IPs culled on clinical confirmation only

IP slaughter > 24hrs after clinical confirmation

29 March IP slaughter< 24 hrs after clinical confirmation

CP slaughter < 48 hrs after clinical confirmation of disease on IP

in ' hot spots'

10 April 24/48 hour cull of IP/CP applied to all affected areas and backdated

to date to 15 March (DEFRA 2001)

 

Definitions: IP- infected premises DC - dangerous contact CP- contiguous premises

sources: Ferguson et al 2001. Keeling et al 2001.

 

For as long as FMD was confined to sheep any delays in slaughter might not result in any more disease transmission other than was already occurring through animal to animal contact. Early on in the heavily infected areas delays in slaughter were significant ( up to 4.5 days). Once cattle became infected any delays in slaughter would almost certainly have resulted in wind borne transmission and a significant rise in the number of secondary infections.

 

 

Howard and Donnelly (2001) reported that in the 1997 Taiwan epidemic and 1967 UK epidemic if all herds had been slaughtered and carcases disposed of on day of diagnosis the number of herds infected would have been reduced by 60% and 25% respectively.

For 2001 UK epidemic, a policy of rapid slaughter of infected premises was only implemented for a short period before it was decided that, as farm infectivity did not appear to increase with time, reducing delays in slaughter would not be sufficient to control the spread of disease, thus, additional culling of contiguous premises was deemed necessary to bring the epidemic under control.

 

 

 

 

 

Farm infectivity

 

The mathematical models formulated by Ferguson et al (2001) and Keeling et al (2001) for UK FMD epidemic assume that farm infectivity is constant and does not increase from infection to slaughter. The latter assumption is based on data from contact tracings carried out by DEFRA at the start of the epidemic which showed that the rate at which secondary infections were generated was approximately constant throughout the infectious period (Keeling et al 2001): that is farm infectivity is constant.

 

Evidence to support the hypothesis that farm infectivity does increase throughout the infectious period

 

 

a) Science

It is known that incubation periods vary between species and between animals of the same species (Howard and Donelly 2001). Unless, all animals on a particular farm are all infected at the same time and their incubation periods are identical there will be a build-up of infection within a farm.

Increases in farm infectivity would almost certainly result in wind borne transmission. As cattle a more susceptible to windborne transmission (Donaldson et al 2001) any increases in farm infectivity are more likely to lead to more secondary infections in cattle than in any other species.

 

 

b) History

 

Both 1967 UK and 1997 Taiwan epidemic demonstrate that in FMD epidemics involving cattle or pigs speed of slaughter is critical to reduce the number of secondary infections.

By implication, for epidemics involving cattle or pigs farm infectivity increases from infection to slaughter, if this were not the case then reductions in slaughter time would have minimal impact on number of secondary infection

In 1997 Taiwan outbreak, once a policy of rapid slaughter was implemented the epidemic was brought under control within 12 days.

 

 

c) 2001 UK FMD epidemic

 

For as long as FMD was confined to sheep, as it was in the early stages of the epidemic, any increase in farm infectivity might only become apparent when cattle became infected. It is interesting to note that Keeling et al (2001) noted that, in 2001 epidemic, cattle were 15 times more susceptible to infection than sheep, this susceptibility has been attributed to an assumed greater contact between cattle and humans/vehicles than would be expected with sheep. Given that cattle are 10 times more susceptible to wind borne spread than sheep perhaps the susceptibilities record by Keeling et al (2001) are an indication that wind borne transmission has been a factor in disease transmission. Wind borne transmission could possibly be considered to be an indication that farm infectivity does increase from infection through to slaughter.

 

From 29 March a policy of rapid slaughter of infected premises was implemented. Given an average infection to report time of 9 days the impact of this policy on daily incidence of infection would not start to take effect until 6 April. The epidemic peaked at a daily incidence of 54 cases, by 2 April daily incidence was 34 cases. However, allowing for misdiagnosis average daily incidence for week begining 30 March may have been 12.

 

The contiguous cull policy for 2001 epidemic did not become national policy until 9 April, though it was implemented in some areas from 2 April, the impact of this policy would start to take affected between 10 April and 17 April. For the week begining 4 April average daily incidence was 32, allowing for misdiagnosis average daily incidence may have been 14.

 

It would appear that a policy of rapid slaughter alone, as implemented from 29 March was starting to bring the epidemic under control, particularly if the effect of misdiagnosis is taken into account. Ferguson et al 2001 have previously demonstrated that if farm infectivity peaked after diagnosis on the infected premises then rapid slaughter of an infected premises alone would be sufficient to bring 2001 UK epidemic under control (Fig A).

 

A more accurate analysis of actual epidemic data, that is one that takes into account misdiagnosed and untested premises, may demonstrate conclusively that rapid slaughter of infected premises was bringing the 2001 epidemic under control. If the latter scenario is accurate then by implication farm infectivity does increase from infection through to slaughter. In addition, increases in farm infectivity offer the most likely explanation for the uncontrolled spread in the hot spot areas and also call in to question the need for a contiguous cull.

 

 

The implications of the incorrect classification of index premises and disease misdiagnosis on actual epidemic data and the contiguous cull

 

 

Incorrect classification of index premises and disease misdiagnosis, if taken into account, will alter the actual data for the epidemic. Once altered the actual data may no longer closely match the output of the mathematical models developed for 2001 UK epidemic.

This mismatch may only be of a significant magnitude in the decay phase of the epidemic. Under such circumstances it could be argued that the mathematical models, still acurately reflect the epidemic from the start to the peak and, therefore, still show that a contiguous cull was essential to bring the epidemic under control.

More detailed information regarding disease misdiagnosis over time and by region together with correct classification of infected premises may show that 9 days after implementation of a rapid slaughter policy the actual drop in daily disease incidence would indicate that the epidemic was under control.

Unfortunately, because the contiguous cull policy was implemented at about the same time as the rapid slaughter policy, it could be argued that the contiguous cull was as important as the rapid slaughter policy in bringing the epidemic under control. It could be argued that the contiguous cull was necessary because animals contiguous to an outbreak were incubating the disease at the time of slaughter

 

Evidence to support the theory that animals contiguous to an outbreak were incubating FMD at the time of slaughter

 

 

a)     blood tests from contiguous premises

 

Though a requirement of EU Directive 85/511/EC there do not appear to be blood tests for

many if any contiguous farms. The information has been requested from DEFRA, the reply indicated that blood tests were taken from animals classed as dangerous contact as far as possible but not from contiguous farms.

 

 

b) blood tests from dangerous contacts

The policy is to blood test 20% of animals within a herd or flock, at slaughter. If evidence of disease is found then the herd or flock will be reclassified as an infected premises.

In this epidemic, over 1 million animals have been slaughtered as dangerous contacts non contiguous, by the very nature of the definition these animals will have been in closer contact to infected animals than most animals on premises which are simply contiguous.

Many of the dangerous contact animals have/ should have been tested, clearly many were found not to be infected: if they were infected they would have been removed from the dangerous contact category to become infected premises.

 

c) contiguous premises not culled

 

Ferguson et al (2001) calculated that for every infected premises (IP) in a typical neighbourhood 16% of the farms would be contiguous to the infected premises. Culling of contiguous farms within 7 days of slaughter on the IP ranged from 5% of the neighbourhood to 10%, in April and May. The latter figures imply that many contiguous farms were not culled, or were not culled within 7 days (if culled more than 7 days after slaughter of the IP, it is likely that there would be some evidence of disease); how many of those farms got FMD ?

 

There is currently no evidence in the public domain to support the theory that animals culled as contiguous culls or as dangerous contacts were incubating the disease at the time of slaughter. There is evidence to suggest that most of these animals ( nearly 2.5 million) were culled unnecessarily.

 

 

 

The implication of farm infectivity assumptions on culling policies

 

The assumption that farm infectivity remained constant from the day after infection up until slaughter (Ferguson et al 2001 & Keeling et al 2001) was the key assumption that effectively implied that a contiguous cull was necessary to bring the epidemic under control.

This assumption is not supported by current scientific knowledge, previous experience (1967 UK and 1997 Taiwan epidemics) or some evidence from this epidemic.

 

The assumption that farm infectivity remained constant was based on data from contact tracings carried out by DEFRA at the start of the epidemic which showed that the rate at which secondary infections were generated was approximately constant throughout the infectious period (Keeling et al 2001): that is farm infectivity was constant.

It is essential that more local and regional investigations are undertaken to establish whether farm infectivity, in this epidemic, was constant.

 

Any evidence of increases in infectivity would conclusively show that rapid slaughter alone of infected premises was sufficient, coupled with movement restrictions and adequate biosecurity (as implemented in this epidemic), to control the 2001 UK FMD epidemic.

 

 

 

 

 

Conclusion

 

The interim conclusions which can be drawn from this report are:

1.The mathematical models developed for the 2001 UK FMD epidemic may not accurately

reflect the reality of the epidemic; the magnitude of this inaccuracy could be quantified if

more actual epidemic data was released by DEFRA.

 

2.There are inaccuracies in epidemic data, currently in the public domain, particularly

regarding disease diagnosis

 

3.The actual situation regarding farm infectivity is crucial in determining disease control

strategies.The assumption that farm infectivity is constant is from day after infection to

slaughter, as used in the mathematical models, is contrary to present scientific

knowledge, previous epidemics and some evidence from this epidemic ( namely the

observation that cattle were 15 times more susceptible to infection than sheep).

It is essential that, at the very least, epidemiological case studies are carried out to

establish whether farm infectivity for all types of infected premises did infact remain

constant, from infection to slaughter, throughout the epidemic. The impact of any

identified increases infectivity should be assessed particularly as increases in infectivity

might explain the uncontrolled spread of disease in the 'hot spots'.

 

4.Given that there is evidence to suggest that many of the 2.5 million animals culled as

dangerous contact or contiguous contacts were not incubating the disease at the time of

slaughter, any evidence to the contrary should be put in to the public domain.

 

 

 

Over reliance on mathematical models, which do not accurately reflect the reality of the 2001 UK epidemic, to explain the evolution of the epidemic will almost certainly mean that any assessment of the success of disease control policies is flawed .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

References

 

Anon (1968) Northumberland Report: The report of the Committee of Inquiry on Foot

and Mouth Disease. London. HMSO.

 

DEFRA (2001) Private correspondance from Marie Chapman Foot and Mouth

Communications Branch.

 

DEFRA (2001) www.defra.gov.uk/footsandmouth.

 

Donaldson A I and Kitching P. (2001) FMD Diagnosis. Letter. The Veterinary Record

148. 640.

 

Donaldson A.I.,Alexandersan S., Sorensen J.H.,& Mikkelsen T. (2001). Relative risks

of uncontrollable (airborne) spread of FMD by different species. The Veterinary

record.148. 602-604.

 

Ferguson N.,Donnelly C. & Anderson R. (2001). The Foot and Mouth Epidemic in Great

Britain: Pattern of spread and impact of interventions. Sciencexpress.

www.sciencexpress.org/12 April 2001/10.1126/science.1061020. pages 1-5.

 

Ferguson N.,Donnelly C. & Anderson R. (2001). Transmission intensity and impact of

control policies on the foot and mouth epidemic in Great Britain. Nature .413.

542-547. Supplementary information was used- available on www.nature.com.

 

Howard S.C. & Donnelly C. (2001). The importance of immediate destruction in epidemics

of foot and mouth disease. Research in Veterinary Science. 69. 189-196.

 

Keeling M.J., Woolhouse M.E.J.,Shaw D.J.,Matthews L.,Chase-Topping M., Haydon D.T,

Cornell S.J.,Kappey J., Wilesmith J. & Grenfell B.T. Dynamics of the 2001 UK Foot and

Mouth Epidemic: Stochastic Dispersal in a Heterogeneous Landscape.

Science.294. 813-817. Supplementary information was used-available on

www.sciencemag.org.

 

Morley E. (2001). Parliamentary Question (PQ 53) Question 24

Written 5/07/01 (on order book 22/06/01).